Split (1)

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

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W4362659518	https://zenodo.org/record/7697453/files/OOPSLA%20Paper%20137%20Artifact.pdf	English	null	Algebro-geometric Algorithms for Template-Based Synthesis of Polynomial Programs	Proceedings of the ACM on programming languages	2,023	cc-by	2,164	Algebro-geometric Algorithms for Template-based Synthesis of Polynomial Programs Amir Kafshdar Goharshady1, S. Hitarth1, Fatemeh Mohammadi2, and Harshit J. Motwani3 Amir Kafshdar Goharshady1, S. Hitarth1, Fatemeh Mohammadi2, and Harshit J. Motwani3 Amir Kafshdar Goharshady1, S. Hitarth1, Fatemeh Mohammadi2, and Harshit J. Motwani3 1 Hong Kong University of Science and Technology, Hong Kong, goharshady@ust.hk, hsinghab@cse.ust.hk 2 KU Leuven, Leuven, Belgium fatemeh.mohammadi@kuleuven.be 3 Ghent University, Ghent, Belgium harshitjitendra.motwani@ugent.be 3 Ghent University, Ghent, Belgium harshitjitendra.motwani@ugent.be 1 Introduction We provide the code of our algorithms for synthesis written in Python3. – The artifact is provided as a Virtual Machine. The virtual machine is available for download at https://drive.google.com/drive/folders/ 16 FduQ7wIeBcqr673EmVxepz08KSktx-?usp=sharing. – The md5sums of the paper 137. ova, the VM of the artifact, is f8abbd6509c54f9f3ae2ee9e7a52e5a6. – We have uploaded the artifact on Zenodo with DOI: https://zenodo.org/ record/7512455. 2.2 Correspondence Table The correspondence between the main aspects of our algorithm pre- sented in the paper and implemented in our code are explained in Table 1. The source code is available on GitHub at https://github.com/hitarths/ polysynth. The source code is available on GitHub at https://github.com/hitarths/ polysynth. The benchmarks we synthesized and reported in our paper are obtained using this code. All the benchmarks mentioned in the paper are available in the source code in various formats as follows: 1. PolySynth: For our algorithms, we write the program template in a C-like language, as explained in the paper. The benchmarks are available in the directory polysynth/benchmarks -polysynth. Our pre-solved results are available in the directory polysynth/Code/Examples. n the directory polysynth/Code/Examples 2. Sketch: We have provided the benchmarks in the format suitable for Sketch in the directory polysynth/benchmarks -sketch. Note that sketch files are only available for the supported examples. 3. Rosette: We have provided the benchmarks in the format suitable for Rosette in the directory polysynth/benchmarks -rosette We have included the command line version for Sketch, Rosette, and our tool to make it easier for reviewers to run these tools on the benchmarks. Moreover, we also provide shell scripts inside the Virtual Machine that allow the reviewers to run these tools for all the benchmarks with a single command and create a CSV file with details about the outputs from these tools. We will explain this in detail later in the document. We have used the following external tools/libraries: Z3 [2], Sketch [3], and Rosette [1] in our codes. The artifact code consists of the following directories. The artifact code consists of the following directories. 1. The directory benchmarks-polysynth contains all the benchmarks in the format that can be passed to our tool PolySynth. 2. The directory benchmarks-rosette contains all the benchmarks that can be passed to Rosette. [1] 3. The directory benchmarks-sketch contains all the benchmarks that can be passed to Sketch [3]. 3. The directory benchmarks-sketch contains all the benchmarks that can be passed to Sketch [3]. 4. The directory Code contains the code for our main algorithms and tool PolySynth. 5. The directory external-tools should contain the executable of the tool Sketch that we use to compare our work with. Please install it manually if you pull the code from Git. We have pre-installed this in our virtual machine. 6. The directory polysynth-outputs contains the output programs and other intermediate files our tool Polysynth creates for all of the benchmarks. 7. The root directory of the repository contains various .sh files that are useful for conveniently running the benchmarks on our tool as well as other external tools. 2.1 Directory Structure of the Code The artifact code consists of the following directories. 3 Getting Started Guide We explain the process of basic setup step by step in the following: We explain the process of basic setup step by step in the following: 1. Download and install the VirtualBox on your system (https://www. virtualbox.org/wiki/Downloads) 2. Setting up the Virtual Machine: Download the VM from the link pro- vided in the introduction 3. Open the paper 137. ova, and this should open the Ubuntu VM in the VirtualBox Description Paper Code Grammar of the polynomial programs Section 2 (Syntax) Code/parser.py function get parser Creating Templates and Translation to SPTS Section 3.1 (Step 1) Code/PolynomialTS.py function graph preprocessor Generating Entailment Con- straints Section 3.1 (Step 2) Code/PolynomialTS.py function get constraint pairs Eliminating Program Vari- ables and Reduction to QP Section 3.1 (Step 3) Code/synthesizer.py function get QP for condition pair Constraint solving with our heuristic Section 3.1 (Step 4) Code/synthesizer.py function solve with abstraction refinement Farkas’ Lemma Theorem 3.5 Code/synthesizer.py lines 191-193 Handelman’s Theorem Theorem 3.9 Code/synthesizer.py lines 137-149, 181-182 Putinar’s Positivstellensatze Theorem 3.12 Code/synthesizer.py lines 88-108, 195-197 Table 1: Paper-to-Code Correspondence Table Description 4. The username for the VM is polysynth, and the password is admin. 4. The username for the VM is polysynth, and the password is admin. 5. Open the terminal and change your directory to /home/polysynth/ polysynth with the command: cd /home/polysynth/polysynth. This is the root directory of our repository. 6. The python packages and other external tools are pre-installed in the VM, so the reviewer does not need to perform any installation by themselves. Yet, we have provided all the required instructions/commands in the Readme.md file present in the root directory of the repository. 5. Open the terminal and change your directory to /home/polysynth/ polysynth with the command: cd /home/polysynth/polysynth. This is the root directory of our repository. 6. The python packages and other external tools are pre-installed in the VM, so the reviewer does not need to perform any installation by themselves. Yet, we have provided all the required instructions/commands in the Readme.md file present in the root directory of the repository. 3.1 Running a single benchmark on the different tools 1. PolySynth: Run the command ./ run_polysynth_one_example.sh We can find the output of the program in the directory ./polysynth - outputs/petter3 which will contain the .smtlib2 files generated using our algorithm, and the .synth file that has the synthesized program. 3. Sketch: Run the command ./ run_sketch_one_example.sh The output of Sketch will show any error that occurred during the synthesis. Note that in many benchmarks, Sketch will return an output but it might be incorrect as Sketch works for fixed-point numbers, and all our programs assume that all variables are real numbers. 3.2 Running all benchmarks on the different tools With the following shell command, one can run any tool on all the supported benchmarks. We have set the timeout to be 60 seconds by default, but one can easily change this in the shell script. 1. PolySynth: Run the command 1. PolySynth: Run the command ./ run_polysynth_all_benchmarks.sh This command will create a CSV file at the root directory polysynth - outputs.csv that mentions all the benchmarks on which our tool worked and failed. Note that timeout might be required to be increased depending on the RAM and other resources provided to the VM. For each benchmark, we generate its own directory polysynth to save all of its outputs. ./ run_polysynth_one_example.sh gi y g 2. Rosette: Run the command ./ run_rosette_one_example.sh This will print the time taken by Rosette to synthesize the program. If the Rosette fails, then it will not print anything. 3. Sketch: Run the command ./ run_sketch_one_example.sh The output of Sketch will show any error that occurred during the synthesis. Note that in many benchmarks, Sketch will return an output but it might be incorrect as Sketch works for fixed-point numbers, and all our programs assume that all variables are real numbers. 3. Sketch: Run the command ./ run_sketch_one_example.sh The output of Sketch will show any error that occurred during the synthesis. Note that in many benchmarks, Sketch will return an output but it might be incorrect as Sketch works for fixed-point numbers, and all our programs assume that all variables are real numbers. ./ run_polysynth_all_benchmarks.sh 2. Rosette: Run the command 2. Rosette: Run the command ./ run_rosette_all_benchmarks.sh This command will generate a CSV file at the root directory rosette - outputs.csv that mentions all the benchmarks on which Rosette worked and failed. 3. Sketch: Run the command ./ run_sketch_all_benchmarks.sh This command will generate a file at the root directory sketch -outputs. csv that mentions the full output of Sketch on each benchmark it ran. Note that we have to manually check whether the program generated by Sketch is valid or not, so we do not generate a CSV like the other cases. The reviewer can manually inspect the output for each benchmark if required to match it with the table in our paper. 3. Sketch: Run the command ./ run_sketch_all_benchmarks.sh 4. Compiling all the results: After running all the ‘./run [polysynth,rosette,sketch] all benchmarks.sh‘, user can run the following command to compile the results of all three tools to one single CSV file named ‘compile-table.csv‘. 4. Compiling all the results: After running all the ‘./run [polysynth,rosette,sketch] all benchmarks.sh‘, user can run the following command to compile the results of all three tools to one single CSV file named ‘compile-table.csv‘. python3 compile_all_results.py The main options for the synthesizer.py are the following two: The main options for the synthesizer.py are the following two: 1. --filename: The path to the template for the polynomial program with holes that has to be synthesized 2. --target: The path to the folder where the output files and synthesized program should be saved A few other technical arguments that can be passed are given in the following. In general, one can ignore setting them and the synthesizer will run using the default configuration. 1. –apply-heuristics: Set it to 0 to disable our abstraction-refinement based heuristics, otherwise 1 (default is 1) 2. --apply -handelman: Set it to 0 to disable applying Handelman theorem to generate QP, otherwise 1 (default is 1) 3. --only -smtlib -files: Set it to 1 to only generate smtlib2 files that can be passed to other solvers. The default value is 0 and the z3 solver will be automatically used to solve the QP and synthesize the program. 4. --h-degree: The h degree is used during the application of the Handelman theorem and Putinar’s Stallensatze. The default value is 2. Please refer to the paper for more details. 4 Step by Step Instructions In this section, we will explain how to run our tool in detail along with other technical details for the interested reader. Input. We have defined the domain-specific language for input to our PolySynth algorithm in the paper. One can check the parser.py file for more detail on the syntax. An example input is as follows: r = 0; while( (x>=(r)), ((a == ( ([(x), 1] + (rr)) - r) ) and (x>=0))) { r 0; while( (x>=(r)), ((a == ( ([(x), 1] + (rr)) - r) ) and (x>=0))) { { x = ([(x,r) ,1]); r = (r + 1); } } @post (((((rr) - r) >= (a-(2r))) and (((rr) - (r)) <= (a)))); } @post (((((rr) - r) >= (a-(2r))) and (((rr) - (r)) <= (a)))); A few important syntactic details are as follows: A few important syntactic details are as follows: 1. Every expression must be fully parenthesized. 1. Every expression must be fully parenthesized. 2. The hole, for example, [(x,r) ,1] represents a symbolic polynomial over the variables x,r of degree 1. 3. The first line real a,x,r defines all the (real) variables that are used in the program. g 4. @pre() defines the pre-condition of the program. 4. @pre() defines the pre-condition of the program. 5. This is followed by the body of the program. 5. This is followed by the body of the program. 6. @post() defines the post-condition of the program. 6. @post() defines the post-condition of the program. To run our algorithm on this input, saved in file closest\_square\_root .c, we will use the following command: python3 synthesizer.py python3 synthesizer.py --filename benchmarks -polysynth/closest_square_root.c --target polysynth -outputs/closest_square_root/ --filename benchmarks -polysynth/closest_square_root.c --target polysynth -outputs/closest_square_root/ arget polysynth -outputs/closest_s The synthesized program is the following: The synthesized program is the following: real a,x,r; @pre(a >= 1); x = (a * 0.5); r = 0; while (x >= r, (a == (((0) (1) +(2)(x) + (r * r)) - r) and x >= 0)) { x = (0) (1) +(1)(x)+(-1)(r); r = (r + 1); r 0; while (x >= r, (a == (((0) (1) +(2)(x) + (r r)) - r) and x >= 0)) { x = (0) (1) +(1)(x)+(-1)(r); } @post ((((r r) - r) >= (a - (2 * r)) and ((r * r) - r) <= a)); As we can see, the hole is replaced by a concrete polynomial. 4.2 Performance of our algorithm and other tools Our tool should usually take less than 60 seconds to run on any benchmark. 2. De Moura, L., Bjørner, N.: Z3: An efficient smt solver. In: Proceedings of the Theory and Practice of Software, 14th International Conference on Tools and Algorithms for the Construction and Analysis of Systems. p. 337–340. TACAS’08/ETAPS’08, Springer-Verlag, Berlin, Heidelberg (2008) p g g, , g ( ) 3. Solar-Lezama, A.: Sketch. https://github.com/asolarlez/sketch-frontend (2020) 1. Rosette. https://emina.github.io/rosette/ (2021) References 1. Rosette. https://emina.github.io/rosette/ (2021)
https://openalex.org/W1990340910	https://www.um.edu.mt/library/oar/bitstream/123456789/45881/1/Smoking_health_professional_student_an_attitudinal_challenge_for_health_promotion_2012.pdf	English	null	Smoking Health Professional Student: An Attitudinal Challenge for Health Promotion?	International journal of environmental research and public health/International journal of environmental research and public health	2,012	cc-by	5,214	Int. J. Environ. Res. Public Health 2012, 9, 2550-2561; doi:10.3390/ijerph9072550 Int. J. Environ. Res. Public Health 2012, 9, 2550-2561; doi:10.3390/ijerph9072550 International Journal of Environmental Research and Public Health ISSN 1660-4601 www.mdpi.com/journal/ijerph International Journal of Environmental Research and Public Health ISSN 1660-4601 www.mdpi.com/journal/ijerph Article 1. Introduction Tobacco use is one of the major preventable causes of premature death and disease in the World. The World Health Organization (WHO) estimates that globally, over 1 billion people currently smoke tobacco [1]. It also attributes approximately 6 million deaths a year to tobacco, and this is expected to rise to around 10 million per year by 2030 [2]. Health professionals are uniquely positioned in that simple, targeted, brief and well-timed interventions on their patients can have an impact on smoking habits, particularly through effective patient counselling [3]. Health professional students equipped with knowledge to provide smoking cessation skills to future patients may also play a crucial role in reducing tobacco use and ultimately reducing smoking-related deaths [4,5]. There is a need for health care facilities and schools to assume a greater responsibility for students’ tobacco education by promoting non-smoking facilities and practices [6]. While physician smoking prevalence in developed countries seems to have decreased over the past decades, this is not the case for several southern European countries and developing countries [7,8], where an expanding body of evidence shows that the prevalence of tobacco smoking is rather high in current health professional students [9–12]. However, few studies have collected standardised information regarding future health professionals’ tobacco use, exposure to second-hand smoke, and training to provide cessation counselling. In fact, smoking issues are often taught in a non- systematic manner and are of limited quality in many countries [13–15]. The WHO and the U.S. Centre for Disease Control and Prevention, have attempted to overcome these limitations by developing and implementing the Global Health Professions Student Survey (GHPSS), which includes surveys of dental, medical, nursing, and pharmacy students’ tobacco habits. Results of the survey have led authors to advocate the introduction of a separate tobacco module in medical schools to counter these worrisome trends [10,16]. Malta is a small Mediterranean island which first participated in the GHPSS in 2010. In 2008 the prevalence of tobacco use of the Maltese adult population stood at 20.4% daily smokers and a further 5.5% as occasional smokers [17]. 1. Introduction Very limited data on the smoking habits of Maltese health professionals/students had been collected previously, although a EUROPREV self-reported survey in 2000 indicated that 12.8% of Maltese GPs were regular smokers [18], and a small study carried out in 2009 on a representative sample of Maltese doctors and medical students (n = 71) indicated a smoking prevalence of around 5% [19]. The aim of this survey was to assess whether smoking habits influenced Maltese health professional students’ attitudes towards tobacco control. Int. J. Environ. Res. Public Health 2012, 9 2551 Int. J. Environ. Res. Public Health 2012, 9 Daniel Cauchi * and Julian Mamo Daniel Cauchi * and Julian Mamo Department of Public Health, Malta Medical School, Mater Dei Hospital, Birkirkara Bypass, Msida MSD 2090, Malta; E-Mail: julian.mamo@um.edu.mt * Author to whom correspondence should be addressed; E-Mail: dcauchi@gmail.com; Tel.: +356-7961-7648; Fax: +356-2340-1304. Received: 25 June 2012; in revised form: 16 July 2012 / Accepted: 16 July 2012 / Published: 23 July 2012 Abstract: Tobacco is a major preventable cause of premature morbidity and mortality. Health professionals are uniquely positioned to provide targeted interventions and should be empowered to provide cessation counselling that influence patient smoking. A cross-sectional national survey was administered to all third year students in four disciplines at the University of Malta. The Global Health Professional Student Survey (GHPSS) questionnaire was distributed to collect standardised demographic, smoking prevalence, behavioural, and attitudinal data. 81.9% completed the questionnaire (n = 173/211). A positive significant association between tobacco smoke exposure at home and current smoking status was identified. Non-smokers regarded anti-tobacco policies more favourably than smokers, being more likely to agree with banning of tobacco sales to adolescents (OR 3.6; 95% CI: 2.5–5.3; p ≤ 0.001); and with a smoking ban in all public places (OR 8.9; 95% CI: 6.1–13.1; p ≤ 0.001). Non-smokers favoured a role for health professionals in promoting smoking cessation (OR 5.1; 95% CI: 3.1–8.5; p ≤ 0.001). Knowledge of antidepressants as tools for smoking cessation was also associated with a perceived role for skilled health professionals in cessation counselling (OR 4.9; 95% CI: 1.8–13.3; p = 0.002). Smoking negatively influences beliefs and attitudes of students toward tobacco control. There is a need to adopt a standard undergraduate curriculum containing comprehensive tobacco prevention and cessation training to improve their effectiveness as role models. 2.1. Design The Global Health Professionals Student Survey (GHPSS)—part of the Global Tobacco Surveillance System—is an international school-based cross sectional survey of third year students 2552 Int. J. Environ. Res. Public Health 2012, 9 pursuing advanced degrees in dentistry, medicine, pharmacy, and nursing [16]. A self-administered, anonymous and validated [20] questionnaire was distributed to collect demographic, prevalence, behavioural and attitudinal data on tobacco use and cessation among health professional students. Fieldwork was carried out between January and May 2010 at the Malta Medical School and the Institute of Health Care. In Malta the medical, dental and pharmacy undergraduate courses last five years, whereas nursing students pursue one of two undergraduate courses of 3 and 4 years of duration, respectively. Malta is unique in that its small size and single acute general teaching hospital allow for a limited number of health care students across all disciplines to graduate each year. Thus, the sample population (n = 211) contained all third year students across all eligible disciplines in Malta, and the results can be used to make important inferences concerning tobacco use and risk behaviours of third year health professional students in the country. Ethical approval was obtained through the University of Malta Research Ethics Committee. The tool included questions on demographic characteristics, smoking habits, use of alternative tobacco products, attitudes and beliefs towards tobacco control activities and the educational training received with regards to smoking and smoking cessation. Prior to questionnaire distribution, all students were informed about the main objectives of the study and provided consent for their voluntary participation. 2.2. Measurement The study provided information on cigarette smoking; exposure to second-hand smoke (SHS) at home and in public places; and whether existing smoking ban policies were enforced. In addition, questions on attitude were asked to assess whether health professionals should be role models for their patients, whether training in smoking cessation techniques should be provided in undergraduate curricula and whether students had ever received formal training on such techniques. For the purposes of the study, ‘smokers’ were defined as those who had smoked cigarettes on one or more days during the previous 30 days, as several reviews suggest that even stable light or occasional smoking can carry substantial adverse health effects [21,22]. 2.3. Statistical Analysis Predictive Analysis SoftWare: PASW 18 (IBM Statistics) was utilized for statistical analysis, whilst weighted prevalence estimates were calculated by the Centre for Disease Control (Atlanta) using Survey Data Analysis: SUDAAN (RTI International) software. Descriptive results are presented as percentages. A finite population correction factor was applied to take into account non-response and used in the variance of the estimates. Univariate analysis was carried out using chi-square testing, with a p value of <0.05 taken as the threshold for statistical significance. All results have a margin of error of ±5% (95% confidence interval [CI]). 3.3. Tobacco Policy Awareness A majority of students (88.7%) were aware of the existence of a non-smoking policy within the hospital grounds. Of these, 64.9% believed that the policy was enforced, while 24.4% disagreed with this view. Only 11.3% of all students surveyed were unaware of the school policy. There was no significant difference between the responses of smokers versus non-smokers with regards to knowledge of the non-smoking policy (p = 0.212) or perception of its enforcement (p = 0.226). 3.2. Smoking Status Of the 65.9% of all respondents who reported experimenting with cigarettes in the past (including one or two puffs), 27.9% were introduced to cigarettes at 11–15 years of age, and another 19% started experimented with smoking at 16–17 years of age. At the time of the survey, 123 (72.8%) were non-smokers, whereas 46 (27.1%) smoked either daily (n = 16, 9.5%) or occasionally (n = 30, 17.8%) during the previous 30 days. There was no significant difference in smoking status between genders (Odds Ratio (OR) 0.96; 95% CI: 0.45–2.1). 3.4. Home Environment and Smoking Status Smokers were significantly more likely to report exposure to second hand smoke within their home (OR 3; 95% CI: 2.27–4.15; p ≤ 0.001) as well as outside the home environment, (OR 2.1; 95% CI: 1.3–3.2; p = 0.001) than were non-smokers. Int. J. Environ. Res. Public Health 2012, 9 Int. J. Environ. Res. Public Health 2012, 9 to participate in the Maltese GHPSS. Student response rates are summarised in Table 1. Among the total population, 73.8% were female and 26.2% were male. Most of the respondents (93%) were aged between 19 and 24 years. Table 1. Health professional student response rate and smoking prevalence (95% CI). Table 1. Health professional student response rate and smoking prevalence (95% CI). Discipline Dental Medical Nursing Pharmacy Total Students (n) 7 77 86 41 211 Respondents 6 59 78 30 173 Response rate (%) 85.7 76.6 90.7 73.2 82.0 % of total sample population (n = 173) 3.5 34.1 45.1 17.3 100 % current smokers (95% CI) 60 (24.8–87.2) 14.3 (10.0–20.0) 40.3 (36.9–43.8) 13.3 (7.9–21.5) 3.2. Smoking Status 3.1. Description of Participants and Smoking Status The study population consisted of 211 students enrolled in their third year of study across the four undergraduate health professional schools. All disciplines and all health professional schools accepted 2553 Int. J. Environ. Res. Public Health 2012, 9 Int. J. Environ. Res. Public Health 2012, 9 Odds Ratios calculated through regression analysis showed that in general, non-smokers were more positive than smokers in their attitudes towards regulation of tobacco sales, enforcement of tobacco-free zones and the perceived role of health professionals in reducing the burden of tobacco. With regards to the perceived role of health professionals in tobacco control and smoking cessation, most students (90%) believed health professionals should receive training on smoking cessation, with non-smokers being more evidently in favour (OR 3.5; 95% CI: 2.3–5.4; p ≤ 0.001). However, not all students seemed to believe in leading by example, as only 65% of all students think that health professionals should be role models, with no statistically significant difference between smokers and non-smokers (p = 0.1). Non-smokers were also more likely to believe that health professionals have a role in giving cessation advice routinely to patients (OR 5.1; 95% CI: 3.1–8.5; p ≤ 0.001), although they were only slightly more positive than smokers in their beliefs regarding the effectiveness of their advice (p = 0.079). Significantly, non-smokers were more likely to believe that patients had a reduced chance of being advised to quit by smoking health professionals (OR 3.5; 95% CI: 2.6–4.8; p ≤ 0.001). There was general agreement that students were taught about the harmful effects of smoking at undergraduate level, with non-smokers showing greater awareness (OR 1.7; 95% CI: 1.1–2.6; p = 0.02), however little attention seems to be given to the psychological basis of smoking, with around half the student population stating that they had not discussed reasons why people smoke (Table 3). Moreover, it was noted that smokers had a heightened awareness about the importance of taking an accurate tobacco history (p ≤ 0.001). The majority of students did not report receiving formal training in smoking cessation thus far in their curricula, and it is not known whether those who replied positively to this question (12.5%) received such training within the framework of their undergraduate studies. Additionally, while the majority had heard of nicotine replacement therapies, only a third of respondents were aware of antidepressant use in cessation programs, and less than two thirds of students had been taught that providing educational quitting materials is important. 3.5. Beliefs and Attitudes towards Smoking by Smoking Status Table 2 summarises students’ attitudes toward tobacco control and smoking according to their smoking status. Although generally, all health professional students were against tobacco promotion irrespective of their own smoking status, non-smokers were more likely than smokers to say that tobacco sales to adolescents should be completely banned (OR 3.6; 95% CI: 2.5–5.3); and to agree with a smoking ban in discotheques, bars and pubs (OR 12.1; 95% CI: 8.4–17.4) in restaurants (OR 3.5; 95% CI: 2.1–5.7) as well as in all public places (OR 8.9; 95% CI: 6.1–13.1). 2554 Int. J. Environ. Res. Public Health 2012, 9 The effect of formal tobacco control education on students’ beliefs about their role as models in providing smoking cessation advice was explored using logistic regression analysis after controlling for current smoking status (Table 4). Students reporting knowledge of antidepressants as tools for stopping smoking were more likely to report that health professionals have a role in smoking cessation (OR 4.9; 95% CI: 1.8–13.3; p = 0.002) and should receive cessation training (OR 2.2; 95% CI: 1.4–3.5; p = 0.001). A slight positive association was found between having knowledge of the adverse effects of smoking and belief that health professionals should routinely give cessation advice, however this association was not statistically significant (OR 2.2; 95% CI: 0.8–6.0; p = 0.123). 555 tobacco control. otal Smokers Non-smokers p value % (n) % (n) % (n) (2-sided) 2 (144) 71.7 (33) 90.2 (111) <0.001 6 (121) 63.0 (29) 74.8 (92) 0.001 (156) 84.8 (39) 95.1 (117) <0.001 7 (133) 45.6 (21) 91.0 112) <0.001 4 (141) 58.7 (27) 92.7 (114) <0.001 0 (152) 80.4 (37) 93.5 (115) 0.012 0 (110) 60.9 (28) 66.7 (82) 0.1 2 (155) 82.2 (37) 96.0 (118) <0.001 9 (152) 82.6 (38) 92.7 (114) <0.001 6 (165) 91.3 (42) 100 (123) <0.001 8 (135) 76.1 (35) 81.3 (100) 0.079 6 (116) 47.8 (22) 76.4 (94) <0.001 ducation. Smokers Non-smokers p value % (n) % (n) (2-sided) 84.7 (39) 90.2 (114) 0.02 56.5 (26) 50.4 (62) 0.10 100 (46) 90.2 (111) <0.001 11.1 (5) 13.1 (16) 0.42 65.2 (30) 61.0 (75) 0.24 95.7 (44) 95.9 (118) 0.85 42.2 (19) 30.9 (38) 0.002 1 1 1 1 1 1 1 1 p val 2-sid <0.0 0.00 <0.0 <0.0 <0.0 0.01 0.1 <0.0 <0.0 <0.0 0.07 <0.0 97. 79. 68. cco al n) 150) (88) 157) (21) 105) 162) (57) uestions regarding attitudes towa 7 9 7 anned? 8 9 6 9 ? 8 9 advice? 7 tients to stop smoking? 6 nse to questions regarding tobacc Total % (n) ? 88.7 (15 52.1 (88 tory? 92.9 (15 on? 12.6 (2 ng materials? 62.1 (10 95.9 (16 ms? 4.1. Main Findings Findings from the Malta GHPSS show that more than a quarter of health professional students are daily or occasional smokers; a rate slightly higher than that (19.3%) found in the corresponding adult Maltese population of the same age [17]. In this study, pharmacy students showed the lowest smoking prevalence. Medical students had a slightly higher prevalence, though this was still far higher than that found among medical doctors in a 2009 study [19]. Additionally, within the specific disciplines, the percentage of current medical student smokers was considerably lower than that of other European countries including Germany (28%), Italy (31.3%) and Poland (33.1%), but similar to that found in Spain (15.7%) [12]. However, smoking prevalence in nursing students was similar to Italy (48.2%) [23] but higher than that in most other participating European countries, including Greece (33.1%), Slovakia (32.2%) and the Czech Republic (32.7%) [11,24]. The number of dental students was considered to be too small to provide meaningful prevalence comparisons with other countries, whereas data on smoking prevalence in pharmacy students in other countries was generally limited. With respect to attitudes, the majority of students seem willing to ‘practice what they preach’, with most non-smokers having satisfactory attitudes towards tobacco control interventions and their potential role model status. On the other hand, among those health professional students that smoke, support for such efforts appears to weaken when these threaten their freedom to smoke in public places. This complements findings which suggest that students’ attitudes towards tobacco control may be linked to their established smoking pattern prior to starting their professional training, which in turn colours their perceptions of tobacco control [25]. Unlike findings from other studies [24] few significant links between previous tobacco education and students’ current attitudes towards tobacco control were elicited. This may be due to the general paucity of systematic training at undergraduate levels, or indeed due to a general lack of experience in practicing their acquired health promotional skills. Additionally, it is of concern that smokers seem dissatisfied with the national smoking ban in public places of entertainment—in place since April 2004, and unwilling to be assigned a role model status which might necessitate stopping smoking themselves. While the majority of smokers had positive attitudes towards policy efforts to control tobacco use, they were significantly less likely to agree to bans in public places when compared to non-smokers. This skepticism is also seen elsewhere [26,27]. Int. J. Environ. Res. Public Health 2012, 9 33.9 (57 56 nt ) 6) of education on health professional students’ beliefs ional aspects ance of ounselling rial Use of antidepressants in smoking cessation Received formal training Nicoti replacem 2–1.2) 4.9 (1.8–13.3) 1 (0.02– 4–1.0) 2.2 (1.4–3.5) 0.7 (0.3– 8–1.4) 1.3 (0.9–1.7) 0.9 (0.6–1.4) –1.7) 1.8 (1.1–3.0) 1.3 (0.6–2.7) 0.5 (0.2– 4–0.8) 1.7 (1.2–2.4) 0.4 (0.2–0.7) bjects to perform the analysis. 2557 Int. J. Environ. Res. Public Health 2012, 9 4.1. Main Findings Furthermore, smokers seemed to disagree with the idea that their smoking status might influence the advice given to patients, and are less willing to conduct health promotion specifically against tobacco. Physician and nurse smoking status and tobacco control counselling is known to positively influence patient smoking habits [4,5] hence developing systematic tobacco cessation training; providing a smoke-free environment and encouraging medical and nursing students to quit is likely to play a significant role in the effectiveness and success of future patient counselling encounters [11,28–31]. The undergraduate health professional courses in Malta do not contain a specific module on tobacco use, but rather relied on separate lectures across the subjects that illustrate the ill effects of smoking on physical health. At the time of the survey, school administrators reported that no structured training in Int. J. Environ. Res. Public Health 2012, 9 2558 smoking cessation was provided to students in the four disciplines, although information on the health effects of smoking and the existence of various cessation methods was provided throughout the undergraduate courses. Knowledge of the latter was seen to impact students’ beliefs about the need for formal training, complementing their perceived role in smoking cessation counselling as future health professionals. In February 2012, the smoking policy at the Malta Medical School and the Institute of Health Care was revised, only permitting smoking strictly outside hospital and school grounds. Free quitting services information and smoking cessation classes are provided to students, qualified professionals and hospital staff who smoke. This is likely to positively contribute to reduction of smoking prevalence in health professional students [28]. 4.2. Limitations The GHPSS is subject to several limitations. Third year students may not have had substantial interaction with patients in a clinical setting, so that these results may not be extrapolated to practicing health professionals. Response rates fluctuated slightly across the different disciplines, and in the case of pharmacy and dental students the result was that no inference to that sub-group could be made given the small numbers involved. This is also reflected in the relatively wide CIs. For this reason also, no intra-disciplinary comparisons in smoking habits and attitudes towards tobacco control could be made. Although the disciplines represented in the GHPSS surveys tend to be on the front line in terms of dispensing cessation advice in Malta, other important health professionals, such as physiotherapists and psychologists, were excluded. In addition, although a reliability study carried out in Italy has determined satisfactory validity and internal consistency of the questionnaire [20], no systematic evaluation at an international level has yet been carried out for GHPSS. However, reliability studies for similar tobacco-related questions in the United States have indicated good test-retest results [32]. Finally, data were extracted from self-reported questionnaires, which do not always provide reliable information [33]. 4.3. What This Study Adds This study establishes the prevalence of smoking among health professionals in Malta today and adds to the evidence that current smoking status directly influences beliefs and attitudes towards tobacco control, despite apparently sufficient knowledge regarding the harmful health effects of smoking. While newly imposed stricter smoking policies within hospital and school grounds are a welcome step in the right direction to discourage tobacco use among all health professional students, this should be combined with formal training in cessation methods. This would reinforce positive attitudes towards tobacco control and increase the effectiveness of counselling skills, particularly since the students themselves seem to be keen on such systematic training. Studies have suggested that training students about the implementation of tobacco cessation techniques and regarding the clinical treatment of tobacco dependence would greatly contribute to a reduction in patient smoking rates [14,24]. The highly significant positive correlations of a current smoking habit with overall negative attitudes to societal efforts at tobacco control and smoking cessation among Maltese health professional students highlights the need for the adoption of a standard curriculum inclusive of comprehensive tobacco prevention and smoking cessation training for these students. This has been Int. J. Environ. Res. Public Health 2012, 9 Int. J. Environ. Res. Public Health 2012, 9 2559 shown to facilitate long-term changes in attitudes and behaviours [34] and would empower them to actively quit smoking themselves, thus increasing the likelihood of performing interventions compared to untrained students [35]. Undergraduate curricula should ideally be revised to enable students to manage smoking dependency with a clear understanding of basic clinical and community-based cessation methods, as successfully done in the USA [36]. A system of regular assessment to prospectively capture health professional students’ changing trends in tobacco use and attitudes towards tobacco control would prove useful in determining the effectiveness of various smoking prevention approaches over time. It is possible that smoking students become smoking professionals who are less effective at tobacco prevention, although their attitudes might change as patient contact increases. Further research is recommended in this regard. Ultimately, investment in the quality of education of health professionals will improve the effectiveness of health professionals [37], reaping significant health benefits for both the professionals themselves and their patients. Acknowledgements This work was supported by the World Health Organization (Grant: NCE: EU ICP FFC 100 XK 08). We would like to thank Wick Warren, Juliette Lee and Veronica Lea from CDC for their help in study design and data management. Conflict of Interest The authors declare no conflict of interest. References 1. Fiore, M.C.; Jaén, C.R.; Baker, T.B.; Bailey, W.C.; Benowitz, N.L.; Curry, S.J.; Dorfman, S.F.; Froelicher, E.S.; Goldstein, M.G.; Healton, C.G.; et al. Treating Tobacco Use and Dependence: 2008 Update; U.S. Department of Health and Human Services, Public Health Service: Rockville, MD, USA, 2008. 2. World Health Organization (WHO). WHO Report on the Global Tobacco Epidemic, 2011; WHO: Geneva, Switzerland, 2011. 2. World Health Organization (WHO). WHO Report on the Global Tobacco Epidemic, 2011; WHO: Geneva, Switzerland, 2011. 3. Lancaster, T.; Stead, L.; Silagy, C.; Sowden, A. Effectiveness of interventions to help people stop smoking: Findings from the cochrane library. BMJ 2000, 321, 355–358. 3. Lancaster, T.; Stead, L.; Silagy, C.; Sowden, A. Effectiveness of interventions to help people stop smoking: Findings from the cochrane library. BMJ 2000, 321, 355–358. 4. Chatkin, J.; Chatkin, G. Learning about smoking during medical school: Are we still missing opportunities? Int. J. Tuberc. Lung Dis. 2009, 13, 429–437. 4. Chatkin, J.; Chatkin, G. Learning about smoking during medical school: Are we still missing opportunities? Int. J. Tuberc. Lung Dis. 2009, 13, 429–437. 5. Raupach, T.; Shahab, L.; Baetzing, S.; Hoffmann, B.; Hasenfuss, G.; West, R.; Andreas, S. Medical students lack basic knowledge about smoking: Findings from two european medical schools. Nicotine Tob. Res. 2009, 11, 92–98. 5. Raupach, T.; Shahab, L.; Baetzing, S.; Hoffmann, B.; Hasenfuss, G.; West, R.; Andreas, S. Medical students lack basic knowledge about smoking: Findings from two european medical schools. Nicotine Tob. Res. 2009, 11, 92–98. 6. Vardavas, C.I.; Bouloukaki, I.; Linardakis, M.K.; Tzilepi, P.; Tzanakis, N.; Kafatos, A.G. Smoke-free hospitals in greece: Personnel perceptions, compliance and smoking habit. Tob. Induc. Dis. 2009, 5, doi:10.1186/1617-9625-5-8. Int. J. Environ. Res. Public Health 2012, 9 2560 7. Sotiropoulos, A.; Gikas, A.; Spanou, E.; Dimitrelos, D.; Karakostas, F.; Skliros, E.; Apostolou, O.; Politakis, P.; Pappas, S. Smoking habits and associated factors among Greek physicians. Public Health 2007, 121, 333–340. 8. Smith, D.R. The historical decline of tobacco smoking among United States physicians: 1949–1984. Tob. Induc. Dis. 2008, 4, doi:10.1186/1671-9625-4-9. 9. Warren, C.W.; Sinha, D.N.; Lee, J.; Lea, V.; Jones, N.; Asma, S. Tobacco use, exposure to secondhand smoke, and cessation counseling training of dental students around the world. J. Dent. Educ. 2011, 75, 385–405. 10. Warren, C.W.; Sinha, D.N.; Lee, J.; Lea, V.; Jones, N.R. References Coggins, C.R.; Murrelle, E.L.; Carchman, R.A.; Heidbreder, C. Light and intermittent cigarette smokers: A review (1989–2009). Psychopharmacology (Berl.) 2009, 207, 343–363. 22. Coggins, C.R.; Murrelle, E.L.; Carchman, R.A.; Heidbreder, C. Light and intermittent cigarette smokers: A review (1989–2009). Psychopharmacology (Berl.) 2009, 207, 343–363. 23. Ficarra, M.G.; Gualano, M.R.; Capizzi, S.; Siliquini, R.; Liguori, G.; Manzoli, L.; Briziarelli, L.; Parlato, A.; Cuccurullo, P.; Bucci, R.; et al. Tobacco use prevalence, knowledge and attitudes among Italian hospital healthcare professionals. Eur. J. Public Health 2011, 21, 29–34. 24. Patelarou, E.; Vardavas, C.I.; Ntzilepi, P.; Warren, C.W.; Barbouni, A.; Kremastinou, J.; Connolly, G.N.; Behrakis, P. Nursing education and beliefs towards tobacco cessation and control: A cross-sectional national survey (GHPSS) among nursing students in Greece. Tob. Induc. Dis. 2011, 9, 4. 25. Rowe, K.; Macleod Clark, J. Why nurses smoke: A review of the literature. Int. J. Nurs. Stud. 2000, 37, 173–181. 26. Ramos, M.C.; Vinagre, S.; Cardoso, M.F. Knowledge, behavior and ethical responsibility of health professionals with regard to smoking. Rev. Port. Cardiol. 2010, 29, 923–946. 27. Ratschen, E.; Britton, J.; Doody, G.A.; Leonardi-Bee, J.; McNeill, A. Tobacco dependence, treatment and smoke-free policies: A survey of mental health professionals’ knowledge and attitudes. Gen. Hosp. Psychiatry 2009, 31, 576–582. 28. Nakashima, M.; Miura, K.; Morikawa, Y.; Nishijo, M.; Nakanishi, Y.; Sakurai, M.; Nakagawa, H. Effect of smoke-free medical school on smoking behavior of medical students. Nihon Koshu Eisei Zasshi 2008, 55, 647–654. 29. Duncan, C.L.; Cummings, S.R.; Hudes, E.S.; Zahnd, E.; Coates, T.J. Quitting smoking: Reasons for quitting and predictors of cessation among medical patients. J. Gen. Intern. Med. 1992, 7, 398–404. 30. Frank, E. StudentJAMA. Physician health and patient care. JAMA 2004, 291, 637. 31. Lenz, B.K. Beliefs, knowledge, and self-efficacy of nursing students regarding tobacco cessation. Am. J. Prev. Med. 2008, 35, S494–S500. 32. Brener, N.D.; Kann, L.; McManus, T.; Kinchen, S.A.; Sundberg, E.C.; Ross, J.G. Reliability of the 1999 youth risk behavior survey questionnaire. J. Adolesc. Health 2002, 31, 336–342. 33. Gorber, S.C.; Schofield-Hurwitz, S.; Hardt, J.; Levasseur, G.; Tremblay, M. The accuracy of self-reported smoking: A systematic review of the relationship between self-reported and cotinine-assessed smoking status. Nicotine Tob. Res. 2009, 11, 12–24. 34. Borrelli, B.; Lee, C.; Novak, S. Is provider training effective? Changes in attitudes towards smoking cessation counseling and counseling behaviors of home health care nurses. Prev. Med. 2008, 46, 358–363. 35. Heath, J.; Crowell, N.A. References Tobacco use, exposure to secondhand smoke, and cessation counseling among medical students: Cross-country data from the global health professions student survey (GHPSS), 2005–2008. BMC Public Health 2011, 11, doi:10.1186/1471-2458-11-72. 11. Warren, C.W.; Sinha, D.N.; Lee, J.; Lea, V.; Jones, N.R. Tobacco use, exposure to secondhand smoke, and training on cessation counseling among nursing students: Cross-country data from the Global Health Professions Student Survey (GHPSS), 2005–2009. Int. J. Environ. Res. Public Health 2009, 6, 2534–2549. 12. La Torre, G.; Kirch, W.; Bes-Rastrollo, M.; Ramos, R.M.; Czaplicki, M.; Gualano, M.R.; Thummler, K.; Ricciardi, W.; Boccia, A. Tobacco use among medical students in Europe: Results of a multicentre study using the global health professions student survey. Public Health 2011, 126, 159–164. 13. Richmond, R.; Taylor, R. Global dissemination of a tobacco curriculum in medical schools. Int. J. Tuberc. Lung Dis. 2006, 10, 750–755. 14. Richmond, R.; Zwar, N.; Taylor, R.; Hunnisett, J.; Hyslop, F. Teaching about tobacco in medical schools: A worldwide study. Drug Alcohol Rev. 2009, 28, 484–497. 15. Geller, A.C.; Brooks, D.R.; Powers, C.A.; Brooks, K.R.; Rigotti, N.A.; Bognar, B.; McIntosh, S.; Zapka, J. Tobacco cessation and prevention practices reported by second and fourth year students at US medical schools. J. Gen. Intern. Med. 2008, 23, 1071–1076. 16. Warren, C.W.; Jones, N.R.; Chauvin, J.; Peruga, A. Tobacco use and cessation counselling: Cross-country. Data from the Global Health Professions Student Survey (GHPSS), 2005–2007. Tob. Control 2008, 17, 238–247. 17. Ministry for Social Policy. European Health Interview Survey (2008)—Summary Statistics; Department of Health Information and Research: G’Mangia, Malta, 2008. 18. Brotons, C.; Bjorkelund, C.; Bulc, M.; Ciurana, R.; Godycki-Cwirko, M.; Jurgova, E.; Kloppe, P.; Lionis, C.; Mierzecki, A.; Pineiro, R.; et al. Prevention and health promotion in clinical practice: The views of general practitioners in europe. Prev. Med. 2005, 40, 595–601. 19. Fenech Magrin, A. Preceived and Actual Cardiovascular Disease Risk Factors, and the Obstacles to Reduce Them, among a Representative Sample of the Current and Future Medical Profession in Malta; University of Malta: San Gwann, Malta, 2009. 20. Gualano, M.R.; Bontempi, C.; Saulle, R.; Ricciardi, W.; La Torre, G. Validation of the global health professions students survey questionnaire in Italy. Ital. J. Public Health 2011, 8, 392–398. 21. Husten, C.G. How should we define light or intermittent smoking? Does it matter? Nicotine Tob. Res. 2009, 11, 111–121. Int. J. Environ. Res. Public Health 2012, 9 2561 22. References Factors influencing intentions to integrate tobacco education among advanced practice nursing faculty. J. Prof. Nurs. 2007, 23, 189–200. 36. Geller, A.C.; Zapka, J.; Brooks, K.R.; Dube, C.; Powers, C.A.; Rigotti, N.; O’Donnell, J.; Ockene, J. Tobacco control competencies for US medical students. Am. J. Public Health 2005, 95, 950–955. 37 Smith D R ; Leggat P A An International review of tobacco smoking among medical students 36. Geller, A.C.; Zapka, J.; Brooks, K.R.; Dube, C.; Powers, C.A.; Rigotti, N.; O’Donnell, J.; Ockene, J. Tobacco control competencies for US medical students. Am. J. Public Health 2005, 95, 950–955. 37. Smith, D.R.; Leggat, P.A. An International review of tobacco smoking among medical students. J. Postgrad. Med. 2007, 53, 55–62. 37. Smith, D.R.; Leggat, P.A. An International review of tobacco smoking among medical students. J. Postgrad. Med. 2007, 53, 55–62. © 2012 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). © 2012 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).
https://openalex.org/W4220717877	https://www.aging-us.com/article/203960/pdf	English	null	Hallmarks of aging-based dual-purpose disease and age-associated targets predicted using PandaOmics AI-powered discovery engine	Aging	2,022	cc-by	16,112	ABSTRACT Aging biology is a promising and burgeoning research area that can yield dual-purpose pathways and protein targets that may impact multiple diseases, while retarding or possibly even reversing age-associated processes. One widely used approach to classify a multiplicity of mechanisms driving the aging process is the hallmarks of aging. In addition to the classic nine hallmarks of aging, processes such as extracellular matrix stiffness, chronic inflammation and activation of retrotransposons are also often considered, given their strong association with aging. In this study, we used a variety of target identification and prioritization techniques offered by the AI-powered PandaOmics platform, to propose a list of promising novel aging-associated targets that may be used for drug discovery. We also propose a list of more classical targets that may be used for drug repurposing within each hallmark of aging. Most of the top targets generated by this comprehensive analysis play a role in inflammation and extracellular matrix stiffness, highlighting the relevance of these processes as therapeutic targets in aging and age-related diseases. Overall, our study reveals both high confidence and novel targets associated with multiple hallmarks of aging and demonstrates application of the PandaOmics platform to target discovery across multiple disease areas. Correspondence to: Alex Zhavoronkov; email: alex@insilico.com Keywords: artificial intelligence, deep learning, drug discovery, multi-omics, target identification Received: January 21, 2022 Accepted: March 6, 2022 Published: March 29, 2022 Copyright: © 2022 Pun et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. AGING 2022, Vol. 14, No. 6 AGING 2022, Vol. 14, No. 6 www.aging-us.com Research Paper Hallmarks of aging-based dual-purpose disease and age-associated targets predicted using PandaOmics AI-powered discovery engine Frank W. Pun1,, Geoffrey Ho Duen Leung1,, Hoi Wing Leung1,, Bonnie Hei Man Liu1, Xi Long1, Ivan V. Ozerov1, Ju Wang1, Feng Ren1, Alexander Aliper1, Evgeny Izumchenko2, Alexey Moskalev3, João Pedro de Magalhães4, Alex Zhavoronkov1,5 1Insilico Medicine Hong Kong Ltd., Hong Kong Science and Technology Park, New Territories, Hong Kong, China 2Department of Medicine, Section of Hematology and Oncology, University of Chicago, Chicago, IL 60637, USA 3School of Systems Biology, George Mason University (GMU), Fairfax, VA 22030, USA 4Integrative Genomics of Ageing Group, Institute of Ageing and Chronic Disease, University of Liverpool, Liverpool L7 8TX, UK Correspondence to: Alex Zhavoronkov; email: alex@insilico.com Keywords: artificial intelligence, deep learning, drug discovery, multi-omics, target identification Received: January 21, 2022 Accepted: March 6, 2022 Published: March 29, 2022 Correspondence to: Alex Zhavoronkov; email: alex@insilico.com Keywords: artificial intelligence, deep learning, drug discovery, multi-omics, target identification Received: January 21, 2022 Accepted: March 6, 2022 Published: March 29, 2022 INTRODUCTION developing interventions that target aging would also impact multiple age-related diseases and result in unprecedented health benefits [1, 2]. Importantly, the stated goal of geroscience is to extend not only lifespan but also health life expectancy, or healthspan [2], enabling wellbeing in older age, the so-called “healthy aging”. Research into longevity pharmacology has exploded in recent years with hundreds of compounds now known to extend lifespan in model organisms [3–7]. Population aging is a key social, economic, and medical challenge on a global scale, which has created a major, growing need to develop interventions that target the aging process. Because age-associated changes in homeostasis are the major risk factors for the most prevalent human diseases (such as neurological, metabolic, fibrotic, and inflammatory conditions), AGING 2475 www.aging-us.com discovery and target prediction in aging [19]. Indeed, a number of studies by our group and others have employed computational and machine learning analysis to identify new candidates in the context of aging and AADs. These approaches have led to the detection of disease-related genes, caloric restriction genes, and longevity drugs [20–24]. The application of AI in the pharmaceutical industry also aims to reduce the tremendous amount of cost and time conventionally needed to discover new therapeutic targets in various diseases. There are multiple philosophies for the formulation of disease hypothesis, prioritization of pathways implicated in a disease, and selection of promising therapeutic targets. Multiple data types can be used for target discovery including text, imaging, and omics. In recent years, machine learning, and especially deep learning technologies, are rapidly increasing in popularity for target discovery. Advanced signaling pathway modeling such as iPANDA [25] and deep neural networks were used to identify promising protein targets driving complex biological processes implicated in cancer and other diseases [26], drug repurposing [27], and geroprotector discovery [28, 29]. Many of these approaches were implemented in PandaOmics, an industrial target discovery engine. Recently, PandaOmics, has successfully identified and nominated novel targets for idiopathic pulmonary fibrosis (IPF) and kidney fibrosis [30–32]. Major challenges remain, however, in translating these findings into humans, and ultimately extending lifespan and healthspan by targeting aging mechanisms. Although the underlying molecular mechanisms of aging remain the subject of debate, several key pathways and processes have been associated with aging processes. INTRODUCTION These have been conceptualized in the hallmarks of aging composed of: Altered intercellular communications, Cellular senescence, Deregulated nutrient signaling, Epigenetic shift, Genomic instability, Impaired proteostasis, Mitochondrial dysfunction, Stem cell exhaustion, Telomere attrition [8]. The hallmarks of aging have been widely used in the field as a starting point for studies, although they are not perfect, and arguably other mechanisms such as extracellular matrix stiffness [9], retrotranspositions [10] and inflammation [11] have also been reported to play an important role in aging. These heterogeneous processes have in turn been associated with age-related diseases. For example, cellular senescence has been associated with pathologies such as cancer, type 2 diabetes, and atherosclerosis [12], as well as pulmonary, neurological, renal, hepatic, infectious, musculoskeletal, and endocrine diseases [13]. At the genetic level, there is a substantial overlap between the genetics of aging and age-associated diseases (AADs) [14]. For example, some known aging-related targets, such as mTOR, AMPK, IGFR, NF-kB, S6K, TGF-β, AT1, Fgf21, FOXO3a, SIRT1, HIF-1, NRF2, and Klotho, may also impact multiple age-associated diseases [1, 8, 15]. Therefore, given that aging is associated with mechanisms that ultimately lead to age-related comorbidities, drugs that act on targets implicated in aging may potentially reduce the severity of gerolavic (from Greek, géros “old man” and epilavís, “harmful”) diseases and preventing multimorbidity [16]. This platform utilizes advanced deep learning models and AI approaches to predict the target genes associated with a given disease through a combination of Omics AI scores, Text-based AI scores, Finance scores, and Key opinion leader (KOL) scores (Figure 1), and is currently being employed in both academic and industry settings. The algorithm also allows the prioritization of protein targets for novelty, confidence, commercial tractability, druggability, safety, and other key properties that drive target selection decisions. The integrated omics database consists of a vast amount of published systems biology data, spanning over 1,500 diseases and 10,000 disease subtypes. The database includes approximately 1.9 trillion data points derived from over 10 million samples with microarrays, RNA sequencing, proteomes, and methylomes, among other data types. PandaOmics’ text database embeds information from over 40 million documents, including patents, grants (that amount to over $2 trillion in funding), publications, clinical trial results, and company reports, among other text-based sources. A substantial percentage of the human clinical trials, including those evaluating investigational anti-aging drugs, fail in Phase II, a phase where efficacy of the drug is tested [17, 18]. RESULTS identification and prioritization techniques offered by PandaOmics and downstream analyses, yielding a list of dual-purpose targets associated with both the aging process and AADs. First, a list of 145 aging-related targets was generated upon the hallmarks of aging assessment. Subsequently, we further narrowed down the number of potential candidates to a total of 9 highly promising therapeutic targets associated with the aging process and AADs based on multiple selection criteria. Discovery of targets implicated in multiple age- associated diseases A combined list of 484 high confidence, 448 medium novel, and 381 highly novel targets were generated from the lists of top-100 targets prioritized by PandaOmics in each of the AADs (Supplementary Figure 1). The top- 100 targets for three levels of novelty settings g y g Figure 1. Workflow of the present study. Thirty-three diseases were separated into either age-associated diseases (AADs) or non-age- associated diseases (NAADs) based on the impact of age on the risk of the disease’s onset. Their corresponding transcriptomic datasets were retrieved from public repositories and processed by PandaOmics. Age bias between case and control groups has been considered during dataset selection. With multiple levels of novelty settings, targets implicated in AADs and NAADs were identified by ‘PandaOmics - target identification’. PandaOmics prioritized targets for one disease and refined the targets based on several flexible druggability filters. The target-disease associations were ranked according to over 20 artificial intelligence and bioinformatics models ranging from Omics AI scores, Text-based AI scores, Finance scores to KOL scores. Target identification was performed independently for each disease. Top-ranked targets shared by both disease categories were regarded as common targets, while targets unique to AADs were defined as age-associated targets (AAD targets). All common targets and AAD targets were subjected to the hallmarks of aging assessment by searching the literature for their evidence in modulating longevity or longevity pathways. To propose potential targets with a dual role in anti-aging and disease treatment, hallmark-associated targets were further evaluated based on their expression profiles across AADs, mechanism of action, and safety. A total of 9 targets were selected, with three levels of novelty. Abbreviation: KOL: Key opinion leader. Figure 1. Workflow of the present study. Thirty-three diseases were separated into either age-associated diseases (AADs) or non-age- associated diseases (NAADs) based on the impact of age on the risk of the disease’s onset. Their corresponding transcriptomic datasets were retrieved from public repositories and processed by PandaOmics. Age bias between case and control groups has been considered during dataset selection. With multiple levels of novelty settings, targets implicated in AADs and NAADs were identified by ‘PandaOmics - target identification’. PandaOmics prioritized targets for one disease and refined the targets based on several flexible druggability filters. The target-disease associations were ranked according to over 20 artificial intelligence and bioinformatics models ranging from Omics AI scores, Text-based AI scores, Finance scores to KOL scores. INTRODUCTION This poor success is in part due to inadequate target choice and the inability to identify a group of patients who will most likely respond to specific agents. This challenge is further complicated by the differences in biological age of the patients, as importance of therapeutic targets varies between the age groups. Hence, identifying potential targets that are implicated in multiple age-associated diseases, and also play a role in the basic biology of aging, may have substantial benefits. In this study, we used PandaOmics to identify a list of potential aging-associated therapeutic targets across various AADs that may be used for drug discovery. We successfully established and validated our unique approach with the application of varied target Given the large number of datasets being generated, data-driven approaches (such as artificial intelligence [AI] and machine learning) are becoming instrumental across various fields in biology, including biomarker AGING 2476 www.aging-us.com Discovery of targets implicated in multiple age- associated diseases Target identification was performed independently for each disease. Top-ranked targets shared by both disease categories were regarded as common targets, while targets unique to AADs were defined as age-associated targets (AAD targets). All common targets and AAD targets were subjected to the hallmarks of aging assessment by searching the literature for their evidence in modulating longevity or longevity pathways. To propose potential targets with a dual role in anti-aging and disease treatment, hallmark-associated targets were further evaluated based on their expression profiles across AADs, mechanism of action, and safety. A total of 9 targets were selected, with three levels of novelty. Abbreviation: KOL: Key opinion leader. AGING 2477 www.aging-us.com Number of comparisons Disease class Age-associated diseases (14 diseases, 87 comparisons) Alzheimer's disease Neurological 12 Amyotrophic lateral sclerosis Neurological 10 Chronic kidney disease Metabolic 7 Chronic obstructive pulmonary disease Inflammatory 6 Cirrhosis of liver Fibrotic 5 Idiopathic Pulmonary Fibrosis Fibrotic 11 Obesity Metabolic 10 Osteoarthritis Inflammatory 5 Osteoporosis Metabolic 2 Parkinson's disease Neurological 4 Primary myelofibrosis Fibrotic 2 Pulmonary arterial hypertension Metabolic 5 Rheumatoid Arthritis Inflammatory 4 Type II diabetes mellitus Metabolic 4 Non-age-associated diseases (19 diseases, 126 comparisons) Acromegaly Metabolic 2 Asthma Inflammatory 13 Bipolar disorder Neurological 4 Celiac disease Inflammatory 3 Crohn's disease Inflammatory 8 Cystic fibrosis Fibrotic 5 Hepatitis, alcoholic Metabolic 3 Hepatitis C virus infection Infectious 2 Huntington's disease Neurological 5 Infectious meningitis Infectious 3 Influenza Infectious 5 Multiple sclerosis Inflammatory 11 Psoriasis Inflammatory 11 Pulmonary tuberculosis Infectious 7 Schizophrenia Neurological 4 Systemic lupus erythematosus Inflammatory 9 Systemic scleroderma Fibrotic 6 Type I diabetes mellitus Metabolic 12 Ulcerative colitis Inflammatory 13 were selected by (1) the occurrence of the target genes in CASP3, VEGFA, and MMP9, which were highly ranked Osteoarthritis CASP3, VEGFA, and MMP9, which were highly ranked in all of the 14 AADs (Figure 2 and Supplementary Figure 2A). LYN was the top gene identified under medium novelty settings, which was also highly implicated in all AADs (Supplementary Figures 2B and 3). For high novelty settings, PPP2CB, CDC34, FES, RHOF, and RAB24 were the top-ranked genes in 12 out of 14 AADs (Supplementary Figures 2C and 4). The top- ranked genes shared by both AADs and NAADs were regarded as common targets, while those genes unique to AADs were defined as age-associated targets, or AAD targets (Venn diagram, Figure 1). Discovery of targets implicated in multiple age- associated diseases Intersecting the two were selected by (1) the occurrence of the target genes in 14 AADs (Supplementary Figure 2) and (2) the average ranking of the target gene in its corresponding disease (Supplementary Table 1). The same approach was applied to non-age-associated diseases (NAADs). Diseases selected for this study and their corresponding disease classes were listed in Table 1. Only the top-100 genes of this combined list from AADs were subjected to the hallmarks of aging assessment by finding their corresponding evidence in modulating longevity or aging pathways in literature and expression analysis. Under high confidence settings, the top genes identified were AGING 2478 www.aging-us.com www.aging-us.com Genes implicated in AADs are associated with the hallmarks of aging Genes implicated in AADs are associated with the hallmarks of aging lists of genes obtained from AADs and NAADs resulted in 42 AAD targets in high confidence setting (Supplementary Figure 5A). The remaining 58 genes were considered as common targets. For medium and high novelty settings, 37 and 29 AAD targets were identified, respectively (Supplementary Figure 5B–5C). In our analysis, 300 genes identified under three different novelty settings were subjected to a literature review (see Materials and Methods) for their association 2. Ranking of the top-100 gene set for AADs under high confidence settings. The ranking of the targets in AADs and are colored in blue-white and red-white thermal scales respectively. High color intensity stands for high ranking. The lowest ranking apped at 100. Targets associated with the hallmark(s) of aging are labeled in green. Abbreviation: COPD: Chronic obstructive nary disease. Figure 2. Ranking of the top-100 gene set for AADs under high confidence settings. The ranking of the targets in AADs and NAADs are colored in blue-white and red-white thermal scales respectively. High color intensity stands for high ranking. The lowest ranking was capped at 100. Targets associated with the hallmark(s) of aging are labeled in green. Abbreviation: COPD: Chronic obstructive pulmonary disease. AGING 2479 www.aging-us.com with the hallmarks of aging. Their corresponding roles in aging processes were summarized in Supplementary Table 2. In total, 145 genes (69 high confidence, 48 medium, and 28 highly novel targets) were linked to at least one aging hallmark (Figure 3, Supplementary Table 2). The most frequently associated aging hallmark was inflammation (n = 48), followed by genomic instability (n = 35), altered intercellular communications (n = 33), mitochondrial dysfunction (n = 32), impaired proteostasis (n = 31) and extracellular matrix stiffness (n = 30). Eighty-six genes (including several well-known aging-associated genes) Figure 3. Targets associated with hallmarks of aging. Age-associated targets and common targets (n = 145) were mapped to th corresponding hallmark(s) of aging based on the literature. For novel targets, their participating pathways were also used for th assessment of their association with the hallmark(s) of aging. The four targets connected to all hallmarks (AKT1, MTOR, SIRT1 and IGF1) ar shown in the inner circle of the plot. The target names are labeled in blue for age-associated targets, and black for common targets. Target annotated as cancer driver genes in the NCG7.0 database are underlined. Figure 3. Targets associated with hallmarks of aging. Genes implicated in AADs are associated with the hallmarks of aging Age-associated targets and common targets (n = 145) were mapped to the corresponding hallmark(s) of aging based on the literature. For novel targets, their participating pathways were also used for the assessment of their association with the hallmark(s) of aging. The four targets connected to all hallmarks (AKT1, MTOR, SIRT1 and IGF1) are shown in the inner circle of the plot. The target names are labeled in blue for age-associated targets, and black for common targets. Targets annotated as cancer driver genes in the NCG7.0 database are underlined. Figure 3. Targets associated with hallmarks of aging. Age-associated targets and common targets (n = 145) were mapped to the corresponding hallmark(s) of aging based on the literature. For novel targets, their participating pathways were also used for the assessment of their association with the hallmark(s) of aging. The four targets connected to all hallmarks (AKT1, MTOR, SIRT1 and IGF1) are shown in the inner circle of the plot. The target names are labeled in blue for age-associated targets, and black for common targets. Targets annotated as cancer driver genes in the NCG7.0 database are underlined. AGING 2480 www.aging-us.com were found to be associated with more than one hallmark. In particular, MTOR, SIRT1, IGF1, and AKT1 were associated with all hallmarks of aging, due to their wide range of interactions with aging-associated pathways. In addition, IGF1R was linked to deregulated nutrient signaling, genomic instability, inflammation, mitochondrial dysfunction, retrotranspositions, and stem cell exhaustion. Whereas HGF was associated with all hallmarks except epigenetic shift, retrotranspositions, and telomere attrition. Moreover, some novel targets were also identified to be associated with multiple hallmarks of aging. For example, MYSM1 was associated with cellular senescence, inflammation, and stem cell exhaustion; KAT6A was associated with cellular senescence, epigenetic shift, and stem cell exhaustion; UBE2E3 was linked to cellular senescence, impaired proteostasis, and stem cell exhaustion; RAB7B was linked to impaired proteostasis, inflammation, and mitochondrial dysfunction; whereas RAB8B and USP2 were related to altered intercellular communications, impaired proteostasis, and mitochondrial dysfunction. Furthermore, a recently proposed hallmark of aging, extracellular matrix stiffness, was associated with 30 target genes identified by PandaOmics in this study, consisting of 18 high confidence (AKT1, CHUK, CASP3, DNMT1, EGFR, FGF2, HGF, IGF1, ITGAV, LOX, MMP1, MMP2, MMP7, MMP9, MTOR, SIRT1, SPP1, TRAF6), 8 medium novel (FAM20C, GALNT1, ITGB5, MMP25, PLOD1, PLOD3, RAB14, TNIK), and 4 highly novel targets (ADAMTS14, ITGA11, RAP2C, RNF14). Validation by intersecting the AI-derived targets with well-known aging-associated genes The significance of the mTOR, the insulin/IGF, and the sirtuin pathways in longevity has been extensively reported, delineating their critical roles in counteracting multiple hallmarks of aging to delay the aging process or to extend lifespan [8, 34, 35]. Remarkedly, Food and Drug Administration (FDA)-approved mTOR inhibitor, rapamycin, was demonstrated to slow down aging and AADs in both preclinical settings and clinical trials [36, 37]. mTOR regulates several hallmarks of aging including nutrient sensing, stem cell exhaustion, proteostasis, and cellular senescence [38]. Upon insulin/IGF receptor activation following the insulin/IGF1 binding, mTOR, a nutrient sensor, regulates cellular functions linked to proliferation, growth, and survival via Akt-mediated activation. Increased insulin sensitivity favored lifespan extension. For example, growth hormone receptor (GHR)- knockout mice showed higher sensitivity to insulin, decelerated senescence, and displayed more phenotypic features related to anti-aging [39]. In addition, activation of SIRT1 suppressed aging by ensuring telomere integrity [40, 41], antagonizing oxidative stress [42, 43], regulating nutrient signaling [8] and maintaining proteostasis [44, 45]. Genes implicated in AADs are associated with the hallmarks of aging Among the 145 genes associated with hallmarks of aging, 55 genes are known cancer drivers (Figure 3) [33], pointing to the aging components underlying cancer pathogenesis. 45 genes (MX2, P2RX1, PRSS23, RAB7B, and RNASE6) were upregulated in all classes; 6 genes were upregulated while 21 were downregulated in 3 disease classes. As described above, upon the hallmarks of aging assessment, 145 genes were considered as potential aging-related targets. Here, these genes were further selected with reference to their expression patterns, and a subset of the candidates was considered as potential dual-purpose targets for subsequent analysis (Supplementary Figure 1). Genes consistently dysregulated in multiple AADs were implicated as potential dual-purpose targets To study the dysregulation state of genes identified under three different novelty settings, their consistency of expression change in each AAD class was summarized in Figure 4. Genes that were consistently dysregulated in two or more disease classes in a unidirectional manner were selected for further analysis. For high confidence targets, 52 genes were selected, of which 10 (CASP3, CXCL10, CXCL12, CYBA, HGF, ITGAM, ITGAV, PLAU, SPP1, and TGFB1) were consistently upregulated while MAPK8 was the only gene that was downregulated in all disease classes; 24 genes were upregulated and 8 were downregulated in 3 disease classes. Forty-four medium novel targets were selected, with 4 genes (CLEC5A, FPR3, ITGB5, and RAB31) and PPM1A being upregulated and downregulated in all disease classes, respectively; 15 genes were upregulated and 10 were downregulated in 3 disease classes. For highly novel targets, 5 of the Our approach utilizing PandaOmics identified a set of well-known aging-associated genes that are part of the mTOR, insulin/IGF and sirtuin family signaling (including IGF1, IGF1R, AKT1, MTOR, and SIRT1), strongly supporting the validity of this promising method for the identification of aging-associated genes. The aging-associated genes listed above were identified as common targets, suggesting their relevance in both aging and other diseases as well as their involvement in multiple signaling networks. It is worth noting that aging genes such as FOXO that did not meet the criteria for druggability (see Materials and Methods) were filtered out. To further evaluate whether our approach could identify aging-related targets with potential clinical relevance, the 100 high confidence targets were compared with a pool of well-known aging-associated genes curated from http://ClinicalTrials.gov (focusing on the treatment AGING 2481 www.aging-us.com MTOR, NR3C1, PDGFRB, SIRT1, SRC and VDR) were identified in the pool of 62 genes procured from the aging trials (expected [Exp] = 1.10, fold = 12.70, p = 1.79E-12, Supplementary Table 3). Twenty- of aging), publication, http://Geroprotectors.org [46] (Supplementary Tables 3–5) and aging gene database, GenAge [47]. Fourteen high confidence targets (ABL1, AR, ESR1, GHR, IGF1, IGF1R, KIT, MAPK14, of aging), publication, http://Geroprotectors.org [46] (Supplementary Tables 3–5) and aging gene database, GenAge [47]. Fourteen high confidence targets (ABL1, AR, ESR1, GHR, IGF1, IGF1R, KIT, MAPK14, Expression of target genes in 4 AAD classes. The consistency of gene dysregulation in each disease class is indica cale, with red standing for upregulation and blue for downregulation. Genes consistently dysregulated in multiple AADs were implicated as potential dual-purpose targets The color intensity indicates the level of consisten nsistently dysregulated (≥60% of comparisons) in 4 AAD classes in a unidirectional manner are shown in the black boxes. Figure 4. Expression of target genes in 4 AAD classes. The consistency of gene dysregulation in each disease class is indicated by the thermal scale, with red standing for upregulation and blue for downregulation. The color intensity indicates the level of consistency. Target genes consistently dysregulated (≥60% of comparisons) in 4 AAD classes in a unidirectional manner are shown in the black boxes. AGING 2482 www.aging-us.com four high confidence targets (AKT1, CASP3, CAT, CHUK, DNMT1, EGFR, HDAC9, IGF1, IGF1R, IL1B, IL6, JAK2, MAPK8, MMP1, MMP2, MMP9, MTOR, PPARA, PTEN, SIRT1, SOD2, TGFB1, TGFBR2 and TNF) were identified in the pool of 48 genes procured from publications (Exp = 0.85, fold = 28.13, p = 1.21E-30, Supplementary Table 4). While seven high confidence targets (CASP1, CASP3, CHUK, ESR1, HSPA5, IKBKB and MTOR) were further identified in the pool of 52 genes procured from geroprotectors (Exp = 0.92, fold = 7.57, p = 3.15E-5, Supplementary Table 5). This significant enrichment might indicate the potential clinical relevance of our AI-derived targets in treating aging-related processes and AADs. Moreover, significant overrepresentation was also observed in 38 high confidence targets that were overlapped with 149 aging-associated genes obtained from the benchmark aging gene database, GenAge (Exp = 2.65, fold = 14.35, p = 1.28E-35), further validating the approach used in this study. Linking the AI-derived targets to aging-associated pathways by pathway enrichment analysis Pathway enrichment analysis was performed on 145 AI-derived targets using Kyoto Encyclopedia of Genes and Genomes (KEGG) PATHWAY Database [48]. Genes were mapped to 225 KEGG pathways, of which 151 were significantly enriched (p < 0.05) (Supplementary Table 6) with 110 of the 145 potential targets. PI3K-AKT signaling pathway (hsa04151), MAPK signaling pathway (hsa04010) and FOXO signaling pathway (hsa04068) were in the top 10 enriched signaling axes known to be associated with aging. Notably, the AI-derived targets crosstalk with multiple key aging-associated pathways, such as those regulated by MAPK, PI3K-AKT and FOXO signaling networks (Figure 5), consequently contributing to modulating apoptosis, autophagy, cell proliferation, cell survival, DNA repair, epigenetic alteration, extracellular matrix organization, inflammation, Figure 5. AI-derived targets crosstalk to aging-associated signaling pathways. Pathway enrichment analysis was performed on our 145 AI-derived targets based on KEGG PATHWAY Database. (A) MAPK signaling pathway (hsa04010), (B) PI3K-AKT signaling pathway (hsa04151) and (C) FOXO signaling pathway (hsa04068) were among the top 10 enriched pathways that were known to be associated with aging. Forty-six AI-derived targets were involved. Target-target interactions were identified in the contexts of pathways and networks retrieved from KEGG PATHWAY Database and literature (Supplementary Table 8). Abbreviation: PAMPs: Pathogen-associated molecular patterns. Figure 5. AI-derived targets crosstalk to aging-associated signaling pathways. Pathway enrichment analysis was performed on our 145 AI-derived targets based on KEGG PATHWAY Database. (A) MAPK signaling pathway (hsa04010), (B) PI3K-AKT signaling pathway (hsa04151) and (C) FOXO signaling pathway (hsa04068) were among the top 10 enriched pathways that were known to be associated with aging. Forty-six AI-derived targets were involved. Target-target interactions were identified in the contexts of pathways and networks retrieved from KEGG PATHWAY Database and literature (Supplementary Table 8). Abbreviation: PAMPs: Pathogen-associated molecular patterns. AGING AGING 2483 www.aging-us.com bone and produced by osteoblasts, osteocytes, other hematopoietic cells, or immune cells [55]. SPP1 was uniformly upregulated in more than 80% comparisons in all four AAD classes, with significantly higher logFC in AADs than in NAADs (p < 0.001, Figure 6B). It was suggested that SPP1 may aggravate neurodegenerative, auto-immune, and inflammatory conditions. For example, SPP1 expressed by fast fatigue-resistant or slow motor neurons contributed to the second-wave neurodegeneration in ALS in vivo [56]. In addition, elevated levels of SPP1 in cerebrospinal fluid of subjects with Parkinson’s disease were associated with more severe motor symptoms. Importantly, SPP1-null mice demonstrated reduced neurodegeneration [57]. Linking the AI-derived targets to aging-associated pathways by pathway enrichment analysis Besides, SPP1 upregulated lysyl oxidase, an enzyme involved in cross-linking insoluble collagen in fibroblasts. An excess of SPP1 was associated with left- ventricular stiffness and systolic dysfunction in patients with chronic heart failure and hypertensive heart disease [58]. The levels of active TGF-beta and MMP-2, two essential fibrogenic signaling mediators, as well as type I collagen expression were significantly attenuated in SPP1-null mice treated with bleomycin, fibrosis inducer, compared to wild-type controls [59]. Taken together, these findings strongly suggest that suppression of SPP1 is a highly potential therapeutic approach for aging and AADs. mitochondrial maintenance, stemness and telomere maintenance (Supplementary Table 2). CXCL12 Aging was associated with elevated levels of proinflammatory cytokines, consequently leading to a decrease in mesenchymal stem cell (MSC) ability to regenerate and differentiate in inflammatory conditions [49]. Despite rescuing oxidative stress-induced hematopoietic stem cell (HSC) damage at the mitochondrial level, C-X-C motif chemokine ligand 12 (CXCL12) acted as a proinflammatory cytokine [50]. CXCL12 was uniformly upregulated in more than 70% comparisons in all four AAD classes i.e., neurological, metabolic, inflammatory, and fibrotic diseases (Figure 6A). In general, the log-fold change (logFC) of CXCL12 in AADs was significantly higher than in NAADs (p < 0.001, Figure 6A). Accumulating evidence demonstrates that CXCL12 upregulation was implicated in AADs including IPF [51], rheumatoid arthritis (RA) [52], and amyotrophic lateral sclerosis (ALS) [53]. Consistent with our findings, upregulation of CXCL12 was suggested to promote migration and proliferation of human lung fibroblast in IPF as well as to enhance monocytes infiltration into the synovial tissue in RA [51, 52]. Treatment with an antagonist of CXCR4, the receptor for CXCL12, extended lifespan, improved motor function, and led to weight loss in ALS in vivo [54]. Aging-associated degenerative diseases such as osteoporosis were linked to dysfunctional stem cell differentiation and a decline in the regenerative capacity of musculoskeletal stem cells, resulting from the secretion of pro-inflammatory cytokines such as CXCL12 [49]. Tinzaparin, a CXCL12 inhibitor, is an FDA- approved drug for the treatment of deep vein thrombosis and pulmonary embolism, which are considered as AADs. Taken together, suppression of CXCL12 is one of the potential therapeutic approaches that may be considered towards slowing down aging-associated processes and preventing the onset of AADs. AI-derived targets demonstrate dual roles in aging and AADs The dual-purpose candidates were selected under the considerations of hallmarks of aging assessment (Supplementary Table 2), expression analysis (Figure 4), ranking calculated by PandaOmics, safety assessment, clinical trial status and druggability, yielding a list of 9 potential candidates (Table 2). Selected promising high confidence targets and novel targets are discussed below. ITGB5 Integrin alpha V beta 5 (ITGB5) encodes a subunit of integrin that can interact with several alpha chains to form a variety of integrin heterodimers. ITGB5 was consistently upregulated in more than 60% comparisons in all four AAD classes, and the logFC of ITGB5 in AADs was significantly higher than NAADs (p < 0.05, Figure 6C). Consistently, ITGB5 was also found to be upregulated in chronic kidney disease and psoriatic arthritis [60, 61]. In particular, ITGB5 was significantly increased in the serum of patients with psoriatic arthritis, a distinct inflammatory arthritis occurring in 30% of psoriasis patients [60]. ITGB5 was suggested as a biomarker for both nonprogressive and progressive kidney diseases [61], and was also one of the genes strongly associated with ischemic heart disease [62]. Moreover, ITGB5 served as a ligand for Cyr61, a molecule stimulating the production of IL-6, which is considered an aging biomarker, via itgav/itgb5/Akt/ NF-κB signaling pathway in RA [63, 64], further supporting its role in various mechanisms underlying multiple AADs. Integrin alpha V beta 5 (ITGB5) encodes a subunit of integrin that can interact with several alpha chains to form a variety of integrin heterodimers. ITGB5 was consistently upregulated in more than 60% comparisons in all four AAD classes, and the logFC of ITGB5 in AADs was significantly higher than NAADs (p < 0.05, Figure 6C). Consistently, ITGB5 was also found to be upregulated in chronic kidney disease and psoriatic arthritis [60, 61]. In particular, ITGB5 was significantly increased in the serum of patients with psoriatic arthritis, a distinct inflammatory arthritis occurring in 30% of psoriasis patients [60]. ITGB5 was suggested as a biomarker for both nonprogressive and progressive kidney diseases [61], and was also one of the genes strongly associated with ischemic heart disease [62]. Moreover, ITGB5 served as a ligand for Cyr61, a molecule stimulating the production of IL-6, which is considered an aging biomarker, via itgav/itgb5/Akt/ NF-κB signaling pathway in RA [63, 64], further supporting its role in various mechanisms underlying multiple AADs. ADAMTS14 Secreted phosphoprotein 1 (SPP1) functions as Th1 cytokine and is a secreted matrix glycoprotein located in A disintegrin and metalloproteinase with thrombospondin motifs 14 (ADAMTS14) cleaves the amino-propeptides AGING 2484 www.aging-us.com Table 2. List of prioritized targets. Table 2. List of prioritized targets. Target1 Protein family Hallmarks of aging Dysregulation in AAD classes Role in aging Drugs in clinical trials Severe toxicity3 Reference High confidence CXCL12 Cytokine Inflammation, Stem cell exhaustion ALL Upregulated CXCL12 is an aging-upregulated gene and a mediator of the crosstalk between vascular cells and many brain cell types (pro-aging; therapy approach: antagonist) Tinzaparin (phase 4) No evidence [94] SPP1 Chemokine Extracellular matrix stiffness, Inflammation, Stem cell exhaustion ALL Upregulated Age-dependent increase in SPP1 levels inhibited skeletal muscle regeneration (pro-aging; therapy approach: antagonist) ASK-8007 (phase 1/2) No evidence [95, 96] NCT00411424 Medium novel ITGB5 Receptor Altered intercellular communications, Extracellular matrix stiffness ALL Upregulated ITGB5 is a TGF-β activator. TGF-β signaling, being downstream of other signals, was shown to repress body size as well as lifespan in vivo (pro-aging; therapy approach: antagonist) Cilengitide (phase 3) No evidence [97] NCT00689221 PPM1A Esterase Deregulated nutrient signaling, Inflammation ALL Downregulated PPM1A stimulated macrophages to produce TNF through TLR4 (anti-aging; therapy approach: agonist) No No evidence; absence in DEG [98] Highly novel RAB7B Hydrolase Impaired proteostasis, Inflammation, Mitochondrial dysfunction ALL Upregulated RAB7B negatively regulated TLR4 signaling in macrophages and autophagic flux as well as prevented inflammation and autophagy upon damage (anti-aging2; therapy approach: agonist) No No evidence; absence in DEG [99] ADAMTS 14 Peptidase Extracellular matrix stiffness Upregulated in neurological and fibrotic diseases ADAMTS14 is responsible for the degradation of ECM collagen. During aging, fibroblast-ECM interactions become disrupted due to the fragmentation of collagen fibrils. Fibroblasts synthesized fewer ECM proteins and more matrix-degrading metalloproteinases (pro-aging; therapy approach: antagonist) No No evidence, absence in DEG [100] KDM7A Oxidoreductase Altered intercellular communications, Genome instability Downregulated in neurological and fibrotic diseases Age-related neural dedifferentiation might contribute to many cognitive abilities decline with age. KDM7A regulated neural differentiation through FGF4, and was associated with Wnt signaling (anti-aging; therapy approach: agonist) No No evidence [101, 102] MYSM1 Peptidase Cellular senescence, Inflammation, Stem cell exhaustion Downregulated in neurological, fibrotic and metabolic diseases MYSM1 functionally reduced cellular senescence and the aging process. MYSM1 deficiency promoted the aging process and decreased lifespan while its overexpression inhibited the aging process and increased lifespan in vivo. ADAMTS14 (anti-aging; therapy approach: agonist) No No evidence [103] MTMR4 Esterase Altered intercellular communications Downregulated in neurological, fibrotic and metabolic diseases Skeletal muscle atrophy accompanies many chronic disease states and normal aging (anti-aging; therapy approach: agonist) No No evidence [104] Note: 1Targets selected for comprehensive target review are in BOLD 2Based on its mechanism of action i e protective role Drugs in clinical trials Severe toxicity3 Reference RAB7B Hydrolase Impaired proteostasis, Inflammation, Mitochondrial dysfunction ALL Upregulated RAB7B negatively regulated TLR4 signaling in macrophages and autophagic flux as well as prevented inflammation and autophagy upon damage (anti-aging2; therapy approach: agonist) No No evidence; absence in DEG [99] ADAMTS 14 Peptidase Extracellular matrix stiffness Upregulated in neurological and fibrotic diseases ADAMTS14 is responsible for the degradation of ECM collagen. During aging, fibroblast-ECM interactions become disrupted due to the fragmentation of collagen fibrils. Fibroblasts synthesized fewer ECM proteins and more matrix-degrading metalloproteinases (pro-aging; therapy approach: antagonist) No No evidence, absence in DEG [100] KDM7A Oxidoreductase Altered intercellular communications, Genome instability Downregulated in neurological and fibrotic diseases Age-related neural dedifferentiation might contribute to many cognitive abilities decline with age. KDM7A regulated neural differentiation through FGF4, and was associated with Wnt signaling (anti-aging; therapy approach: agonist) No No evidence [101, 102] MYSM1 Peptidase Cellular senescence, Inflammation, Stem cell exhaustion Downregulated in neurological, fibrotic and metabolic diseases MYSM1 functionally reduced cellular senescence and the aging process. MYSM1 deficiency promoted the aging process and decreased lifespan while its overexpression inhibited the aging process and increased lifespan in vivo. (anti-aging; therapy approach: agonist) No No evidence [103] MTMR4 Esterase Altered intercellular communications Downregulated in neurological, fibrotic and metabolic diseases Skeletal muscle atrophy accompanies many chronic disease states and normal aging (anti-aging; therapy approach: agonist) No No evidence [104] Note: 1Targets selected for comprehensive target review are in BOLD. 2Based on its mechanism of action i.e., protective role. 3Database of Essential Gene (DEG) is freely accessible from the website http://tubic.tju.edu.cn/deg. Note: 1Targets selected for comprehensive target review are in BOLD. 2Based on its mechanism of action i.e., protective role. 3Database of Essential Gene (DEG) is freely accessible from the website http://tubic.tju.edu.cn/deg. AGING 2485 www.aging-us.com of fibrillar collagens, enabling collagen fibril formation prior to assembly of collagen, a major extracellular matrix (ECM) protein. ADAMTS14 was uniformly upregulated in more than 65% comparisons in neurological and fibrotic diseases. ADAMTS14 The logFC of ADAMTS14 in AADs was also significantly higher than in NAADs (p < 0.01, Figure 6D). Significant upregulation of ADAMTS14 was observed in human osteoarthritis (OA) cartilage, suggesting its involvement in cartilage matrix anabolism [65]. ADAMTS14 was also linked to the susceptibility to aging-related Alzheimer's disease as well as the regulation of immune functions via TGF-beta signaling [66]. ADAMTS14- deficient mice remained healthy, fertile, and displayed normal amino-procollagen processing [67], suggesting that antagonizing ADAMTS14 is unlikely to result in severe toxicity. As such, pharmaceutical inhibition of ADAMTS14 may provide a promising therapeutic approach for aging and AADs. DISCUSSION In recent years, extensive efforts have been applied to generating a wide range of transcriptomic, genomic, proteomic, imaging, methylation, and metagenomic aging-related data. However, analysis of such a massive functions via TGF-beta signaling [66]. ADAMTS14- aging-related data. However, analysis of such a massive Figure 6. Expression of target genes in different diseases. The logFC of gene expression were shown for (A) CXCL12, (B) SPP1, (C) ITGB5, or (D) ADAMTS14 in AADs and NAADs. For each gene, comparisons of the logFC value were conducted between NAAD and each of the AAD classes, with significant difference indicated by asterisks (two-tailed t-test, p < 0.05, p < 0.01, p < 0.001). Figure 6. Expression of target genes in different diseases. The logFC of gene expression were shown for (A) CXCL12, (B) SPP1, (C) ITGB5, or (D) ADAMTS14 in AADs and NAADs. For each gene, comparisons of the logFC value were conducted between NAAD and each of the AAD classes, with significant difference indicated by asterisks (two-tailed t-test, p < 0.05, p < 0.01, p < 0.001). AGING 2486 www.aging-us.com targets with pathways known to be involved in aging such as MTOR, SIRT1, IGF1, and AKT1. Interestingly, the well-known aging-related genes were often the top- ranked targets in both AADs and NAADs, possibly due to their involvement in a wide spectrum of pathways. Among high confidence targets with the most associated hallmarks were IGF1R, HGF, IL6, MMP1, PARP1, SPP1, and ROCK1. Whereas in terms of novel targets, we found MYSM1, KAT6A, UBE2E3, RAB7B, RAB8B, and USP2. The most frequently associated hallmark of our targets is inflammation. Each proposed target is associated with distinct patterns of aging hallmarks, suggesting complex mechanisms underlying the aging process. Nonetheless, targets associated with multiple hallmarks of aging should be considered for further studies. Notably, while some of the targets revealed by our analysis (such as IGF1R, HGF, and KAT6A), are well-characterized drivers of tumorigenesis, others are known tumor suppressors e.g., PTEN, EP300. While these targets may have a theoretic therapeutic potential in AADs setting, modulating these molecules may elevate the risk of cancer development, and they should be excluded from further consideration. amount of data requires tailored computational approaches, capable of providing a detailed overview of the aging process as well as identifying promising targets for delaying aging and treating age-associated diseases. DISCUSSION PandaOmics has several unique advantages with respect to user experience, integrated deep learning-based algorithms, the comprehensive database, and the time machine validation approach [19]. In contrast to other alternatives, PandaOmics platform consists of a diverse set of validated AI analytical models (such as text mining, entity recognition, target ranking, and trend prediction), coupled with the ability to discover novel targets automatically, making this platform unique in the community. While we acknowledge that inclusion and exclusion of AADs can impact the outcome of our analyses, unfortunately, we could not include all AADs into our study due to limited publicly available datasets for some diseases. Given this limitation, the selection of 14 AADs was based on the consideration of whether age is a strong risk factor for the disease’s onset, as well as the availability of public datasets. Cardiovascular diseases were not included in this study due to their common mechanistic root contributing to the insufficient blood supply to multiple organs [68, 69]. Cancers were also excluded, as some of the pathways and mechanisms implicated in tumorigenesis are contradictory to those typically implicated in aging, such as increased cell proliferation [70]. Regarding target selection, some of the aging-associated genes were filtered out due to target family consideration, for example, transcription factors and generic proteins were not included. Furthermore, the current analysis only retrieved transcriptomics data, which, in turn, restricted the depth of analysis. The incorporation of genomic data could bring deeper insights into the shared genetics between aging and aging-associated disorders. Moreover, as aimed to identify dual-purpose targets across aging and multiple AADs, genes that did not meet the dual- purpose were not selected by this approach. It is also worthy to note the trade-off between target novelty and the evidence connecting a target to a disease. The degree of novelty is defined by the volume of related publications, and thus increasing the novelty level will sacrifice the evidence supporting the target’s participation in the disease. Therefore, some of the highly novel targets selected by PandaOmics, were not mapped to any aging hallmarks due to the lack of literature support. Nevertheless, they could be potential aging-related target candidates worthy of further investigation. By further evaluating targets linked to the aging hallmarks and expression changes in AADs, 9 potential candidates were revealed. DISCUSSION Many of these targets play roles in inflammation, which is in line with the view that inflammation is associated with multiple age- related diseases and is an intrinsic and major component of the aging processes [71]. As previous studies have also reported strong overlaps of inflammation-related genes between aging and age-related diseases [14], targeting the immune dysfunction in aging could be a powerful approach for improving healthspan [72]. In addition, several strong candidate targets play roles in ECM remodeling. While this signaling network is not a hallmark of aging, it clearly plays an important role during aging [73]. As such, our findings support the view that ECM can be considered as a hallmark of aging and a promising therapeutic target for developing interventions [9]. Considering the potential targets we selected in the present study, the clinical relevance of CXCL12, SPP1, ITGB5 and ADAMTS14 in neurodegenerative, auto- immune, and inflammatory conditions was demonstrated in AADs. Thus, targeting these genes may have major health and clinical benefits for both aging and AADs. Stromal aging fibroblasts expressed and secreted a higher level of CXCL12 than the young cell [74]. In addition, CXCL12 demonstrated an activating role on mature osteoclast by promoting bone-resorbing activity [75], supporting the observation that CXCL12 plasma level was inversely correlated with bone mineral density [76]. Consistently, with age, the rate of bone By combining genes derived from a variety of AADs, we were able to establish potential targets at different levels of novelty. The subsequent association of these AGING 2487 www.aging-us.com specifically demethylates H3K9me2, H3K27me2 and H4K20me1. Histone-lysine N-methyltransferase EZH2 methylates H3K9me and H3K27me, leading to epigenetic transcriptional repression of the affected gene. Some of these epigenetic regulatory enzyme targets were also involved in modulating aging processes. For example, it was reported that KAT8 might alter the function of ATM, which plays a pro- longevity role [90, 91]. In addition, the activities of DNMT1 and SIRT1 were found to be attenuated during aging, leading to alterations of epigenetic landscape, thereby changing gene expression and promoting aging processes [3]. In aging livers, C/EBPβ-HDAC1 complexes repress E2F-dependent promoters and occupy the promoter of GSK3B, resulting in epigenetic silencing of cell cycle genes and altered GSKβ-cyclin D3 pathways, respectively [92]. DISCUSSION Collectively, the above evidence reveals the involvement of our target in epigenetic regulation of cellular proliferation and development, and suggests the potential mechanism for the involvement of these targets in aging and AADs. Targeting epigenetic regulation may be one of the promising approaches for healthspan-promotion and life-extension [93]. resorption exceeded that of bone formation, leading to bone loss. SPP1 modulated osteoclast differentiation [77], and its levels in the plasma of aged human donors were significantly higher than in young individuals, both in a normal state or upon muscle injury [78]. SPP1 was demonstrated to attenuate the regenerative responses of old muscle stem cells. Neutralization of SPP1 recovered and enhanced the myogenic responses of old muscle stem cells, but failed to induce significant effect in young muscle stem cells, revealing the inhibitory effects of the age-dependent increase in SPP1 level on skeletal muscle regeneration [79]. Besides the age-dependent inflammation and bone loss, with age, collagen fibers became fragmented and stiff [73], disrupting various aspects of homeostasis and affecting healthy function. For example, aged fibroblast-ECM interactions were disrupted due to the fragmentation of collagen fibrils. Such fibroblasts synthesized fewer ECM proteins and more matrix-degrading metalloproteinases [80]. ADAMTS14 participates in degradation of ECM collagen. Aging-related increase in ECM stiffness leads to an imbalance in matrix components as well as deposition and cross-linking of collagen [81]. Other than collagen, fibronectin is also a component of the ECM, where ITGAV:ITGB5 is one of the receptors for fibronectin. Aging is associated with increased stretching of fibronectin fibrils and ECM maturation. ITGB5 was reported as a putative physiologic activator of TGF-β, leading to activation of ECM-bound latent TGF-β1 by traction. Consistently, ITGB5 knockout demonstrated the absence of TGF-β- related phenotype. The most putative TGF-β activators are functionally associated with the ECM [82]. TGF-β signaling, being downstream of other signals, was shown to repress body size as well as lifespan in vivo [83]. Notably, ITGB5 knockdown did not affect the proliferation of human adipose-derived stem cells [84], suggesting minimal cell toxicity induced. Therefore, suppressing ITGB5 may provide new insights for aging treatment. Taken together, inhibition of CXCL12, SPP1, ITGB5 and ADAMTS14 may provide a promising therapeutic approach for aging and AADs. In conclusion, we successfully established an approach to identify potential dual-purpose targets for aging and AADs, enabling biologists and clinicians to further investigate their therapeutic potential in a cost-saving and time-efficient manner for drug discovery. MATERIALS AND METHODS MATERIALS AND METHODS DISCUSSION These promising results underscore the ability of PandaOmics to identify novel targets not only for specific disorders, but across multiple types of diseases. Pathway enrichment analysis for identified targets The 145 genes identified from hallmarks of aging assessment were input to perform pathway enrichment analysis based on the KEGG PATHWAY Database [48] by clusterProfiler in R. Pathways with p < 0.05 were considered significantly enriched. Aging-associated pathways were further selected for visualization. Disease and dataset selection Diseases were selected and classified into either AADs or NAADs based on the consideration of whether age is a strong risk factor for the disease’s onset. To obtain a more aging-oriented result, 14 AADs and 19 NAADs were selected (Table 1). Epigenetic reprogramming is one of the most promising areas in longevity [85]. In line with this notion, 13 of our targets (HDAC1, HDAC9, EP300, KAT6A, KAT8, KDM7A, EZH2, DNMT1, SIRT1, MTOR, IGF1, AKT1, and MYO1C) were associated with the epigenetic shift aging hallmark (Figure 3). Histone modification and DNA methylation are the most studied epigenetic phenomena, and these modifications are accumulating over the life course. Histone deacetylases HDAC1 and HDAC9 are markers of epigenetic transcriptional repression. Whereas histone acetyltransferases EP300, KAT6A, and KAT8 enhance epigenetic transcriptional activation [86–89]. Histone demethylase KDM7A Microarray and RNA-seq datasets for the selected diseases containing case and control samples were retrieved from public repositories and processed by PandaOmics (Supplementary Table 7). A total of 79 and 113 datasets were selected for AADs and NAADs, respectively. For AADs, age information was available in 29 datasets, with 1,223 cases and 819 control samples. For NAADs, age information was available in 35 datasets containing 1,161 cases and 713 control samples. The mean age of cases and controls in AADs was 67.9 (s.d. = 17.50) and 60.91 (s.d. = 21.01), and in NAADs 36.87 (s.d. = 18.29) and 37.20 (s.d. = 19.15), respectively. AGING 2488 www.aging-us.com association with hallmarks of aging (search terms included in Supplementary Table 2). Studies that matched our search terms composed of all hallmarks and keywords of the corresponding pathways were selected for review. Their association with hallmarks of aging was decided based on their biological functions, pathways, and roles in regulating important pathways or genes associated with aging. All genes associated with hallmarks of aging were included, along with their literature evidence and PubMed ID (Supplementary Table 2). Among the genes included, those that are known cancer drivers were annotated by the data of the NCG7.0 database [33]. Identification of targets implicated in multiple diseases For each novelty setting, a list of 100 genes with the highest ranking calculated by PandaOmics was extracted from each disease, generating a combined list of genes from 14 AADs, and another from 19 NAADs. The genes were then prioritized by their (1) descending occurrence, and (2) ascending average ranking across multiple diseases, and those top-100 genes were selected for further analysis. Consequently, these selected genes from AADs were overlapped with those from NAADs to classify the genes into AAD targets and common targets, as exemplified in the Venn diagram (Figure 1). Meta-analysis For each dataset, case and control samples from the same tissue source were selected and compared, resulting in a total of 87 and 126 case-control comparisons for 14 AADs and 19 NAADs, respectively (Table 1). All the case-control comparisons performed for each disease were pooled into a single meta- analysis, yielding a total of 33 meta-analyses for all selected diseases subjected to target identification. Filter settings for target identification Targets were prioritized by PandaOmics (available at https://pandaomics.com/) using its AI hypothesis generation models based on 21 scores from Omics, Text-based, Financial, and KOL categories. Additional filters including Druggability (small molecules, antibodies, safety, novelty), Tissue specificity, Target family, and Development filters were applied to refine the list to satisfy the user’s research goals. In this study, only the genes belonging to the druggable protein class were included. The loss of novelty would be a trade-off for the abundance of evidence connecting a target to a disease. In view of this, a list of target genes in high confidence, medium novelty, and high novelty settings based on the volume of related publications proposed by PandaOmics’ proprietary AI engine, as well as the number of clinical trials they have been involved in was identified. A customized set of scores and filters was applied to obtain a list of genes with the associated final ranking, which represents the strength of association between a gene and the disease, for each novelty setting. Curation of known aging-associated genes The curation of well-known genes associated with aging was based on the genes targeted by the investigated drugs that entered clinical trials with either aging or healthy aging as one of the disease conditions (http://ClinicalTrials.gov, accessed on 30-DEC-2021), publication and geroprotectors (http://Geroprotectors.org, accessed on 17-FEB-2022) [46]. The curated genes were further refined to druggable genes in PandaOmics by (1) applying the filter of a druggable protein class and (2) adjusting the druggability filter (small molecule score ≥1 and safety score ≥1), yielding the final list of 62, 48 and 52 known aging-associated genes from clinical trials (Supplementary Table 3), publication (Supplementary Table 4) and geroprotectors (Supplementary Table 5), Expression levels in age-associated diseases The values of logFC for the genes in each of the 87 case-control comparisons performed for AADs were calculated. Considering the diverse complexity of mechanisms and pathologies in different diseases, we computed the consistency of each gene’s dysregulation state in each of the four disease classes (fibrotic, inflammatory, metabolic, and neurological diseases). Genes that were upregulated or downregulated in 60% or more of case-control comparisons in the disease class were considered as consistently dysregulated. Genes were further investigated provided that they were consistently dysregulated in the same trend in 2 or more disease classes. REFERENCES 1. de Magalhães JP, Wuttke D, Wood SH, Plank M, Vora C. Genome-environment interactions that modulate aging: powerful targets for drug discovery. Pharmacol Rev. 2012; 64:88–101. 1. de Magalhães JP, Wuttke D, Wood SH, Plank M, Vora C. Genome-environment interactions that modulate aging: powerful targets for drug discovery. Pharmacol Rev. 2012; 64:88–101. https://doi.org/10.1124/pr.110.004499 PMID:22090473 2. Kennedy BK, Berger SL, Brunet A, Campisi J, Cuervo AM, Epel ES, Franceschi C, Lithgow GJ, Morimoto RI, Pessin JE, Rando TA, Richardson A, Schadt EE, et al. Geroscience: linking aging to chronic disease. Cell. 2014; 159:709–13. Hallmarks of aging assessment The 300 genes obtained from the three novelty settings consisting of both AAD targets and common targets were subjected to literature review on PubMed for their AGING 2489 www.aging-us.com respectively. In addition, 307 genes associated with aging suggested by the database, GenAge (build 20) [47], were also retrieved and filtered based on the above settings, resulting in 149 genes included for analysis. prepared data curation, methodology, and software. JW and FR participated in the result interpretation. AA designed the study and provided resources. EI and AM reviewed the manuscript. JPM participated in result interpretation and drafted the manuscript. AZ provided conceptualization, reviewed the manuscript, provided resources, and supervised the projects. All authors read and approved the final manuscript. https://doi.org/10.1016/j.cell.2014.10.039 PMID:25417146 3. Moskalev AA, Aliper AM, Smit-McBride Z, Buzdin A, Zhavoronkov A. Genetics and epigenetics of aging and longevity. Cell Cycle. 2014; 13:1063–77. 3. Moskalev AA, Aliper AM, Smit-McBride Z, Buzdin A, Zhavoronkov A. Genetics and epigenetics of aging and longevity. Cell Cycle. 2014; 13:1063–77. https://doi.org/10.4161/cc.28433 PMID:24603410 Abbreviations ADAMTS14: A disintegrin and metalloproteinase with thrombospondin motifs 14; AAD: Age-associated disease; AI: Artificial Intelligence; ALS: Amyotrophic lateral sclerosis; COPD: Chronic obstructive pulmonary disease; CXCR4: C-X-C Motif Chemokine Receptor 4; CXCL12: C-X-C Motif Chemokine Ligand 12; DEG: Database of essential genes; DNMT1: DNA (cytosine- 5)-methyltransferase 1; ECM: Extracellular matrix; FDA: Food and Drug Administration; GHR: Growth Hormone Receptor; HSC: Hematopoietic stem cell; IPF: Idiopathic pulmonary fibrosis; ITGB5: Integrin subunit beta 5; KEGG: Kyoto Encyclopedia of Genes and Genomes; KOL: Key opinion leader; logFC: log-fold change; MMP: Matrix metalloproteinase; MSC: Mesenchymal stem cell; NAAD: Non-age-associated disease; RA: Rheumatoid arthritis; SIRT1: Sirtuin 1; SPP1: Secreted phosphoprotein 1; TLR4: Toll-like receptor 4; OA: Osteoarthritis. Statistical analysis T-test analysis was performed to compare the logFC (two-tailed) for each gene calculated by PandaOmics between AADs and NAADs. The significant level of target enrichment in the pool of curated aging- associated genes or GenAge genes was estimated using the hypergeometric test as: ACKNOWLEDGMENTS We thank Ms. Elizaveta Ekimova and Mr. Mark Grigorenko for their technical assistance with figure design. ( )( ) ( ) 1 0 1 N K r K i n i N i n p − − − = = −∑ CONFLICTS OF INTEREST FP, GL, HL, BL, XL, IO, JW, FR, AA, and AZ are affiliated with Insilico Medicine, a commercial company developing AI solutions for aging research, drug discovery, and longevity medicine. JPM is a consultant to Insilico Medicine. FP, GL, HL, BL, XL, IO, JW, FR, AA, and AZ are affiliated with Insilico Medicine, a commercial company developing AI solutions for aging research, drug discovery, and longevity medicine. JPM is a consultant to Insilico Medicine. where N equals 5,626 which stands for the total number of druggable genes defined in PandaOmics, K represents the number of aging-associated genes in the interested pool, n is the number of identified targets, and r represents the number of genes shared between the interested pool of aging-associated genes and the list of identified targets. FUNDING This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. https://doi.org/10.4161/cc.28433 PMID:24603410 FP performed data analysis, participated in result interpretation and project administration, and drafted the manuscript. GL and HL analyzed data, performed visualization, participated in result interpretation, and drafted the manuscript. BL participated in result interpretation and drafted the manuscript. XL performed the statistical analysis and drafted the manuscript. IO 4. Moskalev A, Chernyagina E, Tsvetkov V, Fedintsev A, Shaposhnikov M, Krut’ko V, Zhavoronkov A, Kennedy BK. Developing criteria for evaluation of geroprotectors as a key stage toward translation to the clinic. Aging Cell. 2016; 15:407–15. AGING 2490 www.aging-us.com 14. Fernandes M, Wan C, Tacutu R, Barardo D, Rajput A, Wang J, Thoppil H, Thornton D, Yang C, Freitas A, de Magalhães JP. Systematic analysis of the gerontome reveals links between aging and age-related diseases. Hum Mol Genet. 2016; 25:4804–18. https://doi.org/10.1093/hmg/ddw307 PMID:28175300 https://doi.org/10.1111/acel.12463 PMID:26970234 5. de Magalhães JP. Longevity pharmacology comes of age. Drug Discov Today. 2021; 26:1559–62. https://doi.org/10.1016/j.drudis.2021.02.015 PMID:33617794 https://doi.org/10.1038/s41573-020-0067-7 https://doi.org/10.1038/s41573-020-0067-7 5. de Magalhães JP. Longevity pharmacology comes of age. Drug Discov Today. 2021; 26:1559–62. https://doi.org/10.1016/j.drudis.2021.02.015 PMID:33617794 6. Partridge L, Fuentealba M, Kennedy BK. The quest to slow ageing through drug discovery. Nat Rev Drug Discov. 2020; 19:513–32. 15. Dugani SB, Hydoub YM, Ayala AP, Reka R, Nayfeh T, Ding JF, McCafferty SN, Alzuabi M, Farwati M, Murad MH, Alsheikh-Ali AA, Mora S. Risk Factors for Premature Myocardial Infarction: A Systematic Review and Meta-analysis of 77 Studies. Mayo Clin Proc Innov Qual Outcomes. 2021; 5:783–94. https://doi.org/10.1016/j.mayocpiqo.2021.03.009 PMID:34401655 https://doi.org/10.1186/s12859-021-04523-8 PMID:34983372 21. Fabris F, Palmer D, Salama KM, de Magalhães JP, Freitas AA. Using deep learning to associate human genes with age-related diseases. Bioinformatics. 2020; 36:2202–8. 13. Gasek NS, Kuchel GA, Kirkland JL, Xu M. Strategies for Targeting Senescent Cells in Human Disease. Nat Aging. 2021; 1:870–9. https://doi.org/10.1038/s43587-021-00121-8 PMID:34841261 https://doi.org/10.1016/j.tem.2022.01.007 PMID:35183431 8. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell. 2013; 153:1194–217. https://doi.org/10.1016/j.cell.2013.05.039 PMID:23746838 8. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell. 2013; 153:1194–217. PMID:32467649 7. Moskalev A, Guvatova Z, Lopes IA, Beckett CW, Kennedy BK, De Magalhaes JP, Makarov AA. Targeting aging mechanisms: pharmacological perspectives. Trends Endocrinol Metab. 2022; 33:266–80. https://doi.org/10.1016/j.tem.2022.01.007 PMID:35183431 16. Zhavoronkov A. Geroprotective and senoremediative strategies to reduce the comorbidity, infection rates, severity, and lethality in gerophilic and gerolavic infections. Aging (Albany NY). 2020; 12:6492–510. https://doi.org/10.18632/aging.102988 PMID:32229705 https://doi.org/10.1016/j.cell.2013.05.039 PMID:23746838 17. Paul SM, Mytelka DS, Dunwiddie CT, Persinger CC, Munos BH, Lindborg SR, Schacht AL. How to improve R&D productivity: the pharmaceutical industry's grand challenge. Nat Rev Drug Discov. 2010; 9:203–14. https://doi.org/10.1038/nrd3078 PMID:20168317 17. Paul SM, Mytelka DS, Dunwiddie CT, Persinger CC, Munos BH, Lindborg SR, Schacht AL. How to improve R&D productivity: the pharmaceutical industry's grand challenge. Nat Rev Drug Discov. 2010; 9:203–14. 9. Fedintsev A, Moskalev A. Stochastic non-enzymatic modification of long-lived macromolecules - A missing hallmark of aging. Ageing Res Rev. 2020; 62:101097. https://doi.org/10.1016/j.arr.2020.101097 PMID:32540391 https://doi.org/10.1038/s41586-021-03542-y PMID:34349292 11. Franceschi C, Garagnani P, Parini P, Giuliani C, Santoro A. Inflammaging: a new immune-metabolic viewpoint for age-related diseases. Nat Rev Endocrinol. 2018; 14:576–90. https://doi.org/10.1038/s41574-018-0059-4 PMID:30046148 PMID:30472217 20. Vega Magdaleno GD, Bespalov V, Zheng Y, Freitas AA, de Magalhaes JP. Machine learning-based predictions of dietary restriction associations across ageing- related genes. BMC Bioinformatics. 2022; 23:10. https://doi.org/10.1186/s12859-021-04523-8 PMID:34983372 20. Vega Magdaleno GD, Bespalov V, Zheng Y, Freitas AA, de Magalhaes JP. Machine learning-based predictions of dietary restriction associations across ageing- related genes. BMC Bioinformatics. 2022; 23:10. 12. Childs BG, Durik M, Baker DJ, van Deursen JM. Cellular senescence in aging and age-related disease: from mechanisms to therapy. Nat Med. 2015; 21:1424–35. https://doi.org/10.1038/nm.4000 PMID:26646499 12. Childs BG, Durik M, Baker DJ, van Deursen JM. Cellular senescence in aging and age-related disease: from mechanisms to therapy. Nat Med. 2015; 21:1424–35. https://doi.org/10.1038/nrd3078 10. Gorbunova V, Seluanov A, Mita P, McKerrow W, Fenyö D, Boeke JD, Linker SB, Gage FH, Kreiling JA, Petrashen AP, Woodham TA, Taylor JR, Helfand SL, Sedivy JM. The role of retrotransposable elements in ageing and age-associated diseases. Nature. 2021; 596:43–53. 18. Arrowsmith J. Trial watch: Phase II failures: 2008- 2010. Nat Rev Drug Discov. 2011; 10:328–9. https://doi.org/10.1038/nrd3439 PMID:21532551 19. Zhavoronkov A, Mamoshina P, Vanhaelen Q, Scheibye- Knudsen M, Moskalev A, Aliper A. Artificial intelligence for aging and longevity research: Recent advances and perspectives. Ageing Res Rev. 2019; 49:49–66. https://doi.org/10.1093/bioinformatics/btz887 PMID:31845988 22. Barardo DG, Newby D, Thornton D, Ghafourian T, de Magalhães JP, Freitas AA. Machine learning for AGING 2491 www.aging-us.com predicting lifespan-extending chemical compounds. Aging (Albany NY). 2017; 9:1721–37. https://doi.org/10.18632/aging.101264 PMID:28783712 30. Hale C. Breaking Big Pharma's AI barrier: Insilico Medicine uncovers novel target, new drug for pulmonary fibrosis in 18 months. Fierce Biotech: USA; 2021. 23. Galkin F, Mamoshina P, Aliper A, de Magalhães JP, Gladyshev VN, Zhavoronkov A. Biohorology and biomarkers of aging: Current state-of-the-art, challenges and opportunities. Ageing Res Rev. 2020; 60:101050. https://doi.org/10.1016/j.arr.2020.101050 PMID:32272169 31. Hale C. Insilico Medicine begins first human trial of its AI-designed drug for pulmonary fibrosis. Fierce Biotech: USA; 2021. 32. Hale C. Insilico Medicine's AI engines continue to churn out new drug candidates, now in kidney fibrosis. Fierce Biotech: USA; 2021. 33. Dressler L, Bortolomeazzi M, Keddar MR, Misetic H, Sartini G, Acha-Sagredo A, Montorsi L, Wijewardhane N, Repana D, Nulsen J, Goldman J, Pollitt M, Davis P, et al. Comparative assessment of genes driving cancer and somatic evolution in non-cancer tissues: an update of the Network of Cancer Genes (NCG) resource. Genome Biol. 2022; 23:35. https://doi.org/10.1186/s13059-022-02607-z PMID:35078504 24. Zhavoronkov A, Bischof E, Lee KF. Artificial intelligence in longevity medicine. Nat Aging. 2021; 1:5–7. https://doi.org/10.1038/s43587-020-00020-4 25. Ozerov IV, Lezhnina KV, Izumchenko E, Artemov AV, Medintsev S, Vanhaelen Q, Aliper A, Vijg J, Osipov AN, Labat I, West MD, Buzdin A, Cantor CR, et al. In silico Pathway Activation Network Decomposition Analysis (iPANDA) as a method for biomarker development. Nat Commun. 2016; 7:13427. https://doi.org/10.1038/ncomms13427 PMID:27848968 https://doi.org/10.18632/oncotarget.23748 27. Aliper A, Plis S, Artemov A, Ulloa A, Mamoshina P, Zhavoronkov A. Deep Learning Applications for Predicting Pharmacological Properties of Drugs and Drug Repurposing Using Transcriptomic Data. Mol Pharm. 2016; 13:2524–30. https://doi.org/10.1021/acs.molpharmaceut.6b00248 PMID:27200455 27. Aliper A, Plis S, Artemov A, Ulloa A, Mamoshina P, Zhavoronkov A. Deep Learning Applications for Predicting Pharmacological Properties of Drugs and Drug Repurposing Using Transcriptomic Data. Mol Pharm. 2016; 13:2524–30. 37. Li J, Kim SG, Blenis J. Rapamycin: one drug, many effects. Cell Metab. 2014; 19:373–9. https://doi.org/10.1016/j.cmet.2014.01.001 PMID:24508508 34. Longo VD, Kennedy BK. Sirtuins in aging and age- related disease. Cell. 2006; 126:257–68. https://doi.org/10.1016/j.cell.2006.07.002 PMID:16873059 34. Longo VD, Kennedy BK. Sirtuins in aging and age- related disease. Cell. 2006; 126:257–68. https://doi.org/10.1016/j.cell.2006.07.002 PMID:16873059 26. West MD, Labat I, Sternberg H, Larocca D, Nasonkin I, Chapman KB, Singh R, Makarev E, Aliper A, Kazennov A, Alekseenko A, Shuvalov N, Cheskidova E, et al. Use of deep neural network ensembles to identify embryonic-fetal transition markers: repression of COX7A1 in embryonic and cancer cells. Oncotarget. 2017; 9:7796–811. https://doi.org/10.18632/oncotarget.23748 PMID:29487692 35. Elibol B, Kilic U. High Levels of SIRT1 Expression as a Protective Mechanism Against Disease-Related Conditions. Front Endocrinol (Lausanne). 2018; 9:614. https://doi.org/10.3389/fendo.2018.00614 PMID:30374331 36. Blagosklonny MV. An anti-aging drug today: from senescence-promoting genes to anti-aging pill. Drug Discov Today. 2007; 12:218–24. https://doi.org/10.1016/j.drudis.2007.01.004 PMID:17331886 38. Papadopoli D, Boulay K, Kazak L, Pollak M, Mallette F, Topisirovic I, Hulea L. mTOR as a central regulator of lifespan and aging. F1000Res. 2019; 8:998. https://doi.org/10.12688/f1000research.17196.1 PMID:31316753 38. Papadopoli D, Boulay K, Kazak L, Pollak M, Mallette F, Topisirovic I, Hulea L. mTOR as a central regulator of lifespan and aging. F1000Res. 2019; 8:998. https://doi.org/10.12688/f1000research.17196.1 PMID:31316753 28. Aliper A, Belikov AV, Garazha A, Jellen L, Artemov A, Suntsova M, Ivanova A, Venkova L, Borisov N, Buzdin A, Mamoshina P, Putin E, Swick AG, et al. In search for geroprotectors: in silico screening and in vitro validation of signalome-level mimetics of young healthy state. Aging (Albany NY). 2016; 8:2127–52. https://doi.org/10.18632/aging.101047 PMID:27677171 39. Arum O, Boparai RK, Saleh JK, Wang F, Dirks AL, Turner JG, Kopchick JJ, Liu JL, Khardori RK, Bartke A. Specific suppression of insulin sensitivity in growth hormone receptor gene-disrupted (GHR-KO) mice attenuates phenotypic features of slow aging. Aging Cell. 2014; 13:981–1000. https://doi.org/10.1111/acel.12262 PMID:25244225 29. Aliper A, Jellen L, Cortese F, Artemov A, Karpinsky- Semper D, Moskalev A, Swick AG, Zhavoronkov A. Towards natural mimetics of metformin and rapamycin. Aging (Albany NY). 2017; 9:2245–68. https://doi.org/10.18632/aging.101319 PMID:29165314 maintenance and augments global homologous recombination. J Cell Biol. 2010; 191:1299–313. https://doi.org/10.1083/jcb.201005160 PMID:21187328 maintenance and augments global homologous recombination. J Cell Biol. 2010; 191:1299–313. https://doi.org/10.1083/jcb.201005160 PMID:21187328 https://doi.org/10.3390/ijms21186884 PMID:32961796 45. Lin YF, Sam J, Evans T. Sirt1 promotes tissue regeneration in zebrafish through regulating the mitochondrial unfolded protein response. iScience. 2021; 24:103118. 53. Andrés-Benito P, Povedano M, Domínguez R, Marco C, Colomina MJ, López-Pérez Ó, Santana I, Baldeiras I, Martínez-Yelámos S, Zerr I, Llorens F, Fernández-Irigoyen J, Santamaría E, Ferrer I. Increased C-X-C Motif Chemokine Ligand 12 Levels in Cerebrospinal Fluid as a Candidate Biomarker in Sporadic Amyotrophic Lateral Sclerosis. Int J Mol Sci. 2020; 21:8680. https://doi.org/10.1172/JCI97831 PMID:30300140 44. Yang YL, Wang PW, Wang FS, Lin HY, Huang YH. miR- 29a Modulates GSK3β/SIRT1-Linked Mitochondrial Proteostatic Stress to Ameliorate Mouse Non-Alcoholic Steatohepatitis. Int J Mol Sci. 2020; 21:6884. 52. Gao B, Sun G, Wang Y, Geng Y, Zhou L, Chen X. microRNA-23 inhibits inflammation to alleviate rheumatoid arthritis via regulating CXCL12. Exp Ther Med. 2021; 21:459. https://doi.org/10.3892/etm.2021.9890 PMID:33777193 https://doi.org/10.1111/acel.12262 PMID:25244225 40. Palacios JA, Herranz D, De Bonis ML, Velasco S, Serrano M, Blasco MA. SIRT1 contributes to telomere AGING 2492 www.aging-us.com https://doi.org/10.1016/j.isci.2021.103118 PMID:34622167 https://doi.org/10.1016/j.isci.2021.103118 PMID:34622167 46. Moskalev A, Chernyagina E, de Magalhães JP, Barardo D, Thoppil H, Shaposhnikov M, Budovsky A, Fraifeld VE, Garazha A, Tsvetkov V, Bronovitsky E, Bogomolov V, Scerbacov A, et al. Geroprotectors.org: a new, structured and curated database of current therapeutic interventions in aging and age-related disease. Aging (Albany NY). 2015; 7:616–28. https://doi.org/10.18632/aging.100799 PMID:26342919 https://doi.org/10.1186/s12974-016-0587-6 55. Zhao H, Chen Q, Alam A, Cui J, Suen KC, Soo AP, Eguchi S, Gu J, Ma D. The role of osteopontin in the progression of solid organ tumour. Cell Death Dis. 2018; 9:356. https://doi.org/10.3390/ijms21228680 PMID:33213069 54. Rabinovich-Nikitin I, Ezra A, Barbiro B, Rabinovich- Toidman P, Solomon B. Chronic administration of AMD3100 increases survival and alleviates pathology in SOD1(G93A) mice model of ALS. J Neuroinflammation. 2016; 13:123. https://doi.org/10.1186/s12974-016-0587-6 PMID:27230771 54. Rabinovich-Nikitin I, Ezra A, Barbiro B, Rabinovich- Toidman P, Solomon B. Chronic administration of AMD3100 increases survival and alleviates pathology in SOD1(G93A) mice model of ALS. J Neuroinflammation. 2016; 13:123. 47. Tacutu R, Thornton D, Johnson E, Budovsky A, Barardo D, Craig T, Diana E, Lehmann G, Toren D, Wang J, Fraifeld VE, de Magalhães JP. Human Ageing Genomic Resources: new and updated databases. Nucleic Acids Res. 2018; 46:D1083–90. https://doi.org/10.1093/nar/gkx1042 PMID:29121237 47. Tacutu R, Thornton D, Johnson E, Budovsky A, Barardo D, Craig T, Diana E, Lehmann G, Toren D, Wang J, Fraifeld VE, de Magalhães JP. Human Ageing Genomic Resources: new and updated databases. Nucleic Acids Res. 2018; 46:D1083–90. https://doi.org/10.1093/nar/28.1.27 PMID:10592173 49. Gilbert W, Bragg R, Elmansi AM, McGee-Lawrence ME, Isales CM, Hamrick MW, Hill WD, Fulzele S. Stromal cell-derived factor-1 (CXCL12) and its role in bone and muscle biology. Cytokine. 2019; 123:154783. https://doi.org/10.1016/j.cyto.2019.154783 PMID:31336263 41. Osum M, Serakinci N. Impact of circadian disruption on health; SIRT1 and Telomeres. DNA Repair (Amst). 2020; 96:102993. https://doi.org/10.1016/j.dnarep.2020.102993 PMID:33038659 42. Meng T, Qin W, Liu B. SIRT1 Antagonizes Oxidative Stress in Diabetic Vascular Complication. Front Endocrinol (Lausanne). 2020; 11:568861. https://doi.org/10.3389/fendo.2020.568861 PMID:33304318 50. Zhang Y, Dépond M, He L, Foudi A, Kwarteng EO, Lauret E, Plo I, Desterke C, Dessen P, Fujii N, Opolon P, Herault O, Solary E, et al. CXCR4/CXCL12 axis counteracts hematopoietic stem cell exhaustion through selective protection against oxidative stress. Sci Rep. 2016; 6:37827. https://doi.org/10.1038/srep37827 PMID:27886253 43. Kim TH, Yang YM, Han CY, Koo JH, Oh H, Kim SS, You BH, Choi YH, Park TS, Lee CH, Kurose H, Noureddin M, Seki E, et al. Gα12 ablation exacerbates liver steatosis and obesity by suppressing USP22/SIRT1-regulated mitochondrial respiration. J Clin Invest. 2018; 128:5587–602. 51. Li F, Xu X, Geng J, Wan X, Dai H. The autocrine CXCR4/CXCL12 axis contributes to lung fibrosis through modulation of lung fibroblast activity. Exp Ther Med. 2020; 19:1844–54. https://doi.org/10.3892/etm.2020.8433 PMID:32104240 https://doi.org/10.1093/nar/gkx1042 PMID:29121237 48. Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000; 28:27–30. https://doi.org/10.1038/s41419-018-0391-6 PMID:29500465 AGING 2493 www.aging-us.com 56. Morisaki Y, Niikura M, Watanabe M, Onishi K, Tanabe S, Moriwaki Y, Okuda T, Ohara S, Murayama S, Takao M, Uchida S, Yamanaka K, Misawa H. Selective Expression of Osteopontin in ALS-resistant Motor Neurons is a Critical Determinant of Late Phase Neurodegeneration Mediated by Matrix Metalloproteinase-9. Sci Rep. 2016; 6:27354. https://doi.org/10.1038/srep27354 PMID:27264390 57. Maetzler W, Berg D, Schalamberidze N, Melms A, Schott K, Mueller JC, Liaw L, Gasser T, Nitsch C. Osteopontin is elevated in Parkinson's disease and its absence leads to reduced neurodegeneration in the MPTP model. Neurobiol Dis. 2007; 25:473–82. https://doi.org/10.1016/j.nbd.2006.10.020 PMID:17188882 65. Yang CY, Chanalaris A, Troeberg L. ADAMTS and ADAM metalloproteinases in osteoarthritis - looking beyond the 'usual suspects'. Osteoarthritis Cartilage. 2017; 25:1000–9. https://doi.org/10.1016/j.joca.2017.02.791 PMID:28216310 65. Yang CY, Chanalaris A, Troeberg L. ADAMTS and ADAM metalloproteinases in osteoarthritis - looking beyond the 'usual suspects'. Osteoarthritis Cartilage. 2017; 25:1000–9. 66. Mukherjee S, Heath L, Preuss C, Jayadev S, Garden GA, Greenwood AK, Sieberts SK, De Jager PL, Ertekin- Taner N, Carter GW, Mangravite LM, Logsdon BA. Molecular estimation of neurodegeneration pseudotime in older brains. Nat Commun. 2020; 11:5781. https://doi.org/10.1038/s41467-020-19622-y PMID:33188183 58. López B, González A, Lindner D, Westermann D, Ravassa S, Beaumont J, Gallego I, Zudaire A, Brugnolaro C, Querejeta R, Larman M, Tschöpe C, Díez J. Osteopontin-mediated myocardial fibrosis in heart failure: a role for lysyl oxidase? Cardiovasc Res. 2013; 99:111–20. https://doi.org/10.1093/cvr/cvt100 PMID:23619422 67. Dupont L, Ehx G, Chantry M, Monseur C, Leduc C, Janssen L, Cataldo D, Thiry M, Jerome C, Thomassin JM, Nusgens B, Dubail J, Baron F, Colige A. Spontaneous atopic dermatitis due to immune dysregulation in mice lacking Adamts2 and 14. Matrix Biol. 2018; 70:140–57. https://doi.org/10.1016/j.matbio.2018.04.002 PMID:29649548 59. Berman JS, Serlin D, Li X, Whitley G, Hayes J, Rishikof DC, Ricupero DA, Liaw L, Goetschkes M, O’Regan AW. Altered bleomycin-induced lung fibrosis in osteopontin-deficient mice. Am J Physiol Lung Cell Mol Physiol. 2004; 286:L1311–8. https://doi.org/10.1152/ajplung.00394.2003 PMID:14977630 60. Cretu D, Liang K, Saraon P, Batruch I, Diamandis EP, Chandran V. Quantitative tandem mass-spectrometry of skin tissue reveals putative psoriatic arthritis biomarkers. Clin Proteomics. 2015; 12:1. https://doi.org/10.1186/1559-0275-12-1 PMID:25678896 68. Dutta P, Courties G, Wei Y, Leuschner F, Gorbatov R, Robbins CS, Iwamoto Y, Thompson B, Carlson AL, Heidt T, Majmudar MD, Lasitschka F, Etzrodt M, et al. Myocardial infarction accelerates atherosclerosis. Nature. 2012; 487:325–9. https://doi.org/10.1038/nature11260 PMID:22763456 68. Dutta P, Courties G, Wei Y, Leuschner F, Gorbatov R, Robbins CS, Iwamoto Y, Thompson B, Carlson AL, Heidt T, Majmudar MD, Lasitschka F, Etzrodt M, et al. Myocardial infarction accelerates atherosclerosis. Nature. 2012; 487:325–9. https://doi.org/10.1016/j.exger.2006.09.006 PMID:17071038 64. Lin J, Zhou Z, Huo R, Xiao L, Ouyang G, Wang L, Sun Y, Shen B, Li D, Li N. Cyr61 induces IL-6 production by fibroblast-like synoviocytes promoting Th17 differentiation in rheumatoid arthritis. J Immunol. 2012; 188:5776–84. https://doi.org/10.4049/jimmunol.1103201 PMID:22547695 https://doi.org/10.1038/nature11260 PMID:22763456 61. Ju W, Eichinger F, Bitzer M, Oh J, McWeeney S, Berthier CC, Shedden K, Cohen CD, Henger A, Krick S, Kopp JB, Stoeckert CJ Jr, Dikman S, et al. Renal gene and protein expression signatures for prediction of kidney disease progression. Am J Pathol. 2009; 174:2073–85. https://doi.org/10.2353/ajpath.2009.080888 PMID:19465643 69. Ross R. Atherosclerosis--an inflammatory disease. N Engl J Med. 1999; 340:115–26. https://doi.org/10.1056/NEJM199901143400207 PMID:9887164 70. Chatsirisupachai K, Palmer D, Ferreira S, de Magalhães JP. A human tissue-specific transcriptomic analysis reveals a complex relationship between aging, cancer, and cellular senescence. Aging Cell. 2019; 18:e13041. https://doi.org/10.1111/acel.13041 PMID:31560156 70. Chatsirisupachai K, Palmer D, Ferreira S, de Magalhães JP. A human tissue-specific transcriptomic analysis reveals a complex relationship between aging, cancer, and cellular senescence. Aging Cell. 2019; 18:e13041. 62. Schooling CM, Huang JV, Zhao JV, Kwok MK, Au Yeung SL, Lin SL. Disconnect Between Genes Associated With Ischemic Heart Disease and Targets of Ischemic Heart Disease Treatments. EBioMedicine. 2018; 28:311–5. https://doi.org/10.1016/j.ebiom.2018.01.015 PMID:29396305 ; https://doi.org/10.1111/acel.13041 PMID:31560156 71. Ferrucci L, Fabbri E. Inflammageing: chronic inflammation in ageing, cardiovascular disease, and frailty. Nat Rev Cardiol. 2018; 15:505–22. 63. Johnson TE. Recent results: biomarkers of aging. Exp Gerontol. 2006; 41:1243–6. AGING 2494 www.aging-us.com 73. Birch HL. Extracellular Matrix and Ageing. Subcell Biochem. 2018; 90:169–90. https://doi.org/10.1007/978-981-13-2835-0_7 PMID:30779010 73. Birch HL. Extracellular Matrix and Ageing. Subcell Biochem. 2018; 90:169–90. https://doi.org/10.1007/978-981-13-2835-0_7 PMID:30779010 82. Munger JS, Sheppard D. Cross talk among TGF-β signaling pathways, integrins, and the extracellular matrix. Cold Spring Harb Perspect Biol. 2011; 3:a005017. https://doi.org/10.1101/cshperspect.a005017 PMID:21900405 74. Begley L, Monteleon C, Shah RB, Macdonald JW, Macoska JA. CXCL12 overexpression and secretion by aging fibroblasts enhance human prostate epithelial proliferation in vitro. Aging Cell. 2005; 4:291–8. https://doi.org/10.1111/j.1474-9726.2005.00173.x PMID:16300481 83. Hirose T, Nakano Y, Nagamatsu Y, Misumi T, Ohta H, Ohshima Y. Cyclic GMP-dependent protein kinase EGL-4 controls body size and lifespan in C elegans. Development. 2003; 130:1089–99. https://doi.org/10.1242/dev.00330 PMID:12571101 75. Grassi F, Cristino S, Toneguzzi S, Piacentini A, Facchini A, Lisignoli G. CXCL12 chemokine up-regulates bone resorption and MMP-9 release by human osteoclasts: CXCL12 levels are increased in synovial and bone tissue of rheumatoid arthritis patients. J Cell Physiol. 2004; 199:244–51. 84. Morandi EM, Verstappen R, Zwierzina ME, Geley S, Pierer G, Ploner C. ITGAV and ITGA5 diversely regulate proliferation and adipogenic differentiation of human adipose derived stem cells. Sci Rep. 2016; 6:28889. https://doi.org/10.1038/srep28889 PMID:27363302 84. Morandi EM, Verstappen R, Zwierzina ME, Geley S, Pierer G, Ploner C. ITGAV and ITGA5 diversely regulate proliferation and adipogenic differentiation of human adipose derived stem cells. Sci Rep. 2016; 6:28889. https://doi.org/10.1038/s41569-018-0064-2 PMID:30065258 80. Cole MA, Quan T, Voorhees JJ, Fisher GJ. Extracellular matrix regulation of fibroblast function: redefining our perspective on skin aging. J Cell Commun Signal. 2018; 12:35–43. https://doi.org/10.1007/s12079-018-0459-1 PMID:29455303 72. Borgoni S, Kudryashova KS, Burka K, de Magalhães JP. Targeting immune dysfunction in aging. Ageing Res Rev. 2021; 70:101410. https://doi.org/10.1016/j.arr.2021.101410 PMID:34280555 81. Deasy SK, Erez N. A glitch in the matrix: organ-specific matrisomes in metastatic niches. Trends Cell Biol. 2022; 32:110–23. https://doi.org/10.1016/j.tcb.2021.08.001 PMID:34479765 https://doi.org/10.1002/jcp.10445 PMID:15040007 76. Carbone LD, Bůžková P, Fink HA, Robbins JA, Bethel M, Hamrick MW, Hill WD. Association of Plasma SDF- 1 with Bone Mineral Density, Body Composition, and Hip Fractures in Older Adults: The Cardiovascular Health Study. Calcif Tissue Int. 2017; 100:599–608. https://doi.org/10.1007/s00223-017-0245-8 PMID:28246930 85. de Magalhães JP, Ocampo A. Cellular reprogramming and the rise of rejuvenation biotech. Trends Biotechnol. 2022. [Epub ahead print]. https://doi.org/10.1016/j.tibtech.2022.01.011 PMID:35190201 85. de Magalhães JP, Ocampo A. Cellular reprogramming and the rise of rejuvenation biotech. Trends Biotechnol. 2022. [Epub ahead print]. https://doi.org/10.1016/j.tibtech.2022.01.011 PMID:35190201 77. Rittling SR, Matsumoto HN, McKee MD, Nanci A, An XR, Novick KE, Kowalski AJ, Noda M, Denhardt DT. Mice lacking osteopontin show normal development and bone structure but display altered osteoclast formation in vitro. J Bone Miner Res. 1998; 13:1101–11. https://doi.org/10.1359/jbmr.1998.13.7.1101 PMID:9661074 77. Rittling SR, Matsumoto HN, McKee MD, Nanci A, An XR, Novick KE, Kowalski AJ, Noda M, Denhardt DT. Mice lacking osteopontin show normal development and bone structure but display altered osteoclast formation in vitro. J Bone Miner Res. 1998; 13:1101–11. 86. Tropberger P, Pott S, Keller C, Kamieniarz-Gdula K, Caron M, Richter F, Li G, Mittler G, Liu ET, Bühler M, Margueron R, Schneider R. Regulation of transcription through acetylation of H3K122 on the lateral surface of the histone octamer. Cell. 2013; 152:859–72. https://doi.org/10.1016/j.cell.2013.01.032 PMID:23415232 78. Lekwuwa M, Choudhary M, Lad EM, Malek G. Osteopontin accumulates in basal deposits of human eyes with age-related macular degeneration and may serve as a biomarker of aging. Mod Pathol. 2022; 35:165–76. https://doi.org/10.1038/s41379-021-00887-7 PMID:34389792 78. Lekwuwa M, Choudhary M, Lad EM, Malek G. Osteopontin accumulates in basal deposits of human eyes with age-related macular degeneration and may serve as a biomarker of aging. Mod Pathol. 2022; 35:165–76. 87. Delvecchio M, Gaucher J, Aguilar-Gurrieri C, Ortega E, Panne D. Structure of the p300 catalytic core and implications for chromatin targeting and HAT regulation. Nat Struct Mol Biol. 2013; 20:1040–6. https://doi.org/10.1038/nsmb.2642 PMID:23934153 https://doi.org/10.1128/MCB.25.12.5292-5305.2005 PMID:15923642 99. Zhan L, Chen S, Li K, Liang D, Zhu X, Liu L, Lu Z, Sun W, Xu E. Autophagosome maturation mediated by Rab7 contributes to neuroprotection of hypoxic preconditioning against global cerebral ischemia in rats. Cell Death Dis. 2017; 8:e2949. https://doi.org/10.1038/cddis.2017.330 PMID:28726776 91. Qian M, Liu Z, Peng L, Tang X, Meng F, Ao Y, Zhou M, Wang M, Cao X, Qin B, Wang Z, Zhou Z, Wang G, et al. Boosting ATM activity alleviates aging and extends lifespan in a mouse model of progeria. Elife. 2018; 7:e34836. 97. Keyes BE, Fuchs E. Stem cells: Aging and transcriptional fingerprints. J Cell Biol. 2018; 217:79–92. https://doi.org/10.1083/jcb.201708099 PMID:29070608 89. Yuan H, Rossetto D, Mellert H, Dang W, Srinivasan M, Johnson J, Hodawadekar S, Ding EC, Speicher K, Abshiru N, Perry R, Wu J, Yang C, et al. MYST protein acetyltransferase activity requires active site lysine autoacetylation. EMBO J. 2012; 31:58–70. 98. Lee B, Song YS, Rhodes C, Goh TS, Roh JS, Jeong H, Park J, Lee HN, Lee SG, Kim S, Kim M, Lee SI, Sohn DH, Robinson WH. Protein phosphatase magnesium-dependent 1A induces inflammation in rheumatoid arthritis. Biochem Biophys Res Commun. 2020; 522:731–5. https://doi.org/10.1016/j.bbrc.2019.11.112 PMID:31791585 90. Gupta A, Sharma GG, Young CS, Agarwal M, Smith ER, Paull TT, Lucchesi JC, Khanna KK, Ludwig T, Pandita TK. Involvement of human MOF in ATM function. Mol Cell Biol. 2005; 25:5292–305. https://doi.org/10.1128/MCB.25.12.5292-5305.2005 PMID:15923642 90. Gupta A, Sharma GG, Young CS, Agarwal M, Smith ER, Paull TT, Lucchesi JC, Khanna KK, Ludwig T, Pandita TK. Involvement of human MOF in ATM function. Mol Cell Biol. 2005; 25:5292–305. https://doi.org/10.1038/s41379-021-00887-7 PMID:34389792 79. Paliwal P, Pishesha N, Wijaya D, Conboy IM. Age dependent increase in the levels of osteopontin inhibits skeletal muscle regeneration. Aging (Albany NY). 2012; 4:553–66. 88. Ogryzko VV, Schiltz RL, Russanova V, Howard BH, Nakatani Y. The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell. 1996; 87:953–9. https://doi.org/10.1016/s0092-8674(00)82001-2 PMID:8945521 https://doi.org/10.18632/aging.100477 PMID:22915705 AGING 2495 www.aging-us.com 97. Keyes BE, Fuchs E. Stem cells: Aging and transcriptional fingerprints. J Cell Biol. 2018; 217:79–92. https://doi.org/10.1083/jcb.201708099 PMID:29070608 93. Pasyukova EG, Vaiserman AM. HDAC inhibitors: A new promising drug class in anti-aging research. Mech Ageing Dev. 2017; 166:6–15. https://doi.org/10.1016/j.mad.2017.08.008 PMID:28843433 94. Tzeng YS, Li H, Kang YL, Chen WC, Cheng WC, Lai DM. Loss of Cxcl12/Sdf-1 in adult mice decreases the quiescent state of hematopoietic stem/progenitor cells and alters the pattern of hematopoietic regeneration after myelosuppression. Blood. 2011; 117:429–39. 102. Yang X, Wang G, Wang Y, Zhou J, Yuan H, Li X, Liu Y, Wang B. Histone demethylase KDM7A reciprocally regulates adipogenic and osteogenic differentiation via regulation of C/EBPα and canonical Wnt signalling. J Cell Mol Med. 2019; 23:2149–62. https://doi.org/10.1111/jcmm.14126 PMID:30614617 https://doi.org/10.7554/eLife.34836 PMID:29717979 100. Bolz H, Ramírez A, von Brederlow B, Kubisch C. Characterization of ADAMTS14, a novel member of the ADAMTS metalloproteinase family. Biochim Biophys Acta. 2001; 1522:221–5. https://doi.org/10.1016/s0167-4781(01)00329-3 PMID:11779638 92. Jin J, Wang GL, Shi X, Darlington GJ, Timchenko NA. The age-associated decline of glycogen synthase kinase 3beta plays a critical role in the inhibition of liver regeneration. Mol Cell Biol. 2009; 29:3867–80. https://doi.org/10.1128/MCB.00456-09 PMID:19398579 101. Zhang Z, Chen B, Zhu Y, Zhang T, Zhang X, Yuan Y, Xu Y. Corrigendum: The Jumonji Domain-Containing Histone Demethylase Homolog 1D/lysine Demethylase 7A (JHDM1D/KDM7A) Is an Epigenetic Activator of RHOJ Transcription in Breast Cancer Cells. Front Cell Dev Biol. 2021; 9:729416. https://doi.org/10.3389/fcell.2021.729416 PMID:34395453 93. Pasyukova EG, Vaiserman AM. HDAC inhibitors: A new promising drug class in anti-aging research. Mech Ageing Dev. 2017; 166:6–15. https://doi.org/10.1016/j.mad.2017.08.008 PMID:28843433 https://doi.org/10.1182/blood-2010-01-266833 PMID:20833981 95. Guidi N, Sacma M, Ständker L, Soller K, Marka G, Eiwen K, Weiss JM, Kirchhoff F, Weil T, Cancelas JA, Florian MC, Geiger H. Osteopontin attenuates aging- associated phenotypes of hematopoietic stem cells. EMBO J. 2017; 36:840–53. https://doi.org/10.15252/embj.201694969 PMID:28254837 103. Tian M, Huang Y, Song Y, Li W, Zhao P, Liu W, Wu K, Wu J. MYSM1 Suppresses Cellular Senescence and the Aging Process to Prolong Lifespan. Adv Sci (Weinh). 2020; 7:2001950. https://doi.org/10.1002/advs.202001950 PMID:33240758 104. Sheffield DA, Jepsen MR, Feeney SJ, Bertucci MC, Sriratana A, Naughtin MJ, Dyson JM, Coppel RL, Mitchell CA. The myotubularin MTMR4 regulates phagosomal phosphatidylinositol 3-phosphate turnover and phagocytosis. J Biol Chem. 2019; 294:16684–97. https://doi.org/10.1074/jbc.RA119.009133 PMID:31543504 96. Hashimoto M, Sun D, Rittling SR, Denhardt DT, Young W. Osteopontin-deficient mice exhibit less inflammation, greater tissue damage, and impaired locomotor recovery from spinal cord injury compared with wild-type controls. J Neurosci. 2007; 27:3603–11. https://doi.org/10.1523/JNEUROSCI.4805-06.2007 PMID:17392476 96. Hashimoto M, Sun D, Rittling SR, Denhardt DT, Young W. Osteopontin-deficient mice exhibit less inflammation, greater tissue damage, and impaired locomotor recovery from spinal cord injury compared with wild-type controls. J Neurosci. 2007; 27:3603–11. https://doi.org/10.1523/JNEUROSCI.4805-06.2007 PMID:17392476 AGING 2496 www.aging-us.com SUPPLEMENTARY MATERIALS Supplementary Figures Supplementary Figures Supplementary Figure 1. Flowchart of the selection of dual-purpose targets from the 14 AADs. Upon target identifications from the 14 AADs, 484 high confidence (HC) targets, 448 medium novel (MN) targets and 381 highly novel (HN) targets were identified by PandaOmics. Targets ranked as top 100 for each novelty (total 300 targets) were subjected to the hallmarks of aging assessment by searching the literature for their evidence in modulating longevity or longevity pathways, and consistency in dysregulated expression across disease classes. A total of 145 targets including 69 HC targets, 48 MN targets and 28 HN targets were associated with aging hallmarks whereas 52 HC targets, 44 MN targets and 45 HN targets were consistently dysregulated in two or more disease classes in a unidirectional manner. Potential dual-purpose candidates were selected with reference to both the hallmarks of aging assessment and expression analysis. Supplementary Figure 1. Flowchart of the selection of dual-purpose targets from the 14 AADs. Upon target identifications from the 14 AADs, 484 high confidence (HC) targets, 448 medium novel (MN) targets and 381 highly novel (HN) targets were identified by PandaOmics. https://doi.org/10.1182/blood-2010-01-266833 PMID:20833981 Targets ranked as top 100 for each novelty (total 300 targets) were subjected to the hallmarks of aging assessment by searching the literature for their evidence in modulating longevity or longevity pathways, and consistency in dysregulated expression across disease classes. A total of 145 targets including 69 HC targets, 48 MN targets and 28 HN targets were associated with aging hallmarks whereas 52 HC targets, 44 MN targets and 45 HN targets were consistently dysregulated in two or more disease classes in a unidirectional manner. Potential dual-purpose candidates were selected with reference to both the hallmarks of aging assessment and expression analysis. AGING 2497 www.aging-us.com www.aging-us.com plementary Figure 2. Occurrence of the top-100 targets in the 14 AADs. The y-axis indicates the percentage of disea ch the target was highly ranked (≤100) under (A) high confidence, (B) medium novel, and (C) highly novel filter settings. The targets highest percentages are exemplified above the horizontal dashed lines with their occurrence percentages shown in brackets. AAD red according to their disease classes. Supplementary Figure 2. Occurrence of the top-100 targets in the 14 AADs. The y-axis indicates the percentage of diseases in which the target was highly ranked (≤100) under (A) high confidence, (B) medium novel, and (C) highly novel filter settings. The targets with the highest percentages are exemplified above the horizontal dashed lines with their occurrence percentages shown in brackets. AADs are colored according to their disease classes. AGING 2498 www.aging-us.com ntary Figure 3. Ranking of the top-100 gene set for AADs under medium novelty settings. The ranking of the targets NAADs were colored in blue-white and red-white thermal scales respectively. High color intensity stands for high rankings. The ing was capped at 100. Targets associated with the hallmark(s) of aging are labeled in green. AADs and NAADs are colored o their disease classes. Supplementary Figure 3. Ranking of the top-100 gene set for AADs under medium novelty settings. The ranking of the targets in AADs or NAADs were colored in blue-white and red-white thermal scales respectively. High color intensity stands for high rankings. The lowest ranking was capped at 100. Targets associated with the hallmark(s) of aging are labeled in green. AADs and NAADs are colored according to their disease classes. AGING 2499 www.aging-us.com ary Figure 4. Ranking of the top-100 gene set for AADs under high novelty settings. https://doi.org/10.1182/blood-2010-01-266833 PMID:20833981 The ranking of the targets i Ds were colored blue-white and red-white thermal scales respectively. High color intensity stands for high rankings. Th was capped at 100. Targets associated with the hallmark(s) of aging are labeled in green. AADs and NAADs are colored eir disease classes. Supplementary Figure 4. Ranking of the top-100 gene set for AADs under high novelty settings. The ranking of the targets in AADs or NAADs were colored blue-white and red-white thermal scales respectively. High color intensity stands for high rankings. The lowest ranking was capped at 100. Targets associated with the hallmark(s) of aging are labeled in green. AADs and NAADs are colored according to their disease classes. AGING 2500 www.aging-us.com Supplementary Figure 5. Overlapping between the two sets of top-100 genes from AADs and NAADs. Top-ranked targets shared by both AAD and NAAD categories were regarded as common targets, while targets unique to AADs were defined as AAD targets under (A) high confidence, (B) medium novelty, and (C) high novelty filter settings. Supplementary Figure 5. Overlapping between the two sets of top-100 genes from AADs and NAADs. Top-ranked targets shared by both AAD and NAAD categories were regarded as common targets, while targets unique to AADs were defined as AAD targets under (A) high confidence, (B) medium novelty, and (C) high novelty filter settings. AGING 2501 www.aging-us.com www.aging-us.com Supplementary Tables Supplementary Tables Please browse Full Text version to see the data of Supplementary Tables 1, 2, 6 and 7. Supplementary Table 1. Top-100 genes for each filter setting. Supplementary Table 2. Genes associated with hallmarks of aging. Supplementary Table 3. Overlapping of high confidence targets with the pool of curated aging-associated genes from clinical trials. Gene1 Aging clinical trial2 Druggability3 Target family4 Top-100 target5 ABL1 DASATINIB (NCT04994561, NCT04946383) 2,0,2,0 Tyrosine kinase ✓ AR TESTOSTERONE (NCT00182871, NCT00309855, NCT00680797, NCT02203656, NCT02679274, NCT02990533) 2,0,2,0 Nuclear receptor ✓ ESR1 CLIMARA (NCT00220454, NCT02042196) 2,2,2,0 Nuclear receptor ✓ GHR GROWTH HORMONE RELEASING HORMONE (GHRH) (NCT01410799) 2,2,2,0 Immunoglobulin ✓ IGF1 ORALLY ACTIVE GROWTH HORMONE SECRETAGOGUE (MK-677) (NCT00474279) 2,2,2,0 Growth factor ✓ IGF1R INSULIN-LIKE GROWTH FACTOR 1 (NCT03932162) 2,2,2,0 Receptor kinase ✓ KIT DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Receptor kinase ✓ MAPK14 DASATINIB (NCT04994561, NCT04946383) 2,0,1,0 CMGC kinase ✓ MTOR RAPAMYCIN (NCT02874924, NCT04488601, NCT04742777) 2,0,2,0 Protein kinase ✓ NR3C1 METHYLPREDNISOLONE (NCT03529929) 2,0,2,0 Nuclear receptor ✓ PDGFRB DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Receptor kinase ✓ SIRT1 RESVERATROL (NCT02523274) 2,0,2,0 Acyltransferase ✓ SRC DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase ✓ VDR NT-020 (NCT01963767) 2,0,2,0 Nuclear receptor ✓ ACE PERINDOPRIL (NCT03295734) 2,2,2,0 Glycosylase AGTR1 CANDESARTAN (NCT00605072) 2,2,2,0 GPCR AMY2A ACARBOSE (NCT02865499, NCT02953093) 2,0,2,0 Glycosylase BCR DASATINIB (NCT04994561, NCT04946383) 2,2,1,0 Protein kinase BST1 NICOTINAMIDE MONONUCLEOTIDE (NCT04823260) 2,2,2,0 Glycosylase BTK DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase CHRNA3 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,1,0 Ion channel CHRNA4 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,2,0 Ion channel CSK DASATINIB (NCT04994561, NCT04946383) 2,0,2,0 Tyrosine kinase EPHA2 DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Receptor kinase EPHA5 DASATINIB (NCT04994561, NCT04946383) 2,2,1,0 Receptor kinase EPHB4 DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Receptor kinase ETFDH METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 2,0,1,0 Oxidoreductase FGR DASATINIB (NCT04994561, NCT04946383) 2,1,2,2 Tyrosine kinase FKBP1A RAPAMYCIN (NCT02874924, NCT04488601, NCT04742777) 2,0,2,0 Isomerase FRK DASATINIB (NCT04994561, NCT04946383) 2,0,1,2 Tyrosine kinase Please browse Full Text version to see the data of Supplementary Tables 1, 2, 6 and 7. Supplementary Table 2. Genes associated with hallmarks of aging. Supplementary Tables of high confidence targets with the pool of curated aging-associated genes AGING 2502 www.aging-us.com FYN DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase GAA ACARBOSE (NCT02865499, NCT02953093) 2,0,2,0 Glycosylase GABRA1 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GABRA2 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GABRA3 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,1 Ion channel GABRB1 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,1,1 Ion channel GABRG2 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GHSR ORALLY ACTIVE GROWTH HORMONE SECRETAGOGUE (MK-677) (NCT00474279) 2,0,2,0 GPCR GPD1 METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 1,0,1,0 Oxidoreductase GPD2 METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 2,0,1,0 Oxidoreductase LCK DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase MGAM ACARBOSE (NCT02865499, NCT02953093) 2,0,2,0 Glycosylase NR3C2 TESTOSTERONE (NCT00182871) 2,0,2,0 Nuclear receptor NR4A3 DASATINIB (NCT04994561, NCT04946383) 1,0,1,0 Nuclear receptor OXTR OXYTOCIN NASAL SPRAY (NCT03119610) 2,2,2,0 GPCR PPAT DASATINIB (NCT04994561, NCT04946383) 2,0,2,0 Glycosyltransferase PTGS1 DICLOFENAC (NCT03072485) 2,0,2,0 Oxidoreductase PTGS2 DICLOFENAC (NCT03072485) 2,2,2,0 Oxidoreductase SI ACARBOSE (NCT02865499, NCT02953093) 2,0,1,0 Glycosylase SRD5A1 DUTASTERIDE (NCT00309855) 2,0,2,0 Oxidoreductase SRD5A2 DUTASTERIDE (NCT00309855) 2,0,1,0 Oxidoreductase SRMS DASATINIB (NCT04994561, NCT04946383) 2,0,1,0 Tyrosine kinase TERT AAV-HTERT (NCT04133649) 2,2,2,0 Transferase YES1 DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase ABL2 DASATINIB (NCT04994561, NCT04946383) 1,0,1,1 Tyrosine kinase CHRNB2 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,2,1 Ion channel CHRNB4 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,2,1 Ion channel SRD5A3 DUTASTERIDE (NCT00309855) 2,2,1,1 Oxidoreductase BLK DASATINIB (NCT04994561, NCT04946383) 2,0,2,1 Tyrosine kinase HCK* DASATINIB (NCT04994561, NCT04946383) 2,2,2,1 Tyrosine kinase LYN* DASATINIB (NCT04994561, NCT04946383) 2,2,2,1 Tyrosine kinase PRKAB1^ METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 2,0,1,2 Protein kinase 1Curated pool of aging associated genes (genes identified as medium novel targets were marked with asterisks and highly n FYN DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase GAA ACARBOSE (NCT02865499, NCT02953093) 2,0,2,0 Glycosylase GABRA1 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GABRA2 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GABRA3 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,1 Ion channel GABRB1 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,1,1 Ion channel GABRG2 ZOLPIDEM (NCT00383357, NCT03657212) 2,0,2,0 Ion channel GHSR ORALLY ACTIVE GROWTH HORMONE SECRETAGOGUE (MK-677) (NCT00474279) 2,0,2,0 GPCR GPD1 METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 1,0,1,0 Oxidoreductase GPD2 METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 2,0,1,0 Oxidoreductase LCK DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase MGAM ACARBOSE (NCT02865499, NCT02953093) 2,0,2,0 Glycosylase NR3C2 TESTOSTERONE (NCT00182871) 2,0,2,0 Nuclear receptor NR4A3 DASATINIB (NCT04994561, NCT04946383) 1,0,1,0 Nuclear receptor OXTR OXYTOCIN NASAL SPRAY (NCT03119610) 2,2,2,0 GPCR PPAT DASATINIB (NCT04994561, NCT04946383) 2,0,2,0 Glycosyltransferase PTGS1 DICLOFENAC (NCT03072485) 2,0,2,0 Oxidoreductase PTGS2 DICLOFENAC (NCT03072485) 2,2,2,0 Oxidoreductase SI ACARBOSE (NCT02865499, NCT02953093) 2,0,1,0 Glycosylase SRD5A1 DUTASTERIDE (NCT00309855) 2,0,2,0 Oxidoreductase SRD5A2 DUTASTERIDE (NCT00309855) 2,0,1,0 Oxidoreductase SRMS DASATINIB (NCT04994561, NCT04946383) 2,0,1,0 Tyrosine kinase TERT AAV-HTERT (NCT04133649) 2,2,2,0 Transferase YES1 DASATINIB (NCT04994561, NCT04946383) 2,2,2,0 Tyrosine kinase ABL2 DASATINIB (NCT04994561, NCT04946383) 1,0,1,1 Tyrosine kinase CHRNB2 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,2,1 Ion channel CHRNB4 NICOTINE PATCH, ORAL MECAMYLAMINE, PLACEBO (NCT03408574) 2,0,2,1 Ion channel SRD5A3 DUTASTERIDE (NCT00309855) 2,2,1,1 Oxidoreductase BLK* DASATINIB (NCT04994561, NCT04946383) 2,0,2,1 Tyrosine kinase HCK* DASATINIB (NCT04994561, NCT04946383) 2,2,2,1 Tyrosine kinase LYN* DASATINIB (NCT04994561, NCT04946383) 2,2,2,1 Tyrosine kinase PRKAB1^ METFORMIN (NCT01765946, NCT02308228, NCT02432287, NCT03072485, NCT03309007, NCT03451006, NCT03713801, NCT04264897) 2,0,1,2 Protein kinase 1C t d l f i i t d ( id tifi d di l t t k d ith t i k d hi hl 1Curated pool of aging-associated genes (genes identified as medium novel targets were marked with asterisks, and highly novel targets with arrow heads) Sources of curation was http://ClinicalTrials.gov aging drug targets. Supplementary Tables 2Example of drug investigated in aging clinical trials with clinical trial ID shown in parenthesis. Target-drug association were manually curated. 3Druggability scores defined in PandaOmics (small molecules, antibodies, safety, novelty). 4Druggable gene classes defined in PandaOmics. 5Genes identified as top-100 high- confidence targets are marked with ticks. AGING 2503 www.aging-us.com Supplementary Table 4. Overlapping of high confidence targets with the pool of curated aging associated gen from publication. Supplementary Tables Genes1 Gene Name2 Druggability3 Target family4 Top-100 target5 AKT1 AKT serine/threonine kinase 1 2,0,2,0 AGC kinase ✓ CASP3 caspase 3 2,0,2,0 Peptidase ✓ CAT catalase 2,0,2,0 Oxidoreductase ✓ CHUK component of inhibitor of nuclear factor kappa B kinase complex 2,0,2,0 Protein kinase ✓ DNMT1 DNA methyltransferase 1 2,0,1,0 Methyltransferase ✓ EGFR epidermal growth factor receptor 2,2,2,0 Receptor kinase ✓ HDAC9 histone deacetylase 9 2,0,1,0 Hydrolase ✓ IGF1 insulin like growth factor 1 2,2,2,0 Growth factor ✓ IGF1R insulin like growth factor 1 receptor 2,2,2,0 Receptor kinase ✓ IL1B interleukin 1 beta 2,2,2,0 Interleukin ✓ IL6 interleukin 6 2,2,2,0 Interleukin ✓ JAK2 Janus kinase 2 2,0,2,0 Tyrosine kinase ✓ MAPK8 mitogen-activated protein kinase 8 2,0,2,0 CMGC kinase ✓ MMP1 matrix metallopeptidase 1 2,2,2,0 Peptidase ✓ MMP2 matrix metallopeptidase 2 2,2,2,0 Peptidase ✓ MMP9 matrix metallopeptidase 9 2,2,2,0 Peptidase ✓ MTOR mechanistic target of rapamycin kinase 2,0,2,0 Protein kinase ✓ PPARA peroxisome proliferator activated receptor alpha 2,0,2,0 Nuclear receptor ✓ PTEN phosphatase and tensin homolog 1,1,1,0 Esterase ✓ SIRT1 sirtuin 1 2,0,2,0 Acyltransferase ✓ SOD2 superoxide dismutase 2 1,0,1,0 Oxidoreductase ✓ TGFB1 transforming growth factor beta 1 2,2,2,0 Growth factor ✓ TGFBR2 transforming growth factor beta receptor 2 2,2,2,0 Receptor kinase ✓ TNF tumor necrosis factor 2,2,2,0 Tumor necrosis factor ✓ ABO ABO, alpha 1-3-N-acetylgalactosaminyltransferase and alpha 1-3-galactosyltransferase 1,0,1,0 Glycosyltransferase AKT2 AKT serine/threonine kinase 2 2,0,2,0 AGC kinase BMP1 bone morphogenetic protein 1 1,0,1,0 Peptidase GZMB granzyme B 2,0,2,0 Peptidase HAS1 hyaluronan synthase 1 1,2,1,0 Glycosyltransferase HAS2 hyaluronan synthase 2 1,0,1,0 Glycosyltransferase HDAC11 histone deacetylase 11 2,2,2,0 Hydrolase HDAC4 histone deacetylase 4 2,0,2,0 Hydrolase IL15 interleukin 15 2,2,2,0 Interleukin MME membrane metalloendopeptidase 2,2,2,0 Peptidase MMP13 matrix metallopeptidase 13 2,2,2,0 Peptidase MT-ND2 mitochondrially encoded NADH:ubiquinone oxidoreductase core subunit 2 2,0,1,0 Translocase NOX4 NADPH oxidase 4 2,0,2,0 Oxidoreductase PPIA peptidylprolyl isomerase A 2,0,2,0 Isomerase PRDX2 peroxiredoxin 2 1,0,1,0 Oxidoreductase PRKCD protein kinase C delta 2,0,2,0 AGC kinase AGING 2504 www.aging-us.com PTGS2 prostaglandin-endoperoxide synthase 2 2,2,2,0 Oxidoreductase SIRT6 sirtuin 6 2,0,1,0 Acyltransferase SOD1 superoxide dismutase 1 2,2,2,0 Oxidoreductase TYR tyrosinase 2,0,2,0 Oxidoreductase WNK2 WNK lysine deficient protein kinase 2 1,0,1,0 Protein kinase XDH xanthine dehydrogenase 2,0,2,0 Oxidoreductase AOPEP aminopeptidase O (putative) 2,0,1,2 Peptidase CLOCK^ clock circadian regulator 1,0,1,2 Acyltransferase 1Curated pool of aging-associated genes (genes identified as medium novel targets are marked with asterisks, and highly novel targets with arrow heads). Supplementary Tables 2These genes were associated with aging or skin aging with reference to a publicity database. 3Druggability scores defined in PandaOmics (small molecules, antibodies, safety, novelty). 4Druggable gene classes defined in PandaOmics. 5Genes identified as top-100 high confidence targets were marked with ticks. 1Curated pool of aging-associated genes (genes identified as medium novel targets are marked with asterisks, and highly novel targets with arrow heads). 2These genes were associated with aging or skin aging with reference to a publicity database. 3Druggability scores defined in PandaOmics (small molecules, antibodies, safety, novelty). 4Druggable gene classes defined in PandaOmics. 5Genes identified as top-100 high confidence targets were marked with ticks. Supplementary Table 5. Overlapping of high confidence targets with the pool of curated aging-associated genes from geroprotectors. Genes1 Geroprotectors2 Druggability3 Target family4 Top-100 target5 CASP1 Aspirin 2,0,2,0 Peptidase ✓ CASP3 Aspirin 2,0,2,0 Peptidase ✓ CHUK N-acetyl-L-cysteine 2,0,2,0 Protein kinase ✓ ESR1 17-A-Estradiol; Melatonin 2,2,2,0 Nuclear receptor ✓ HSPA5 Aspirin 2,2,2,0 Hydrolase ✓ IKBKB Aspirin; N-acetyl-L-cysteine 2,0,2,0 Protein kinase ✓ MTOR Rapamycin 2,0,2,0 Protein kinase ✓ ACE Enalapril 2,2,2,0 Glycosylase ACY1 N-acetyl-L-cysteine 2,0,2,0 Hydrolase ADRB1 Metoprolol; Nebivolol 2,2,2,0 GPCR ADRB2 Metoprolol; Nebivolol 2,0,2,0 GPCR ADRB3 Nebivolol 2,2,2,0 GPCR AKR1C1 Aspirin 1,0,1,0 Oxidoreductase ALOX5 Nordihydroguaiaretic Acid 2,0,2,0 Oxidoreductase AMY2A Acarbose 2,0,2,0 Glycosylase ASMT Melatonin 1,0,1,0 Methyltransferase CHRNA4 17-A-Estradiol 2,0,2,0 Ion channel CKB Creatine 2,0,1,0 Unclassified kinase CKM Creatine 2,0,1,0 Non-protein kinase EDNRA Aspirin 2,2,2,0 GPCR EPX Melatonin 1,0,1,0 Oxidoreductase ESR2 17-A-Estradiol 2,0,2,0 Nuclear receptor ETFDH Metformin 2,0,1,0 Oxidoreductase GAA Acarbose 2,0,2,0 Glycosylase GAMT Creatine 2,0,1,0 Methyltransferase GPD1 Metformin 1,0,1,0 Oxidoreductase GPER1 17-A-Estradiol 1,0,1,0 GPCR GRIN1 N-acetyl-L-cysteine 2,0,1,0 Ion channel GRIN2A N-acetyl-L-cysteine 2,0,1,0 Ion channel GRIN2B N-acetyl-L-cysteine 2,0,2,0 Ion channel Overlapping of high confidence targets with the pool of curated aging-associated genes Supplementary Table 5. Overlapping of high confidence targets with the pool of curated aging-associated genes from geroprotectors. tary Table 5. Overlapping of high confidence targets with the pool of curated aging-ass Supplementary Table 5. Overlapping of high confidence targets with the pool of curate from geroprotectors. Supplementary Tables AGING 2505 GRIN2D N-acetyl-L-cysteine 2,0,1,0 Ion channel GRIN3A N-acetyl-L-cysteine 2,0,1,0 Ion channel GSS N-acetyl-L-cysteine 2,0,1,0 Ligase IFNG D-Glucosamine 2,2,2,0 Interferon MAOA Deprenyl or Selegiline 2,0,2,0 Oxidoreductase MAOB Deprenyl or Selegiline 2,0,2,0 Oxidoreductase MGAM Acarbose 2,0,2,0 Glycosylase MPO Melatonin 2,0,2,0 Oxidoreductase MTNR1A Melatonin 2,0,2,0 GPCR MTNR1B Melatonin 2,0,2,0 GPCR NEU1 Aspirin 1,0,1,0 Glycosylase NPY2R Cysteamine 2,0,2,0 GPCR NQO2 Melatonin 2,0,2,0 Oxidoreductase NR1I2 17-A-Estradiol 2,0,2,0 Nuclear receptor PTGS1 Aspirin 2,0,2,0 Oxidoreductase PTGS2 Aspirin 2,2,2,0 Oxidoreductase RPS6KA3 Aspirin 2,0,2,0 AGC kinase SI Acarbose 2,0,1,0 Glycosylase CKMT1A Creatine 2,0,1,1 Unclassified kinase CKMT2 Creatine 2,0,1,1 Unclassified kinase RORB Melatonin 2,0,2,1 Nuclear receptor PRKAB1^ Metformin 2,0,1,2 Protein kinase 1Curated pool of geroprotector-associated genes (genes identified as medium novel targets are marked with asterisks, and highly novel targets with arrow heads). 2These geroprotectors were (1) approved drugs for human use and (2) investigated for antiaging effects using human or animal models (with reference to http://geroprotectors.org). Target-drug association were manually curated. 3Druggability scores defined in PandaOmics (small molecules, antibodies, safety, novelty). 4Druggable gene classes defined in PandaOmics. 5Genes identified as top-100 high confidence targets were marked with ticks. Supplementary Table 6. Pathway enrichment analysis based on 145 targets associated with the hallmarks of aging. pplementary Table 6. Pathway enrichment analysis based on 145 targets associated with t Supplementary Table 7. List of AAD and NAAD datasets analyzed. Supplementary Table 8. The identification of target-target interactions. Targets interactions1 Reference (PMID) ROCK1-c-Myc 30613282 FOXO-SIRT3 27686535 mTOR-STAT3 26697523 mTOR-TFEB 30120233 mTOR-PPARG 27901044 MAPK14-c-Myc 10623602 MAPK8-c-Myc 10623602 MAPK-CREB 30214393 MAPK-NRF2 31221142 KDM7A-Catenin beta-1 30614617 KDM7A-Catenin beta-1 32214833 1Targets interactions were identified outside the context of pathways with KEGG pathway database. Reference (PMID) AGING 2506 2506
https://openalex.org/W2011039105	https://europepmc.org/articles/pmc1971437?pdf=render	English	null	Immunohistochemical detection of major histocompatibility complex antigens and quantitative analysis of tumour-infiltrating mononuclear cells in renal cell cancer	British journal of cancer	1,990	cc-by	5,838	Immunohistochemical detection of major histocompatibility complex antigens and quantitative analysis of tumour-infiltrating mononuclear cells in renal cell cancer (1987) found only two cases of RCC positive for class I antigen and one positive for class II antigen among 10 cases. In the present study, we examined the expression of HLA-A, B, C, P2m and HLA-DR, DQ, DP, and also the population of TIM using immunoperoxidase staining in a larger number of cases of RCC. Renal cell carcinoma (RCC), which accounts for about 90% of tumours originating from the renal parenchyma, has char- acteristics unique among malignant tumours. First, more than 50 cases of spontaneous regression have been reported (Freed et al., 1977), an incidence that is remarkably higher than that seen in other malignant tumours. Second, RCC shows a relatively high response to adoptive immunotherapy with lymphokine-activated killer (LAK) cells (Rosenberg et al., 1987) and also to certain cytokines such as interferon alpha (Krown et al., 1987). It is also known that tumour- infiltrating lymphocytes (TIL) obtained from RCC tissue are able to lyse autologous tumour cells after culture with IL-2 (Belldegrum et al., 1988). In the light of these findings it seems that the immune system influences the behaviour of RCC cells in vivo. Major histocompatibility complex (MHC) class I antigens, composed of highly polymorphic glycoproteins associated with beta-2 microglobulin (p2m), are expressed on virtually all nucleated cells (Daar et al., 1984; Natali et al., 1984). MHC class I molecules has been shown to act as restriction elements for the lysis of target cells by cytotoxic T lympho- cytes (CTL) (Zinkernagel & Doherty, 1979). In a murine system, it was demonstrated that loss or reduction of class I molecules on tumour cells decreased their susceptibility to lysis by CTL (Bernards et al., 1983). Some reports have also described remarkable reduction of class I antigens in poorly differentiated tumours (Momburg et al., 1986; Moller et al., 1987), a highly malignant type of human tumour (van den Ingh et al., 1987) and also in tumour cell lines (Doyle et al., 1985). MHC class II molecules are known to function as restriction molecules for the provision antigen fragments to helper T cells by antigen-presenting cells (Benacerraf et al., 1981) and also to be responsible for allograft rejection. Immunohistochemical detection of major histocompatibility complex antigens and quantitative analysis of tumour-infiltrating mononuclear cells in renal cell cancer Y. Tomital"2, T. Nishiyama2, M. Fujiwara' & S. Sato2 t of 'Immunology and 2Urology, Niigata University, School of Medicine, Asahimachi 1, Niigata 951, Japan. Department of 'Immunology and 2Urology, Niigata University, School of Medicine, Asahimachi 1, Niig Summary In order to investigate the anti-tumour immune responsiveness of patients with renal cell cancer (RCC), we examined 30 such patients for the degree of expression of major histocompatibility complex (MHC) class I and class II antigens on RCC and the populations of tumour-infiltrating mononuclear cells (TIM). Normal renal tubular cells expressed class I but not class II antigens. Most of the tumour cells expressed class I antigens in 25 (83%) cases, but the proportion of such cells was reduced in five cases, three of which were of granular cell type histologically. Class II antigens were detected in all specimens with class I positivity. Various numbers of TIM were detected in 25 cases, being composed mainly of T cells and a smaller number of macrophages. Examination for the phenotype of T cells showed that CD8-positive cells were the dominant population. B cells were not detected. Quantitative analysis revealed that the numbers of TIM were signi- ficantly lower in cases showing class I reduction than in those with normal class I expression. Therefore, it was clear that class I antigens were preserved in RCC cells in most cases. Furthermore, a higher rate of reduction of class I antigens was observed in cases of granular cell type, which has been reported to have a worse prognosis than the clear cell type. The present data suggest that degree of the expression of MHC class I antigen on RCC might influence the host immune responsiveness against it. reduction of these antigens and tumour malignancy (Mom- burg et al., 1987) whereas others have shown that an increase of the antigens is related to tumour progression (Broker et al., 1985). Thus, MHC antigens are considered to play an important role in allowing RCC to escape the host's immune reaction, so that it seems worthwhile to investigate MHC class I and II expression and the characterisation of tumour- infiltrating mononuclear cells (TIM) in RCC. In relation to this aspect, Natali et al. (1984) detected class I antigens in nine of ten cases of RCC. On the other hand, Heinemann et al. Immunohistochemical detection of major histocompatibility complex antigens and quantitative analysis of tumour-infiltrating mononuclear cells in renal cell cancer Some kinds of class II-positive tumour cell line are reported to stimulate proliferation of alloreactive T cells in the mixed lymphocyte reaction (Fossate et al., 1984; Sakai et al., 1987), and to induce cytotoxic T cells against class II antigens (Pfizenmaier et al., 1985). However, in situ studies on class II antigens of human tumour cells have produced rather confus- ing results. Some studies have indicated a correlation between Tissue specimens Specimens were obtained from 30 patients (18 males and 12 females) who had undergone nephrectomy for RCC between October 1987 and July 1989. None of the patients had received chemotherapeutic or immunomodulatory agents, or irradiation preoperatively. Furthermore, there was no evi- dence of urinary tract infection in these patients. The mean age at the time of surgery was 59.5 years, with range of 34-78 years. Normal kidney tissues were collected from the unaffected portion of the removed kidney. All the tissue samples were embedded in OCT compound (Miles Labora- tories, Naperville, IL, USA) after being rinsed in PBS, and then snap-frozen in isopentane precooled in dry ice-acetone. These blocks were stored at - 80°C until 5 gm frozen sections were cut in a cryostat. C Macmillan Press Ltd., 1990 1^ Macmillan Press Ltd., 1990 C Macmillan Press Ltd., 1990 1^ Macmillan Press Ltd., 1990 Br. .1. Cancer (1990), 62, 354-359 Br. J. Cancer (1990), 62, 354-359 Correspondence: Y. Tomita, Department of Immunology, Niigata University, School of Medicine, Asahimachi 1, Niigata 951, Japan. Received 3 January 1990; and in revised form 13 March 1990. Immunoperoxidase staining Immunoperoxidase staining was performed using the strep- tavidin-biotin bridge technique (Bonnard et al., 1984). Serial sections prepared in a cryostat were air-dried for 30 min and fixed in cold acetone for 10 min. After rehydration with PBS, the sections were incubated in PBS containing 20% normal sheep serum (Antibodies Inc., Davis, CA, USA) for 30 min and endogenous biotin was blocked using an Endogenous Biotin Blocking Kit (Vector Laboratories, Burlingame, CA, USA). The sections were then incubated with mouse mono- clonal antibodies for 60 min followed by incubation with biotinylated sheep anti-mouse immunoglobulin (Amersham International, Amersham, Bucks, UK) diluted 1:100, contain- ing 20% human type AB serum (Biological Speciality Co., Lansdale, PA, USA). Subsequently, they were incubated with streptavidin peroxidase (Amersham) diluted 1:200 for 45 min. Each step was followed by washing in PBS with three changes of buffer. Table I Clinical features and histopathological diagnosis of RCC patients No. l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Sex F M M M M M M M M F M M F F M M M F M F F M F F M M F F F M Age 58 60 76 62 62 56 34 78 51 65 59 36 38 56 55 70 54 51 58 48 65 53 64 71 60 71 76 53 68 76 TNM classification T2N0M0 T2NoMo T2NoMo T2NoMo T2NOMO T2NoMo T2NoMo T2NoMo T2NoM, T2NOM0 T2NoMo T2NoMo T2NoMo T2NOMO T2NOM, T3NXMI T2NOM, T2NoMo T3NoMo T3NoMO T2NoMo T2NoMo T3NXMj T2NoMo T2NoMo T3NoMo T3NoMo T2NoMo T3NOMO T3NoMo Grade 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Histological typea mixed clear granular clear clear clear clear mixed clear ciear clear granular clear clear clear clear granular clear mixed granular granular clear mixed granular clear clear clear clear mixed clear Status" NED (22) NED (21) NED (19) NED (19) NED (18) NED (18) NED (17) NED (5) Alive (14) NED (13) NED (13) NED (13) NED (12) NED (11) NED (10) Died Alive (9) NED (9) NED (8) NED (5) NED (5) NED (4) Died NED (3) NED (2) NED (2) NED (2) NED (2) NED (I) NED (1) aClear, clear cell type; granular, granular cell type; mixed, mixed cell type. Immunoperoxidase staining bNED, no evidence of disease; alive, alive with disease; died, died by other cause. Figures in parentheses were following up months after operation. Table I Clinical features and histopathological diagnosis of RCC patients g Finally the sections were immersed in 0.05% diaminoben- zidine (Sigma Chemical Co., St Louis, MO, USA) and 0.01% HO, in 0.05 M Tris HCI buffer for 3-5 min to visualise the reaction products. After washing in tap-water, some speci- mens were counterstained with Mayer's haematoxylin and mounted with Eukitt (0. Kuldler, Freiburg, FRG) after de- hydration in a graded ethanol series and xylene. As negative controls for MHC antigen staining, serial sections of tumour tissue were stained with the same subclass of monoclonal antibodies against a variety of immune cells as described previously. As positive controls for class I antigens, the staining patterns of endothelial cells, fibroblasts and macro- phages were checked, and for those of class II antigens endothelial cells and dendritic cells were examined. Histological examinations Histological examination was performed on haematoxylin and eosin (H&E)-stained tissue sections. Cancer tissues were histologically graded according to the General Rule for Clinical and Pathological Studies on Renal Cell Carcinoma (Japanese Urological Association, The Japanese Pathological Society and Japan Radiological Society, 1983). The grading MHC ANTIGENS IN RENAL CELL CANCER 355 scored according to the sum of anti-Leul- and anti-LeuM3- positive cells as follows: score 0, none; score 1, occasional (less than 10 cells per field); score 2, mild (10-49); score 3, moderate (50-99); score 4, high (more than 100). system includes low-grade (Gl), moderate-grade (G2) and high-grade (G3) categories according to the degree of atypia of the tumour cells. Histological stage was determined according to the TNM classification of malignant tumours (UICC, 1987). Histopathological classification and clinicalfeatures of RCC Histopathological classification and clinical features were examined in each case of RCC and the results are shown in Table I. Tumour cells were estimated for histological type and classified into clear cell, granular cell and mixed cell types. On the basis of the TNM classification, group T2 included 22 cases and group T3a eight cases. The pN cate- gory consisted of 27 pN0 cases and three pNX cases. There were five cases with distant metastasis. Pathological examina- tion of metastatic sites was performed in only two cases, and three cases had lung lesions which were strongly suspected to be metastases of RCC on clinical grounds. grounds. Although the follow-up periods after surgery were too short for evaluation of prognosis, the patients have been followed from 1 to 22 months, and 26 patients are currently alive with no evidence of the disease, two are alive with disease and two have died of other causes. MHC antigens expression on normal kidney tissue Before the examination of RCC tissue, we examined the staining patterns of normal kidney tissue for MHC class I and II antigens. Class I antigens were expressed on virtually all cells comprising the renal tubules. The collecting ducts were stained rather more weakly than glomerular cells and endothelial cells. Class II antigens were expressed on glome- rular cells but could not be detected on renal tubular cells in this study. Monoclonal antibodies Monoclonal antibodies used in this study were as follows: W6/32 (IgG2,) against a monomorphic determinant on the heavy chain of MHC class I antigens associated with ,2m (Dako Japan Co., Kyoto, Japan) and SRL-2 (IgG,) against P2m (Serotec Co., Blackthorn, Bicester, Bucks., UK). For the detection of class II antigens, L243 (IgG2,) against a mono- morphic determinant of HLA-DR, SK10 (IgG1) against a common polymorphic determinant of HLA-DQ, and B7/21 (IgGj) against a monomorphic determinant of HLA-DP were used. Anti-Leul (IgGu) against pan-T cell (CD5), anti-Leu2a (IgG,) against cytotoxic/suppressor T cells (CD8), anti-Leu3a (IgG,) against helper/inducer T cells (CD4), anti-Leul2 (IgG,) against B cells and anti-LeuM3 (IgG2b) against macro- phages were used for evaluating the TIM phenotypes. These monoclonal antibodies against class II antigens and immune cells were purchased from Becton Dickinson, Mountain View, CA, USA. The optimal dilution of each antibody was determined by staining two specimens of adenoid vegetation resected surgically and three specimens of lymph nodes obtained at nephrectomy for RCC. Histopathological classification and clinicalfeatures of RCC Histopathological classification and clinicalfeatures of RCC Infiltrating mononuclear cells in RCC Various numbers of mononuclear cells had infiltrated the tumours (Figure 3a-d). These cells were often scattered with- in the tumours but some remarkable perivascular infiltration was also seen. The degree of infiltration was evaluated quan- titatively as described in Materials and methods. These infiltrating cells were composed of T cells and a smaller number of macrophages, but B cells were not detected. Upon phenotyping of the infiltrating T cells, CD8-positive (killer/ suppressor) T cells were predominant in 20 of 23 cases with a score of more than 2. CD4-positive (helper/inducer) T cells were predominant in only three cases. Macrophages were detected in 17 cases (Figure 3e) but were outnumbered by T cells. These results are summarised in Table II. were predominant in only three cases. Macrophages were detected in 17 cases (Figure 3e) but were outnumbered by T cells. These results are summarised in Table II. Correlations between MHC antigen expression and the degree of cellular infiltration were examined for each MHC antigen. Significant correlation was noticed between class I antigen reduction and a decrease in the number of infiltrating cells (P<0.01, x2 = 10.77, Table IV). Correlations between MHC antigen expression and the degree of cellular infiltration were examined for each MHC antigen. Significant correlation was noticed between class I antigen reduction and a decrease in the number of infiltrating cells (P<0.01, x2 = 10.77, Table IV). Table II Immunohistochemical staining for MHC antigens and tumour infiltrating mononuclear cells Approximate % ofpositive cells No. MHC antigen expression of RCC As can be seen in Table II, most of the tumour cells were positive for class I antigens and P2m, showing greater inten- sity of staining than the renal tubular cells (Figure 1). In five cases, reduced expression (less than 30% of cells positive) of class I antigens was observed (Figure 2). Cases showing reduced class I antigen expression accounted for 50% cases in the granular cell group but only 11% of those in the clear cell group; this difference was significant (X2 = 4.441, Table III). ) Class II antigen expression was more variable than class I antigen, but surprisingly, class II antigens which could not be detected on normal renal tubular cells were detected in 28 cases tested for DR antigen staining, 24 for DP and 23 for DQ. However, the number of class II antigen-positive cells was lower than that of class I antigen-positive cells except in one case. Two cases without DR staining were also negative for DP or DQ. As to the correlation between DP and DQ expression, an equal number or more DP-positive cells exist- ed in 25 cases, and more DQ-positive cells existed in the other five cases. No relationship was found between class II antigen expression and the clinical or histopathological features of RCC. In summary, MHC antigens were expressed on RCC with a hierarchy of positivity of class I antigens, HLA, DR, DP and DQ. Figure 1 a, H&E stained clear cell type of RCC (patient 11). b, Stained with W6/32. Most of the tumour cells are positive. Bar = 75 pM. Evaluation ofstaining After reaction with either anti-class I or II antibodies, the tumour tissue showed various staining patterns. The degree of positive staining of tumour cells, which were distinguish- able from non-tumour cells, was expressed as the approxi- mate percentage of positive cells. For quantitative analysis of TIM, five fields were selected randomly and TIM were counted in serial sections at a magnification of x 100 using a microscope equipped with a graticule (0.25 mm square, Olympus, Tokyo). For comparison between the amount of TIM and clinical and histopathological factors, TIM were aClear, clear cell type; granular, granular cell type; mixed, mixed cell type. bNED, no evidence of disease; alive, alive with disease; died, died by other cause. Figures in parentheses were following up months after operation. 356 Y. TOMITA et al. 356 Y. TOMITA et al. 356 Y. TOMITA et al. Figure 1 a, H&E stained clear cell type of RCC (patient 11). b, Stained with W6/32. Most of the tumour cells are positive. Bar = 75 pM. MHC antigen expression of RCC Infiltrating mononuclear cells in RCC HLA-A,B,C 1 100 2 100 3 100 4 100 5 100 6 30 7 20 8 100 9 100 10 100 11 100 12 5 13 100 14 100 15 100 16 100 17 80 18 100 19 100 20 20 21 90 22 100 23 90 24 25 25 100 26 100 27 100 28 100 29 100 30 100 B2M 100 100 100 100 100 20 20 100 100 100 100 5 100 100 100 100 80 100 100 20 90 100 90 30 100 100 100 100 100 100 DR 50 80 80 100 50 5 10 10 90 80 90 0 90 80 90 90 80 90 60 90 30 90 90 30 100 80 100 100 100 70 DQ 10 60 30 90 10 0 0 10 5 30 10 0 30 30 80 50 10 40 10 5 5 50 70 5 50 30 10 20 30 30 DP 20 80 30 70 10 0 0 10 60 S 20 0 70 50 90 40 20 10 10 5 5 80 90 10 90 30 90 40 60 30 Number ofpositive cells Leul 2a 3a 12 M3 351 271 152 0 159 248 153 150 0 39 832 528 432 0 0 631 486 172 0 0 36 25 22 0 0 0 0 0 0 0 233 157 143 0 0 332 227 96 0 30 847 570 246 0 60 153 135 86 0 42 567 327 132 0 91 0 0 0 0 0 297 137 176 0 105 872 550 196 0 0 355 247 40 0 80 495 270 259 0 0 238 75 101 0 0 28 17 0 0 93 49 43 0 0 42 0 0 0 0 0 0 0 0 0 0 817 243 574 0 164 526 585 96 0 70 0 0 0 0 0 132 106 25 0 83 136 97 29 0 16 238 147 65 0 0 105 77 42 0 55 46 16 10 0 84 0 0 0 0 0 Ratio of Leu3a/2a 0.56 0.98 0.82 0.35 0.91 0.42 0.43 0.64 0.40 1.28 0.36 0.16 0.96 1.35 2.36 0.16 0.24 0.30 0.44 0.55 0.63 TIM score" 4 3 4 4 1 0 2 3 4 2 4 0 4 4 3 3 2 2 2 1 0 4 4 0 2 2 2 2 2 0 'These cases divided into five categories according to the number ofTIM, as described in Materials and methods. Infiltrating mononuclear cells in RCC Table II Immunohistochemical staining for MHC antigens and tumour infiltrating mononuclear cells A i t % f iti ll b f i i ll MHC ANTIGENS IN RENAL CELL CANCER 357 Figure 2 a, H&E stained granular cell type of RCC (patient 12). b, Stained with W6/32. Tumour cells showed severe reduction of class I antigens. Note the intense staining of endothelial cells of vessels in the tumour. Bar = 75 gxM. related to tumour growth in vivo (Tanaka et al., 1985) and susceptibility to lysis by CTL (Bernards et al., 1983). In our studies on RCC, class I antigens on the tumour cells seemed to be preserved to a greater extent than in other types of cancer, which might be advantageous for host's immune system since CTL lyse tumour cells in a class I-restricted manner. This might explain the higher rate of spontaneous regression in cases of RCC. On the other hand, susceptibility of tumour cells to natural killer cells, which are considered to be the main effector cells preventing tumour metastasis (Waner et al., 1982), is decreased in proportion to increased expression of class I antigens on tumour cells (Piontek et al., 1985). This fact seems to contradict the susceptibility of RCC to CTL lysis, although it might contribute to the propor- tionally greater percentage of metastasis of RCC among all other carcinomas (Mostofi & Davis, 1984). ( , ) Previous reports showed that reduced expression of class I antigens was inversely correlated with the degree of differ- entiation in some types of tumour (Momburg et al., 1986; Moller et al., 1987). Although no relationships with grade and TNM classification, or with the age and sex of the patient were found, a lower degree of expression was observed in the granular cell type than in the clear or the mixed cell type. This result is intriguing because the granular cell type has often been reported to have a worse prognosis than the clear cell type (Murphy & Mostofi, 1965). yp ( p y , ) Class II antigens, which could not be detected on normal renal tubular cells in this study, were variably expressed in all tumour specimens that expressed class I antigens simultan- eously. In other types of tumour the correlation between class II antigen expression and malignancy were diverse. Table III Correlation between class I antigens expression and histological cell type Class I antigens expression on tumour cells Cell type Normally expresseda Reducedb Clear cell type 17 (89) 2 (11) Granular cell type 3 (50) 3 (50)c Mixed cell type 5 (100) 0 (0) Figures in parentheses are % in each group. aMore than 90% of positive cells. bLess then 30% of positive cells. CP< 0.05 compared with clear cell group (X2 = 4.441). Recently, TIL were demonstrated to have stronger ability to lyse tumour cells than lymphokine-activated killer cells (Rosenberg et al., 1986) and they have been used for adop- tive immunotherapy (Kradin et al., 1989). TIL of RCC have been examined to ascertain their effect, and they were reported to have a potential to lyse autologous tumour cells after culture with interleukin-2 (Belldegrum et al., 1988). It was also reported that TIM, which were composed of T cells and macrophages, frequently infiltrated into RCC (Heinemann et al., 1987). In the present study, lymphocytes and a smaller number of macrophages were shown to infiltrate into RCC tissue in various patterns, and it was also demonstrated that lymphocytes consisting of T cells and CD8-positive cells were the dominant population in 20 out of 23 cases. In the remaining seven cases, we were unable to detect more than 10 TIM per field. Interestingly, in four of these seven cases with a smaller number of TIM, the tumours showed reduced expression of class I antigens and the numbers of TIM were significantly lower in all cases with class I reduction than in those showing normal class I expres- sion. These results suggest that the expression of MHC class I antigen on RCC might influence lymphocyte infiltration into the tumour. Table IV Correlation between class I antigen expression and the degree of mononuclear cell infiltration Table IV Correlation between class I antigen expression and the degree of mononuclear cell infiltration Class I antigens expression on tumour cells TIM score Normally expresseda Reducedb 2-4 22 1 P<0.01, x2 = 10.77 Oorl 3 4 aMore than 90% of positive cells. bLess than 30% of positive cells. Table IV Correlation between class I antigen expression and the degree of mononuclear cell infiltration Table IV Correlation between class I antigen expression and the degree of mononuclear cell infiltration Infiltrating mononuclear cells in RCC B cell lymphoma shows reduced class II antigen expression in accordance with dedifferentiation (Momburg et al., 1987) whereas class II antigen expression increases with disease progression in malignant melanoma (Brocker et al., 1985). In the present study, no correlation was found between HLA- DR, DQ, DP expressions and clinical and histopathological features, and it was concluded that MHC antigens were expressed with the hierarchy: class I antigens, HLA-DR, DP, DQ. This hierarchy was also reported in a study of gastric carcinoma (Sakai et al., 1987), although its significance was not clear. Figure 2 a, H&E stained granular cell type of RCC (patient 12). b, Stained with W6/32. Tumour cells showed severe reduction of class I antigens. Note the intense staining of endothelial cells of vessels in the tumour. Bar = 75 gxM. Table III Correlation between class I antigens expression and histological cell type Discussion Figure 3 Immunoperoxidase staining for TIM in RCC. a, H&E stained clear cell type of RCC (patient 16). b, Stained with anti-Leul1 (CD5). Positive cells infiltrated perivascular area and are scattered within the tumour mass. Stained with anti-Leu2a (CD8) c, and anti-Leu3a (CD4) d, on serial section. e, Stained with anti-LeuM3 for patient 13. Macrophages were scattered within the tumour mass. Bar =75 iM. The authors thank Dr T Tanikawa (First Department of Pathology, Niigata University School of Medicine) for his advice on tumour pathology, Drs Y. Matsumoto, Y. Ikarashi (Department of Immunology, Niigata University School of Medicine), Drs Y. Sakata, S. Komatsubara, Y. Kitamura, M. Watanabe (Niigata Cancer Center, Niigata), Dr M. Hiraiwa (Koseiren Sanjo General Hospital,, Sanjo), Dr T. Ando (Tsubame Rosai Hospital, Tsubame) and Dr T. Watanabe (Sado General Hospital, Sado) for their assis- tance and useful advice. class I antigen expression in the granular cell type, which is reported to have a worse prognosis than the clear cell type (Murphy et al., 1965), is intriguing. TIM were significantly fewer in cases showing class I reduction than in those with normal class I expression. Since our studies on TIM sub- populations showed that CD8-positive T cells predominantly infiltrated in most cases, the degree of expression of MHC class I antigen on cancer cells is considered to influence the host immune responsiveness against RCC. class I antigen expression in the granular cell type, which is reported to have a worse prognosis than the clear cell type (Murphy et al., 1965), is intriguing. TIM were significantly fewer in cases showing class I reduction than in those with normal class I expression. Since our studies on TIM sub- populations showed that CD8-positive T cells predominantly infiltrated in most cases, the degree of expression of MHC class I antigen on cancer cells is considered to influence the host immune responsiveness against RCC. Discussion In the present study, we analysed MHC antigen expression of RCC cells in 30 cases. Class I antigens were detected on most of the tumour cells in 25 of the 30 cases. Our results were very similar to those of Natali et al. (1984), who reported that nine out of ten cases of RCC expressed class I antigens. In contrast, Heinemann et al. (1987) reported the detection of MHC class I antigen in only two out of ten cases of RCC. Although it is not possible to reconcile these different results, the reason may have been differences in the monoclonal antibodies or staining procedures used. In this study, class I antigens were found on most of the RCC cells in 25 out of 30 cases, and the intensity of expres- sion was comparable with that seen in renal tubular cells. These results indicate that class I antigen expression is more preserved in RCC cells compared with other types of cancer. This would seem advantageous for the host's immune system against the tumour cells, since CTL are known to lyse class I-positive tumour cells, and this might be related to the higher rate of spontaneous regression in cases of RCC (Freed et al., 1977). Furthermore, the greater degree of reduction of It has been shown that CTL need to recognise MHC class I molecules in order to lyse target cells (Zinkernagel & Doherty, 1979). In this connection, it is interesting that the degree of expression of MHC class I antigens is closely Y. TOMITA et al. 358 Y. TOMITA et al. ,,\|,.fe,~' ... V r~~ 4 A~~~~ 11\|1EiS1i' S51\|2. ...": *1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~lv 4,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~N. Figure 3 Immunoperoxidase staining for TIM in RCC. a, H&E stained clear cell type of RCC (patient 16). b, Stained with anti-Leul1 (CD5). Positive cells infiltrated perivascular area and are scattered within the tumour mass. Stained with anti-Leu2a (CD8) c, and anti-Leu3a (CD4) d, on serial section. e, Stained with anti-LeuM3 for patient 13. Macrophages were scattered within the tumour mass. Bar =75 iM. Figure 3 Immunoperoxidase staining for TIM in RCC. a, H&E stained clear cell type of RCC (patient 16). b, Stained with anti-Leul1 (CD5). Positive cells infiltrated perivascular area and are scattered within the tumour mass. Stained with anti-Leu2a (CD8) c, and anti-Leu3a (CD4) d, on serial section. e, Stained with anti-LeuM3 for patient 13. Macrophages were scattered within the tumour mass. Bar =75 iM. Figure 3 Immunoperoxidase staining for TIM in RCC. a, H&E stained clear cell type of RCC (patient 16). b, Stained with anti-Leul1 (CD5). Positive cells infiltrated perivascular area and are scattered within the tumour mass. Stained with anti-Leu2a (CD8) c, and anti-Leu3a (CD4) d, on serial section. e, Stained with anti-LeuM3 for patient 13. Macrophages were scattered within the tumour mass. Bar =75 iM. MHC ANTIGENS IN RENAL CELL CANCER 359 MHC ANTIGENS IN RENAL CELL CANCER HEINEMANN, D., SMITH, P.J.B. & SYMES, M.O. (1987). Expression of histocompatibility antigens and characterization of mononuclear cell infiltrates in human renal cell carcinomas. Br. J. Cancer, 56, 433. PIONTEK, G.E., TANIGUCHI, K., LIUNGGREN, H. & 4 others (1985). Yak-I MHC class I variants reveal an association between decreased NK sensitivity and increased H-2 expression after interferon treatment or in vivo passage. J. Immunol., 135, 4281. JAPANESE UROLOGICAL ASSOCIATION, THE JAPANESE PATHO- LOGICAL SOCIETY AND JAPAN RADIOLOGICAL SOCIETY (1983). General Rules for Clinical and Pathological Studies on Renal Cell Cancer. Kanehara: Tokyo. p g ROSENBERG, S.A., SPIESS, P. & LAFRENIERE, R.A. (1986). A new approach to the adoptive immunotherapy of cancer with tumor- infiltrating lymphocytes. Science, 23, 1318. g y p y , , ROSENBERG, S.A., LOTZE, M.T., MUUL, L.M. & 10 others (1987). A progress report on the treatment of 157 patients with advanced cancer using lymphokine-activated killer cells and interleukin-2 or high dose interleukin-2 alone. N. Engl. J. Med., 316, 889. y KRADIN, R.L., KURNICK, J.T., LAZARUS, D.S. & 8 others (1989). Tumor-infiltrating lymphocytes and interleukin-2 in treatment of advanced cancer. Lancet, i, 577. , , KROWN, S.E. (1987). Interferon treatment of renal cell carcinoma: current status and future prospects. Cancer, 59, 647. SAKAI, K., TAKIGUCHI, M., MORI, S. & 5 others (1987). Expression and function of class II antigens on gastric carcinoma cells and gastric epithelia: differential expression of DR, DQ and DP antigens. J. Natl Cancer Inst., 79, 923. p p , MOLLER, P., HERRMANN, B., MOLDENHAUER, G. & MOMBURG, F. (1987). Defective expression of MHC class I antigens is frequent in B-cell lymphomas of high-grade malignancy. Int. J. Cancer, 40, 32. g TANAKA, K., TSSELBACHER, K.J., KHOURY, G. & JAY, G. (1985). Reversal of oncogenesis by the expression of major histocom- patibility complex class I gene. Science, 228, 26. MOMBURG, F., DEGENER, T., BACCHUS, E., MOLDENHAUER, G., HAMMERLING, G.J. & MOLLER, P. (1986). Loss of HLA-A,B,C and de novo expression of HLA-D in colorectal cancer. Int. J. Cancer, 37, 179. p y p g , , UICC (1987). TNM Classification of Malignant Tumours, 4th edn. UICC: Geneva. VAN DEN INGH, H.F., REUTER, D.J. III, GRIFFIOEN, G., VAN MUIJEN, G.N.P. & FRRONE, S. (1987). HLA antigens in colorectal tumours - low expression of HLA class I antigens in mucinous colorectal carcinomas. Br. J. Cancer, 55, 125. References DAAR, A.S., FUGGLE, S.V., FABRE, J.W., TING, A. & MORRIS, P.J. (1984). The detailed distribution of HLA-A, B, C antigens in normal human organs. Transplantation, 38, 287. BELLDEGRUM, A., MUUL, L.M. & ROSENBERG, S.A. (1988). Interleukin 2 expanded tumor infiltrating lymphocytes in human renal cell cancer: isolation, characterization and antitumor activity. Cancer Res., 48, 206. g p DOYLE, A., MARTIN, W.J., FUNA, K. & 8 others (1985). Markedly decreased expression of class I histocompatibility antigens, pro- tein and mRNA in human small cell lung cancer. J. Exp. Med., 161, 1135. y BENACERRAF, B. (1988). Role of MHC gene products in immune regulation. Science, 212, 1229. g BERNARDS, R., SCHRIER, P.I., HOUWELLING, A. & 4 others (1983). Tumorigenecity of cells transformed by adenovirus type 12 by evasion of T-cell immunity. Nature, 305, 776. FOSSATI, G., TARAMELLI, D., BALSARI, A., BOGDANOVICH, G., ANDREOLA, S. & PARMIANI, G. (1984). Primary but not meta- static human melanomas expressing DR antigens stimulate auto- logous lymphocytes. Int. J. Cancer, 33, 591. BONNARD, C., DEPERMASTER, D.S. & KRAEHENBUHL, J.-P. (1984). The streptavidin-biotin bridge technique: application in light and electron microscope immunocytochemistry. In Immunolabelling for Electron Microscopy, Polak, J.M. & Varndell, I.M. (eds) p. 95. Elsevier: Amsterdam. g y p y FREED, S.Z., HALPERIN, J.P. & GORDON, M. (1977). Idiopathic regression of Metastasis from renal cell carcinoma. J. Urol., 113, 538. BROCKER, E.B., STUER, L., BRUGGEN, J., REUTER, D.J., MACHER, E. & SORG, C. (1985). Phenotypic dynamics of tumor progression in human malignant melanoma. Int. J. Cancer, 36, 29. MHC ANTIGENS IN RENAL CELL CANCER , , MOMBURG, F., HERRMANN, B., MOLDENHAUER, G. & MOLLER, P. (1987). B-cell lymphomas of high grade malignancy frequently lack HLA-DR, -DP and -DQ antigens and associated invariant chain. Int. J. Cancer, 40, 598. , , WANER, J.F. & DENNERT, G. (1982). Effects of a cloned cell line with NK activity on bone marrow transplants, tumour develop- ment and metastasis in vivo. Nature, 300, 31. , MOSTOFI, F.K. & DAVIS, C.J. JR (1984). Pathology of tumors of the kidney. In Cancer of the Kidney, Javadpour, N. (ed.), p. 15. Thieme-Stratton: New York. ZINKERNAGEL, R.M. & DOHERTY, P.C. (1979). MHC-restricted cytotoxic T cells: studies on the biological role of polymorphic major transplantation antigens determining T cell restriction specificity, function and responsiveness. Adv. Immunol., 27, 51. MURPHY, G.P. & MOSTOFI, F.K. (1965). The significance of cytoplas- mic granularity in the prognosis of renal cell carcinoma. J. Urol., 94, 48. NATALI, P.G., BIOGOTTI, A., NICOTRA, M.R., VIORA, M., MAN- FREDI, D. & FERRONE, S. (1984). Distribution of human class I (HLA-A,B,C) histocompatibility antigens in normal and malig- nant tissues of nonlymphoid origin. Cancer Res., 44, 4679. y p g , , PFIZENMAIER, K., BARTSCH, H., SCHEURICH, P. & 4 others (1985). Differential gamma-interferon response of human colon carcin- oma cells: inhibition of proliferation and modulation of immuno- genicity as independent effects of gamma-interferon on tumour cell growth. Cancer Res., 45, 3503.
https://openalex.org/W2799369791	https://locus.ufv.br//bitstream/123456789/19830/1/artigo.pdf	Portuguese	null	Indicadores de insegurança alimentar e nutricional associados à anemia ferropriva em crianças brasileiras: uma revisão sistemática	Ciência & Saúde Coletiva	2,018	cc-by	5,069	1 Departamento de Nutrição e Saúde, Centro de Ciências Biológicas e da Saúde, Universidade Federal de Viçosa. Av. Peter Henry Rolfs s/n, Campus Universitário. 36570- 000 Viçosa MG Brasil. herciliopaulino@gmail.com 2 Universidade Federal do Rio de Janeiro. Macaé RJ Brasil. 3 Departamento de Nutrição, Universidade Federal de Sergipe. São Cristóvão SE Brasil. Indicadores de insegurança alimentar e nutricional associados à anemia ferropriva em crianças brasileiras: uma revisão sistemática Indicadores de insegurança alimentar e nutricional associados à anemia ferropriva em crianças brasileiras: uma revisão sistemática REVISÃO REVIEW REVISÃO REVIEW Food and nutrition insecurity indicators associated with iron deficiency anemia in Brazilian children: a systematic review Hercilio Paulino André 1 Naiara Sperandio 2 Renata Lopes de Siqueira 3 Sylvia do Carmo Castro Franceschini 1 Silvia Eloiza Priore 1 Abstract This study aimed to review food and nutrition insecurity indicators associated with iron deficiency anemia in Brazilian children be low 5 years. We searched in electronic databases (SciELO, Lilacs, and Medline) and selected stud ies by titles, abstracts and full-text reading. Of the 1,023 studies analyzed, 11 fit the inclusion criteria. The results of the studies evidenced that iron deficiency anemia in Brazilian children was associated with sociodemographic and health in dicators (male, age below 24 months, children of adolescent mothers, respiratory infections, diar rhea, low maternal schooling, parents’ working conditions, nursery time, lack of basic sanitation, maternal anemia, lack of ferrous sulfate use by the mother and/or child and late onset of prena tal care), nutritional indicators (low birth weight, diet characteristics, such as the habit of milk con sumption close to meals, low exclusive and full breastfeeding time) and economic indicators (low per capita income). The food and nutrition inse curity analyzed in this study from the perspective of different indicators is associated with iron defi ciency anemia in children under 5 years in Brazil. Key words Iron deficiency anemia, Iron deficien cy, Food and nutrition security, Children Resumo Objetivou-se revisar os indicadores de insegurança alimentar e nutricional associados à anemia ferropriva em crianças brasileiras menores de 5 anos. Realizou-se busca em bases eletrônicas (ScieLO, Lilacs, Medline), com seleção dos estu dos pelos títulos, resumos e pela leitura na íntegra. Dos 1023 trabalhos analisados, 11 contemplaram os critérios de inclusão. DOI: 10.1590/1413-81232018234.16012016 DOI: 10.1590/1413-81232018234.16012016 Indicadores de insegurança alimentar e nutricional associados à anemia ferropriva em crianças brasileiras: uma revisão sistemática Os resultados dos estudos retrataram que a anemia ferropriva em crianças brasileiras associou-se aos indicadores sociodemo gráficos e de saúde (sexo masculino, idade inferior aos 24 meses, filhos de mães adolescentes, infec ções respiratórias, diarreias, baixa escolaridade materna, condição de trabalho dos pais, tempo de creche, ausência de saneamento básico, anemia materna, não uso de sulfato ferroso pela mãe e/ou criança e início tardio do pré-natal), indicadores nutricionais (baixo peso ao nascer, características da dieta como, hábito de ingerir leite próximo dos horários das refeições, baixo tempo de aleitamen to materno exclusivo e total) e econômicos (bai xa renda per capita). A insegurança alimentar e nutricional, analisada no presente estudo sob a óptica de diferentes indicadores, está associada à ocorrência de anemia ferropriva em crianças me nores de 5 anos no Brasil. Palavras-chave Anemia ferropriva, Deficiên cia de ferro, Segurança alimentar e nutricional, Crianças Palavras-chave Anemia ferropriva, Deficiên cia de ferro, Segurança alimentar e nutricional, Crianças 1160 André HP et al. 116 André HP et al. 1160 Introdução nutricional são utilizados para caracterização de situações de violação desse direito, ou seja, da in segurança alimentar e nutricional8. A deficiência de micronutrientes é um impor tante problema de saúde pública, especialmente, em países em desenvolvimento. Segundo a Or ganização Mundial da Saúde (OMS) aproxima damente 2 bilhões de pessoas no mundo sofrem de fome oculta, que é a deficiência subclínica de micronutrientes, sendo os principais vitamina A, ferro, zinco e iodo1. Alguns indicadores como a baixa renda fa miliar per capita, baixa escolaridade, principal mente materna, maior número de filhos, elevada densidade de morador por cômodo, precárias condições de acesso a serviços públicos, como sa neamento básico e energia elétrica, consumo ali mentar inadequado, quanti e qualitativamente, dentre outros, caracterizam situações de insegu rança alimentar e nutricional que predispõem ao risco de desenvolvimento de doenças carenciais, dentre elas a anemia ferropriva9,10. A carência de ferro atinge todas as células do organismo humano e ocorre em três estágios, sendo o primeiro a depleção do estoque de fer ro, seguido pela eritropoiese ferro deficiente até a ocorrência da anemia ferropriva, caracterizada pela redução dos níveis de hemoglobina. A defi ciência de ferro e a anemia ferropriva resultam do desequilíbrio no balanço entre a quantidade de ferro biodisponível absorvido na dieta e a ne cessidade do mineral no organismo2. Apesar da anemia ferropriva ser um proble ma de saúde pública democraticamente distribu ído entre as diferentes classes socioeconômicas, situações que caracterizam um quadro de inse gurança alimentar e nutricional podem favorecer e contribuir para o surgimento dessa doença9. Diante do exposto, o objetivo deste artigo foi revisar os indicadores de insegurança alimentar e nutricional associados a anemia ferropriva em crianças brasileiras menores de 5 anos de idade. A anemia ferropriva é um distúrbio nutricio nal que compromete o sistema imunológico pre judicando o crescimento e desenvolvimento da criança3. O público infantil constitui um grupo vulnerável a deficiência de ferro devido a demanda aumentada desse mineral em função da intensa ve locidade de crescimento. Além disso, alguns fatores negativos da alimentação na infância podem au mentar essa vulnerabilidade, como por exemplo, consumo insuficiente de alimentos fontes de ferro (carne de boi, figado, frango, peixe e vegetais ver des escuros)2,3 e ingestão de leite de vaca e cabra antes dos primeiros seis meses de vida, que além dos baixos teores de ferro, podem ocasionar san gramento gastrointestinal e gerar perda de sangue nas fezes4. Resultados ram da análise dos títulos e resumos. Nos casos em que não houve concordância entre os dois avaliadores, um terceiro autor analisou os artigos. Os 11 estudos selecionados refletem a relação dos indicadores de insegurança alimentar e nutricio nal com a ocorrência da anemia ferropriva, sendo que em todos eles a anemia ferropriva associou-se a algum indicador sociodemográfico e de saúde; em quatro observou-se associaçao com indicado res econômicos, e em sete com nutricionais. Para a elaboração da revisão sistemática, primeiramente realizou-se a busca por palavras chaves nas bases de dados descritas anteriormen te, identificando 1023 estudos publicados no pe ríodo de interesse. A etapa seguinte constou da seleção e revisão dos estudos, avaliando primei ramente os títulos, na qual foram excluidos 972 trabalhos, sendo que 569 fugiram do tema desta revisão: 81 por serem estudos repetidos e 322 por ter sido conduzido em população não brasileira. Os indicadores sociodemográficos e de saúde, retratados pelos estudos, que apresentaram asso ciação (p < 0,05) com a ocorrência da anemia ferropriva foram: idade inferior a 24 meses12-16, idade materna inferior a 20 anos17,18, criança do sexo masculino17,19, número de moradores no domicílio2,15,20, baixa escolaridade materna13,15,20, área geográfica13, ausência de casa própria19, pre sença de infecções respiratórias e diarreias2,19, condições de trabalho dos pais21, tempo de cre che2, ausência de saneamento básico2,15, presença de anemia materna19 (Quadro 1). A partir da leitura dos resumos dos 51 estu dos restantes, excluíu-se 31 porque não analisa ram associação com indicadores de insegurança alimentar e nutricional. Foram analisados na íntegra 20 artigos, sendo que, desses, 9 foram excluidos por tratarem de crianças maiores de 5 anos de idade. Portanto, 11 estudos contempla ram os critérios de inclusão e foram utilizados nesta revisão sistemática (Figura 1). Em relação aos indicadores econômicos as sociados com a ocorrência de anemia ferropriva, Figura 1. Etapas da elaboração da revisão sistemática. Metodologia Para elaboração deste artigo realizou-se busca sistemática nas bases de dados eletrônicas SciE LO, Lilacs, Medline. Foram incluídos estudos pu blicados nos últimos 11 anos – a partir do ano 2004 – uma vez que foi nesse ano que se iniciou a fortificação com composto de ferro e ácido fólico, com vistas a atender um dos objetivos presentes no Plano Nacional de Segurança Alimentar e Nu tricional, que se refere a prevenção de carências nutricionais11. As palavras chaves usadas foram: aleitamento materno, indicadores nutricionais, carência de ferro, fome oculta, anemia materna, assim como seus respectivos vocábulos em inglês (breastfeeding, nutritional indicators, Iron defi ciency, hidden hunger, maternal anemia). A anemia no mundo acomete aproximada mente 1,620 milhões de indivíduos, sendo que a ocorrência por deficiência de ferro é 2,5 vezes maior5,6. No Brasil, segundo dados da Pesquisa Nacional de Demografia e Saúde da Mulher e da Criança7, a prevalência de anemia em menores de 5 anos foi de 20,9%, sendo que as maiores prevalên cias foram observadas nas regiões Sudeste e Nor deste do país (22,6% e 25,5%, respectivamente). Incluiu-se nesta revisão sistemática artigos orginais, conduzidos no Brasil, que relacionavam anemia ferropriva em crianças brasileiras, meno res de cinco anos, a possíveis indicadores de in segurança alimentar e nutricional, categorizados em: econômicos, nutricionais sociodemográficos e de saúde. Foram excluidos artigos de revisão, monografias, dissertações, teses, capítulos de li vros, além dos estudos realizados com crianças de outros países. A Segurança Alimentar e Nutricional (SAN) refere-se à garantia do acesso a alimentação ade quada e saudável. É um conceito multidimensio nal que perpassa o campo da produção, disponi bilidade e acesso a alimentos, adequadas condi ções de saúde, educação, moradia e saneamento básico8. Sendo assim, indicadores de múltiplas vulnerabilidades, relacionados ao acesso, consu mo e aproveitamento biológico dos alimentos, das condições sociais, econômicas e de estado Os artigos foram sistematicamente revisados, sendo que, inicialmente, dois autores participa 1161 Resultados Busca nas bases eletrónicas Palavras-chave: Indicadores nutricionais, carência de ferro, fome oculta, anemia materna e aleitamento materno Estudos identificados (n = 1023) Lilacs (n = 528 ou 51,61%) Scielo (n = 136 ou 13,29%) Medline (n = 359 ou 35,09%) Critérios de inclusão: estudos originais com população brasileira; Publicados a partir do ano 2004 ; Referentes a anemia ferropriva relacionada a indicadores de (in) segurança alimentar e nutricional Excluídos (n = 972) Com base na leitura do título (n=569 ou 58,53%) Repetidos (n = 81 ou 8,3%) *Realizada em população não brasileira (n = 322 ou 33,12%) Selecionados para leitura do resumo (n = 51) Lidos na integra (n = 20) Excluídos na leitura do resumo (n = 31), pois não analisaram associação com indicadores de (in) segurança alimentar e nutricional Excluídos por contemplarem crianças maiores de 5 anos de idade (n = 9) Incluídos na revisão (n = 11) Estudos identificados (n = 1023) Lilacs (n = 528 ou 51,61%) Scielo (n = 136 ou 13,29%) Medline (n = 359 ou 35,09%) Selecionados para leitura do resumo (n = 51) Incluídos na revisão (n = 11) Figura 1. Etapas da elaboração da revisão sistemática. Figura 1. Etapas da elaboração da revisão sistemática. 1162 André HP et al. Q 116 André HP et al. 1162 André HP et al. Quadro 1. Resumo dos estudos referentes a indicadores de insegurança alimentar e nutricional associados a anemia ferropriva em crianças menores de 5 anos. Referências Metodologia dos Estudos Resultados Título Tipo de estudo Local Avaliação do EN de ferro Anemia Indicadores de insegurança alimentar e nutricional Neves et al., 200512 Prevalência e fatores associados à deficiência de ferro em lactentes (n = 365) Transversal Belém-Pará Hemoglobina Ferritina Prevalência de anemia ferropriva e deficiência de ferro: 55,1 e 15,3% respectivamente A anemia ferropriva e deficiência de ferro apresentaram associação (p < 0,05) com indicadores sociodemográficos (lactentes de 6 a 24 meses) e econômicos (renda per capita menor que ½ salário mínimo). Spinelli et al., 200517 Fatores de risco para anemia em crianças (n = 2715) Transversal Estado do Paraná, Rio Grande do Sul, Minas Gerais, Rio de Janeiro, São Paulo e Mato Grosso (Multicêntrico). Hemoglobina Prevalência de anemia: 65,45%. Houve associação (p < 0,05) entre anemia e indicadores sociodemográficos (idade materna inferior a 20 anos, sexo masculino) e nutricionais (falta de aleitamento materno ou estar em aleitamento misto, peso ao nascer < 2500g). Resultados Oliveira et al., 200613 Concentração de hemoglobina e anemia em crianças: fatores socioeconômicos e de consumo alimentar (n = 746) Transversal Recife- Pernambuco Hemoglobina Prevalência de anemia: 40,6%. Anemia apresentou associação (p < 0,05) com indicadores sociodemográficos (idade da criança, área geográfica, escolaridade materna), econômicos (renda per capita menor que ½ salário mínimo) e nutricionais (ingestão de leite de vaca na dieta). continua em menores de cincos anos, foi unânime entre os estudos a baixa renda per capita17,20,21 (Quadro 1). Quanto aos indicadores nutricionais destaca- se ausência ou baixo tempo de aleitamento ma terno total17,20,21 e de aleitamento materno exclu sivo18, o baixo peso ao nascer17,19, introdução pre coce dos alimentos20, consumo de leite próximo das refeições13,20, o não uso de sulfato ferroso pela mãe e/ou criança e o início tardio do pré-natal21 (Quadro 1) mãe e/ou criança e o início tardio do pré-natal21 (Quadro 1) em menores de cincos anos, foi unânime entre os estudos a baixa renda per capita17,20,21 (Quadro 1). Quanto aos indicadores nutricionais destaca- se ausência ou baixo tempo de aleitamento ma terno total17,20,21 e de aleitamento materno exclu sivo18, o baixo peso ao nascer17,19, introdução pre coce dos alimentos20, consumo de leite próximo das refeições13,20, o não uso de sulfato ferroso pela Discussão A ocorrência da anemia por deficiência de ferro é um dos maiores problemas de saúde pública 1163 Quadro 1. continuação continua Referências Metodologia dos Estudos Resultados Título Tipo de estudo Local Avaliação do EN de ferro Anemia Indicadores de insegurança alimentar e nutricional Netto et al., 200620 Prevalência e fatores associados à anemia e deficiência de ferro em crianças (n = 101) Transversal Viçosa- Minas Gerais Hemoglobina Ferritina Prevalências de anemia, deficiência de ferro:30,1 e 38,4 respectivamente. Os indicadores sociodemográficos (número de moradores no mesmo domicilio, escolaridade materna) e nutricionais ( idade de introdução de sucos e/ou frutas, e tempo de aleitamento materno total, consumo de leite próximo das refeições) associaram-se (p < 0,05) aos baixos níveis de hemoglobina e deficiência de ferro. Vieira et al., 200714 Avaliação do estado nutricional de ferro e anemia em crianças (n = 162) Transversal Recife- Pernambuco Hemoglobina Ferritina Prevalência de anemia: 55,6%; deficiência de ferro: 30,8%. Anemia e deficiência de ferro foram associados (p < 0,05), com indicadores sociodemográficos (idade inferior aos 24 meses). Konstantyner et al., 200918 Riscos isolados e agregados de anemia em crianças frequentadoras de berçários de creches (n = 482). Transversal Creches públicas- São Paulo Hemoglobina Prevalência de anemia: 43,6%. Anemia ferropriva teve associação (p = < 0,05), com indicadores sociodemográficos (menor idade materna), econômicos (renda per capita menor que ½ salário mínimo) e nutricionais (aleitamento materno exclusivo inferior a 2 meses). Netto et al., 201121 Fatores associados à anemia em lactentes nascidos a termo e sem baixo peso (n = 104) Transversal Viçosa- Minas Gerais Hemoglobina Ferritina Prevalência de anemia: 26% A anemia dos lactentes se associou (p < 0,05) com indicadores nutricionais (não uso de composto ferroso no pós-parto pela mãe ou pela criança, início tardio do pré-natal, aleitamento materno predominante) e sociodemográficos (condição de trabalho dos pais). Quadro 1. continuação Em relação aos fatores sociodemográficos, citados pelos estudos desta revisão, destacam- se idade inferior a 24 meses, o número elevado de moradores nos domicílios e a escolaridade materna2,12-15,20. Outros estudos9,22 encontraram, além desses fatores sociodemográficos, a associa ção da anemia ferropriva a menor idade e escola ridade materna. no mundo, e destaca-se como a principal carên cia nutricional em função dos efeitos negativos à saúde. As crianças menores de 5 anos estão entre os grupos vulneráveis, devido a demandas au mentadas para o crescimento e desenvolvimento, característicos dessa fase16,17. Discussão Dentre as principais consequências da anemia ferropriva destacam-se déficit no desenvolvimento psicomotor, na fun ção congnitiva e maior suscetibilidade à infecc ções7,16,17. Situações de múltiplas vulnerabilidades, como o número elevado de moradores no domi 116 André HP et al. 1164 Referências Metodologia dos Estudos Resultados Titulo Tipo de estudo Local Avaliação do EN de ferro Anemia Indicadores de insegurança alimentar e nutricional Rodrigues et al., 20112 Deficiência de ferro, prevalência de anemia e fatores associados em crianças (n = 256) Transversal Cascavel- Paraná Hemoglobina, volume corpuscular médio, ferro sérico e eosinófilos. Prevalência da anemia foi de 29,7%, sendo 77,3% da baixa concentração de ferro. A anemia e à deficiência de ferro associou-se (p < 0,05) a indicadores sociodemográficos (doenças frequentes na família, condições de moradia, tempo de creche, número de moradores no domicílio e falta de saneamento básico). Leal et al., 201115 Prevalência da anemia e fatores associados em crianças (n = 1403) Transversal Recife- Pernambuco Hemoglobina Prevalência de anemia: 32,8% Indicadores sociodemográficos (escolaridade e anemia materna, número de crianças no mesmo domicílio, tratamento da água, idade da criança) tiveram associação (p < 0,05) com anemia. Castro et al., 201119 Anemia e deficiência de ferro em pré- escolares da Amazônia (n = 624) Transversal Acre- Amazônia Hemoglobina ferritina e receptor solúvel de Transferrina plasmática. Prevalências de anemia, anemia ferropriva e deficiência de ferro: 30,6; 20,9 e 43,5% respectivamente. Anemia, anemia ferropriva e deficiência de ferro apresentaram associação (p < 0,05), com indicadores sociodemográficos (sexo masculino, não possuir casa própria, ocorrência de infeções respiratórias e diarreias) e nutricionais (baixo peso ao nascer). Lisboa et al., 201516 Prevalência de anemia ferropriva em crianças (n = 725) Transversal Minas-Gerais Hemoglobina Prevalência de anemia:37,4%, crianças de 6 à 24 meses: 43%. Anemia associou-se (p < 0,05) com indicadores sociodemográficos (não frequentar creche, lactentes entre 6 a 24 meses). EN- Estado nutricional. ç renda está relacionada a presença de condições de moradia e saneamento básico, importantes para o aproveitamento biológico dos nutrientes presentes nos alimentos. Sendo assim, situações de baixo rendimento estão diretamente relacio nadas a dois importantes determinantes da SAN: o acesso e o aproveitamento dos alimentos pelo organismo, que podem condicionar a situações de insegurança, especialmente nos casos que a baixa renda está presente conjuntamente com outros indicadores citados nesta revisão. EN- Estado nutricional. Discussão cílio, menor escolaridade materna, menor ren da mensal per capita e menor poder aquisitivo, influenciam e dificultam as condições de acesso a alimentação adequada e saudável o que pode favorecer a ocorrência de carências nutricionais, como a anemia ferropriva23. A associação da anemia ferropriva com a me nor idade materna, especialmente em relação a gestantes adolescentes, pode ser atribuída ao fato da menor experiência para cuidado com os fi lhos (vínculo mãe-filho), reflexo, na maioria dos casos, da falta de conhecimento ou orientação adequada durante o pré-natal, que em algumas situações nem é realizado adequadamente9,22. Em relação a asssociação observada da ane mia com a idade menor que 24 meses acredita-se que o risco de desenvolver anemia ferropriva nes se grupo etário seja devido ao crescimento acele rado característico dessa fase, acompanhado de indicadores nutricionais, como à dieta de tran sição, que geralmente é composta por alimentos O baixo rendimento monetário per capita, foi o indicador econômico mais citado pelos estu dos que associou-se a presença da anemia12,13,18. Além de possibilitar o acesso à alimentação, a 1165 com baixa biodisponibilidade de ferro, baixa pre valência do aleitamento materno, além de ocor rência de infecções respiratórias e diarreias9,12,24. Acredita-se que a situação de insegurança ali mentar e nutricional, analisada sob a perspectiva de diferentes indicadores que abarquem a mul tideterminação envolvida com o conceito brasi leiro de SAN, é influenciada pelas desigualdades relativas ao sistema social e econômico excluden te32, sendo a pobreza e as iniquidades sociais fato res determinantes desse fenômeno33,34. Os indicadores nutricionais de insegurança alimentar e nutricional, retratados nesta revisão, resumiram-se nas condições de nascimento (bai xo peso ao nascer), ao aleitamento materno e in trodução precoce de alimentos complementares. A maioria dos estudos, analisados nesta re visão sistemática, que avaliaram indicadores de insegurança alimentar e nutricional, associados a anemia ferropriva em crianças brasileiras meno res de cinco anos, são observacionais transversais. Esse tipo de delineamento epidemiológico invia biliza o estabelecimento de relações causais, o que constitui uma limitação desta revisão sistemática e ressalta a importância de realização de estudos longitudinais envolvendo os determinantes da anemia ferropriva em crianças brasileiras. Em estudo nacional referente a fatores associa dos a anemia em crianças brasileiras de 6 a 12 me ses, os autores retratam prevalência de anemia de 65,45%, e associação (p < 0,05) dessa com o baixo peso ao nascer e prematuridade12. Conclusão A presença da insegurança alimentar e nutricio nal, analisada sob a óptica de diferentes indica dores, está relacionada com a ocorrência da ane mia ferropriva em crianças brasileiras menores de 5 anos. As condições de insegurança avaliada segundo os indicadores supracitados indicam a necessidade de investimentos em melhorias das condições de vida, assim como a necessidade de estímulo ao aleitamento materno e introdução adequada da alimentação complementar. Estudos20,26-29 referentes a fatores associados à anemia e deficiência de ferro em crianças, retrata ram associação (p < 0,05) com outros indicadores nutricionais como, por exemplo, idade da intro dução de sucos e/ou frutas, consumo de leite pró ximo das refeições e tempo de aleitamento mater no total, sendo que as maiores prevalências foram em crianças menores de 24 meses de idade. É ne cessário destacar que é difícil o estabelecimento de valores críticos de hemoglobina como ponto de corte em crianças menores de 6 meses de vida, devido as rápidas mudanças de concentração des se indicador bioquímico nessa fase da vida30. Nesta revisão observou-se que a anemia ferropriva associou-se aos indicadores sociode mográficos e de saúde (sexo masculino, idade inferior aos 24 meses, não frequentar creche, fi lhos de mães adolescentes, número elevados de moradores no mesmo domicílio, infecções respi ratórias, diarreias, baixa escolaridade materna), nutricionais (baixo peso ao nascer, característi cas da dieta, “hábito de ingerir leite próximo dos horários das refeições” e introdução precoce de alimentação complementar) e econômicos (bai xa renda per capita) que refletem a determinação social dessa carência. O consumo de leite de vaca fluido foi um dos principais determinantes da anemia no primeiro ano de vida; a caseína e as proteínas do soro do leite de vaca, que constituem a fração proteica da maioria das fórmulas lácteas, e os alimentos infantis industrializados têm influência negativa sobre a absorção do ferro, que se agravam pelas necessidades nutricionais aumentadas em função do crescimento acelerado da criança24. O leite e seus derivados, como iogurte e queijo, possuem cálcio e ao serem consumidos durante ou proxi mo das refeições inibem a absorção do ferro31. Discussão Uma das possí veis explicações dessa associação foi em relação às baixas reservas de ferro ao nascer, devido princi palmente a prematuridade e baixo peso, e à maior demanda desse mineral para o crescimento21. Em um estudo referente a presença da ane mia e deficiência de ferro em pré-escolares da Amazônia Ocidental brasileira os autores obser varam prevalências de anemia, anemia ferropriva e deficiência de ferro de 30,6%, 20,9% e 43,5% respectivamente, e associação (p < 0,05) com o baixo peso ao nascer e o sexo masculino19. A asso ciação entre anemia ferropriva e sexo masculino está relacionada ao maior ganho do peso, ao au mento da atividade da eritropoiese na vida fetal, às menores reservas, maiores perdas intestinais e menor absorção do ferro, observado nos meni nos em relação às meninas25. Referências Referências HP André trabalhou na concepção, redação final e revisão; N Sperandio, RL Siqueira, SCC Fran ceschini e SE Priore na redação final e revisão crítica. 1. Organización Mundial de la Salud (OMS). Documento final de la Segunda Conferencia Internacional sobre Nu trición: Declaracion de Roma sobre la Nutrición. Roma: OMS; 2014. 2. Rodrigues VC, Mendes BD, Gozzi A, Sandrini F, San tana RG, Matioli G. Deficiência de ferro, prevalência de anemia e fatores associados em crianças de creches públicas do oeste do Paraná, Brasil. Rev Nutr 2011; 24(3):407-420. 3. World Health Organization (WHO). Growth Reference 5–19 Years. Genava: WHO; 2007. [acessado 2015 Jul 15] Disponível em http: // who.org.int/growthref/who.pdf. 3. World Health Organization (WHO). Growth Reference 5–19 Years. Genava: WHO; 2007. [acessado 2015 Jul 15] Disponível em http: // who.org.int/growthref/who.pdf. Disponível em http: // who.org.int/growthref/who.pdf. 4. Janus J, Moerschel, SK. Evaluation of anemia in chil dren. Am. Fam. Physician 2010; 81(12):1462-1471. 5. Oliveira TSC, Lamounier JA, Alves CRL, Capanema FD, Rocha DS, Silva MC. Anemia entre pré-escolares – um problema de saúde pública em Belo Horizonte, Brasil. Cien Saude Colet 2014; 19(1):59-66. 6. World Health Organization (WHO). Worldwide preva lence of anemia 1993-2005: WHO global data base on anemia. Geneva: WHO; 2008. 7. Brasil. Ministério da Saúde (MS). Pesquisa Nacional de Demografia e Saúde da Criança e da Mulher – PNDS 2006: Dimensões do processo reprodutivo e da saúde da criança. Brasília: MS; 2006. 8. Brasil. Lei n° 11.346, de 15 de setembro de 2006. Lei Or gânica de Segurança Alimentar e Nutricional. Dispõe sobre a Criação do Sistema Nacional de Segurança Ali mentar e Nutricional–SISAN com vistas em assegurar o direito humano à alimentação adequada e dá outras providências. Diário Oficial da União 2006; 18 set. 9. Oliveira JS, Lira PIC, Maia SR, Sequeira LAS, Amorim RCA, Batista Filho M. Insegurança alimentar e estado nutricional de crianças de Gameleira, zona da mata do Nordeste brasileiro. Rev Bras Saúde Matern Infan 2010; 10(2):237-245. 10. Morais DC, Dutra LV, Franceschini SCC, Priore SE. Insegurança alimentar e indicadores antropométricos, dietéticos e sociais em estudos brasileiros: uma revisão sistemática. Cien Saude Colet 2014; 19(5):1475-1488. 11. Brasil. Ministério do Desenvolvimento Social e Com bate à Fome (MDS). Secretaria Nacional de Segurança Alimentar e Nutricional – SESAN. Câmara Interminis terial de Segurança Alimentar e Nutricional – CAISAN. Brasília: MDS; 2011. 12. Neves MBP, Silva EMK, Morais MB. Conclusão Para garantia da SAN se faz necessário a ado ção de medidas intersetoriais que incidam sobre os múltiplos determinantes envolvidos com esse direito, especialmente aqueles relacionados ao acesso a alimentação adequada e saudável e ao aproveitamento biológico dos alimentos, que diretamente relacionam-se com as carências de micronutrientes, que acometem parcela signifi cativa da população, especialmente as crianças. 1166 André HP et al. 1166 116 André HP et al. Referências Prevalência e fato res associados à deficiência de ferro em lactentes aten didos em um centro de saúde-escola em Belém, Pará, Brasil. Cad Saude Publica 2005; 21(6):1911-1918. 13. Oliveira MAA, Osório MM, Raposo MCF. Concentra ção de hemoglobina e anemia em crianças no Estado de Pernambuco, Brasil: fatores socioeconômicos e de consumo alimentar associados. Cad Saude Publica 2006; 22(10):2169-2178. 14. Vieira ACF, Diniz AS, Cabral PC, Oliveira RS, Lóla MMF, Silva SMM, Kolsteren P. Avaliação do estado nutricional de ferro e anemia em crianças menores de 5 anos de creches públicas. Jornal de Pediatria 2007; 83(4):370-376. 1167 15. Leal LP, Filho MB, Lira PIC, Figueiroa JN, Osório MM. Prevalência da anemia e fatores associados em crian ças de 6 a 59 meses de Pernambuco. Rev Saude Publica 2011; 45(3):457-466. 29. Reis MCG, Nakano AMS, Silva IA, Gomes FA, Perei ra MJB. Prevalence of Anemia in Children Three to 12 Months Old in a Health Service in Ribeirão Preto, São Paulo, Brazil. Rev Latino-Am Enfermagem 2010; 18(4):792-799. 16. Lisboa MBMC, Oliveira EO, Lamounier JA, Silva CAM, Freitas RN. Prevalência de anemia ferropriva em crian ças menores de 60 meses: estudo de base populacio nal no Estado de Minas Gerais, Brasil. Rev Nutr 2015; 28(2):121-131. 30. Szarfarc SC, Souza SB, Furumoto RAV, Brunken GS, Assis AMO, Gaudenzi EM, Silva RCR, Souza JMP. Con centração de hemoglobina em crianças do nascimento até um ano de vida. Cad Saude Publica 2004; 20(1):266- 274. 17. Spinelli MGN, Marchioni DML, Souza JMP, Souza SB, Szarfarc SC. Fatores de risco para anemia em crian ças de 6 a 12 meses no Brasil. J Public Health 2005; 17(2):84-91. 31. Sociedade Brasileira de Pediatria (SBP). Anemia fer ropriva em lactentes: revisão com foco em prevenção. Departamento Científico de Nutrologia. São Paulo: SBP; 2012. 18. Konstantyner T, Taddei JAAC, Oliveira MN, Palma D, Colugnati FAB. Isolated and combined risks for anemia in children attending the nurseries of daycare centers. Jornal de Pediatria 2009; 85(3):209-216. 32. Barroso GS, Sichieri R, Salles-Costa R. Fatores asso ciados ao déficit nutricional em crianças residentes em uma área de prevalência elevada de inseguran ça alimentar. Rev Brasileira de Epidemiologia 2008; 11(3):484-494. 19. Castro TG, Nunes MS, Conde WL, Muniz PT, Cardoso MA. Anemia e deficiência de ferro em pré-escolares da Amazônia Ocidental brasileira: prevalência e fatores associados. Cad Saude Publica 2011; 27(1):131-142. 33. Sicoli JL. Artigo apresentado em 14/03/2016 Aprovado em 11/07/2016 Versão final apresentada em 13/07/2016 Referências Pactuando conceitos fundamentais para a construção de um sistema de monitoramento da SAN. Instituto Pólis, São Paulo, 2005. [acessado 2012 Jul 25]. Disponível em: http://www.polis.org.br/download/65. pdf. 20. Netto MP, Priore SE, Sant’ana HMP, Peluzio MCG, Sa barense CM, Silva DG, Franceschini SCC. Prevalência e fatores associados à anemia e deficiência de ferro em crianças de 18 a 24 meses. Arch Latinoam Nutr 2006; 56(3):229-236. 34. Segall-Corrêa AM, Marin-Leon L, Helito H, Pérez-Es camilla R, Santos LMP, Paes-Sousa R. Transferência de renda e segurança alimentar no Brasil: análise dos da dos nacionais. Rev Nutr 2008; 21(Supl.):39-51. 21. Netto MP, Rocha DS, Franceschini SCC, Lamounier JA. Fatores associados à anemia em lactentes nascidos a termo e sem baixo peso. Rev Assoc Med Bras 2011; 57(5):550-558. 22. Araújo TS, Muniz PT, Cardoso MA, Oliveira CSM. Anemia em crianças de 6 a 59 meses e fatores associa dos no Município de Jordão, Estado do Acre, Brasil. Cad Saude Publica 2011; 27(5):1008-1020. 23. Vieira RCS, Ferreira HS, Costa ACS, Moura FA, Florên cio TAMT, Torres ZMC. Prevalência e fatores de risco para anemia em crianças pré-escolares, Alagoas, Brasil. Rev Bras Saúde Matern Infant 2010; 10(1):107-116. 24. World Health Organization (WHO). Haemoglobin con centrations for the diagnosis of anaemia and assessment of severity. Vitamin and mineral nutrition information system. Geneva: WHO; 2011. 25. Domellof M, Lonnerdal B, Dewey KG, Cohen RJ, Rive ra LL, Hernell O. Sex differences in iron status during infancy. Pediatrics 2002; 110(3):545-552. 26. Gondim SSR, Diniz AS, Souto RA, Bezerra RGS, Albu querque EC, Paiva AA. Magnitude, tendência temporal e fatores associados à anemia em crianças do Estado da Paraíba. Rev Saude Publica 2012; 46(4):649-656. 27. Souto TE, Oliveira MN, Casoy F, Machado EHS, Julia no Y, Gouvêa LC. Anemia e renda per capita familiar de crianças frequentadoras da creche do Centro Edu cacional Unificado Cidade Dutra, no Município de São Paulo. Rev Paul Pediatr 2007; 25(2):161-166. 28. Assunção MCF, Santos IS, Barros AJD, Gigante DP, Victora CG. Anemia em menores de seis anos: estudo de base populacional em Pelotas, RS. Rev Saude Publica 2007; 41(3):328-335. BY CC Este é um artigo publicado em acesso aberto sob uma licença Creative Commons
https://openalex.org/W2993389209	https://europepmc.org/articles/pmc6906669?pdf=render	English	null	Public perception of carbon capture and storage: A state-of-the-art overview	Heliyon	2,019	cc-by	28,420	A R T I C L E I N F O Keywords: Public perception Carbon capture and storage CCS Review Sustainable development Natural resource economics Willingness-to-Pay Stakeholder analysis Decision sciences Well-being Public opinion Environmental change Environmental science Energy Carbon capture and storage (CCS) is a technology enabling to use fossil fuels in a sustainable way. Therefore, it attracts much attention from the industrial sector, government authorities and scientiﬁc community. However, public awareness of the technology is extremely low, and the studies of the lay people's opinion have been launched only during the last decade. Taking into account the role of public support during the implementation of CCS projects, the authors would like to present herein their review of materials on this subject published during 2002–2018 (135 articles). As part of our review, we determined 9 key aspects forming the public perception of CCS. For each of the key aspects, we summarized the available results of the studies. Apart from that, we compared the CCS current status in different countries and provided a number of reasons for involving new countries into the ﬁght against global warming. This work shows that most attention is devoted to CO2 storage; whereas its capture and transportation are poorly studied in terms of public perception. Wider development is required for the methodology enabling a transition from global rhetoric concerning global warming issues to the implementation of particular projects, namely, CCS. The issues related to public awareness of CCS are studied rather thoroughly, but no recommendations are provided regarding the establishment of an optimal database for the lay people. Numerous assessments of general public perception have been carried out. However little attention was paid to the regions with active projects, namely, to the factors considered the most important by the local public, and how actual project results meet their expectations. Therefore, despite an extensive scientiﬁc base developed over 17 years, further studies should be aimed at ﬁlling the existing gaps. This will enable to improve CCS attractiveness for the public, including the cases when it is compared with alternative low-carbon technologies. * Corresponding author. E-mail address: pscvetkov@yandex.ru (P. Tcvetkov). Pavel Tcvetkov a,, Alexey Cherepovitsyn b, Sergey Fedoseev c Pavel Tcvetkov a,, Alexey Cherepovitsyn b, Sergey Fedoseev c a Department of Informatics and Computer Technologies, Saint Petersburg Mining University, Saint Petersburg, Vasilevskiy Island, 21st Line, 2, Russia b Department of Organization and Management, Saint Petersburg Mining University, Saint Petersburg, Vasilevskiy Island, 21st Line, 2, Russia c Luzin Institute for Economic Studies – Subdivision of the Federal Research Centre "Kola Science Centre of the Russian Academy of Sciences", Apatity, Fersman St., 24a, Russia Heliyon 5 (2019) e02845 Heliyon 5 (2019) e02845 p // g/ /j y Received 30 September 2018; Received in revised form 4 May 2019; Accepted 8 November 2019 2405-8440/© 2019 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). P. Tcvetkov et al. Assuming the fact that the CCS projects are efﬁcient, the experience of early countries that adopted the technology shows that negative public perception could be one of the barriers for its large-scale implementation, as experts or politicians usually have a neutral or positive opinion [4]. This is reasonable not only for CCS but also for all environmental tech- nologies in general, as humans that are the source of pollution. Industrial operations themselves would not produce such a negative impact on the environment if the persons taking decisions strived to ﬁnd a balance between economic efﬁciency and environmental safety, which is one of the fundamental principles of sustainable development. Table 1. Structure of this article. Subsection Content and explanation Awareness In this subsection, we analyze the role of awareness in the CCS public perception. The subsection describes factors impacting on information sharing process, and possible ways of public awareness improvement. Public awareness of CCS implies the existence of fair knowledge about the nature of the technology, the causes, and consequences of its use, its strengths, and weaknesses, as well as beneﬁts and risks. Knowledge Since CCS is not a thoroughly studied technology, we consider the problems of providing the necessary knowledge on its nature to the public. By knowledge, we mean the public ability to understand available information about global warming and climate mitigation technologies. NIMBY CCS is analyzed in terms of its susceptibility to the NIMBY (Not In My Backyard) effect, which may be deﬁned as “social rejection of facilities, infrastructure, and services location, which are socially necessary but have a negative connotation” [11]. Beneﬁts and risks perception The key factors inﬂuencing on beneﬁts and risks perception are described, and the relation between this perception and public attitude towards CCS is analyzed. By beneﬁts/risks perception we mean the subjective judgment that people make about the characteristics and signiﬁcance of consequences (positive or negative, respectively) for themselves and their environment. Socio-demographic factors The subsection determines the role of socio- demographic factors in CCS perception development. Taking into account the speciﬁcs of large-scale environmental projects, as well as the strong dependence of CCS project implementation on the mood of local public, this section considers the following aspects related to the social and demographic characteristics of the population: age, gender, education level, religion, expectations and values of people, as well as mentality and cultural speciﬁc. P. Tcvetkov et al. Willingness to pay for CCS Here we review the papers containing an assessment of public willingness to pay for energy rates growth due to the implementation of environmentally friendly technologies. Trust The subsection content can be described as follows: “trust is a psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another” [12]. Acceptance and Preferences between Technologies Comparative analysis of public preferences related to the development of low-carbon technology packages, including CCS, or when several separate technologies are compared. Governmental Policy and Interaction between Stakeholders Analysis of the state policy inﬂuence on the CCS perception, and the role of individual stakeholders and their associations in public relations. Cross-Country Outlook Comparative analysis of the CCS status in different countries. Today, despite the fact that the CCS projects are implemented in different countries, the available scientiﬁc background is focused on two issues: studies of the CCS public perception, sometimes, in the regions where no pilot projects are implemented, but the public interest exists [5]; and global discussion related to the development of environmental technologies, namely CCS. At the same time, there is no connection be- tween these two research groups, which would enable the transition from a global rhetoric to practice [6]. 5–7 years ago, CCS was a young technology, and the scientists had to rely on the achievements in the ﬁeld of public perception of more mature technologies (mainly, nuclear energy [7, 8]). Now the CCS technology has enough scientiﬁc background. Besides, until the present time, social studies, with some exceptions, were based on a predictive approach to the CCS public perception assessment. Today there is a long overdue need in the development of proactive social studies in this ﬁeld focused on the justiﬁcation of approaches providing an objective knowledge and creating a fair image of CCS technologies for public [9], including the countries, where a CCS project is only prepared for implementation. Based on the above mentioned, we consider it logical to step back and consolidate available knowledge in this ﬁeld. The subsection determines the role of socio- demographic factors in CCS perception development. 2. Materials and methods Google Scholar (www.scholar.google.com) and Science Direct (www .sciencedirect.com). The inclusion of materials from Google Scholar (not included in the Science Direct) is explained by the fact that today many authoritative scientiﬁc institutions admit that high-quality research can be published in little-known journals, see [14]. Besides, the Science Direct indexes the most, but not all of the authoritative sci- entiﬁc journals. However, it should be noted that almost 90% of the selected materials were available in both databases. P. Tcvetkov et al. Taking into account the speciﬁcs of large-scale environmental projects, as well as the strong dependence of CCS project implementation on the mood of local public, this section considers the following aspects related to the social and demographic characteristics of the population: age, gender, education level, religion, expectations and values of people, as well as mentality and cultural speciﬁc. Therefore, the purpose of this study is to formulate the main princi- ples of the CCS public perception development based on the global experience in the technologies perception assessment and extend it with Russian point of view on this problem for further implementation. Practically speaking, this will enable to develop a system of proactive public relations for balancing interests of all stakeholders, to achieve higher project efﬁciency and to minimize protest risk after the project startup, caused by misconceptions of locals [10]. Here we review the papers containing an assessment of public willingness to pay for energy rates growth due to the implementation of environmentally friendly technologies. The following part of this article includes 4 sections. Section 2 de- scribes the selection of articles for the review, and distribution of sci- entiﬁc papers by various characteristics. Section 3 includes 10 subsections (Table 1) each of which is devoted to a separate group of factors that have the greatest impact on the public perception of CCS. The deﬁnition of these groups was carried out on the basis of preliminary analysis of the studies' results on the assessment of various factors impact on the public perception of CCS (see Appendix 1, column “Aim of the research”). Comparability and generalization of the results of these studies was possible because the key ideas underlying in most of them are interconnected. Section 4 has a similar structure to Section 3 and includes general outlook; and summary on each aspect of the CCS public perception. Section 5 highlights concluding remarks of the study and further research directions. They will be based on the results obtained herein. 1. Introduction large-scale implementation was out of the question due to insufﬁcient knowledge, speciﬁc risks, capital intensity, inadequate regulatory and legal framework, and absence of efﬁcient mechanisms for carbon markets management. Global warming is a widely known problem discussed since 1960 after publishing of the Manua Loa Observatory's monitoring results, Hawaii [1]. During the past half of the century, several solutions were proposed, one of them is the implementation of the carbon capture and storage (CCS) technologies. The CCS technology involves carbon capturing at industrial facilities (gas and coal-ﬁred power plants, cement plants, etc.), and its storage in geological reservoirs (depleted oil ﬁelds, saline formations, coal beds), or further use in production (Carbon Capture and Utilization - CCU). Even though several countries (for example, China, USA, Australia) have managed to overcome such negative factors; today, implementation of the CCS projects slows down due to insufﬁcient support from the government [3]. Therefore, it is important to involve new countries in the CCS technologies studies, and their capability for a large-scale imple- mentation of such projects, to restrain annual growth of CO2 emissions. This especially relates to the leading producers of CO2 emissions, such as Russia, where the CCS projects are not considered, even in scientiﬁc papers. During recent years, the CCS projects have shown that they can be economically viable, providing certain conditions are created and they can reduce CO2 emissions [2]. Although, just a while ago, their Heliyon 5 (2019) e02845 P. Tcvetkov et al. 2.2. Studies distribution by sources, countries and years This study reviews 135 articles related to the public perception of CCS technology. Nine factors (Table 1) that inﬂuence public perception of CCS were identiﬁed for the analysis. Table 2 shows how many of the reviewed articles consider these factors. The distribution of factors by the articles is shown in Appendix 2. Despite such negative impression created by the information on closed projects, in most studies awareness is considered a basic condition for the CCS attractiveness improvement [19]. For example, characteristics such as “nature-like”, which is widely used now in scientiﬁc literature and mass media, can have a positive impact on public perception [20]. In general, this literature review enables us to conclude that the efﬁciency of infor- mation distribution depends on several factors, which are usually reason- able regardless of the speciﬁc characteristics of the target audience. The largest number of articles devoted to the inﬂuence of knowledge about the nature of CCS on the public perception, which is fair for almost all countries, because it is still little-known technology. The least number of references is accounted for NIMBY reaction and WTP (willingness to pay). On the one hand, this indicates the least degree of study of these issues. On the other hand, it is necessary to understand that these factors are part of the beneﬁts and risks perception and they may simply not be highlighted in the reviewed articles. First of all, it is worth emphasizing that the efﬁciency of information distribution among the local public depends both on the level of trust in the project stakeholders [21], and the policy they are implementing (see Section 3.9); and on the quality of the presented information [22]. However, one should remember that the quantity of information is not equivalent to its quality. According to [23], the provision of additional information has a positive inﬂuence on the CCS public perception; but if we overload people with information, we create a distorted interpreta- tion of the technology risks and advantages. Therefore, the distributed public information should be thoroughly selected, and the materials based only on emotional components should be avoided [19]. Scientiﬁ- cally based facts proved with the world's practice should be preferred [24]. In General, during 17 years, numerous materials were published in the ﬁeld. This proves a high interest in this issue in the scientiﬁc com- munity. Table 3 shows the distribution of the articles by countries and years. 2.2. Studies distribution by sources, countries and years Active studies on the CCS public perception were launched during the last decade, and are implemented until now. One should also remember that 2018 is covered only partially herein. Cross-country research was presented as a separate group divided into countries as shown in Table 4. Table 5 shows the distribution of the articles by sources. Most of the articles are concentrated in several journals. Firstly, this is the Energy Procedia as it publishes the results of the largest international conference “Greenhouse Gas Technologies”. Secondly, this is the International Journal of Greenhouse Gas Technologies, which, as the name implies, specializes in greenhouse gas emissions. Other journals contain only a few articles on the subject, including 21 of 35 journals containing only one article. Nevertheless, complete ignorance of emotional drivers of the CCS perception can be a mistake [25], as people are not always ready to assess rationally the advantages and disadvantages of a certain solution [26]. Sometimes it is necessary to show the best practices implemented by the leading countries [27]. According to [28], the most efﬁcient sources of information on the CCUS projects are brochures describing project experience. The brochures' effectiveness can be explained not only by the avail- ability of information on active projects but also by the fact that they include a lot of graphics enabling people to understand complex concepts used in the CCS technology description [29]. For example, the paper [30] notes that storage is a special chain of CCS technology, which should be explained with illustrations, for instance, a picture of a sponge. This will help to avoid a wrong understanding of CO2 pumping into an under- ground reservoir, which is not similar to a balloon's inﬂation. However, we should remember that the ability to improve information perception with graphics is limited, and overloading of materials with illustrations can harm the perception of the text. Figure 1 shows the distribution of the studies between three key el- ements of CCS: capture, transport, and storage. Many social studies are focused on the technology in general, and even in such studies, much attention is paid to CO2 storage, which is a sole subject of 18 scientiﬁc works. Table 2. The number of articles relating to the factors considered. 3.1. Awareness Almost all modern studies of the CCS public perception highlight poor public awareness of global warming issues and CCS [15]; although the level of general awareness has grown during the recent decades. There- fore, the ﬁrst public reaction is usually negative [16]; however the same can be said about any little-known technology characterized by certain risks [17]. It should be noted that even a detail search in the abovementioned databases does not enable us to conclude that this study includes a complete review of all scientiﬁc works corresponding to the above re- quirements. In addition, the information was collected from the selected articles, and structured (Sections 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9 and 3.10) on the basis of the available experience in the ﬁeld rather than stringent rules. Nevertheless, we believe that our review enables to analyze the most authoritative and important publications in this ﬁeld, and achieve the goal of our study. Recognizing this, the results show that in some regions the level of CCS public awareness is lower as compared with alternative green energy technologies. According to [18], in Australia 77% of respondents know about CCS, in the Netherlands — 84%, in Canada — 61%, in Scotland — 36%. It should be highlighted that a high public awareness in the Netherlands and Australia is explained by different reasons. In Australia, this is due to mass distribution of information on successful projects; whereas, in the Netherlands, this results from failure of the Barendrecht project, and the government's prohibition of the on-shore CO2 storages. This means that awareness does not show public acceptance of a project. 3. Results 3. Results combined with public, perception, involvement, social, acceptance, communication, stakeholder, awareness. Selection of materials was based on the following mandatory requirements: language — English, scientiﬁc ﬁelds — social or economic studies, type — articles, review, reports. In total, 135 studies were selected. They are listed in the refer- ences among other materials added during the references development stage, and in Appendix 1. 2.1. Studies selection According to [13], this study includes the following steps: search and selection of articles, data collection, literature review arrangement, reference list development, literature review writing. Search for the literature on the CCS public perception were limited with the time interval of 2002–2018. This period covers all the history of such studies development. The articles were searched for in the databases For binary search, the following keywords were used: CCS, carbon capture and storage, CO2 geological storage, CCU, CO2 utilization; 2 Heliyon 5 (2019) e02845 Heliyon 5 (2019) e02845 P. Tcvetkov et al. 2.2. Studies distribution by sources, countries and years Country 2002–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16 2017–18 Australia 1 2 3 1 1 1 Canada 1 2 1 1 China 1 2 1 1 Finland 1 3 France 1 1 Germany 1 3 2 1 5 Greece 1 Italy 1 Japan 1 1 1 1 1 1 Netherlands 4 6 5 1 Norway Poland 1 1 Romania 1 Spain 1 2 1 Sweden 1 1 Switzerland 2 5 2 UK 1 1 2 1 7 1 6 US 1 1 5 2 3 Vietnam 1 Singapore 1 Cross-country 2 1 2 6 7 4 1 opinion control. However, we should pay much attention to the content of the provided materials, their relevance, level of trust in their source, and methods of the information delivery [35]. For example, such activ- ities as press releases raise signiﬁcantly the level of local public aware- ness and interest [36]. Table 4. Distribution of the articles between countries in the cross-country sec- tion of Table 3. N Country Number of references 1 UK 12 2 Netherlands 9 3 Germany 8 4 Spain 6 5 Norway 5 6 Poland 6 7 Finland 4 8 Greece 4 9 Italy 4 10 Romania 4 11 Sweden 4 12 Belgium 3 13 France 3 14 Japan 3 15 US 3 16 Denmark 3 17 Czech Republic 2 18 Bulgaria 2 19 Australia 2 20 Canada 2 Table 4. Distribution of the articles between countries in the cross-country sec- tion of Table 3. N Country Number of references 1 UK 12 2 Netherlands 9 3 Germany 8 4 Spain 6 5 Norway 5 6 Poland 6 7 Finland 4 8 Greece 4 9 Italy 4 10 Romania 4 11 Sweden 4 12 Belgium 3 13 France 3 14 Japan 3 15 US 3 16 Denmark 3 17 Czech Republic 2 18 Bulgaria 2 19 Australia 2 20 Canada 2 Another important issue is the assessment of public awareness of CCS, which is mainly based on questionnaires both for national and for local scale. Differences at these scales can be seen in the way the surveys are conducted. At the local level, it is possible to organize focus groups and workshops, where paper-and-pencil and face-to-face interviews could be conducted. At the national scale, as a rule, online tests and telephone surveys are used, which are carried out by specialized agencies. 2.2. Studies distribution by sources, countries and years Obtained results are a valu- able source of information that allows predicting future public opinion about CCS, for example, with using Theory of Planned Behavior. 2.2. Studies distribution by sources, countries and years Knowledge The CCS understanding depends much on a public understanding of the global warming issues, CO2 emissions growth, and potential alter- natives of the emissions reduction, as well as the potential of such ac- tivities for the economy [37]. Misunderstanding of the CO2 emissions concept creates the wrong idea about CCS technologies [38]. In addition, the level of background knowledge and awareness of the CCS technology, before its public discussion involving more detailed information and various points of view, does not always show a proba- bility of its approval or disapproval [39]. Nevertheless, when a person decides on the CCS approval/disapproval, a general level of knowledge and awareness plays an important role [40]. At the same time, an organization of public discussions should include distribution of information concerning the technologies fundamentals, which can be described later, among as many members of the local public as possible [41]. This will enable to minimize the negative impact of the contradictory information from the Internet. We should also bear in mind the public's limited attention [30] resulting from a tremendous amount of information daily received. Therefore, the experts should make addi- tional efforts to raise public interest in familiarization with the material. Table 3. Distribution of the articles between countries in 2002–2018 years. Country 2002–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16 2017–18 Australia 1 2 3 1 1 1 Canada 1 2 1 1 China 1 2 1 1 Finland 1 3 France 1 1 Germany 1 3 2 1 5 Greece 1 Italy 1 Japan 1 1 1 1 1 1 Netherlands 4 6 5 1 Norway Poland 1 1 Romania 1 Spain 1 2 1 Sweden 1 1 Switzerland 2 5 2 UK 1 1 2 1 7 1 6 US 1 1 5 2 3 Vietnam 1 Singapore 1 Cross-country 2 1 2 6 7 4 1 Table 4. Distribution of the articles between countries in the cross-country sec- tion of Table 3. N Country Number of references 1 UK 12 2 Netherlands 9 3 Germany 8 4 Spain 6 5 Norway 5 6 Poland 6 7 Finland 4 8 Greece 4 9 Italy 4 10 Romania 4 11 Sweden 4 12 Belgium 3 13 France 3 14 Japan 3 15 US 3 16 Denmark 3 17 Czech Republic 2 18 Bulgaria 2 19 Australia 2 20 Canada 2 Table 3. Distribution of the articles between countries in 2002–2018 years. 2.2. Studies distribution by sources, countries and years Factor Number of studies Factor Number of studies Knowledge 85 Socio-demographic factors 52 Acceptance of CCS and preference between technologies 83 Trust 42 Governmental Policy and Interaction between Stakeholders 81 Awareness 41 NIMBY 38 Beneﬁts and Risks Perception 79 Willingness to pay 12 To minimize the risk of information overload, it should be reasonable to divide the information into separate parts and present them to the public one by one. However, if such activity is implemented in the form of training, then ensuring a required coverage can be a problem. A usu- ally better understanding of the material is ensured with a small number of participants. Although there are successful examples of large groups training [31]. Nevertheless, for the purposes of public awareness raising 3 Heliyon 5 (2019) e02845 P. Tcvetkov et al. at a national scale, such an approach will not be enough. This means that various information distribution methods should be used [32] with mass media support. Mass media have a strong inﬂuence on the perception of any new tendencies and technologies, and their relationship with the existing problems, although the degree of such inﬂuence depends on many factors [33]. In [34] we see that the opinion of one of the respondent groups has become negative when they read an article describing the risks and un- certainties of the IPCC Special Report related to these technologies. On opinion control. However, we should pay much attention to the content of the provided materials, their relevance, level of trust in their source, and methods of the information delivery [35]. For example, such activ- ities as press releases raise signiﬁcantly the level of local public aware- ness and interest [36]. Another important issue is the assessment of public awareness of CCS, which is mainly based on questionnaires both for national and for local scale. Differences at these scales can be seen in the way the surveys are conducted. At the local level, it is possible to organize focus groups and workshops, where paper-and-pencil and face-to-face interviews could be conducted. At the national scale, as a rule, online tests and telephone surveys are used, which are carried out by specialized agencies. Obtained results are a valu- able source of information that allows predicting future public opinion about CCS, for example, with using Theory of Planned Behavior. 3.2. 3.2. Knowledge The CCS understanding depends much on a public understanding of the global warming issues, CO2 emissions growth, and potential alter- natives of the emissions reduction, as well as the potential of such ac- tivities for the economy [37]. Misunderstanding of the CO2 emissions concept creates the wrong idea about CCS technologies [38]. In addition, the level of background knowledge and awareness of the CCS technology, before its public discussion involving more detailed information and various points of view, does not always show a proba- bility of its approval or disapproval [39]. Nevertheless, when a person decides on the CCS approval/disapproval, a general level of knowledge and awareness plays an important role [40]. at a national scale, such an approach will not be enough. This means that various information distribution methods should be used [32] with mass media support. At the same time, an organization of public discussions should include distribution of information concerning the technologies fundamentals, which can be described later, among as many members of the local public as possible [41]. This will enable to minimize the negative impact of the contradictory information from the Internet. We should also bear in mind the public's limited attention [30] resulting from a tremendous amount of information daily received. Therefore, the experts should make addi- tional efforts to raise public interest in familiarization with the material. Without it, opinion polls will not produce the required results, as polling of the respondents who do not have minimum knowledge of the subject is not a reliable source of data [42]. Mass media have a strong inﬂuence on the perception of any new tendencies and technologies, and their relationship with the existing problems, although the degree of such inﬂuence depends on many factors [33]. In [34] we see that the opinion of one of the respondent groups has become negative when they read an article describing the risks and un- certainties of the IPCC Special Report related to these technologies. On the other hand, the second respondent group read the same information in the form of an information booklet, and their opinion has not become worse. This means that mass media is an efﬁcient method of public 4 Heliyon 5 (2019) e02845 P. Tcvetkov et al. Table 5. Distribution of articles between sources. 3.2. Knowledge Source 2002–2004 2005–2006 2007–2008 2009–2010 2011–2012 2013–2014 2015–2016 2017–2018 Total Energy Procedia 1 8 5 13 5 8 40 International Journal of Greenhouse Gas Control 1 3 8 8 4 3 2 29 Energy policy 1 2 2 4 9 Environmental Science & Technology 1 1 1 2 5 Climate Policy 2 1 3 Greenhouse Gases: Science and Technology 3 3 Applied energy 1 1 2 Energy Research & Social Science 1 1 2 Mitigation and Adaptation Strategies for Global Change 1 1 2 Risk Analysis 1 1 2 Risk Analysis: An International Journal 2 2 The Journal of Environmental Psychology 1 1 2 Renewable and Sustainable Energy Reviews 2 2 AGH Drilling Oil Gas 1 1 Australian Journal of Emerging Technologies and Society 1 1 Emory Law Journal 1 1 Energy 1 1 Energy & Environment 1 1 Energy & Environmental Science 1 1 Environmental Modeling & Assessment 1 1 Environmental Research Letters 1 1 Frontiers in Energy Research 1 1 GeoJournal 1 1 Human and Ecological Risk Assessment: An International Journal 1 1 International Journal of Global Environmental Issues 1 1 Journal of cleaner production 1 1 Journal of CO2 Utilization 1 1 Journal of Environmental Planning and Management 1 1 Journal of Experimental Psychology: Applied 1 1 Marine Policy 1 1 Sustainable production and consumption 1 1 Technology in Society 1 1 The Journal of Experimental Psychology: Applied 1 1 Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy 1 1 Other sources 3 1 2 1 1 3 11 Journal of Experimental Psychology: Applied carbon technology alternatives, their relation to the global warming issues [22]. If the target audience does not have the necessary knowledge about CCS, it is important to provide the information, which enables to look at such projects from different angles, including brief reviews of low- For instance, despite a detailed description of CCS, the information given in [43] was not exhaustive for people unaware of the low-carbon technologies, as this information showed only the opinion of NGOs, which disapproved the CCS implementation in Quebec. Several other materials show the same situation. Figure 1. Distribution of the articles by technological stages. On the other hand, today, there are no CCS studies that enable to deﬁnitely determine a complete list of information materials required for the description of the whole situation and alternatives. 3.3. Not in/under my back yard The research compared the public attitude to the CCS in general, and to the implementation of certain projects near their cities in partic- ular. The results showed that most respondents understand much better the risks of local projects. According to [9], despite a high CCS approval rating, the same situ- ation exists in China, where 48.4% of respondents would prefer if CO2 storage was located more than 100 km from their home, 23.9% of re- spondents approve the location of CO2 storage within a radius of 100 km, and only 3.5% approve a radius of 10 km. Even in Australia, the world leader in the CCS technology develop- ment, the studies [50, 51] showed that 42% of 1273 respondents would be concerned if CO2 storage was located near their city, and the re- spondents (41%) also added that the CCS is a temporary solution for greenhouse gas emissions. Only 21% of respondents were conﬁdent in the technologies safety and strict control of the projects. Besides, there are examples where local incidents have a negative impact on the public perception of CCS. For example, the explosion of a gas pipeline in Belgium in 2004 increased public unrest in relation to the reliability of the CCS process chain, namely a transportation component [65]. A similar situation is observed in Japan, where until 2011 people were quite loyal to the offshore CCS and more concerned about on-shore projects. However, after the accident at the Fukushima NPP, people began to look more negatively at prospects of the on-shore and off-shore CO2 storage due to possible leakages caused by earthquakes [66], although the overall CCS perception has changed insigniﬁcantly. In fact, the CCS risks are not signiﬁcantly higher than those of other new tech- nologies [4]; however, due to insufﬁcient development of the technology itself and the novelty of the geological CO2 storage technologies, which have appeared in some countries only recently, such risk generate there the same concerns that were observed in the pioneering countries at the beginning of the century [67]. The analysis enables to conclude that there are no standard solutions guaranteeing the elimination of the NIMBY effect. Raising of public awareness, improvement of the technology image, provision of addi- tional high economic incentives, and other similar methods can produce an opposite reaction and a suspicion that the stakeholders hide some important information on the real project risks. 3.3. Not in/under my back yard Beneﬁts and risks are two sides of the same coin from the economic point of view. However, when we talk about social phenomena, such as public perception of CCS, factors impacting and depending on beneﬁts and risks perception are different, although, for example, trust in stake- holders can impact on both of them [56, 57]. Besides, the strength of their impact on CCS perception is also different. According to [58, 59] and many other studies, beneﬁts perception have a stronger impact on CCS acceptance, than risks perception, regardless of a serious concern about the immaturity of a storage technology [60]. The NIMBY effect is produced by various factors, including the known risks, values and a sense of unfairness. The effect is a new concept for the CCS technologies, and, in general, it is a normal response to a potential hazard located near a permanent residential area. The paper [48] shows that in Japan people were rather tolerant or neutral towards the CCS ideas and its role in the governmental programs on CO2 emission reduction until the moment when it comes to imple- mentation of particular projects (before the accident at the Fukushima NPP). This is how the CCS comes in practice, and the risks and the public concerns become real. On the one hand, this means that it is necessary to focus on a description of positive aspects of the technology as they have the greatest impact on the public. On the other hand, it requires a detailed study of the information base related to the risks of CCS technologies, which are still not known completely. This uncertainty, along with insufﬁcient knowledge about the physical-chemical properties of carbon dioxide [61], strengthens a negative image of the CCS, which raises public con- cerns [62] and increases the likelihood of a protest potential [63]. Similar ideas are described in [34], where it is stated that despite the prevailing impact of the project efﬁciency on general acceptance, such factors as the risk of CO2 leakage from the storage facility change the attitude to CCS technology. Similar conclusions were drawn in [64], which states that it is the process risks that are of primary importance for the public. The same is proved in [49], but in relation to the cities of Alkmaar and Bergen. P. Tcvetkov et al. Heliyon 5 (2019) e02845 According to [45], education in the CCS and global warming issues can be effective only if it is implemented before a person develops his own opinion on these issues. If a person has a certain idea of the CCS issues, in some cases, it will be rather difﬁcult to change it [34, 46]; but this is possible providing the necessary communication channels and approaches to the target audience are found [39, 47]. Nevertheless, it is believed that a change in the initial opinion about the CCS from negative to neutral, or positive, which occurs after such training, can be driven by a focus on certain CCS advantages, rather than a general description of the situation [42]. reaction is increased when the local public does not approve their local authorities, and their project implementation policy. Eventually, this can lead to the loss of trust in other project stakeholders [55] and refusal to support the local project. Whereas in Russia, speaking about a large-scale implementation of CCS-EOR, the NIMBY reaction is not such as acute issue due to the following factors. Firstly, there are quite many gas pipelines here now, and some of them are laid through the cities. Additionally, a heat supply system consists of pipelines. In other words, if people do not protest against existing similar infrastructure, then, probably, the risk of protest against CO2 pipeline will be minimal. Secondly, according to global practice, the most negative NIMBY reaction appears in the case of CO2 storage near residential areas. The most prospective oil ﬁelds for the CCS projects implementation are located in Siberia, far away from both large cities, and small settlements. Development of knowledge about CCS, global warming issues and alternative low-carbon technologies is a multifaceted objective, which can be complicated by false hopes and beliefs based on superﬁcial knowledge on the subject. That is why stakeholders should use a pro- active approach towards the provision of information on CCS develop- ment issues. 3.2. Knowledge Nevertheless, it is important to provide at least the data from various stakeholders, and on several alternative technologies to the respondents. This will enable to get reliable perception results. However, information overload can lead to the lay people's misunderstanding of the key issues [23, 44]. We can improve the understandability of the CCS materials by making compar- isons with clear and obvious things, or natural events [34]. The difﬁculty of objective CCS knowledge presentation can be explained by the fact that the technology has not been studied enough, for example, the issue of CO2 underground behavior at extremely high pressures and long-term migration. This can be a reason for misunder- standing. One of the most commonly held misconceptions is an idea that CO2 pumping into the underground reservoir is similar to balloon inﬂation [30]. Figure 1. Distribution of the articles by technological stages. 5 P. Tcvetkov et al. 3.3. Not in/under my back yard Nevertheless, not everybody has the NIMBY reaction, and under certain conditions, it is possible to ﬁnd a compromise. According to [52], the NIMBY reaction can be mitigated by changing of separate CCS pro- cess stages. For example, respondents prefer the construction of a biogas-ﬁred plant in their cities rather than a gas-ﬁred plant. However, the reluctance to have CO2 storage and a pipeline system nearby remains. The results are similar to [53], where, however, different results were obtained for a gas pipeline system, which did not bring a protest. Contrary to the signiﬁcant role of technical factors, the study [57] shows a signiﬁcant impact of socio-cultural factors on the beneﬁts and risks perception. And, in spite of the fact that the number of such factors can be very high, and they can vary depending on the region assessed, an attempt was made in [68] to identify the key ones: According to [54], despite a rather high percentage of the NIMBY reaction in Indiana, the respondents’ opinion was mitigated by a detail description of economic beneﬁts from the project implementation near their residence, for example, job growth, ﬁscal loosening, and other economic incentives. At the same time, the information on such risks as CO2 leakage, induced seismicity, explosions, and groundwater contam- ination increased the NIMBY reaction dramatically. In addition, the - “risks perception: uncertainty avoidance, determined as “the extent to which members of society feel uncomfortable with uncertain, un- known, ambiguous, or unstructured situations” and society's short- term or long-term orientation; 6 P. Tcvetkov et al. Heliyon 5 (2019) e02845 - beneﬁt perception: uncertainty avoidance, long-term orientation, and inequality of power distribution in society (power distance)”. - beneﬁt perception: uncertainty avoidance, long-term orientation, and inequality of power distribution in society (power distance)”. development of approaches to various religious groups, which is espe- cially important for Russia, where there are representatives of many re- ligions, and the predominant religions are Orthodoxy and Islam. The study [30] implemented in Switzerland also notes that, under certain conditions, concern about socio-economic factors has a stronger impact on the risks perception than technical factors. Concerning the analysis of socio-economic factors, we would like to highlight the work [69], where the ﬁrst attempt was made to reconcile the economic pa- rameters of the CCS implementation and public perception. The authors managed to achieve two important results. 3.6. Willingness to pay for CCS According to [47], in terms of the CCS public perception, an impor- tant role is played by personal perception, expectations, and values of people, rather than care of the national economy. In other words, during interaction with the public, we should, ﬁrst of all, take into account their mentality [74, 75] and cultural speciﬁc [73], and draw parallels between their personal needs and CCS global impact on climate change, see [76]. This enables us to draw a conclusion conﬁrmed in [77] that interaction with the public should be carried out after a detailed study of their in- ternal organization, speciﬁc motivating factors, expectations, and goals. Theoretically speaking, users of electricity may not care about what resources it is generated from, as it is a product with stable quantitative and qualitative characteristics, which often remain permanent. Addi- tionally, in some regions, large power plants are located far from any cities, which enables to minimize their effect on the environment. However, there is a number of evidence of altruistic behavior of the lay people, so-called "Willingness to pay for CCS", i.e. voluntary consent of the public to electricity and/or heat rates increase for implementation of the CCS projects, in case of capturing CO2 at power generation facilities. An objective fact is that at the stage of demonstration projects, the CCS should be implemented with substantial ﬁnancial support from the state without charging people. However, large-scale distribution of the CCS technologies is closely associated with the public willingness to in- crease electricity rates, which can adversely affect the perception of the technology [88]. It should be noted that the age, contrary to expectations, is not a factor determining the CCS perception [50, 73]; although, in some earlier studies, the perception of innovative energy technologies was associated with the age of respondents [78]. This can be explained by a popular opinion that a human conservatism increases with age. In general, it is obvious that people of different age groups have different thinking, and different approaches are required to inﬂuence them [79]. 3.6. Willingness to pay for CCS According to [89], in Germany, respondents were ready to pay 15.9% more for electricity in case of a 10% increase in the CCS power, and 26.3% more in case of a 10% increase in green energy; although these results differ from the studies conducted earlier [90], which showed that 25% of respondents are willing to pay 1–5% more for energy produced from environmentally friendly sources, 16% of respondents are willing to pay 6–10% more. A signiﬁcant proportion of the respondents either are not ready to pay more (33%) or could not give an exact answer (15%). We should also bear in mind that men are more likely to perceive new technologies and participate in their public discussion than women, which is conﬁrmed by numerous studies [50]. Additionally, men have a more tolerant perception of the risks where the economic potential exists, while women are more concerned about safety [79]. The same is true for CCS technology, as shown in the national survey of UK residents [80] and some other studies. Similar ideas are outlined in [81], which shows that masculine societies are much more focused on economic growth than feminine nations, which increases negative impact on the environment. These results enable us to state that cultural features play important role in the CCS perception, even when it comes to expert assessments [50]; however, one cannot draw conclusions basing only on such factors. The [91] shows that, based on altruistic considerations, the Japanese are ready to pay a certain price for electricity from certain sources. For 1% growth of the renewables share in their energy consumption, they are willing to pay 11 yen, thermal power with CCS — 4 yen, and 1% reduction of nuclear energy share — 14 yen. Religion has also a certain inﬂuence on the CCS perception. For example, according to [82], atheists have the most favorable attitude to the CCS as they do not believe in the afterlife. Christians are somewhat less loyal to CCS technologies; however, they are ready to support them if they have a positive effect on human well-being. Muslims demonstrate the most problematic CCS perception due to the peculiarities of their religious beliefs. In general, both Christians and Muslims have a low perception of the global warming issues and greenhouse gas emissions driven by their belief in the afterlife and divine intervention. 3.3. Not in/under my back yard Firstly, they showed that the CCS projects implementation could inﬂuence on the public (country) welfare. Secondly, they showed that a possibility to use CO2 could positively inﬂuence the national industry. In other words, only detail analysis of the technology risks and beneﬁts will enable to take a balanced decision [70]. Religion determines a person's position in society and also determines its place in global trends. However, individual responsibility for reducing CO2 emissions is a little-discussed issue [83]. Although a lot of people understand the importance of energy conservation, recycling, and other environmental activities, there are clear problems with the establishment of a relationship between global warming and the daily life of a person, according to [84, 85]. Quite the opposite results are shown in the paper [86], where the discussion about the perception of geological CO2 storage, on the con- trary, led to people's descriptions of their personal impact on the envi- ronment, and of the fact that the need for CO2 storage is associated with not enough sustainable ways of our life. On the one hand, such differ- ences can be explained by the difference in the methods of discussions. On the other hand, the results, probably, were inﬂuenced by certain national factors, since [86] is a cross-country survey, however, there is no evidence of such dependence. 3.5. Socio-demographic factors Social sciences helped considerably to widen our understanding of the CCS projects risks. This is associated with different points of view on such risks existing among scientists, technical experts and community, especially people that live in the regions where projects are planned [71], including the impact of social, cultural and demographic factors [72, 73]. Perception of global warming issues, understanding of a human role in this process and development of an objective view of the low-carbon technologies prospects, including CCS, depend on the education of re- spondents [50, 64]. Therefore, probably, we should consider the imple- mentation of an educational strategy for sustainable development that begins in school, and could be a part of a national "green" policy [87]. The socio-demographic group of factors is extremely large and, depending on the features of a particular group of people, the impact of individual factors on their perception of an idea can change. Despite this, this section describes some examples of the key socio-demographic fac- tors inﬂuence on the perception of the CCS technologies. 3.6. Willingness to pay for CCS The extremely limited number of publications in this ﬁeld complicates the According to the research [24] conducted in the UK, 90.3% of re- spondents after the workshop on low-carbon energy technologies (88.6% before the workshop) are willing to pay maximum 50 pounds per quarter (2.2% of their monthly salary). In general, another research conducted in the UK [80] shows that people tend to be more negative about technol- ogies that increase the electricity rates, which, along with certain CCS risks, is a signiﬁcant barrier. As for Russia, it is safe to say that willingness to pay will be much lower, because the average salary in the country is 39,000 rubles, which is equivalent to 446 pounds as of August 23, 2018. In this regard, it is 7 P. Tcvetkov et al. Heliyon 5 (2019) e02845 important to determine whether lay people are willing to allocate enough funds for large-scale implementation of projects, which largely depends on different technical, political and economic factors [92]. perception can be explained by the fact that it is economically unsound for the industry to honestly and openly provide information, which poses a potential obstacle to its activities. Whereas the NGOs' purpose is to maximally reduce a human impact on the environment regardless of economic effects. Such results can also be useful in case of freedom in choosing an electricity supplier and can be an effective tool for an energy company to gain consumers’ trust. In Russia, such a situation is possible in the central and eastern parts of the country, where there is no developed network of gas pipelines and there are various options for energy supply. At the same time, the public trust implies its involvement in the implementation of the CCS project at the earliest stages [105], and interaction with other stakeholders should include, but not limited to, risk communications, as conﬁrmed by a number of projects in Canada and the USA [106]. This can have a positive effect on the public trust in the project stakeholders, although it does not guarantee support of the CCS technologies [107]. We should also note that competence trust in the CCS-related issues is an important factor [101, 108], which requires the provision of certain proofs from stakeholders, and selection of such ex- perts, whose opinion the public will consider [65]. 3.8. Acceptance and preferences between technologies The studies of the public perception of CCS are aimed to identify the factors that inﬂuence its acceptance or rejection. According to [109], acceptance by itself is a ‘behavior that enables, supports or promotes an energy technology, in contrast, to open and expressed resistance to it, while acceptability is referred to as an attitude or evaluative judgment towards an energy technology’. However, the paper [110] shows that the terms “acceptance to” and “support for” are not equivalents in terms of environmental technologies. The former means a passive form of a technology approval (for example, approval of the corresponding research ﬁnancing [111]), while the latter implies active support of project implementation. This enables us to conclude that positive results of acceptance assessment mean not so much public support as its will- ingness not to protest against a project. Such a situation can be rather shaky, and result in mass protests, especially, if people are aware of the technology risks, and do not have enough information about its beneﬁts [63]. In view of this, we can conclude that there is a need for a proactive policy in the ﬁeld of stakeholders’ interaction with the local public, which should be launched before project implementation begins [99]. At the same time, in the context of globalization and information avail- ability, this policy should be transparent and reliable, as any negative consequences of the project implementation in any country will imme- diately become available to the public [100]. The experience obtained in Barendrecht also shows that the choice of trusted organizations also important. Such organizations can mediate between the local public and unknown project stakeholders. This will increase the level of the people's trust in the information received, and establish feedback for timely response to misunderstanding [38]. Similar results were obtained in the study [50] conducted in Australia. It showed that in all groups of respondents the Commonwealth Scientiﬁc and In- dustrial Research Organization (CSIRO) had the highest level of trust in comparison with other organizations (68.6%). The Australians also tend to trust scientists who are not working for the government, and the Na- tional Government by itself had only 20.4% of trust. Development of a protest potential can originate from both the essence of the technology itself and the absence of alternative options that can be offered to people. 3.6. Willingness to pay for CCS Trust in experts in case of CCS technologies is much more important than in case of alter- native low-carbon technologies, as the CCS effect is not apparent immediately; therefore, most lay people have to rely on the specialists’ opinion on a positive long-term effect [64]. Willingness to pay can strongly correlate with willingness to accept [93], which determines the importance of this aspect of public perception for the purposes of this study. If we have information about a potential project cost and the level of average salary in the region of the project implementation, we can preliminarily assess the prospects of the project approval by the local public. 3.7. Trust Trust is an effective tool for the popularization of CCS, but only if the public understands the goals of the industry, NGOs and government authorities [49, 94] under the project, and these goals are not conﬂicting [63]. Otherwise, people will rely on their own judgments, philosophy [95], knowledge and experience [96], ignoring information from other sources. As for the CCS implementation in countries where public awareness of such projects is minimal, it should be noted that, unlike the ﬁrst CCS projects, today we have already experience of such projects imple- mentation in the world practice. This can increase the attractiveness of the technology; however, it requires trust in the project stakeholders, who can adequately describe their experience, and prove that they can ensure fulﬁllment of their promises under the project [95]. The research [97], based on the analysis of the opinions of 811 re- spondents from Barendrecht, shows that local residents had a negative attitude towards the current CCS project, as the town council and the respected local activist group "CO2isNee" could inﬂuence on the project implementation only to a small extent. Whereas the level of trust demonstrated by the local public in relation to the main project stake- holders (national government and Shell) was much lower. Partially, this resulted from the fact that they did not have enough experience of suc- cessful joint interaction, the importance of which was highlighted in [98]. 3.9. Governmental Policy and Interaction between stakeholders 3.9. Governmental Policy and Interaction between stakeholders Being an innovative technology, the CCS can be developed only with support from the state, which will enable these technologies to establish at the market [124, 125]. Comparison of the CCS projects public perception in 4 USA states [126] showed that a conﬁdent governmental policy of greenhouse gas emissions reduction and provision of informa- tion on the problem importance to the public enabled to receive a more favorable attitude from the citizens during the survey, in spite of the fact that they expressed their concerns about economic, technical and polit- ical risks. The paper [127] also notes that the state policy in the ﬁeld of CCS development should be developed for a long term, as its presentation as an interim measure can negatively affect its perception, both by the public and other stakeholders. In the absence of such policy, even the interest in the implementation of the projects expressed by numerous stakeholders cannot produce the expected positive result. The paper [53] shows that the type of CO2 storage selected under the CCS project is also important for people. Thus, CO2 storage in depleted oil ﬁelds is more preferable in comparison with storage in saline for- mations. This fact enables us to consider the option of providing eco- nomic incentives to local public funded with the proﬁt from the implementation of the project. This approach can slightly increase a payback period of the projects, but, at the same time, it will increase the likelihood of support from local residents. There are also examples of a reaction when the CCS idea itself is supported by the respondents; however, when it comes to particular options of its implementation, for example, as part of gas and coal power plants, public preferences can change to negation [119]. We should note that the opposite situation can also happen. The paper [120] describes a signiﬁcant work done to improve the public perception of CCS under the Otway Project. According to the authors, one of the reasons for public approval is its favorable attitude to the development of the gas infrastructure. According to [104], there is little difference in perception of the global warming issues by the CCS stakeholders from different industries and organizations. P. Tcvetkov et al. Heliyon 5 (2019) e02845 Nevertheless, there are a number of factors that are still relevant for comparison of the CCS and renewable energy and cause public concern. According to [88], a high proportion of respondents from different parts of the EU fear that the CCS development can lead to budget cuts and delay in the renewable energy development, which is more preferable [117]. It should be noted that the world scientiﬁc literature does not have any reliable evidence conﬁrming such substitution, as these technologies are not interchangeable, although they are both aimed at the energy sector greening. However, it is possible to redistribute funds in case of refusal from the CCS technologies. In a number of studies, we can see also public concerns about the fact that the CCS will not enable to solve the problems of climate change and energy [16], which can be neither proven or completely refuted at the current stage of their development. Despite a number of methodological issues, such studies are up-to- date and required for determination of public preferences. They are also useful for creating a general image required for some people to understand the CCS technologies [122]; otherwise, they will search for information about alternative uses of CO2 in unreliable, and often inconsistent Internet sources [62]. During the review of scientiﬁc papers related to the comparison of CCS with alternative low-carbon technologies, we would like to note that in early studies the technologies are widely compared with the nuclear power because they have similar risk levels. Today scientiﬁc papers do not use this single comparison. Firstly, this is explained by the fact that the CO2 storage phase alone arouses serious concerns due to insufﬁcient knowledge, although the CCS has not yet led to cataclysms, unlike nu- clear power plants [123]; therefore discussions about their risks are much more of theoretical nature. Capture, transportation and beneﬁcial use of CO2 are generally perceived by the public either neutrally or positively. Another study [118] conducted in China shows that alternative en- ergy is slightly more preferable than the CCS. P. Tcvetkov et al. At the same time, the re- spondents believe that CCS perception can be improved in comparison with the perception of renewable energy, if the government, industry, and NGOs take the following actions: strengthen international coopera- tion, organize public events and training workshops, develop regulatory framework for the industry control, and develop a system of incentives for various stakeholder groups. Secondly, over the past 10 years, much progress has been made in the studies related to technology safety, and a lot of experts, whose opinion the lay people consider, have appeared. Despite the fact that up to now the information on the CCS projects safety is not exhaustive, this enabled to develop a scientiﬁc framework, which, to a certain extent, overlaps other environmental technologies, but remains independent. Although there are some exceptions, for example, the paper [80] compares the perception of wind energy, nuclear energy, and CCS. And, while a decided preference for wind energy over the other alternatives was quite predictable, the choice of nuclear energy as a safer technology than the CCS was unexpected. Another area of studies is the assessment of the CCS perception with account to changes in conditions of individual process stages imple- mentation [113]. For example, the paper [53] shows that coal-ﬁred power plants are less preferable than biomass power plants, or capture at industrial plants. Another paper [52] shows that using a biogas-ﬁred plant as a capture facility is more preferable than using a gas-ﬁred plant. In general, public preferences related to capturing facilities are explained by not so much the CCS technology issues as the problems of the existing energy infrastructure. Nevertheless, this factor also impacts on general and local acceptance of the CCS projects. 3.8. Acceptance and preferences between technologies In this regard, the studies aimed to compare the degree of support for various low-carbon technologies become important. For example, the paper [112] shows that having known ad- vantages and disadvantages of the CCS, the respondents are inclined to compile a portfolio of several technologies, including the CCS, rather than choose one particular low-carbon technology. At the same time, in such energy portfolios, the CCS can take both a small [113] and a sig- niﬁcant share [112]. Denial of the fact that people demonstrate a higher level of trust in some organizations, and lower — in the others can lead to a denial of the proposed idea and technology importance [97]. In addition, in countries with a developed system of environmental NGOs, people can trust them more than the government and companies. According to [101], people consider NGOs the nature protectors and people's servants, unlike com- panies, whose motives are not always clear to them [102]. Therefore, development of a public relations strategy needs the involvement of as many experts from the non-government sector as possible [64]. Many papers [9, 114, 115] note that renewable energy is better perceived by the public than the CCS, even after a detailed study of the technologies. However, most of these studies were carried out at the initial stage of the CCS establishment, when the world practice did not have enough project experience. For example, a survey of citizens in the USA, UK, Sweden, and Japan [116] showed that among a signiﬁcant number of alternative low-carbon energy technologies, the CCS has one of the lowest priorities along with the nuclear power, although the number of overt opponents of the CCS is slightly lower. Today these re- sults are not unambiguous and are refuted by a number of local and national studies. The related ideas are also described in [101], which shows that the level of public trust in the environmental arguments of the industry representatives is signiﬁcantly lower in comparison with NGOs. How- ever, the opposite is also true: the industry's arguments about the eco- nomic aspects of the project implementation are perceived better by the community than the NGOs' ones [103, 104]. Logically speaking, this 8 P. Tcvetkov et al. P. Tcvetkov et al. 3.9. Governmental Policy and Interaction between stakeholders These features adversely affect the CCS perception, although there is no apparent confrontation against this technology. The paper [9] shows that a large part of the respondents expresses a favorable opinion regarding the development of the state CCS support policy. Despite the absence of explicit public support, the Chinese government plans to develop new large CCS projects. According to [134], the public discussions based on the equality of votes and freedom of expression have a positive inﬂuence on the perception of a particular technology and enable to formulate a package of measures that will increase its attractiveness, as well as assess uncer- tainty and some risks related to the technology implementation [62]. At the same time, if government authorities, as well as the neighboring countries, implement an active policy in this area, the issue also takes a political context [135], which can strengthen public trust due to the importance of collaborative decisions. Australia is one of the world leaders in the development and pro- motion of CCS technologies [141]. This resulted from the governmental support and other stakeholders’ interest in the implementation of the projects, especially those involving integration with the oil industry [142]. In Russia, the government is actively pursuing a policy of import substitution due to the imposed sanctions. In particular, it covers also renewable energy, where a regulatory and legal framework was devel- oped to enable renewables to occupy a dominant position on the energy market. However, due to insufﬁcient interaction between individual stakeholders, Russia lacks production capacities to achieve these goals, and the public is not involved in the implementation of clean energy projects. Such factors challenge the effectiveness of clean energy long- term development [136]. Assessment of the expert opinion on the CCS perception in Spain carried out in [40] showed that this technology is favorably perceived an intermediate step in solving the global warming issues, regardless of a slight concern about the reliability of storage and costs of capture. The experience of CIUDEN's CCS project in Spain is one of the good examples of a proactive stakeholder policy accompanied by interaction with the local community. The paper [143] formulates a number of key factors, which enabled to achieve a high level of public support, for example, a highly qualiﬁed team, community engagement plan, identiﬁcation of the local community needs, etc. 3.10. Cross-country outlook The difference in the national context of CCS perception is formed under the inﬂuence of numerous factors, ranging from geographical location and ending with the experience of the public interaction with the state and large energy companies. Accordingly, the main objective of this section is to review similarities and differences in the trends of the CCS public perception. In contrast to successful examples, there are a number of countries, where CCS has uncertain or minimal prospects due to the difference between the stakeholders’ opinions. For example, in Finland, there is no a strong opposition against CCS, but there is no public support either as some stakeholders are not interested in the technologies application due to questionable indicators of their ﬁnancial efﬁciency [72]. Additionally, there is no adequate regulatory framework that could increase the attractiveness of CCS technologies [144], which have certain prospects for country development. The CCS perception can have a pronounced national context. For example, in some countries of Western Europe [26], and in the USA [137], people are quite loyal to the seabed CO2 storage. The same situ- ation is observed in the Nordic Region, where the CCS on-shore projects failed under the public inﬂuence [135]. On the other hand, in North America, the situation is the opposite, and people prefer on-shore storage facilities. In Sweden and Denmark, the governmental authorities demonstrate considerable uncertainty towards these issues, and public perception remains poorly studied. However, both countries have a certain potential for the implementation of the CCS process stages [135, 145]. Another example of different CCS perception due to national differ- ences is the NGO's of Norway and Germany. The former actively support CCS as an efﬁcient method of the ﬁght against global warming. The latter criticize these technologies because of potential risks and low efﬁciency of common storage methods [72]. The same situation is observed in Scotland, where the main CCS stakeholders are skeptical about the CO2-EOR projects, as unattractive and temporary [146]. At the same time, the Scottish people are loyal to almost any kind of activity [18], which is also observed in Romania, where, according to [39], even after a provision of negative information about the CCS, the technology perception has improved signiﬁcantly. The factors determining risk and beneﬁt perception also differ. 3.9. Governmental Policy and Interaction between stakeholders Heliyon 5 (2019) e02845 impact on the risk and beneﬁt perception. Despite various impacting factors determined by the national context of technology development, the technology perception is favorable in Canada, providing certain government support [139], and in Switzerland. elements of the project [126]. The public perception directly depends on the effectiveness of interaction with stakeholders, and how they share responsibility [131], and also whether they have similar expectations about the potential effect of the project implementation [108]. For example, the paper [132] notes a clear skepticism of the local community about the company building a CO2 pipeline, as its impact on the resi- dents’ life has not been clariﬁed. Nevertheless, this disapproval could be mitigated by providing the local people with comprehensive information on the fulﬁlled safety measures, and a detailed description of the impact that has the gas pipeline construction on living conditions. In Canada, this can be explained by the fact that one of the drivers of the national economic development is fossil fuels, especially in such provinces as Alberta and Saskatchewan. In other words, people have already got used to the presence of mining companies and transport systems, which is a positive factor for the public support of CCS devel- opment. This fact is conﬁrmed in [140], which shows that the CCS is perceived more loyally in those regions where mining companies have been actively operating already, in particular, oil and gas companies. On the one hand, this is a way to improve the environmental situation in the region, and, on the other hand, it will increase the economic efﬁciency of the industry in the region. In addition, according to [101], the public tends to trust political decisions if they are sure that all stakeholders had an opportunity to express their opinion on the project, and their interests were not restricted. Public discussions of the CCS technologies enable to show a socio-economic and technical nature of such projects [133], as well as the fact that these technologies are one of the possible steps to reframing the society's energy policy and a long-term transition to clean energy [84]. China produces the largest portion of the global CO2 emissions and has a number of CCS projects on its territory. At the same time, the paper [64] notes that a low level of awareness and limited attention to envi- ronmental issues characterize the lay people of China. 3.9. Governmental Policy and Interaction between stakeholders In general, we can say that all these factors constitute a part of a detailed project implementation plan that takes into account the diverse needs of the local community. 3.9. Governmental Policy and Interaction between stakeholders The survey included 142 respondents (North America, Europe, Japan) from the power industry, oil and gas companies, NGOs, government and educational organizations, as well as a number of other carbon-intensive industries. The results showed that global warming is- sues and CO2 emissions growth are urgent global problems, which is difﬁcult to solve with the available technologies. At the same time, most respondents believe that such environmental technologies as CCS will ﬁnd ways for a large-scale expansion during the next 10–20 years due to the presence of important drivers of their development. The same situ- ation was observed in the ﬁeld of renewable energy at the end of the last century when the cost of renewables was much higher than the cost of traditional energy. However, active state support in a number of coun- tries enabled to reduce the renewable energy generation cost multifold, and achieve a large-scale implementation [128]. Additionally, in recent years, the studies related to CCS and CCU comparison have begun to develop. According to [26], people tend to prefer beneﬁcial use of CO2, rather than its storage in sub-seabed or saline formations. However, in the context of socio-economic studies, the issues of effective CO2 usage should be considered with due caution. Firstly, CO2-EOR, despite the "storage" stage, implies an economic effect obtained by increased mining rate, which is not always communicated to the respondents. Secondly, practical implementation of many CO2 usage alternatives, such as methanol or fuel production, is impossible under current conditions due to the immaturity of the technologies and extremely competitive market. For instance, despite a small number of respondents and the university-based nature, the paper [121] determines some results ob- tained for the public perception of possible alternatives to geological CO2 storage. Thus, the most preferred way of CO2 usage is methanol pro- duction, whereas the most efﬁcient process chain of CCS-EOR is perceived as one of the worst alternatives, second only to the CCS without the beneﬁcial use of CO2. Consequently, one of the key challenges to improving public perception of the CCS is the consolidation of the government, industry and NGOs' efforts [129]. However, an organization of such interaction based on the principles of transparency and openness [130] is a much more labor-intensive process than the interaction of one of the stake- holders with the local public, due to different points of view on separate 9 9 P. Tcvetkov et al. 4.2. Awareness The approach to public awareness raising should be based on a detailed preliminary analysis, which will enable to identify speciﬁc socio- economic characteristics of the public, their motives, and factors that can arouse interest in new knowledge. The studies deﬁne public awareness as one of the key factors providing an objective CCS perception and deter- mining the degree of understanding of the technology. Nevertheless, the Nevertheless, the papers [125, 153, 154, 155] show that the perception of the German local community can depend on the objectives of a CCS project. Comparison of the two projects in Germany showed that when a project is implemented by a scientiﬁc institution that does not proﬁt from the project (Ketzin), the public trust is higher than in the case when a project is implemented by an energy company (Vattenfall at Beeskow). At the same time, it is noted that in the second case, the public was not sufﬁciently informed about the details of the project, and did not have any opportunities to inﬂuence its implementation. It should also be noted that in Germany people are quite loyal to the technologies for the production of various products from CO2 (CCUS) [156], which makes them similar to other EU countries, for example, the UK [157]. Table 6. The frequency of simultaneous presence of two factors in one article. Aw Kn NIMBY BR SD Will Tr Acc Aw Kn 39 NIMBY 17 27 BR 29 55 32 SD 15 30 14 30 Will 6 7 5 9 2 Tr 13 30 19 37 25 5 Acc 31 61 24 50 25 11 27 GovS 24 43 24 46 31 9 38 44 Aw – Awareness, Kn – Knowledge, BR – Beneﬁts and Risks Perception, SD – Socio-demographic factors, Will – Willingness to pay, Tr – Trust, Acc – Accep- tance of CCS and preference between technologies, GovS – Governmental Policy and Interaction between Stakeholders. Table 6. The frequency of simultaneous presence of two factors in one article. 4.1. General review In the Netherlands, after the cancellation of the Shell Barendrecht project in 2010, which faced apparent and unexpected public confron- tation, the CCS perception can be described as negative. Besides, the national government is pursuing an active policy aimed to restrain pro- jects of geological CO2 storage. Almost all of the reviewed articles consider two or more of the above- mentioned factors, except 3 articles [3, 45, 69], that consider only one factor. Table 6 shows the frequency of combinations of two factors: green (40 times and more), yellow (20–40 times), and red (less than 20 times). The smallest number of pairs is observed with WTP. However, this is not due to the isolation of this factor, but to a small number of articles in this area (Table 2). In Japan, the CCS perception is, to a certain extent, related to the experience of nuclear power use. Recent catastrophes have strengthened the negative perception of CCS, and increased the attractiveness of renewable energy for the local public, as well as affected the perception of CCS projects. The public opinion about the prospects of on-shore and off-shore CO2 storage became more negative due to possible leakages caused by earthquakes; although the CCS ideas themselves do not pro- duce a negative reaction [66]. All reviewed articles could be divided into Qualitative (47 articles), Quantitative (83 articles) and studies with combined analysis (5 articles). The most commonly used data collection methods (Table 7) in these articles are online surveys (including one online focus group [62]), in- terviews and organization of various sessions (mostly focus group dis- cussions – 14 articles). Among the countries, where CCS has faced apparent confrontation from the majority of stakeholders, including the public, Germany has the most noticeable experience. Such experience is described in many pub- lications, which show that the German socio-economic and political sit- uation itself is unfavorable for the CCS implementation [149]. This situation results from a popular view that CO2 emissions can be reduced by means of alternative energy, and also from a special political position of the coal industry [72]. The most common research methodologies (Table 8) are descriptive statistics and various types of parametric analysis (mostly regression analysis – 33 studies). Only in three studies [69, 150, 163], modeling elements based on ecological-economic indicators are used. 4.1. General review It should be noted that the TPB (Theory of Planned Behavior) is a base for a signiﬁ- cant number of Quantitative studies related to the inﬂuence of various factors on public perception. However, this fact clearly stated only in 11 articles. Some studies have also noted that large-scale development of CCS projects is economically unreasonable in the country [150]. The paper [39] also shows that whatever positive or negative information is pro- vided to the respondents, they can change their opinion to the worse. This happened to the respondents from Germany, but not from other EU countries. As a result of the analysis, a scheme (Figure 2) reﬂecting the main linkages between the considered groups of factors and their relationship with the public perception of CCS was drawn up. The following sub- sections discuss the ﬁndings for each of the mentioned groups. When comparing the perception of onshore and offshore CO2 storage [151], many respondents from Germany could not choose an acceptable alternative, as they have a negative attitude to CO2 storage in general. At the same time, the Germans have a rather neutral opinion about the use of pipelines for CO2 transportation; although it is necessary to take into account the needs and socio-demographic characteristics of the local public in a particular area due to the NIMBY reaction [152]. 3.10. Cross-country outlook Ac- cording to the study [138] conducted in Switzerland, the key factors are socio-economic factors related to the unsustainable nature of individual stages of the CCS technology. Probably, this is determined by the national internal policy and energy strategy, which is aimed at the reduction of the nuclear power share. As part of this process, the transition to gas-ﬁred power plants can be an intermediate stage of sustainable development, which involves also CCS. In Poland, the local public demonstrates a positive reaction to the CCS projects, and expects a positive effect from their implementation, both for people and for their region in general [41, 107], despite some skepticism about the location of CO2 storages on their territory due to certain risks of leakage. But, despite the public approval, CCS projects in Poland are In contrast to [138], according to the study [57] conducted in Canada, level of trust in the state authorities, companies and NGOs has maximum 10 P. Tcvetkov et al. Heliyon 5 (2019) e02845 not implemented, as there is no effective governmental support for this process. Taking into account different status of the CCS projects in different countries [158], we believe that an important factor of their further development is international cooperation, which would enable to combine efforts creating favorable conditions for the projects and adopt successful experience of other countries, for example, Australia [159]. Such cooperation can also appear to be an efﬁcient tool for the devel- opment of the ideas for the environmental technologies introduction, and communicating them to the general public. The prospects for the CCS development in Vietnam are connected with the necessity to create ﬁnancial incentives for the technologies implementation, for example, preferential taxation for land use, devel- opment of an integrated environmental policy, and availability of inter- national cooperation and support from the countries experienced in the implementation of such projects [147]. In France, CCS support is insigniﬁcant [148], although there is no apparent confrontation against this technology. We would rather say that people are suspicious of CCS. This results from insufﬁcient knowledge about separate process stages, and a great number of risks. 4. Summary 4.3. Knowledge Promotion of the public understanding in the ﬁeld of sustainable development, including global warming and the technologies used to ﬁght it (for example, CCS), has a dual character. On the one hand, it enables involve more people in an open discussion, and, therefore, re- view potential risks and effects of CCS more thoroughly. On the other hand, this enables us to launch reframing of an environmentally balanced society development, which will include not only assessment of indus- trial environmental projects, but also an understanding of the individual responsibility of every person. At the same time, it is necessary to have a clear idea of the sources, where the public knowledge originates from. If people do not have an opportunity to rely on expert opinion about the quality of a material, they will turn to the Internet, where information can be inconsistent and unreliable. As a result, during the project implementation, the stake- holders will have to ﬁght against false judgments, rather than develop the required public knowledge, which seems to be a much more labor- intensive activity. Table 8. Methodologies distribution. Methodology Type of analysis Number of studies Qualitative Quantitative Combined Case study 10 5 1 16 Review 5 0 0 5 Ecology-economical modeling 0 3 0 3 PESTEL analysis 0 1 0 1 Non-parametric analyses (Wilcoxon tests, Mann-Whitney U tests, Kruskal-Wallis tests, chi-squared test, Friedman test, component, and structure analysis) 0 13 0 13 Descriptive statistics (frequencies, means, standard deviations, correlations) 0 68 4 72 Parametric analyses (t-test, ANOVA, regression analysis, cluster analysis) 0 44 2 46 When we talk about a speciﬁc project and the local public, the lack of reliable knowledge about CCS can become, on the one hand, an instru- ment of opposition in the ﬁght against the project. On the other hand, it will allow unscrupulous stakeholders to deceive the local public in relation to key aspects of the project. Thus, dissemination of knowledge about the nature of technology among the local public is necessary to prevent possible conﬂicts. 4.2. Awareness Method of data collection Type of analysis Total Qualitative Quantitative Combined Survey (not speciﬁed or traditional paper-and-pencil questionnaire) 1 27 1 29 Mail survey 0 7 0 7 Information-choice questionnaire 2 11 0 13 Online survey 2 28 1 31 Interviews 17 17 2 36 Telephone survey 3 8 0 11 Media analysis 4 3 0 7 Various sessions (workshops, seminars, panels, focus group) 12 15 3 30 Theoretical (including reviews and case studies) 22 0 0 22 Table 8. Methodologies distribution. Methodology Type of analysis Number of studies Qualitative Quantitative Combined Case study 10 5 1 16 Review 5 0 0 5 Ecology-economical modeling 0 3 0 3 PESTEL analysis 0 1 0 1 Non-parametric analyses (Wilcoxon tests, Mann-Whitney U tests, Kruskal-Wallis tests, chi-squared test, Friedman test, component, and structure analysis) 0 13 0 13 Descriptive statistics (frequencies, means, standard deviations, correlations) 0 68 4 72 Parametric analyses (t-test, ANOVA, regression analysis, cluster analysis) 0 44 2 46 tool for them. In this case, the public can be misled, which will enable to achieve some short-term goals, but, in the long term, this can lead to open protests. To avoid potential misinterpretation of the facts, we consider it necessary to involve stakeholders who are independent from each other, and who will not allow a distortion of the real picture. Table 7. Methods of data collection distribution. Method of data collection Type of analysis Total Qualitative Quantitative Combined Survey (not speciﬁed or traditional paper-and-pencil questionnaire) 1 27 1 29 Mail survey 0 7 0 7 Information-choice questionnaire 2 11 0 13 Online survey 2 28 1 31 Interviews 17 17 2 36 Telephone survey 3 8 0 11 Media analysis 4 3 0 7 Various sessions (workshops, seminars, panels, focus group) 12 15 3 30 Theoretical (including reviews and case studies) 22 0 0 22 Table 7. Methods of data collection distribution. 4.2. Awareness Aw Kn NIMBY BR SD Will Tr Acc Aw Kn 39 NIMBY 17 27 BR 29 55 32 SD 15 30 14 30 Will 6 7 5 9 2 Tr 13 30 19 37 25 5 Acc 31 61 24 50 25 11 27 GovS 24 43 24 46 31 9 38 44 Aw Awareness Kn Knowledge BR Beneﬁts and Risks Perception SD Perhaps, there is a deﬁnite correlation between a sharply negative perception of various innovative technologies by the Germans (CCS, gene technology etc.), which have a number of uncertain risks. How- ever, at the moment, there is no reliable conﬁrmation of this correlation in the world scientiﬁc literature. At the same time, active resistance of the German stakeholders to the CCS projects provokes an opposite re- action in the scientiﬁc community, which can be seen in numerous publications. Aw – Awareness, Kn – Knowledge, BR – Beneﬁts and Risks Perception, SD – Socio-demographic factors, Will – Willingness to pay, Tr – Trust, Acc – Accep- tance of CCS and preference between technologies, GovS – Governmental Policy and Interaction between Stakeholders. 11 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. Table 7. Methods of data collection distribution. Method of data collection Type of analysis Total Qualitative Quantitative Combined Survey (not speciﬁed or traditional paper-and-pencil questionnaire) 1 27 1 29 Mail survey 0 7 0 7 Information-choice questionnaire 2 11 0 13 Online survey 2 28 1 31 Interviews 17 17 2 36 Telephone survey 3 8 0 11 Media analysis 4 3 0 7 Various sessions (workshops, seminars, panels, focus group) 12 15 3 30 Theoretical (including reviews and case studies) 22 0 0 22 Table 8. Methodologies distribution. Methodology Type of analysis Number of studies Qualitative Quantitative Combined Case study 10 5 1 16 Review 5 0 0 5 Ecology-economical modeling 0 3 0 3 PESTEL analysis 0 1 0 1 Non-parametric analyses (Wilcoxon tests, Mann-Whitney U tests, Kruskal-Wallis tests, chi-squared test, Friedman test, component, and structure analysis) 0 13 0 13 Descriptive statistics (frequencies, means, standard deviations, correlations) 0 68 4 72 Parametric analyses (t-test, ANOVA, regression analysis, cluster analysis) 0 44 2 46 Table 7. Methods of data collection distribution. 4.4. Not in/under my back yard The NIMBY effect is a natural reaction of a person to the unknown, in particular, to unfamiliar technologies such as CCS, which still remain poorly studied, especially, in the area of geological carbon dioxide stor- age. There are no methods eliminating this effect completely; however, it is possible to mitigate the negative perception of lay people, provided that it is identiﬁed in the early pre-project stages. This necessity is related to the explicit preferences of the local public for various process solutions at all stages of the CCS production chain. In addition, the NIMBY reaction can be reduced by provision of the most complete information about the existing scientiﬁc background does not enable to determine the exact scope of information about CCS that would be exhaustive. This is important, as both insufﬁcient and excessive information can lead to misunderstanding of the CCS fundamental principles. In addition, a certain misinterpretation of the facts, which some project stakeholders may be interested in, can appear a more efﬁcient Figure 2. Main linkages between studied groups of factors and public perception of CCS. Figure 2. Main linkages between studied groups of factors and public perception of CCS. 12 P. Tcvetkov et al. Heliyon 5 (2019) e02845 Table 10. Most mentioned risks. Risks Number of studies CO2 leakage, migration (from storage or pipeline) or explosion 48 Disposal of CO2 may cause seismic activity 29 Environmental impact (underground, marine environment) 31 CCS cannot achieve the goals of reducing CO2 emissions because of the lack of effectiveness and lack of facilities for storage. This is just a temporary solution that supports the use of fossil fuels. 34 Possible destruction of facilities due to the lack of stakeholders' responsibility. 9 Loss of land due to the construction of infrastructure. 9 measures taken to ensure the safety of the local public, as well as the development of measures aimed to stimulate the local public. 4.6. Socio-demographic factors It seems that trust is a central element determining a positive perception of CCS technologies. It is the public trust in stakeholders that determined their readiness to consider a project implementation option. This can be observed even in the experience of Germany, where the CCS ideas themselves cause a negative public reaction. At the same time, earning of public trust is an extremely long and labor-intensive process, which largely depends on the experience of the interaction between lay people and the project stakeholders. In addition, in the case of the negative experience, it will be difﬁcult to change the public opinion for the better. For example, such a situation exists in the relationships be- tween the local public in some regions of Russia and local governmental authorities. Socio-demographic factors determine a worldview of a person, his position in the context of global problems, understanding of the impor- tance of individual responsibility for environmental protection. Most of these factors cannot be controlled and managed, and therefore, other CCS project stakeholders have to adapt their work methods to a speciﬁc target audience. At the same time, in most cases, a detailed study of its char- acteristics enables to determine potentially efﬁcient options of interac- tion between stakeholders. To prepare the acceptable methods of community engagement, we need to collect a considerable amount of information to characterize the target audience. We cannot prefer one information collection method Table 9. Mentions of risks by groups. Risks Number of studies Beneﬁts Number of studies Risk for the society 40 Beneﬁts for the society 44 Risk for personal safety 42 Beneﬁts for oneself 17 Risk for the environment 51 Beneﬁts for the environment 53 Risk for future generations/long-term sustainability 16 For future generations 9 General risks 70 General beneﬁts 34 Table 9. Mentions of risks by groups. Risks Number of studies Beneﬁts Number of studies Risk for the society 40 Beneﬁts for the society 44 Risk for personal safety 42 Beneﬁts for oneself 17 Risk for the environment 51 Beneﬁts for the environment 53 Risk for future generations/long-term sustainability 16 For future generations 9 General risks 70 General beneﬁts 34 4.7. Willingness to pay for CCS Among the speciﬁc CCS risks (Table 10), the most frequently mentioned is the risk of CO2 leakage. The second most frequently mentioned risk is the risk of failure to implement plans to reduce CO2 emissions. This risk includes one dual point – support for fossil fuels. This point is perceived both in terms of risks (continued use of environmen- tally harmful fuels) and in terms of beneﬁts (the possibility of using cheaper energy without a signiﬁcant impact on the environment). The least mentioned risks are usually considered in articles on the perception of CCS by local residents of the regions, where the projects are planned to be implemented. In the coming decades, the growth of energy rates is an inevitable trend, both in case of renewable energy development, and expansion of fossil fuels use with or without CCS. Nevertheless, the willingness to support technology that will accelerate the growth of energy rates is a signiﬁcant barrier to large-scale implementation of CCS. This problem already exists in countries aware of CCS and is expected in those coun- tries, where CCS is only at the initial stage of development. Some studies show that a positive perception of the technology en- ables rising energy rates for the public on a voluntary basis; however, such growth is extremely limited and only possible in a few countries. In addition, the willingness to pay for the technology implementation is closely related to the quality of life; therefore, it is necessary not only to increase the technology attractiveness but also develop adequate socio- economic incentives. Despite the results of this review, we should bear in mind that the expected beneﬁts and the most important risks can differ by countries, regions, cities and even social classes. A certain impact on the risks and beneﬁts perception can be ensured by a proper presentation of the technology, which will provide the public with fair information on the measures to be taken to assure its safety, and the beneﬁts for the regions, where the projects will be implemented. 4.5. Beneﬁts and risks perception It is the risks and beneﬁts perception that underlies approval/disap- proval of the CCS technology. At the same time, the perception of tech- nology is more inﬂuenced by the perception of the beneﬁts, whereas the perception of the risks is an indicator of a protest potential. This means that it is necessary to develop a single and considered policy governing all stakeholders’ interaction with the public, which will allow to equally inﬂuence both of these factors. All risks and beneﬁts discussed in the reviewed articles can be divided into ﬁve groups (Table 9). With regard to the beneﬁts of technology, the environmental beneﬁts of reducing greenhouse gas emissions are most often mentioned. In addition, a large number of articles highlight the beneﬁts for society and for the project area (44 studies), such as job creation, the attraction of investments, etc. The study of risks is more common than beneﬁts, due to the lack of knowledge about the technol- ogy and the lack of intrinsic public knowledge about its essence. At the same time, 70 articles refer to risks as consequences of CCS imple- mentation but do not specify their nature. over another, as the local public can include people of different ages, confessions, and worldviews. Nevertheless, the involvement of people, whom the local public trusts, can have a certain contribution to the collection of the necessary amount of reliable information. It is necessary to keep in mind that collection of socio-demographic data about the target audience is a normal pre-interaction stage, because other stake- holders need to know, who will be their partner in the project. 4.11. Cross-country outlook However, while the possibility of national companies’ efﬁciency in- crease will be positively perceived by the public, this is not obvious for an environmental component of such projects. Usually, environmental pollution is a local problem of the regions; therefore, the same opinion can also be expressed regarding greenhouse gas emissions [164]. In particular, this situation is typical for Russia, which occupies a huge territory, where people are rather poorly informed about the events occurring at the other end of the country and even in neighboring regions [165]. At the national level, there is a large huge number of factors that can inﬂuence the public perception of technology, its approval or willingness to protest against its implementation. At the same time, according to our review, these factors are difﬁcult to reveal as they largely depend on the speciﬁc features of a region to be assessed, the mentality of the local public, actions of local and regional authorities, and other things. Even when several projects within the EU are assessed, we have to admit that the key factors of their performance are not exhaustive, and can have different importance when considering projects in other countries. Nevertheless, the main groups of factors remain the same: "beneﬁts", "costs/risks", "climate change". In addition, Russia needs to develop a compensation mechanism for the public and other stakeholders' risks at the expense of excess revenues from oil production; as the companies do not consider such risky and long-term projects, and they require state support in the form of socio- economic mechanisms that require co-ﬁnancing, insurance and/or implementation of a risk management system for the projects [147]. For this purpose, it is necessary to develop appropriate legislation in the ﬁeld of CCS, as well as a general environmental policy [166] deﬁning the responsibilities of each stakeholder. These measures will enable to pro- tect the stakeholders' interests, and ensure ﬁnancial security. The The necessity to identify speciﬁc factors suggests that the best starting point for promotion of the CCS technologies is to study socio-economic characteristics of the target audience, their views, and knowledge about the role of environmental projects, for example, by means of online surveys that showed good results in the international practice. This initial step will allow to ﬁnd ways of achieving the balance of interests, which should be further discussed during dialogues and various sessions. 4.10. Governmental Policy and Interaction between stakeholders The state environmental policy plays a deﬁning role in the efﬁciency of further CCS development, as any innovative technology at the stage of its development needs substantial support. The experience of some countries shows that CCS can be actively implemented only with a long- term development strategy, which, among other things, determines a stakeholders’ engagement procedure and their responsibilities. In terms of the geographical distribution of CCS public perception research, this review will form the basis of the ﬁrst studies in the ﬁeld of the CCS technologies promotion in Russia. Therefore, preparation of a sound plan for the development and promotion of the CCS projects is a relevant and up-to-date objective. A positive feature of the Russian CCS projects development is the presence of a signiﬁcant number of depleted oil ﬁelds located far from residential areas, which can be used for CO2-EOR. Production of addi- tional oil volumes can have a positive effect on public perception [65]. In addition, a preliminary assessment of the CCS-EOR projects ﬁnancial performance in Russia showed that they can be economically efﬁcient under the current conditions [163]. For the public, successful implementation of the CCS projects largely depends on a well-coordinated work of all stakeholders with due consideration of their opinions. In addition, the arguments for lay people in favor of the CCS project implementation, which can cover ecological, technical, or economic issues, should be expressed by those stakeholders, whose goals the public approves, and whose opinion the public trusts. 4.11. Cross-country outlook On the one hand, it will enable to determine the necessary measures increasing a general level of awareness; on the other hand, it will enable to demon- strate the authorities’ interest in the public opinion, and, consequently, improve the public trust in one of the key CCS stakeholders at the initial stage — the state. For example, one can use elements of the approaches to the Social Site Characterization proposed in [160] (main principles of public participation), and [161] (approach to the CCS project management). Table 11. Key barriers for CCS implementation. Barriers for CCS implementation Number of studies Lack of public knowledge about CCS, misconceptions. 95 Lack of or poor communication strategy 82 Competition between alternative technologies 70 Lack of long-term policy of CCS implementation 57 Controversial economic efﬁciency, capital-intensity, weak market-based mechanism 55 Not enough studied the long-term effects of the technology 52 Lack of trust in some stakeholders 54 NIMBY reaction 38 Site selection and project design without taking into account the speciﬁc of locals 34 Appearance of protest potential due to negative public perception 25 Increase in price of energy 22 4.9. Acceptance and preferences between technologies Objectively, in most countries, the CCS projects are less preferable as compared with renewable energy. However, the history of renewable energy development shows that its large-scale implementation started only as a result of balanced and aggressive marketing policy, and huge state support in a number of countries. As for CCS, in a number of studies, the respondents note that such supporting factors are not provided for these technologies. We can say that in countries where the ideas of a positive impact of the renewable energy development, conﬁrmed by a number of successful projects, has been already established; CCS is 13 P. Tcvetkov et al. Heliyon 5 (2019) e02845 perceived as a competing technology, and its support is equivalent to a delay in the renewable energy development. area. As such goals, it is possible to set the necessity of overcoming key barriers for the CCS implementation, which were identiﬁed in this re- view (Table 11). However, it should be noted that the renewables cannot completely replace fossil fuels in the near future, given the current industry growth rates. This fact is rarely mentioned in the information materials for lay people. Additionally, CCS is not a direct competitor to renewable energy, as the purpose of this technology is to increase the environmental safety of fossil fuels, rather than replace them. Thus, on the one hand, it is necessary to eliminate the public misconception opposing these tech- nologies. On the other hand, we should show the spheres of inﬂuence and contribution of each technology in the environmentally sustainable development of the society. To overcome these barriers, it is necessary not only to ﬁnd new ways of cooperation between stakeholders, but also to improve our knowledge about the possible consequences of this technology, to provide this new knowledge to the public, to create suitable economic conditions and to determine the necessity of CCS projects by comparison with other low- carbon alternatives. At this step, it is necessary to mention a signiﬁcant gap in our knowledge about alternative options of CO2 sequestration, such as CCUS and CCU, which also should be taken into account during the assessment of alternatives. The differences between these options should be widely discussed and investigated, because they have different CO2 management principles and organizational features [162]. These differences could also have a signiﬁcant inﬂuence on a public perception of different CO2 sequestration options. Funding statement The research was carried out within research school «Rational subsoil use» in the Department of Organization and Management; with the The ﬁrst author/author Year Sample (number of respondents) Level Focus Method of data collection** Aim of the research Anderson C. [120] 2012 20 Local Project experience, public acceptance Interviews, telephone survey to explain community acceptance of CCS through a human and social capital analysis, and through that analysis assess the Otway public participation process. Anghel S. [46] 2017 1002 General Public perception, communication ICQ to collect and analyze data on Romanian public awareness and knowledge of climate change, energy policy and CCS in general and furthermore on public awareness and knowledge about local demo initiatives and existing CCS information material and campaigns. Arning K. [156] 2017 232 General Public perception, risk perception Focus groups, online survey to explore public perception of carbon dioxide utilization technologies in Germany; to conceptualize carbon dioxide utilization risk perception and acceptance. Ashworth P. [114] 2009 41 Local Public perception Workshops to understand public perception to climate change and low-emission technologies and how to engage communities on these topics. Ashworth P. [31] 2009 305 General Effectiveness of large group dialogues Workshops to explore Australian society's acceptance of energy technologies; to assess the effectiveness of dialogue with large groups for informing knowledge and changing attitudes of low emission energy technologies. Ashworth P. [71] 2015 N/A General Public communications Qualitative, literature review to collect results related to public communication on CCS Ashworth P. [105] 2012 N/A Local, cross- country Comparison of projects experience Interviews, telephone survey to identify factors that contributed to successful project deployment, as well as to assess lessons learned about various communication and engagement practices. Ashworth P. [99] 2009 N/A General Communication activities Qualitative, online survey, telephone survey to synthesize the range of communication activities that have been planned or implemented since 2002 in Australia and internationally, and examines the strengths and weaknesses of these activities. Ashworth P. [18] 2013 374 Cross- country Public perception Workshops, questionnaire to explore how international context may have impacted on the results of engaging the general public on issues related to climate change, energy technologies, and the overall shift towards a low carbon society; to discuss arised differences and the implications for policy makers and research developers. Billson M. Declarations The authors declare no conﬂict of interest. 5. Conclusion and further research The analysis showed that the public is an important stakeholder of CCS projects, the opinion and needs of which should be taken into account. In this regard, it is necessary to improve the mechanism of interaction between stakeholders in order to ﬁnd compromises at all stages of the project, starting with the planning of the site selection. Such improvements require an interdisciplinary approach and the identiﬁcation of key goals for further development of research in this 14 Heliyon 5 (2019) e02845 P. Tcvetkov et al. ﬁnancial support of the grant of the Russian Science Foundation (Project No. 18-18-00210, «Development of assessment methodology of public efﬁciency of projects devoted to carbon dioxide sequestration»). Saint- Petersburg Mining University. importance of such measures is determined by a dual nature of stake- holders (industry and NGOs) in the CCS projects. Firstly, they are inter- ested in the project's implementation from their professional point of view. Secondly, they are the most competent experts in their disciplines, which can have a positive effect on the CCS public perception in Russia [35]. Additional information All authors listed have signiﬁcantly contributed to the development and the writing of this article. No additional information is available for this paper. Appendix 1 Appendix 1 Funding statement [128] 2017 N/A Cross- country Policy Qualitative to show the relevance of state support for CCS projects implementation ( ti d t ) (continued on next page) 15 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. (continued) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research Boyd A.D. [139] 2017 1471 General Public perception, EOR Online and telephone survey to examine descriptive statistics to understand public perceptions of CCS and applied regression models to assess how risk perceptions, perspectives of climate change and trust in government relate to the support for or opposition to CCS development and funding for the technology. Bradbury J. [76] 2009 N/A General- Local Public perception Focus group and interviews to discuss ﬁndings from the joint review of the focus groups and the potential lessons for research and application to CCS deployment. Braun C. [149] 2017 3526 General Preferences between technologies, public perception Online survey to compare public perception in Germany of three speciﬁc measures: solar radiation management via stratospheric sulphate injection, large-scale afforestation, and carbon capture and sub-seabed storage. Breukers S. [77] 2015 15 General, cross- country Public engagement Qualitative, interviews to improve understanding of how project developers view and practice engagement and communication. Brunsting S. [160] 2013 1850 Cross- country Social site characterization Focus conferences, paper-and-pencil questionnaire, telephone interviews, media analysis to present the results of social site characterisation and public participation activities at two prospective CCS sites in Poland and Scotland. Brunsting S. [94] 2011 N/A General Public perception, communication Qualitative, Case study to explore the differences among chosen case studies to develop communications exercises (or even a part of consultation policy) for future CCS projects. Budinis S. [163] 2018 N/A General Prospects of CCS Modeling to identify and review potential CCS barriers, with a focus on CCS costs; Buhr K. [44] 2014 N/A General Communication approaches, public engagement Qualitative to explore assumptions made about senders and receivers of information when involving the public in CCS communication and how these assumptions relate to different communication objectives. Carley S.R. [47] 2012 1001 Local Public perception Telephone survey, mail survey, interviews to examine early public impressions of CCS in a coal-intensive state, Indiana. Chaudhry R. (continued on next page) Funding statement [126] 2013 84 General Energy policy, stakeholders' perception Qualitative, Interviews to assess variation in the state-level energy context for CCS development by exploring energy policy stakeholders' perceptions of CCS in four geographically and demographically diverse states. Chen Z.-A. [9] 2014 679 General Public perception Questionnaire to assess public understanding of the climate sciences, society's knowledge and acceptance of low emission technologies, public interests and concerns about the positive and negative impacts of CCS technology, and public attitudes towards CCS policies supported by the government. Cherry T. L. [111] 2014 674 General Public perception, opposition Telephone interviews to provide insights to the origins of public opposition that can impede the adoption of low-carbon technologies by investigating how perceptions are shaped by local economic interests and individual cultural worldviews. de Best- Waldhober M. [119] 2012 971 General Preferences between technologies Mail survey, ICQ to measure informed opinions regard- ing CCS in comparison with other CO2 emission reduction options by combining valid and well-balanced information with a large sample that is representative of the Dutch public. de Best- Waldhober M. [113] 2008 1322 General Public perception, acceptance Information Choice Questionnaire to analyze the awareness and perception of the Dutch general public regarding CCS. de Best- Waldhober M. [33] 2012 846 General Public perception Questionnaire, interviews, media analysis to enhance insight into currently held beliefs and awareness among the general public about CCS; to investigate the role of the media as a vehicle for knowledge transfer. de Bruin W.B. [45] 2014 891 General Public perception, impression formation Online survey to learn more about how people respond to a validated educational communication about CCS. (continued on next page) to examine descriptive statistics to understand public perceptions of CCS and applied regression models to assess how risk perceptions, perspectives of climate change and trust in government relate to the support for or opposition to CCS development and funding for the technology. 16 16 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. (continued) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research De Coninck H. [141] 2009 N/A General International cooperation Qualitative to show the relevance of strong international cooperation on CCS demonstration. Desbarats J. Funding statement [125] 2010 N/A General Review Qualitative, Case study to describe the results of the European project NearCO2 ﬁrst phase, which focuses on lessons learned from CCS and analogous developments in recent years. Dowd A.-M. [38] 2014 2470 Cross- country Public perception, knowledge Online survey to address the gap around identifying what is the public's knowledge of CO2 properties, sources, uses and effects; to examine the relationship of existing knowledge on perceptions of CO2 and CCS; to explore the effect of information provision on knowledge and opinions of both CO2 and CCS. Duan H. [87] 2010 534 General Awareness, attitudes towards technology, determinants of acceptance Mail survey, online survey, interviews to explore the public's perspectives on the development of CCS in China. Duetschke E. [53] 2014 1830 General Public perception, preferences between CCS system elements Online survey to investigate the relevance of different speciﬁcations of the three main steps of CCS on the public perception of CCS as well as possible interactional effects between the speciﬁcations. Duetschke E. [155]. 2011 13 Local Comparison of projects experience Qualitative, Case study, Interviews to analyze and compare projects' properties, communication strategies, public perception, local context and history; to identify factors that contributed to the respective positive or negative reaction of the local public. Duetschke E. [37] 2016 1830 General Public perception, preferences between CCS system elements Online survey to examine the public perception of CCS in more detail by looking into different options within the CCS chain, i.e. for the three elements capture, transport and storage. Einsiedel E.F. [84] 2012 82 Local Public deliberations Workshop, questionnaire to examine citizens' views on climate change and a number of energy systems, with a speciﬁc focus on the use of carbon capture and storage (CCS) as a technology to address greenhouse gas emissions. Fischedick M. [4] 2009 232 General Public acceptance, stakeholders' opinion Media analysis, interviews to understand the relevance of technical and non-technical aspects of CCS in terms of social acceptance. Fleishman L.A. [112] 2010 60 General Preferences between technologies Workshops, ICQ to examine people's informed decisions about electricity-generating technologies. Gough C. [132] 2014 19 Local Public perception of CO2 pipeline Focus group to assess individuals' understanding of CO2 and identify their existing perceptions of it; to explore perceptions of risk and key areas of concern with respect to pipeline transportation of CO2 for the purposes of CCS. Gough C. (continued on next page) Funding statement [127] 2010 31 General Prospects of CCS in UK Qualitative, workshop to present the results of the workshop aimed at formulation of CCS long-term roadmap in UK taking into account opinion of wide range of stakeholders. Gough C. [36] 2017 Around 32 (incl. 10 stakeholders) General Social license Workshops, interviews, media analysis to summarise results from empirical research with the broad aim of exploring societal responses to CO2 storage, framed around the concept of social license to operate. Gough C. [95] 2018 12 interviews General Public perception Focus group, interviews to explore the social context for CO2 storage in the UK; to assess potential social responses to subsurface injection and site monitoring approaches; to identify signiﬁcant factors in establishing a social license in the context of CCS and in particular offshore CO2 storage in the UK. Gough C. [67] 2002 19 General Public perception Focus group to explore public reaction about burying CO2 under the sea. Ha-Duong M. [148] 2009 1076 General Public perception ICQ to explore awareness about CCS in France, and the degree of approval of or opposition to (continued on next page) 17 Heliyon 5 (2019) e02845 P. Tcvetkov et al. (continued) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research the variability of this opinion relative to the provision of information; to explore the variability of this opinion relative to the semantics used to describe the technology. Hansson A. [145] 2005 12 General Stakeholders' perception Interviews to examine the attitudes of Swedish politicians, scientists, NGOs and industry regarding CCS, i.e. actors who possess knowledge about CCS today and will inﬂuence the public opinion of tomorrow. Haug J.K. [135] 2016 N/A Cross- country Local acceptance Qualitative to assess the Nordic situation with regard to carbon capture and storage (CCS) deployment at the local level. Hope A.L.B. [82] 2014 20 General Role of religion in attitude towards CCS Focus groups, questionnaire to explore potential differences between the Muslim, Christian and secular participants in terms of pro-environmental values and beliefs; to explore how attitudes to CCS and climate change were shaped by religious beliefs. Howell R. [24] 2013 99 General Public perception Workshop to present the results of a large group process conducted in Edinburgh, Scotland investigating public perceptions of climate change and low-carbon energy technologies, speciﬁcally CCS. Itaoka K. Funding statement [34] 2009 2490 General Public perception Online survey to investigate the extent of recognition and latent social acceptance on global warming mitigation measures including CCS, as well as the kind of factors that would inﬂuence their views. Itaoka K. [91] 2016 548 General Public preferences between low-carbon technologies Door-to-door survey to examine the rationale for policy parity of basic low carbon power sources and the consumer preference for those power sources. Itaoka K. [66] 2014 1251 General Dependence between public perception and natural, and technogenic accidents Online survey to measure the inﬂuence of the large earthquakes and nuclear plant accidents on public perception of CCS. Johnsson F. [104] 2009 142 Cross- country Stakeholder perceptions Questionnaire to identify, study, and address non-technical issues associated with CCS from fossil-ﬁred plants in the energy sector, and to provide guidance to decision makers. Jones C.R. [121] 2014 16 General- Local Public perception of CO2 utilization technologies Focus group, ICQ to design and test a methodology for investigating public perceptions of CDU; to elucidate new understanding of people's attitudes towards the technology. Jouvet P.-A. [69] 2014 0 General Social acceptance, balance between tax and pollutions Modeling to determine, from the social point of view, simultaneously the amount of production as well as the optimal allocation of CO2 emissions between the atmosphere and underground storage sites. Kaiser M. [107] 2014 1006 General Public engagement, public perception Interviews, media analysis, focus group to analyze the local public perception of CCS among citizens and stakeholders; to inform community representatives and the local public about CCS technology and to involve them in the planning process for the prospective CCS project. Karayannis V. [92] 2014 N/A General Public perception, economic aspects Qualitative to discuss recent socio-economic aspects of CCS technologies. Karimi F. [3] 2017 19 Cross- country CCS policy Interviews to deﬁne temporal features (i.e. frame, timing, tempo, and duration) for policy making and deployment of large-scale CCS projects. Karimi F. [68] 2018 13901 Cross- country Public perception, cross- cultural differences Based on Eurobarometer to explain the importance and role of cross- cultural differences and the reaction of people in different countries towards the technology vis-a-vis the other factors and demonstrate how those differences operate. Karimi F. (continued on next page) Funding statement [72] 2015 19 General Experts' risk perception Interviews, Case study to contribute to the risk governance of CCS by investigating the concerns of experts about CCS and the role of socio-cultural factors in their risk perception. (continued on next page) to investigate the extent of recognition and latent social acceptance on global warming mitigation measures including CCS, as well as the kind of factors that would inﬂuence their views. (continued on next page) 18 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. (continued) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research Karimi F. [81] 2016 13901 Cross- country Inﬂuence of socio-cultural factors on risks and beneﬁts perception Based on Eurobarometer to explore how do cultural structures of a society affect beneﬁt and risk perception of CCS and what extent is the reaction of the public to implementation of the technology predictable in a crosscultural comparative framework. Karimi F. [73] 2014 13901 Cross- country Relation between cultural factors and risks perception Based on Eurobarometer to explore how do cultural structures of a society affect risk perception of CCS and what extent is the reaction of the public to implementation of the technology predictable. Klass A.B. [124] 2008 0 General Liability Qualitative to create a potential framework to address liability and funding issues associated with the long-term storage of CO2 in connection with CCS. Kraeusel J. [89] 2012 130 General Public acceptance, willingness to pay Online survey to explore social acceptance and willingness to pay for the Carbon Capture and Storage technology in Germany. Krause R.M. [54] 2014 1001 General Public perception, NIMBY Telephone–mail–telephone survey to examine how the closeness of a hypothetical CCS facility to individuals' communities inﬂuences their acceptance of it. Kubota H. [19] 2017 23612 General Public perception Online survey to analyze the attitudes and perception of CCS and thermal power generation through internet questionnaire surveys, and to provide appropriate information to promote public understanding and decision-making for introducing CCS technology for thermal power plants. Li Q. [118] 2014 679 General Public perception Questionnaire to explore public perception of CCUS in China. Li Q. [28] 2017 570 General Public perception Questionnaire to investigate the public awareness and understanding of the environmental impact and management of CCUS technology. Lupion M. (continued on next page) Funding statement Tcvetkov et al. evolve and get shaped in the social context comprising of the professionally involved actors, and how opinion formation of lay citizens and that of professionally involved actors interact. Offermann- van Heek J. [122] 2018 137 General Trust, public perception Interviews, online survey to investigate the connection between trust in CCU companies and the acceptance of innovative CCU products. Oltra C. [153] 2012 51–69 Local, cross- country Comparison of projects experience Qualitative, case study to examine the development of public reactions in relation to ﬁve European CO2 storage projects; to identify any lessons that may be learned for the future. Oltra C. [20] 2012 500 General Public perception Online survey to analyse how additional information on CCS affects individuals' reactions to CCS. Oltra C. [16] 2010 52 General Public perception Focus group to analyze the lay understandings and perceptions of CCS technologies and projects in Spain. Palmgren C. [137] 2004 144 General Public perception Questionnaire, interviews to explore likely public perceptions in the United States of CO2 disposal in deep rock formations and the ocean. Perdan S. [157] 2017 1213 General Public perception, awareness Online questionnaire to establish the extent of people's awareness and acceptance of CCUS and to elicit the importance they put on different sustainability issues relevant to CCUS. Pietzner K. [39] 2011 6168 Cross- country Public perception, awareness Questionnaire to summarise the results of public perception and awareness surveys in six European countries - Germany, Greece, the Netherlands, Norway, Romania and the United Kingdom (UK). Pihkola H. [144] 2017 0 General Sustainability of CCS Qualitative to discuss the sustainability of CCS technologies from a cross-disciplinary point of view. Prangnell M. [140] 2013 N/A General Public communications Qualitative, Case study to describe key aspects of CCS image crisis. Reiner D. [115] 2006 4009 Cross- country Public perception Paper-and-pencil survey, online survey to compare public attitudes in the United States, United Kingdom, Sweden and Japan towards key questions of energy and the environment, with particular emphasis on attitudes towards carbon capture and storage (CCS). Reiner D.M. [116] 2006 4009 Cross- country Preferences between technologies Online survey, questionnaire, telephone survey to explore difference in CCS perception, compared to other low carbon technologies, between countries. Riesch H. Funding statement [143] 2013 N/A Local Project experience Qualitative, Case study to describe the integral communication plan and public outreach strategy designed and implemented in the areas of inﬂuence of CIUDEN's large facilities on CCS. Mabon L. [26] 2013 23 General Public perception, offshore storage Qualitative, Interviews to challenge arguments that, due to the greater distances from centres of population, it will be ‘easier’ to garner public and stakeholder support for offshore CO2 storage than onshore. Mabon L. [86] 2013 72 Cross- country Public perception Qualitative, Interviews to illustrate how publics and stakeholders often evaluate the geological storage of carbon dioxide in terms of its relation to their broader world views, rather than purely in terms of the perceived techno-scientiﬁc risks of the technology. Mabon L. [146] 2015 N/A General Stakeholders' perception, policy Focus groups to overview the key aspects of CO2-EOR stakeholders interaction and perception of such projects. Malone E.L. [42] 2010 N/A General Stakeholders involvement Qualitative to discuss the issues involved in providing information as part of the CCS survey, maintaining that such information is never unbiased and thus tends to produce pseudo opinions that reﬂect the pollster's or researcher's bias. Markusson N. [2] 2012 N/A General Social dynamics of technology Qualitative to explore the role of social sciences in the development of CCS Midden C.J.H. [49] 2009 112 Local Trust, risk perception ICQ to analyze the role of trust and risk perception in attitudes formation towards CO2 storage. Miller E. [50] 2007 1273 General Socio-demographic differences, trust, knowledge Online survey to explore the extent to which socio- demographic characteristics inﬂuence knowledge, trust, risk perception and acceptance of CCS. Miller E. [51] 2008 1273 General Public perception Online survey to provide a benchmark of perceptions and initial reactions to geosequestration technologies in Australia; to explore some principles for the to illustrate how publics and stakeholders often evaluate the geological storage of carbon dioxide in terms of its relation to their broader world views, rather than purely in terms of the perceived techno-scientiﬁc risks of the technology. to overview the key aspects of CO2-EOR stakeholders interaction and perception of such projects. to discuss the issues involved in providing information as part of the CCS survey, maintaining that such information is never unbiased and thus tends to produce pseudo opinions that reﬂect the pollster's or researcher's bias. (continued on next page) 19 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Funding statement [62] 2013 942 Cross- country Public perception Online focus group discussion to discuss online focus groups as a deliberative method in experimental and perhaps consultative contexts; to show the role of anchoring and associative reasoning in the development of public opinion of CCS; to discuss the managing public-facing energy messaging in an age of public access to online information. Rychlicki S. [41] 2015 90 General Public perception, social acceptance Questionnaire to explore public sentiment associated with using CCS and CO2-EOR technologies in Poland. Sacuta N. [106] 2017 0 General Projects experience Qualitative, Case study to examine the public outreach enacted for three different CO2 injection projects to identify differences and similarities in the strategies employed for public dissemination of information. Sala R. [40] 2011 97 General Stakeholders' perception, social acceptance Online survey to report an empirical analysis of stakeholder perceptions on the risks, challenges and barriers facing CCS deployment in Spain. (continued on next page) (continued on next page) (continued on next page) 20 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. ( ) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research Schumann D. [70] 2012 N/A General Public acceptance Qualitative to overview methods of CCS acceptance research. Schumann D. [152] 2017 1000 General Public perception, pipelines Computer-aided telephone interviews to investigate the public perception of CO2 pipelines among the German public. Schumann D. [151] 2014 2003 General Public perception Interviews to investigate and compare the public perception of CO2 offshore storage, CO2 onshore storage and CO2 transport via pipeline in Germany nationwide and in two coastal regions. Seigo S.L. [29] 2013 30 General The role of illustrations Interviews to take a closer look at what constitutes a good illustration of CCS and how illustrations can impact perception of the technology. Seigo S.L. [57] 2014 1510 General- Local Risk and beneﬁt perception Online survey to explore if there are differences in terms of risk and beneﬁt perceptions of CCS between regions with different stages of CCS deployment. Seigo S.L. [75] 2011 200 General Public perception, communication Online survey to investigate the inﬂuence of information about monitoring measures at CO2 storage sites on laypeople's perceptions of CCS. Selma L. [5] 2014 N/A General Review Qualitative, Literature review to review and analyze public perception research. Shackley S. (continued on next page) Information-choice questionnaire, focus group Funding statement [22] 2005 212 General Public perception Panel discussion to explore public perceptions of carbon dioxide capture and storage, both when ﬁrst presented with the idea and when more background information is provided; to explore perceptions of the key risks and concerns surrounding CCS and what information, policies and processes would make CCS more and less acceptable to the public. Shackley S. [88] 2008 512 Cross- country Stakeholders' perception, public perception Questionnaire to analyze social acceptability on the part of both the lay public and stakeholders; to examine the acceptability of CO2 capture and geological storage within the Clean Development Mechanism of the Kyoto Protocol. Sharma S. [142] 2006 0 General- Local Project experience Qualitative, Case study to show an experience of solving a number of regulatory, organisational and social challenges which were occurred within the Otway Basin CCS Pilot Project. Sharp J.D. [74] 2009 1972 General Public perception of beneﬁts and risks, level of support Online survey to investigate the public's perceptions of the beneﬁts and risks of CCS, the likely determinants of public opinion, and overall support for the use of CCS. Stephens J.C. [79] 2009 100 General Public perception, learning Seminar, ICQ to explore stakeholders' perceptions of the risks and beneﬁts of CCS technology, and how those perceptions changed with additional information provided by CCS technology experts. Ter Mors E. [21] 2010 220 General Stakeholders collaboration ICQ to examine whether people expect more balanced information from diverging collaborating stakeholders than from individual stakeholders; to analyze social acceptability on the part of both the lay public and stakeholders; to examine the acceptability of CO2 capture and geological storage within the Clean Development Mechanism of the Kyoto Protocol. to show an experience of solving a number of regulatory, organisational and social challenges which were occurred within the Otway Basin CCS Pilot Project. to investigate the public's perceptions of the beneﬁts and risks of CCS, the likely determinants of public opinion, and overall support for the use of CCS. 21 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. (continued) The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection** Aim of the research Ter Mors E. Funding statement [129] 2009 N/A General Public information Qualitative, ICQ to examine whether public information would be more effective (i.e., perceived to be of greater value) when multiple stakeholders communicate information about CCS in collaboration instead of doing so separately. Terwel B.W. [103] 2008 393 General Trust Online survey to explore how organizational motives and organizational communications affect public trust in these organizations. Terwel B.W. [134] 2010 203 General Group voice, trustworthiness ICQ to examine whether group voice affects people's perceptions of the trustworthiness of the political decision maker; to examine whether perceived trustworthiness of the political decision maker inﬂuences acceptance of people's decision. Terwel B.W. [97] 2012 811 Local Public perception, opposition Telephone interviews to determine how widespread the local resistance against the proposed CCS project actually was at this point of time (it was before Barendrecht project rejection); to determine to what extent the CCS project was an important issue for the people of Barendrecht; to explain the local public's attitudes toward the CCS project. Terwel B.W. [101] 2010 148 General Trust, public perception Questionnaire to highlight public trust in CCS stakeholders as an inﬂuential factor of public acceptance. Terwel B.W. [55] 2012 205 General- Local Public perception, NIMBY Questionnaire to determine whether the psychological structure of initial attitudes towards plans for CO2 storage differs for people living in the direct vicinity of a proposed CO2 storage location and people who do not. Terwel B.W. [60] 2009 148 General Public acceptance, trust Questionnaire to examine the inﬂuence of competence- based and integrity-based trust on public acceptance. Toikka A. [133] 2014 25 General Social and political issues of CCS Interviews, workshops, media analysis, based of Eurobarometer to map societal issues of CCS adoption based on multiple data sets from two research projects in Finland, looking at social, cultural, and political issues. Tokushige K. [56] 2007 423 General Public perception, acceptance ICQ to analyze and evaluate through a factor analysis how the general public perceives the CO2 geological storage, what factors are crucial to their acceptance of the storage, and what kinds of information would inﬂuence the acceptance and its factors. Tokushige K. Funding statement [58] 2006 267 General Risk and beneﬁt perception, acceptance Questionnaire to analyze and evaluate how the general public perceives CO2 geological storage technology among other global warming mitigation technologies and what factors are crucial for its acceptance; to analyze and evaluate what kind of information would inﬂuence public acceptance and notiﬁcation. Trinh H.A.N. [147] 2015 16 General Experts' perception Qualitative, Interviews to summarize expert opinions regarding crucial factors that may inﬂuence Vietnam's future use of carbon capture and storage. Upham P. [117] 2011 56 Cross- country Public perception Questionnaire to determine European public perception of CCS. Upham P. [65] 2011 N/A Cross- country Public perception Focus groups, questionnaire to describe the methods and results of six focus groups in different countries, aimed at developing communication strategies and media that are designed to brieﬂy convey to stakeholders and the public the advantages and risks of CO2 capture and storage. van Alphen K. [35] 2007 N/A General Public perception, mass media inﬂuence Workshops, media analysis, interviews to describe an extensive study on the acceptance of CCS by stakeholders in the Netherlands; to explore the inﬂuence of the way the Dutch press perceives and portrays CCS on the acceptance of CCS by the lay public. (continued on next page) to analyze and evaluate through a factor analysis how the general public perceives the CO2 geological storage, what factors are crucial to their acceptance of the storage, and what kinds of information would inﬂuence the acceptance and its factors. (continued on next page) (continued on next page) 22 Heliyon 5 (2019) e02845 P. Tcvetkov et al. P. Tcvetkov et al. The ﬁrst author/author Year Sample (number of respondents) Level* Focus Method of data collection Aim of the research van Os H.W.A. [131] 2013 55 General- Local Public perception, role of stakeholders, NIMBY Interviews to explore the role of stakeholders responsibilities in the process of CCS project implementation. Vercelli S. [15] 2013 N/A General Review Qualitative, Literature review to offer an overview of research bodies and provide useful criteria for its exploitation with regard to the interaction between information provision and public perception of CCS. Vercelli S. [83] 2009 13 school classes, 13 researchers General Public perception, culture Interviews, classes to explore social representations and cultural models that could facilitate or hinder the necessary decisions for the implementation of Carbon Capture and Storage. V€ogele S. Funding statement General level includes surveys of all scale in one country, theoretical studies and reports. Questionnaire means paper-and-pencil questionnaire. Funding statement [150] 2018 0 General Comparison of low carbon alternatives Multi-criteria analysis, scenario assessment to highlight reasons for CCS support descending in Germany and other European countries. Wade S. [161] 2011 N/A General Social site characterization Qualitative to advance understanding of the subsurface and the technical, social and legal aspects of CCS. Wallquist L. [138] 2010 654 General Beneﬁt and risk perception Mail survey to quantify laypeople's perception of critical aspects of CCS and to examine their impact on perceived beneﬁts and perceived risks. Wallquist L. [52] 2011 139 General Preferences between CCS system elements, NIMBY Online survey to examine public preferences for the characteristics of the elements capture, transport, and storage in combination. Wallquist L. [63] 2012 769 General Trust, convictions, protest potential Mail questionnaire to examine the roles of trust and convictions for public attitude towards CCS to test a model that explains the public's protest potential against CCS. Wallquist L. [23] 2011 297 General Beneﬁt and risk perception Mail survey to examine whether comprehensive information about CCS can have any inﬂuence on perceived risks and beneﬁts. Wallquist L. [61] 2009 16 General Public perception Interviews to study how laypeople perceive CCS and which cognitions they hold with respect to this technique. Wallquist L. [30] 2011 63 General Public perception, knowledge Questionnaire to examine antecedents of risk and beneﬁt perception of CCS by means. Weber V. [158] 2018 N/A General Review Qualitative to review CCS Directive of the European Union. Wong-Parodi G. [98] 2009 14 Local Public perception Focus group to explore the views of communities that may be directly impacted by the siting of CCS. Wong-Parodi G. [59] 2011 59 General- Local Public perception Interviews to understand how to inﬂuence citizens' attitudes toward CCS in regions with signiﬁcant dependence on energy sector. Yang L. [64] 2016 349 General Trust, public perception Online survey, interviews to explore the factors affecting public acceptance of CCS technologies in China. Yu H. [80] 2018 2080 General Preferences between technologies Online questionnaire to study public attitudes towards different low-carbon energy technologies, using nuclear power, CCS and wind energy as examples. * Local level includes studies, based on areas with already started process of CCS project execution. General level includes surveys of all scale in one country, theoretical studies and reports. * Local level includes studies, based on areas with already started process of CCS project execution. Appendix 2. * Reference Kn Acc BR GovS SD Tr Aw NIMBY Will Anghel S. [46] Ashworth P. [18] Ashworth P. [31] Duetschke E. [53] (continued on next page) Reference Kn Acc BR GovS SD Tr Aw NIMBY Will Anghel S. [46] Ashworth P. [18] Ashworth P. [31] Duetschke E. [53] (continued on next page) 23 Reference Kn Acc BR GovS SD Tr Aw NIMBY Will Anghel S. [46] Ashworth P. [18] Ashworth P. [31] Duetschke E. [53] (continued on next page) 23 P. Tcvetkov et al. Heliyon 5 (2019) e02845 (continued) Reference Kn Acc BR GovS SD Tr Aw NIMBY Will Ter Mors E. [32] þ þ Fleishman L.A. [112] Hansson A. [145] Rychlicki S. [41] Schumann D. [151] Wallquist L. [23] þ þ þ Breukers S. [77] Brunsting S. [94] Buhr K. [44] Gough C. [95] þ þ þ Karayannis V. [92] Prangnell M. [140] Riesch H. [62] Yang L. [64] þ þ þ þ þ Ashworth P. [105] Mabon L. [86] Sharma S. [142] þ þ Chen Z.-A. [9] Li Q. [118] Wallquist L. [61] þ þ þ þ þ þ de Best-Waldhober M. [119] Itaoka K. [34] Jones C.R. [121] þ þ þ Gough C. [127] Karimi F. [3] Trinh H.A.N. [147] þ Lupion M. [143] Wallquist L. [63] Wong-Parodi G. [59] þ þ þ þ Ter Mors E. [129] Terwel B.W. [101] Terwel B.W. [103] þ þ Anderson C. [120] Oltra C. [153] þ þ þ þ þ Ashworth P. [99] Shackley S. [22] þ þ þ þ þ Boyd A.D. [139] Bradbury J. [76] þ þ þ þ þ Carley S.R. [47] Yu H. [80] þ þ þ þ þ Cherry T. L. [111] Mabon L. [26] þ þ þ þ de Best-Waldhober M. [113] Kraeusel J. [89] þ þ þ þ þ de Best-Waldhober M. [33] Ha-Duong M. [148] þ þ þ þ de Coninck H. [141] Reiner D.M. [116] þ þ þ þ Desbarats J. [125] Terwel B.W. [97] þ þ þ þ þ þ þ þ Duetschke E. [155] Gough C. [36] þ þ þ Duetschke E. [37] Pietzner K. [39] þ þ þ Einsiedel E.F. [84] Tokushige K. [56] þ þ þ þ þ þ þ Mabon L. [146] Pihkola H. [144] þ þ Midden C.J.H. [49] (continued on next page) P. Tcvetkov et al. Heliyon 5 (2019) e02845 (continued on next page) 24 Heliyon 5 (2019) e02845 P. Tcvetkov et al. Appendix 2. * (continued) Reference Kn Acc BR GovS SD Tr Aw NIMBY Will Offermann-van Heek J. [122] þ þ Perdan S. [157] Wong-Parodi G. [98] þ þ þ þ Sacuta N. [106] Shackley S. [88] þ þ þ Arning K. [156] þ þ Ashworth P. [114] þ þ Billson M. [128] þ þ þ Braun C. [149] þ þ þ þ Brunsting S. [160] þ þ þ þ þ þ Budinis S. [163] þ þ Chaudhry R. [126] þ þ þ þ de Bruin W.B. [45] þ Dowd A.-M. [38] þ þ þ þ þ þ Duan H. [87] þ þ þ þ þ Fischedick M. [4] þ þ þ Gough C. [132] þ þ þ þ þ Gough C. [67] þ þ þ þ Haug J.K. [135] þ þ þ þ þ Hope A.L.B. [82] þ þ þ þ Howell R. [24] þ þ þ þ þ þ Itaoka K. [66] þ þ Itaoka K. [91] þ þ Johnsson F. [104] þ þ þ Jouvet P.-A. [69] þ Kaiser M. [107] þ þ þ þ Karimi F. [68] þ þ þ þ þ þ Karimi F. [72] þ þ þ þ þ þ Karimi F. [73] þ þ þ þ Karimi F. [81] þ þ þ Klass A.B. [124] þ þ Krause R.M. [54] þ þ Kubota H. [19] þ þ þ þ Li Q. [28] þ þ þ þ þ Malone E.L. [42] þ þ þ þ þ Markusson N. [2] þ þ þ þ þ Miller E. [50] þ þ þ þ þ Miller E. [51] þ þ þ þ þ þ Moutenet J.-P. [43] þ þ þ þ þ Nicole M.A. [108] þ þ þ þ þ Oltra C. [16] þ þ þ þ Oltra C. [20] þ þ þ þ Palmgren C. [137] þ þ þ þ þ þ Perdan S. [157] þ þ þ þ Reiner D. [115] þ þ þ þ Sala R. [40] þ þ þ þ Seigo S.L. [29] þ þ þ Seigo S.L. [57] þ þ þ þ þ þ Seigo S.L. [75] þ þ þ Sharp J.D. [74] þ þ þ þ þ þ þ Stephens J.C. [79] þ þ þ þ Ter Mors E. [21] þ þ þ Terwel B.W. [134] þ þ þ Terwel B.W. [55] þ þ þ þ þ þ Terwel B.W. [60] þ þ þ Toikka A. [133] þ þ þ Tokushige K. References [1] P. Tcvetkov, A. Cherepovitsyn, Prospects of CCS projects implementation in Russia: environmental protection and economic opportunities, J. Ecol. Eng. 17 (Issue 2) (2016) 24–32. [25] P. Slovic, M.L. Finucane, E. Peters, D.G. MacGregor, The affect heuristic, Eur. J. Oper. Res. 177 (2007) 1333–1352. [26] L. Mabon, S. Shackley, N. Bower-Bir, Perceptions of sub-seabed carbon dioxide storage in Scotland and implications for policy: a qualitative study, Mar. Policy 45 (2014) 9–15. [2] N. Markusson, S. Shackley, et al. (Eds.), The Social Dynamic of Carbon Capture and Storage: Understanding CCS Representations, Governance and Innovation, Routledge, Taylor and Francis Group, EarthScan, 2012. [27] G. Gigerenzer, Why heuristics work, Perspect. Psychol. Sci. 3 (1) (2008) 20–29 [28] Q Li G Li G L L C Li B C i Z A Ch A i l f bl [3] F. Karimi, Timescapes of CCS projects: is deferring projects and policies just kicking the can down the road? Energy Procedia 114 (2017) 7317–7325. [28] Q. Li, G. Liu, G. Leamon, L.C. Liu, B. Cai, Z.A. Chen, A national survey of public awareness of the environmental impact and management of CCUS technology in China, Energy Procedia 114 (2017) 7237–7244. [4] M. Fischedick, K. Pietzner, N. Supersberger, A. Esken, W. Kuckshinrichs, P. Zapp, E. Gruber, Stakeholder Acceptance of Carbon Capture and Storage in Germany, 2009. gy [29] S. L'Orange Seigo, S. Dohle, M. Siegrist, The effect of ﬁgures in CCS communication Int J Greenh Gas Contr 16 (2013) 83 90 [29] S. L'Orange Seigo, S. Dohle, M. Siegrist, The effect of ﬁgures in CCS communication, Int. J. Greenh. Gas Contr. 16 (2013) 83–90. [5] L. Selma, O. Seigo, S. Dohle, M. Siegrist, Public perception of carbon capture and storage (CCS): a review, Renew. Sustain. Energy Rev. 38 (2014) 848–863. [30] L. Wallquist, V.H. Visschers, M. Siegrist, Antecedents of risk and beneﬁt perception of CCS, Energy Procedia 4 (2011) 6288–6291. [6] E. Boyd, Governing the Clean Development Mechanism: global rhetoric versus local realities in carbon sequestration projects, Environ. Plan. 41 (10) (2009) 2380–2395. [31] P. Ashworth, S. Carr-Cornish, N. Boughen, K. Thambimuthu, Engaging the public on carbon dioxide capture and storage: does a large group process work? Energy Procedia 1 (1) (2009) 4765–4773. [7] V.H.M. Visschers, C. Keller, M. Siegrist, Climate change beneﬁts and energy supply beneﬁts as determinants of acceptance of nuclear power stations: investigating an explanatory model, Energy Policy 39 (2011) 3621–3629. [32] E. References ter Mors, B.W. Terwel, D.D. Daamen, D.M. Reiner, D. Schumann, S. Anghel, et al., A comparison of techniques used to collect informed public opinions about CCS: opinion quality after focus group discussions versus information-choice questionnaires, Int. J. Greenh. Gas Contr. 18 (2013) 256–263. [8] V.H.M. Visschers, M. Siegrist, Fair play in energy policy decisions: procedural fairness, outcome fairness and acceptance of the decision to rebuild nuclear power plants, Energy Policy 46 (2012) 292–300. [33] M. de Best-Waldhober, S. Brunsting, M. Paukovic, Public concepts of CCS: understanding of the Dutch general public and its reﬂection in the media, Int. J. Greenh. Gas Contr. 11 (2012) S139–S147. [9] Z.A. Chen, Q. Li, L.C. Liu, X. Zhang, L. Kuang, L. Jia, G. Liu, A large national survey of public perceptions of CCS technology in China, Appl. Energy 158 (2015) 366–377. [34] K. Itaoka, Y. Okuda, A. Saito, M. Akai, Inﬂuential information and factors for social acceptance of CCS: the 2nd round survey of public opinion in Japan, Energy Procedia 1 (1) (2009) 4803–4810. [10] J.J. Cohen, J. Reichl, M. Schmidthaler, Re-focussing research efforts on the public acceptance of energy infrastructure: a critical review, Energy 76 (2014) 4–9. [35] K. Van Alphen, Q.V.V. tot Voorst, M.P. Hekkert, R.E. Smits, Societal acceptance of carbon capture and storage technologies, Energy Policy 35 (8) (2007) 4368–4380. [11] E. Pol, A. Di Masso, A. Castrechini, M.R. Bonet, T. Vidal, Psychological parameters to understand and manage the NIMBY effect, Revue Europeenne de Psychologie Appliquee/Eur. Rev. Appl. Psychol. 56 (1) (2006) 43–51. [36] C. Gough, R. Cunningham, S. Mander, Societal responses to CO2 storage in the UK: media, stakeholder and public perspectives, Energy Procedia 114 (2017) 7310–7316. [12] D.M. Rousseau, S.B. Sitkin, R.S. Burt, C. Camerer, Not so different after all: a cross- discipline view of trust, Acad. Manag. Rev. 23 (1998) 393–404. [37] E. Dütschke, K. Wohlfarth, S. H€oller, P. Viebahn, D. Schumann, K. Pietzner, Differences in the public perception of CCS in Germany depending on CO2 source, transport option and storage location, Int. J. Greenh. Gas Contr. 53 (2016) 149–159. [13] J. Rowley, F. Slack, Conducting a literature review, Manag. Res. News 27 (6) (2004) 31–39. [14] Statement by three national academies (Academie des Sciences, Leopoldina and Royal Society) on good practice in the evaluation of researchers and research programmes. URL: http://www.academie-sciences.fr/pdf/rapport/avis111217.p df (accessed: 14.09.2018). [38] A.M. Dowd, K. Itaoka, P. Ashworth, A. Saito, M. References de Best-Waldhober, Investigating the link between knowledge and perception of CO2 and CCS: an international study, Int. J. Greenh. Gas Contr. 28 (2014) 79–87. [15] S. Vercelli, J. Anderlucci, R. Memoli, N. Battisti, L. Mabon, S. Lombardi, Informing people about CCS: a review of social research studies, Energy Procedia 37 (2013) 7464–7473. [39] K. Pietzner, D. Schumann, S.D. Tvedt, R. Næss, D.M. Reiner, S. Anghel, et al., Public Awareness and Perceptions of Carbon Dioxide Capture and Storage (CCS): Insights from Surveys Administered to Representative Samples in Six European Countries, 2011. [16] C. Oltra, R. Sala, R. Sola, M. Di Masso, G. Rowe, Lay perceptions of carbon capture and storage technology, Int. J. Greenh. Gas Contr. 4 (4) (2010) 698–706. [40] R. Sala, C. Oltra, Experts’ attitudes towards CCS technologies in Spain, Int. J. Greenh. Gas Contr. 5 (5) (2011) 1339–1345. [17] A. Prades, T. Horlick-Jones, C. Oltra, R. Sola`, Lay perceptions of nuclear fusion: multiple modes of understanding, Sci. Public Policy 35 (2) (2008) 95–105. [41] S. Rychlicki, P. Kosowski, J. Wartak, M. Solecki, Social acceptance for CO2-EOR and CCS projects based on survey conducted in southeastern Poland, AGH Drilling, Oil, Gas 32 (2015). [18] P. Ashworth, E. Einsiedel, R. Howell, S. Brunsting, N. Boughen, A. Boyd, et al., Public preferences to CCS: how does it change across countries? Energy Procedia 37 (2013) 7410–7418. [42] E.L. Malone, J.J. Dooley, J.A. Bradbury, Moving from misinformation derived from public attitude surveys on carbon dioxide capture and storage towards realistic stakeholder involvement, Int. J. Greenh. Gas Contr. 4 (2) (2010) 419–425. [19] H. Kubota, A. Shimota, How should information about CCS be shared with the Japanese public? Energy Procedia 114 (2017) 7205–7211. l l h ﬂ f f d d l ' [20] C. Oltra, R. Sala, A. Boso, The inﬂuence of information on individuals' reactions to CCS technologies: results from experimental online survey research, Greenh. Gases: Sci. Technol. 2 (3) (2012) 209–215. [43] J.-P. Moutenet, K. Bedard, M. Malo, Public awareness and opinion on CCS in the province of Quebec, Canada, Greenh. Gases Sci. Technol. 2 (2012) 126–135. [21] E. ter Mors, M.W.H. Weenig, N. Ellemers, D.D.L. Daamen, Effective communication about complex environmental issues: perceived quality of information about carbon dioxide capture and storage (CCS) depends on stakeholder collaboration, J. Environ. Psychol. 30 (2010) 347–357. [44] K. Buhr, V. Appendix 2. * [58] þ þ þ Upham P. [117] þ þ þ þ þ þ þ þ Upham P. [65] þ þ þ þ þ (continued on next page) 25 25 (continued) Reference Kn Acc BR GovS SD Tr Aw NIMBY Will van Alphen K. [35] þ þ þ þ van Os H.W.A. [131] þ þ þ Vercelli S. [83] þ þ þ þ V€ogele S. [150] þ þ þ þ þ Wade S. [161] þ þ þ Wallquist L. [138] þ þ þ þ þ Wallquist L. [30] þ þ þ Wallquist L. [52] þ þ þ þ Kn – Knowledge, Acc – Acceptance of CCS and preference between technologies, BR - Beneﬁts and Risks perception, GovS - Governmental Policy and Interaction between Stakeholders, SD - Socio-demographic factors, Tr – Trust, Aw – Awareness, Will - Willingness to pay. P. Tcvetkov et al. Heliyon 5 (2019) e02845 Heliyon 5 (2019) e02845 P. Tcvetkov et al. Kn – Knowledge, Acc – Acceptance of CCS and preference between technologies, BR - Beneﬁts and Risks perception, GovS - Governmental Policy and Interaction between Stakeholders, SD - Socio-demographic factors, Tr – Trust, Aw – Awareness, Will - Willingness to pay. [24] R. Howell, S. Shackley, L. Mabon, P. Ashworth, T. Jeanneret, Engaging the public with low-carbon energy technologies: results from a Scottish large group process, Energy Policy 66 (2014) 496–506. [24] R. Howell, S. Shackley, L. Mabon, P. Ashworth, T. Jeanneret, Engaging the public with low-carbon energy technologies: results from a Scottish large group process, Energy Policy 66 (2014) 496–506. References Duan, The public perspective of carbon capture and storage for CO 2 emission reductions in China, Energy Policy 38 (2010) 5281–5289. [58] K. Tokushige, K. Akimoto, T. Tomoda, Public acceptance and risk-beneﬁt perception of CO 2 geological storage for global warming mitigation in Japan, Mitig. Adapt. Strategies Glob. Change 12 (2007) 1237–1251. [88] S. Shackley, D. Reiner, P. Upham, H. de Coninck, G. Sigurthorsson, J. Anderson, The acceptability of CO2 capture and storage (CCS) in Europe: an assessment of the key determining factors: Part 2. The social acceptability of CCS and the wider impacts and repercussions of its implementation, Int. J. Greenh. Gas Contr. 3 (3) (2009) 344–356. Mitig. Adapt. Strategies Glob. Change 12 (2007) 1237–1251. [59] G. Wong-Parodi, H. Dowlatabadi, T. McDaniels, X. Isha, Inﬂuencing attitudes toward carbon capture and sequestration: a social marketing approach, Environ. Sci. Technol. 45 (2011) 6743–6751. [89] J. Kraeusel, D. M€ost, Carbon Capture and Storage on its way to large-scale deployment: social acceptance and willingness to pay in Germany, Energy Policy 49 (2012) 642–651. [60] B.W. Terwel, F. Harinck, N. Ellemers, D.D. Daamen, Competence-based and integrity-based trust as predictors of acceptance of carbon dioxide capture and storage (CCS), Risk Anal.: Int. J. 29 (8) (2009) 1129–1140. [90] S. Eurobarometer, Europeans’ attitudes towards climate change, Eur. Comm. 29 (2009) 30. [61] L. Wallquist, V.H. Visschers, M. Siegrist, Lay concepts on CCS deployment in Switzerland based on qualitative interviews, Int. J. Greenh. Gas Contr. 3 (5) (2009) 652–657. [91] K. Itaoka, A. Saito, M. Akai, Policy parity for CCS?–Public preference on low carbon electricity, Energy Procedia 114 (2017) 7573–7580. [62] H. Riesch, C. Oltra, A. Lis, P. Upham, M. Pol, Internet-based public debate of CCS: lessons from online focus groups in Poland and Spain, Energy Policy 56 (2013) 693–702. [92] V. Karayannis, G. Charalampides, E. Lakioti, Socio-economic aspects of CCS technologies, Procedia Economics and Finance 14 (2014) 295–302. [93] W.M. Hanemann, Willingness to pay and willingness to accept: how much can they differ? Am. Econ. Rev. 81 (3) (1991) 635–647. [63] L. Wallquist, V.H. Visschers, S. Dohle, M. Siegrist, The role of convictions and trust for public protest potential in the case of carbon dioxide capture and storage (CCS), Hum. Ecol. Risk Assess. Int. J. 18 (4) (2012) 919–932. y [94] S. Brunsting, J. Desbarats, M. de Best-Waldhober, E. Duetschke, C. Oltra, h i h h bli d h i f i i [94] S. References Wibeck, Communication approaches for carbon capture and storage: underlying assumptions of limited versus extensive public engagement, Energy Res. Soc. Sci. 3 (2014) 5–12. [45] W. Bruine de Bruin, G. Wong-Parodi, The role of initial affective impressions in responses to educational communications: the case of carbon capture and sequestration (CCS), J. Exp. Psychol. Appl. 20 (2) (2014) 126. [22] S. Shackley, C. McLachlan, C. Gough, The Public Perceptions of Carbon Capture and Storage, Tyndall Centre for Climate Change Research, Manchester, UK, 2004. [23] L. Wallquist, V.H. Visschers, S. Dohle, M. Siegrist, Adapting communication to the public's intuitive understanding of CCS, Greenhouse Gases: Sci. Technol. 1 (1) (2011) 83–91. [46] S. Anghel, Impact of CCS communication on the general and local public in Romania-oltenia region, Energy Procedia 114 (2017) 7343–7347. 26 P. Tcvetkov et al. Heliyon 5 (2019) e02845 [47] S.R. Carley, R.M. Krause, D.C. Warren, J.A. Rupp, J.D. Graham, Early public impressions of terrestrial carbon capture and storage in a coal-intensive state, Environ. Sci. Technol. 46 (2012) 7086–7093. [77] S. Breukers, P. Upham, Organisational aspects of public engagement in European energy infrastructure planning: the case of early-stage CCS projects, J. Environ. Plan. Manag. 58 (2) (2015) 252–269. g [78] B.C. Farhar, C.T. Unseld, R. Vories, R. Crews, Public opinion about energy, Annu. Rev. Energy 5 (1980) 141–172. [48] K. Itaoka, A. Saito, Akai M Public acceptance of CO2 capture and storage technology: a survey of public opinion to explore inﬂuential factors, in: Paper Presented for the 7th International Conference on Greenhouse Gas Control Technologies. Canada, Vancouver, September, 5–9, 2004. [79] J.C. Stephens, J. Bielicki, G.M. Rand, Learning about carbon capture and storage: changing stakeholder perceptions with expert information, Energy Procedia 1 (1) (2009) 4655–4663. Technologies. Canada, Vancouver, September, 5–9, 2004. [49] C.J.H. Midden, N.M.A. Huijts, The role of trust in the affective evaluation of novel risks: the case of CO 2 storage, Risk Anal. 29 (2009) 743–751. [80] H. Yu, D. Reiner, H. Chen, Z. Mi, A Comparison of Public Preferences for Different Low-Carbon Energy Technologies: Support for CCS, Nuclear and Wind Energy in the United Kingdom, 2018. [50] E. Miller, L. Bell, L. Buys, Public understanding of carbon sequestration in Australia: socio-demographic predictors of knowledge, engagement and trust, Aust. J. Emerg. Technol. Soc. 5 (2007) 15–33. [81] F. Karimi, A. Toikka, J.I. Hukkinen, Comparative socio-cultural analysis of risk perception of carbon capture and storage in the European union, Energy Res. References Soc. Sci. 21 (2016) 114–122. [51] E. Miller, J. Summerville, L. Buys, L. Bell, Initial public perceptions of carbon geosequestration: implications for engagement and environmental risk communication strategies, Int. J. Glob Environ. Issues (2008) 147–164. [82] A.L. Hope, C.R. Jones, The impact of religious faith on attitudes to environmental issues and Carbon Capture and Storage (CCS) technologies: a mixed methods study, Technol. Soc. 38 (2014) 48–59. [52] L. Wallquist, S.L.O. Seigo, V.H. Visschers, M. Siegrist, Public acceptance of CCS system elements: a conjoint measurement, Int. J. Greenh. Gas Contr. 6 (2012) 77–83. [83] S. Vercelli, S. Lombardi, CCS as part of a global cultural development for environmentally sustainable energy production, Energy Procedia 1 (1) (2009) 4835–4841. [53] E. Duetschke, D. Schumann, K. Pietzner, K. Wohlfarth, S. H€oller, Does it Make a Difference to the Public where CO2 Comes from and where it Is Stored?: an Experimental Approach to Enhance Understanding of CCS Perceptions, 2014. l h ( d ) [84] E.F. Einsiedel, A.D. Boyd, J. Medlock, P. Ashworth, Assessing socio-technical mindsets: public deliberations on carbon capture and storage in the context of energy sources and climate change, Energy Policy 53 (2013) 149–158. [54] R.M. Krause, S.R. Carley, D.C. Warren, J.A. Rupp, J.D. Graham, “Not in (or under) my backyard”: geographic proximity and public acceptance of carbon capture and storage facilities, Risk Anal. 34 (3) (2014) 529–540. [85] J. Barnett, K. Burningham, G. Walker, N. Cass, Imagined publics and engagement around renewable energy technologies in the UK, Public Underst. Sci. 21 (2012) 36–50. g [55] B.W. Terwel, D.D.L. Daamen, Initial public reactions to carbon capture and storage (CCS): differentiating general and local views, Clim. Policy 12 (2012) 288–300. [ 6] k hi Aki d bli i h f [86] L. Mabon, S. Vercelli, S. Shackley, J. Anderlucci, N. Battisti, C. Franzese, K. Boot, ‘Tell me what you think about the geological storage of carbon dioxide’: towards a fuller understanding of public perceptions of CCS, Energy Procedia 37 (2013) 7444–7453. [56] K. Tokushige, K. Akimoto, T. Tomoda, Public perceptions on the acceptance of geological storage of carbon dioxide and information inﬂuencing the acceptance, Int. J. Greenh. Gas Contr. 1 (1) (2007) 101–112. [57] S.L.O. Seigo, J. Arvai, S. Dohle, M. Siegrist, Predictors of risk and beneﬁt perception of carbon capture and storage (CCS) in regions with different stages of deployment, Int. J. Greenh. Gas Contr. 25 (2014) 23–32. [87] H. References Brunsting, J. Desbarats, M. de Best-Waldhober, E. Duetschke, C. Oltra, P. Upham, H. Riesch, The public and CCS: the importance of communication and participation in the context of local realities, Energy Procedia 4 (2011) 6241–6247. P. Upham, H. Riesch, The public and CCS: the importance of communication and participation in the context of local realities, Energy Procedia 4 (2011) 6241–6247. [64] L. Yang, X. Zhang, K.J. McAlinden, The effect of trust on people's acceptance of CCS (carbon capture and storage) technologies: evidence from a survey in the People's Republic of China, Energy 96 (2016) 69–79. [95] C. Gough, R. Cunningham, S. Mander, Understanding key elements in establishing a social license for CCS: an empirical approach, Int. J. Greenh. Gas Contr. 68 (2018) 16–25. [65] P. Upham, T. Roberts, Public perceptions of CCS in context: results of NearCO2 focus groups in the UK, Belgium, The Netherlands, Germany, Spain and Poland, Energy Procedia 4 (2011) 6338–6344. [96] N.M.A. Huijts, E.J.E. Molin, L. Steg, Psychological factors inﬂuencing sustainable energy technology acceptance: a review-based comprehensive framework, Renew. Sustain. Energy Rev. 16 (2012) 525–531. [66] K. Itaoka, A. Saito, A.M. Dowd, M. de Best Waldhober, P. Ashworth, Inﬂuence of the large earthquake and nuclear plant accident on perception of CCS, Energy Procedia 63 (2014) 7133–7140. [97] B.W. Terwel, E. ter Mors, D.D.L. Daamen, It's not only about safety: beliefs and attitudes of 811 local residents regarding a ccs project in Barendrecht, Int. J. Greenh. Gas Contr. 9 (2012) 41–51. [67] C. Gough, I. Taylor, S. Shackley, Burying carbon under the sea: an initial exploration of public opinions, Energy Environ. 13 (6) (2002) 883–900. [98] G. Wong-Parodi, I. Ray, Community perceptions of carbon sequestration: insights from California, Environ. Res. Lett. 4 (3) (2009), 034002. [68] F. Karimi, A. Toikka, General public reactions to carbon capture and storage: does culture matter? Int. J. Greenh. Gas Contr. 70 (2018) 193–201. [99] P. Ashworth, N. Boughen, M. Mayhew, F. Millar, An integrated roadmap of communication activities around carbon capture and storage in Australia and beyond, Energy Procedia 1 (1) (2009) 4749–4756. [69] P.A. Jouvet, M. Renner, Social acceptance and optimal pollution: CCS or tax? Environ. Model. Assess. 20 (4) (2015) 285–302. [70] D. Schumann, Public acceptance, in: W. Kuckshinrichs, J.-F. Hake (Eds.), Carbon Capture, Storage and Use – Technical, Economic, Environmental and Societal Perspectives, Springer International Publishing, Cham, 2014. [100] K. B€ackstrand, J. Meadowcroft, M. References Tcvetkov, Overview of the prospects for developing a renewable energy in Russia, in: Proceedings of 2017 International Conference on Green Energy and Applications, ICGEA, 2017, pp. 113–118. [107] M. Kaiser, R. Zimmer, S. Brunsting, J. Mastop, M. Pol, Development of CCS projects in Poland. How to communicate with the local public? Energy Procedia 51 (2014) 267–273. [137] C.R. Palmgren, G.M. Morgan, W. Bruine de Bruin, D.W. Keith, Initial public perceptions of deep geological and oceanic disposal of carbon dioxide, Environ. Sci. Technol. 38 (2004) 6441–6450. [108] N.M. Huijts, C.J. Midden, A.L. Meijnders, Social acceptance of carbon dioxide storage, Energy Policy 35 (5) (2007) 2780–2789. [138] L. Wallquist, V.H.M. Visschers, M. Siegrist, Impact of knowledge and misconceptions on beneﬁt and risk perception of ccs, Environ. Sci. Technol. 44 (2010) 6557–6562. [109] N.M.A. Huijts, E.J.E. Molin, L. Steg, Psychological factors inﬂuencing sustainable energy technology acceptance: a review-based comprehensive framework, Renew. Sustain. Energy Rev. 16 (1) (2012) 525–531. [139] A.D. Boyd, J.D. Hmielowski, P. David, Public perceptions of carbon capture and storage in Canada: results of a national survey, Int. J. Greenh. Gas Contr. 67 (2017) 1–9. [110] S. Batel, P. Devine-Wright, T. Tangeland, Social acceptance of low carbon energy and associated infrastructures: a critical discussion, Energy Policy 58 (2013) 1–5. [140] M. Prangnell, Communications for Carbon Capture and Storage: Identifying the Beneﬁts, Managing Risk and Maintaining the Trust of Stakeholders, Global CCS Institute, Canberra, 2013. [111] T.L. Cherry, J.H. García, S. Kallbekken, A. Torvanger, The development and deployment of low-carbon energy technologies: the role of economic interests and cultural worldviews on public support, Energy Policy 68 (2014) 562–566. [ ] l i h i f d bli f f [141] H. de Coninck, J.C. Stephens, B. Metz, Global learning on carbon capture and storage: a call for strong international cooperation on CCS demonstration, Energy Policy 37 (6) (2009) 2161–2165. [112] L.A. Fleishman, W.B. De Bruin, M.G. Morgan, Informed public preferences for electricity portfolios with CCS and other low-carbon technologies, Risk Anal.: Int. J. 30 (9) (2010) 1399–1410. y [142] S. Sharma, P. Cook, S. Robinson, C. Anderson, Regulatory challenges and managing public perception in planning a geological storage pilot project in Australia, Int. J. Greenh. Gas Contr. 1 (2) (2007) 247–252. [113] M. de Best-Waldhober, D. Daamen, A. Faaij, Informed and uninformed public opinions on CO 2 capture and storage technologies in The Netherlands, Int. J. Greenh. Gas Contr. 3 (2009) 322–332. References Oppenheimer, The Politics and Policy of Carbon Capture and Storage: Framing an Emergent Technology, 2011. [71] P. Ashworth, S. Wade, D. Reiner, X. Liang, Developments in public communications on CCS, Int. J. Greenh. Gas Contr. 40 (2015) 449–458. [101] B.W. Terwel, F. Harinck, N. Ellemers, D.D. Daamen, Going beyond the properties of CO2 capture and storage (CCS) technology: how trust in stakeholders affects public acceptance of CCS, Int. J. Greenh. Gas Contr. 5 (2) (2011) 181–188. [72] F. Karimi, N. Komendantova, Understanding experts’ views and risk perceptions on carbon capture and storage in three European countries, Geojournal 82 (1) (2017) 185–200. [102] G. De Vries, B.W. Terwel, N. Ellemers, D.D. Daamen, Sustainability or proﬁtability? How communicated motives for environmental policy affect public perceptions of corporate greenwashing, Corp. Soc. Responsib. Environ. Manag. 22 (3) (2015) 142–154. [73] F. Karimi, A. Toikka, The relation between cultural structures and risk perception: how does social acceptance of carbon capture and storage emerge? Energy Procedia 63 (2014) 7087–7095. [103] B.W. Terwel, F. Harinck, N. Ellemers, D.D.L. Daamen, How organizational motives and communications affect public trust in organizations: the case of carbon dioxide capture and storage, J. Environ. Psychol. 29 (2009) 290–299. [74] J.D. Sharp, M.K. Jaccard, D.W. Keith, Anticipating public attitudes toward underground CO 2 storage, Int. J. Greenh. Gas Contr. 3 (2009) 641–651. ll hl f [104] F. Johnsson, D. Reiner, K. Itaoka, H. Herzog, Stakeholder attitudes on carbon capture and storage—an international comparison, Int. J. Greenh. Gas Contr. 4 (2) (2010) 410–418. [75] S. L'Orange Seigo, L. Wallquist, S. Dohle, M. Siegrist, Communication of CCS monitoring activities may not have a reassuring effect on the public, Int. J. Greenh. Gas Contr. 5 (2011) 1674–1679. [76] J. Bradbury, I. Ray, T. Peterson, S. Wade, G. Wong-Parodi, A. Feldpausch, The role of social factors in shaping public perceptions of CCS: results of multi-state focus group interviews in the US, Energy Procedia 1 (1) (2009) 4665–4672. [105] P. Ashworth, J. Bradbury, S. Wade, C.F.J.Y. Feenstra, S. Greenberg, G. Hund, et al., What's in store: lessons from implementing CCS, Int. J. Greenh. Gas Contr. 9 (2012) 402–409. 27 P. Tcvetkov et al. Heliyon 5 (2019) e02845 [106] N. Sacuta, D. Daly, B. Botnen, K. Worth, Communicating about the geological storage of carbon dioxide–comparing public outreach for CO2 EOR and saline storage projects, Energy Procedia 114 (2017) 7245–7259. [136] A. Cherepovitsyn, P. References [143] M. Lupion, A. Perez, F. Torrecilla, B. Merino, Lessons learned from the public perception and engagement strategy-experiences in CIUDEN's CCS facilities in Spain, Energy Procedia 37 (2013) 7369–7379. [114] P. Ashworth, A. Pisarski, K. Thambimuthu, Public acceptance of carbon dioxide capture and storage in a proposed demonstration area, Proc. Inst. Mech. Eng. Part A J Power Energy 223 (2009) 299–304. [144] H. Pihkola, E. Tsupari, M. Kojo, L. Kujanp€a€a, M. Nissil€a, L. Sokka, K. Behm, Integrated sustainability assessment of CCS–identifying non-technical barriers and drivers for CCS implementation in Finland, Energy Procedia 114 (2017) 7625–7637. gy [115] D. Reiner, T. Curry, M. de Figueiredo, H. Herzog, S. Ansolabehere, K. Itaoka, et al., An International Comparison of Public Attitudes towards Carbon Capture and Storage Technologies, NTNU [2006], 2006. http://www.ghgt8.no. [145] A. Hansson, M. Bryngelsson, Attitudes Regarding CO2 Capture and Storage from a Swedish Perspective, 2005. [116] D.M. Reiner, T.E. Curry, M.A. De Figueiredo, H.J. Herzog, S.D. Ansolabehere, K. Itaoka, et al., American exceptionalism? Similarities and differences in national attitudes toward energy policy and global warming, Environ. Sci. Technol. 40 (2006) 2093–2098. [146] L. Mabon, C. Littlecott, CO₂-EOR Stakeholder Perceptions and Policy Responses, SCCS, 2015. [117] P. Upham, T. Roberts, Public perceptions of CCS: emergent themes in pan- European focus groups and implications for communications, Int. J. Greenh. Gas Contr. 5 (5) (2011) 1359–1367. [147] H.A. Nguyen-Trinh, M. Ha-Duong, Perspective of CO2 capture & storage (CCS) development in Vietnam: results from expert interviews, Int. J. Greenh. Gas Contr. 37 (2015) 220–227. [148] M. Ha-Duong, A. Nadaï, A.S. Campos, A survey on the public perception of CCS in France, Int. J. Greenh. Gas Contr. 3 (5) (2009) 633–640. [118] Q. Li, L.C. Liu, Z.A. Chen, X. Zhang, L. Jia, G. Liu, A survey of public perception of CCUS in China, Energy Procedia 63 (2014) 7019–7023. [149] C. Braun, C. Merk, G. P€onitzsch, K. Rehdanz, U. Schmidt, Public perception of climate engineering and carbon capture and storage in Germany: survey evidence, Clim. Policy 18 (4) (2018) 471–484. [119] M. de Best-Waldhober, D. Daamen, A.R. Ramirez, A. Faaij, C. Hendriks, E. de Visser, Informed public opinion in The Netherlands: evaluation of CO 2 capture and storage technologies in comparison with other CO 2 mitigation options, Int. J. Greenh. Gas Contr. 10 (2012) 169–180. [150] S. V€ogele, D. Rübbelke, P. Mayer, W. Kuckshinrichs, Germany’s “No” to carbon capture and storage: just a question of lacking acceptance? References Haszeldine, A roadmap for carbon capture and storage in the UK, Int. J. Greenh. Gas Contr. 4 (1) (2010) 1–12. [159] R. Wüstenhagen, M. Wolsink, M.J. Bürer, Social acceptance of renewable energy innovation: an introduction to the concept, Energy Policy 35 (2007) 2683–2691. l h kl l l [128] M. Billson, M. Pourkashanian, The evolution of European CCS policy, Energy Procedia 114 (2017) 5659–5662. [129] E. ter Mors, M.W. Weenig, N. Ellemers, D.D. Daamen, M. de Best-Waldhober, Public information: on why and when multiple information sources are more effective than single information sources in communication about CCS, Energy Procedia 1 (1) (2009) 4715–4718. [160] S. Brunsting, M. Pol, J. Mastop, M. Kaiser, R. Zimmer, S. Shackley, et al., Social Site Characterisation for CO2 storage operations to inform public engagement in Poland and Scotland, Energy Procedia 37 (2013) 7327–7336. [161] S. Wade, S. Greenberg, Social site characterisation: from concept to application, in: A Review of Relevant Social Science Literature and a Toolkit for Social Site Characterisation, 2011, p. 80. [130] C. Gross, Community perspectives of wind energy in Australia: the application of a justice and community fairness framework to increase social acceptance, Energy Policy 35 (2007) 2727–2736. [162] P. Tcvetkov, A. Cherepovitsyn, S. Fedoseev, The changing role of CO2 in the transition to a circular economy: review of carbon sequestration projects, Sustainability 11 (2019) 5834. [131] H.W. van Os, R. Herber, B. Scholtens, Not Under Our Back Yards? A case study of social acceptance of the Northern Netherlands CCS initiative, Renew. Sustain. Energy Rev. 30 (2014) 923–942. [163] A. Cherepovitsyn, S. Fedoseev, P. Tcvetkov, K. Sidorova, A. Kraslawski, Potential of Russian regions to implement CO2-enhanced oil recovery, Energies 11 (6) (2018) 1528. [132] C. Gough, S. Mander, Public perceptions of CO2 transportation in pipelines, Energy Policy 70 (2014) 106–114. [164] S. Budinis, S. Krevor, N.M. Dowell, N. Brandon, A. Hawkes, An assessment of CCS costs, barriers and potential, Energy Strategy Rev. 22 (2018) 61–81. [133] A.I. Toikka, M. Kojo, L. Kainiemi, What is the socio-political scaffolding CCS needs to thrive? Energy Procedia 63 (2014) 7119–7124. [134] B.W. Terwel, F. Harinck, N. Ellemers, D.D.L. Daamen, Voice in political decision making: the effect of group voice on perceived trustworthiness of decision makers and subsequent acceptance of decisions, J. Exp. Psychol. Appl. 16 (2010) 173–186. [165] A. Kozlov, A. Teslya, S. References Appl. Energy 214 (2018) 205–218. [120] C. Anderson, J. Schirmer, N. Abjorensen, Exploring CCS community acceptance and public participation from a human and social capital perspective, Mitig. Adapt. Strategies Glob. Change 17 (2012) 687–706. [151] D. Schumann, E. Duetschke, K. Pietzner, Public Perception of CO2 Offshore Storage in Germany: Regional Differences and Determinants, 2014. h bl f l d [121] C.R. Jones, R.L. Radford, K. Armstrong, P. Styring, What a waste! Assessing public perceptions of Carbon Dioxide Utilisation technology, J. CO2 Util. 7 (2014) 51–54. [152] D. Schumann, Public perception of CO2 pipelines, Energy Procedia 114 (2017) 7356–7366. [122] J. van Heek, K. Arning, A. Linzenich, M. Zieﬂe, Trust and distrust in Carbon Capture and Utilization industry as relevant factors for the acceptance of carbon- based products, Front. Energy Res. 6 (2018) 73. [153] C. Oltra, P. Upham, H. Riesch, A. Boso, S. Brunsting, E. Dütschke, et al., Public responses to CO 2 storage sites: lessons from ﬁve European cases, Energy Environ. 23 (2012) 227–248. [123] L. Sj€oberg, Perceived risk and tampering with nature, J. Risk Res. 3 (2000) 353–367. [154] T. Lockwood, Public Outreach Approaches for Carbon Capture and Storage Projects, 2017. [124] A.B. Klass, E.J. Wilson, Climate change and carbon sequestration: assessing a liability regime for long-term storage of carbon dioxide, Emory LJ. 58 (2008) 103. [155] E. Dütschke, What drives local public acceptance–comparing two cases from Germany, Energy Procedia 4 (2011) 6234–6240. [125] J. Desbarats, P. Upham, H. Riesch, D. Reiner, S. Brunsting, M. de Best-Waldhober, et al., Review of the public participation practices for CCS and non-CCS projects in Europe, in: Report of the FP7 project “NearCO2, 11, 2010. [156] K. Arning, J. van Heek, M. Zieﬂe, Risk perception and acceptance of CDU consumer products in Germany, Energy Procedia 114 (2017) 7186–7196. [126] R. Chaudhry, M. Fischlein, J. Larson, D.M. Hall, T.R. Peterson, E.J. Wilson, J.C. Stephens, Policy stakeholders' perceptions of carbon capture and storage: a comparison of four US States, J. Clean. Prod. 52 (2013) 21–32. [157] S. Perdan, C.R. Jones, A. Azapagic, Public awareness and acceptance of carbon capture and utilisation in the UK, Sustain. Prod. Consum. 10 (2017) 74–84. b l b l d d l h [158] V. Weber, Uncertain liability and stagnating CCS deployment in the European union: is it the member states’ turn? Review of European, Comp. Int. Environ. Law 27 (2) (2018) 153–161. [127] C. Gough, S. Mander, S. References Chernogorskii, Game theory model of state investment into territories of advanced development in the regions of mineral resources specialization, J. Min. Inst. 234 (6) (2019) 673. territories of advanced development in the regions of mineral resources specialization, J. Min. Inst. 234 (6) (2019) 673. [166] R.F. Van Voorhees, Crediting carbon dioxide storage associated with enhanced oil recovery, Energy Procedia 114 (2017) 7659–7666. [135] J.K. Haug, P. Stigson, Local acceptance and communication as crucial elements for realizing CCS in the Nordic region, Energy Procedia 86 (2016) 315–323. 28
https://openalex.org/W2789170144	https://periodicals.karazin.ua/socecongeo/article/download/10242/9761	Ukrainian	null	Impact of the of geographical space's length on the distribution of international tourist flows	Časopis socìalʹno-ekonomìčnoï geografìï	2,017	cc-by	6,584	2017 2017 Часопис соціально-економічної географії випуск 23(2) УДК 911.9 (477) Александр Король. ВЛИЯНИЕ ПРОТЯЖЕННОСТИ ГЕОГР РАСПРЕДЕЛЕНИЕ МЕЖДУНАРОДНЫХ ТУРИСТИЧЕСКИХ ПОТОКОВ В статье выдвинуто и проверено предположение, что такое свойство географического пространства, как протяжен- ность, влияет на распределение международных туристические потоков посредством расстояния между странами происхо- ждения и назначение туристов, а также через длину общей сухопутной границы, которая во многих случаях определяет объемы туристического обмена между странами-соседями. В результате анализа структуры международного туризма от- дельных стран выяснен характер влияния протяженности географического пространства на географию распределения меж- дународных туристических потоков. В частности, выявлена закономерность, согласно которой доля прибытий или отбытий в структуре международного туризма некой страны может быть любой, но, как правило, не превышает значения, которое с расстоянием уменьшается по экспоненте. Т.е., установлено, что расстояние не определяет, а ограничивает интенсивность туристического обмена между странами. Также установлено, что около 50% международного туристического обмена про- исходит между странами-соседями, а доля страны-соседа в структуре туристических прибытий или отбытий часто согласо- вывается с протяженностью общей сухопутной границы. Ключевые слова: международный туризм, въездные туристические потоки, туристические прибытия, выездные ту- ристические потоки, туристические отбытия, расстояние между границами, протяженность общей границы. ВПЛИВ ПРОТЯЖНОСТІ ГЕОГРАФІЧНОГО ПРОСТОРУ НА РОЗПОДІЛ МІЖНАРОДНИХ ТУРИСТИЧНИХ ПОТОКІВ У статті висунуте та перевірене припущення, що така властивість географічного простору, як протяжність, впливає на розподіл міжнародних туристичні потоків через відстані між країнами походження і призначення туристів, а також через довжину спільного кодону суходолом, яка у багатьох випадках визначає обсяги туристичного обміну між країнами- сусідами. За результатами аналізу структури міжнародного туризму окремих країн з’ясований характер впливу протяжності географічного простору на географію розподілу міжнародних туристичних потоків. Зокрема, виявлена закономірність, за якою частка прибуттів або вибуттів у структурі міжнародного туризму певної країни може бути будь-якою, але, зазвичай, не перевищує значення, яке із відстанню зменшується за експонентою. Тобто, встановлено, що відстань не визначає, а обмежує інтенсивність туристичного обміну між країнами. Також встановлено, що близько 50% міжнародного туристичного обімну відбувається між країнами-сусідами, а частка країни-сусіда в структурі туристичних прибуттів або вибуттів часто узгоджується з протяжністю спільного кордону суходолом. у у р р у у Ключові слова: міжнародний туризм, в’їзні туристичні потоки, туристичні прибуття, виїзні туристичні потоки, туристичні вибуття, відстань між кордонами, протяжність спільного кордону. _________________ © Король О., 2017 Александр Король. ВЛИЯНИЕ ПРОТЯЖЕННОСТИ ГЕОГРАФИЧЕСКОГО ПРОСТРАНСТВА НА РАСПРЕДЕЛЕНИЕ МЕЖДУНАРОДНЫХ ТУРИСТИЧЕСКИХ ПОТОКОВ Александр Король. ВЛИЯНИЕ ПРОТЯЖЕННОСТИ ГЕОГРАФИЧЕСКОГО ПРОСТРАНСТВА НА РАСПРЕДЕЛЕНИЕ МЕЖДУНАРОДНЫХ ТУРИСТИЧЕСКИХ ПОТОКОВ Oleksandr Korol. IMPACT OF THE OF GEOGRAPHICAL SPACE'S LENGTH ON THE DISTRIBUTION OF INTERNATIONAL TOURIST FLOWS The article suggests and verifies the assumption that such a property of geographical space as the length affects the distribu- tion of international tourist flows due to the distance between the countries of origin and destination of tourists, as well as the length of the common border by land, which in many cases determines the volume of tourist exchange between the countries-neighbors. As a result of the analysis of the structure of international tourism by countries, the nature of the influence of the length of geographical space on the geography of the distribution of international tourist flows has been determined. In particular, a pattern has been found in which the share of arrivals or departures in the structure of international tourism of a certain country can be any, but usually does not exceed the value that decreases from distance by the exponent. So, it is defined that distance does not determine, but limits the intensity of tourist exchange between countries. It is also estimated that around 50% of international tourist exchange is between neighboring countries, and the share of a neighboring country in the structure of tourist arrivals or departures is often consistent with the length of the common border by land. In addition, the cases that not comply with this pattern are explained. Exceptions are usually related to countries with an “overseas” geographical location, including are in the part of the world where they are surrounded by mentally different countries. At the same time, each of them near or far has a mentally close large country, from where arrive a much more tourists than usual for such a distance. Exceptions for international tourist flows in the neighborhood are usually associated with the distortion of the isot- ropy of the geographical space, primarily in the border areas, in particular due to the uneven distribution of the population in the countries of origin or tourist potential in the destinations. international tourism, inbound tourist flows, tourist arrivals, outbound tourist flows, tourist departures, distance length of a common border. Keywords: international tourism, inbound tourist flows, tourist arrivals, outbound tourist flows, tourist de between borders, length of a common border. дослідження. що сприяє встановленню й розширенню міжнарод- них культурних зв’язків між країнами, зміцненню дружби й співробітництва народів світу. Зважаючи на залучення до міжнародного туризму великої кількості людей (більше 1 млрд.), він справляє знач- ний вплив на різні сфери суспільних відносин. Тому видається важливим з’ясування того, наскільки ту- ристичний обмін між країнами є інтенсивним Актуальність теми дослідження. Міжнародний туризм посідає особливе місце в зовнішньоекономічних зв’язках. Він є важливим стимулом розвитку міжнародної торгівлі, сприяє розширенню й активізації епізодичних міграцій між країнами. Більше того, значення туризму не можна оцінювати тільки економічними вигодами. Він є однією з найактивніших форм спілкування людей, теми _________________ © Король О., 2017 _________________________________ DOI: 10.26565/2076-1333-2017-23-14 98 Часопис соціально-економічної географії 2017 випуск 23(2) залежно від відстані між ними або протяжності спільного кордону. країни-походження, яке проживає поближче до кор- дону, мало б частіше відвідувати прикордонні регіони сусідньої країни [5]. Виходячи з цього, метою роботи є з’ясування впливу протяжності географічного простору на розподіл міжнародних туристичних потоків між ок- ремим країнами, зокрема виявлення залежності част- ки прибуттів або вибуттів у структурі міжнародного туризму від відстані між країнами походження та призначення туристів, а для сусідніх країн – від дов- жини спільного кордону. Отже, така властивість географічного простору, як протяжність, впливає на розподіл міжнародних туристичні потоків через відстані між країнами по- ходження і призначення туристів, а також через протяжність спільного кодону суходолом, яка у ба- гатьох випадках визначає обсяги туристичного обміну між країнами-сусідами. Цей розподіл проявляється в частоті поїздок до тієї чи іншої дестинації, що можна побачити в структурі тури- стичних прибуттів або вибуттів певної країни. Методика дослідження побудована на використанні статистичної бази та методології Всесвітньої туристичної організації, також застосо- вувалися методи математичної статистики, у т.ч. кореляційно-регресійний аналіз і критерій χ2, який називають ще критерієм незалежності, узгодженості. У міжнародному туризмі виділяють дві його форми – іноземний (в’їзний) і зарубіжний (виїзний), які відрізняються напрямом туристського потоку. Один і той самий турист буде класифікований як іноземний в країні призначення і зарубіжний для країни походження. Очікується, що протяжність географічного простору впливатиме на ці дві форми міжнародного туризму обопільно. Виходячи з цього, аналіз впливу протяжності географічного простору на розподіл туристичних потоків варто робити як для іноземного, так і зарубіжного туризму, що лише подвоїть репрезентативність отриманих результатів. Аналіз останніх досліджень і публікацій. Зва- жаючи на поставлену мету, на передній план вихо- дить географічна сутність туризму, як форми міграції населення, що ґрунтується на подоланні геопросто- ру, пов’язана з територією та її відмінностями «від місця до місця». В цьому контексті саме географам належить значний доробок у дослідженнях туризму, висвітлений у працях В.С. Преображенського, Ю.О. Вєдєніна, І.В. Зорина, М.А. Ананьєва, Аналіз останніх досліджень і публікацій. Зва- жаючи на поставлену мету, на передній план вихо- дить географічна сутність туризму, як форми міграції населення, що ґрунтується на подоланні геопросто- ру, пов’язана з територією та її відмінностями «від місця до місця». В цьому контексті саме географам належить значний доробок у дослідженнях туризму, висвітлений у працях В.С. Преображенського, Ю.О. Вєдєніна, І.В. Зорина, М.А. Ананьєва, Н.П. Зачиняева, Н.С. Фальковича, Є.А. Котлярова та інших дослідників радянських часів, а в Україні – М.П. Крачила, О.О. Любіцевої, О.О. Бейдика. теми З-поміж останніх вагомих робіт присвячених міжнародному туризму можна виділити праці А.Ю. Александрової (2002 р.) [1], О.О. Любіцевої (2003 р.) [4] тощо. У цих роботах охарактеризовані основні поняття міжнародного туризму, проаналізована географія туристського попиту за регіонами світу, відображені останні тенденції та процеси глобалізації в цій галузі. Проте, деякі аспек- ти міжнародного туризму залишаються невисвітленими, зокрема й той, якому присвячене це дослідження. Відстані між країнами та розподіл зарубіжних (виїзних) туристичних потоків. Найперше, слід оцінити вплив відстані від країни походження туристів до окремих дестинацій на їхні частки в структурі зарубіжного туризму. Іншими словами доцільно з’ясувати, чи частіше відвідується дестинація, якщо вона ближче знаходиться. Для цьо- го необхідно зіставити частки вибуттів із країни по- ходження до тієї чи іншої країни призначення туристів і відстані між ними. Для спрощення обчис- лень обиралася найменша відстань між їхніми кор- донами. «Нульову» відстань, наприклад, мали країни-сусіди зі спільним кордоном суходолом. Та- кий підхід видається цілком обґрунтованим, адже за стандартами UNWTO при дослідженні міжнародного туризму надається перевага обліку на кордоні, тобто формально туристичний потік розглядається від кор- дону до кордону [3]. Виклад основного матеріалу. Вплив протяжності географічного простору на туризм проявляється, насамперед, у необхідності долати відстані. Усі просторові переміщення туристів здійснюються з використанням транспортних засобів і пов’язані з витратами грошей і часу. Тому відстань перетворюється на чинник, який впливає на форму- вання туристичний потоків до тієї чи іншої дестинації, адже, чим вона ближча, тим частіше мо- же відвідуватися, а від цього залежить розподіл ту- ристичних потоків у географічному просторі, зокре- ма між окремими країнами. Отже, дані про відстані до кордонів дестинацій і частки вибуттів до них для обраних 23-ох країн по- ходження туристів були зведені в один масив, що складався із 249 випадків, у кожному з яких зіставлялися відсоток вибуттів і відстань [8]. Отримані результати засвідчили відсутність функціональної залежності між цими двома ознака- ми (див. рис. 1). Однак у розподілі «точок» на рисунку 1 спостерігається певна закономірність – зі зростанням відстані зменшується дисперсія частки вибуттів, а починаючи від 4,5 тис. км спостереження, за винят- ком «викидів», мають значення менші за 3%. Також можна побачити, що більше 90% випадків розташо- вуються в межах криволінійної трапеції, яка згори обмежується графіком експоненційної функції: У випадку із туристичними потоками до сусідніх країн протяжність географічного простору проявляється через довжину спільного кордону су- ходолом. Такі поїздки часто охоплюють прикордонні території та мають характер «дифузії». Тому, за інших рівних умов, обсяги туристичного обімну між країнами сусідами узгоджуються з протяжністю спільного кордону (суходолом). теми Ця закономірність пояснюється тим, що туристичні потоки до сусідів можуть підпадати під статистичну теорією переміщень Хегерстранда, згідно з якою населення У випадку із туристичними потоками до сусідніх країн протяжність географічного простору проявляється через довжину спільного кордону су- ходолом. Такі поїздки часто охоплюють прикордонні території та мають характер «дифузії». Тому, за інших рівних умов, обсяги туристичного обімну між країнами сусідами узгоджуються з протяжністю спільного кордону (суходолом). Ця закономірність пояснюється тим, що туристичні потоки до сусідів можуть підпадати під статистичну теорією переміщень Хегерстранда, згідно з якою населення у = 23,26exp(-0,0003x); у = 23,26exp(-0,0003x); де: у – це відсоток зарубіжних вибуттів; де: у – це відсоток зарубіжних вибуттів; х – відстань між кордонами. х – відстань між кордонами. 99 Часопис соціально-економічної географії Часопис соціально-економічної географії 2017 випуск 23(2) їхніми кордонами. Данні щодо обраних 25-и країн були зведені в один масив, що складався зі 140 випадків, у кожному з яких зіставлялися частка прибуттів і відстань [7]. Отримані результати засвідчили відсутність функціональної залежності між цими двома ознаками (див. рис. 2). Відстані між країнами та розподіл іноземних (в’їзних) туристичних потоків. Для з’ясування того, чи матиме країна походження туристів більшу част- ку в структурі іноземного туризму країни призна- чення, якщо вона ближче до неї знаходиться, були зіставлені частки прибуттів і мінімальні відстані між Рис. 1. Частка вибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Рис. 1. Частка вибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Водночас, так само як й у випадку з виїзним ту- ризмом, у розподілі «точок» на рисунку 2 спостерігається певна закономірність. Зокрема, зі зростанням відстані зменшується дисперсія частки прибуттів, а починаючи від 4 тис. км усі випадки, за винятком «викидів», мають значення менші за 10%. Також можна побачити, що близько 90% спостере- жень розташовуються в межах криволінійної трапеції, яка згори обмежується графіком експоненційної функції. Схожа картина спостерігалася при проведенні подібного аналізу для виїзного туризму. Це підтверджує припущення про те, що протяжність географічного простору однаково впливає як на іноземні (в’їзні), так і на зарубіжні туристичні потоки. Виходячи з цього, можна зробити висновок, що як для іноземних (в’їзних), так і зарубіжних (виїзних) туристичних потоків відстань є не детермінуючим, а обмежуючим чинником. Частка прибуттів або вибуттів не залежить безпосередньо від відстані між країнам, а, зазвичай, не перевищує значення експоненційної функції, в якої аргументом є відстань, інакше кажучи знаходиться в межах окресленої нею криволінійної трапеції (див. рис. 3). де: у – це відсоток іноземних прибуттів; теми Видається важливим розглянути «викиди», що найвище відхилилися від графіка експоненційної функції, як для іноземних (в’їзних), так і зарубіжних (виїзних) туристичних потоків [7, 8]. В усіх екстре- мальних випадках спостерігається експоненційна залежність в їхньому розташуванні, що лише підтве- рджує вплив відстані на інтенсивність туристичного обміну між країнами. Водночас, значне зміщення догори уявної кривої, яку вони утворюють, свідчить про дію певних чинників (див. рис. 1, рис. 2). Для перевірки обопільного характеру впливу відстані на інтенсивність туристичного обміну між країнами, застосуємо формулу експоненційної кривої, яка була отримана для зарубіжного (виїзного) туризму. При накладанні її на подібний графік розподілу прибуттів для іноземного (в’їзного) туриз- му, задля задовільної апроксимації із «верхніми» випадками, за винятком викидів, криву цієї функції довелося лише «підняти» на п’ять одиниць. Отже, нова функція мала такий самий експоненційний ха- рактер та відрізнялася від старої лише на «+5»: На «нульовій» відстані виділяються два випадки – це туристичні потоки Сінгапур → Малайзія для зарубіжного туризму та США → Канада для іноземного. Положення сусіда, тісні історичні зв’язки між Сінгапуром і Малайзією та відсутність мовного бар’єру зумовили те, що на цей напрямок припадало 74% вибуттів. Переважно це були короткотривалі подорожі на вихідних задля походів у = 23,26exp(-0,0003x) + 5 де: у – це відсоток іноземних прибуттів; 100 Часопис соціально-економічної географії 2017 випуск 23(2) півдні вона має протяжний кордон суходолом зі США, причому він є «прозорим» – американцям не потрібна канадська віза. Крім того, цьому сприяє відсутність мовного бар’єру та культурна близькість. Американські туристи переважно здійснюють часті поїздки автомобілем, в основному, задля відпочинку в сільській місцевості, також відвідують у Канаді з екскурсіями великі міста. по магазинах до сусіднього малайзійського міста Джохор-Бару, з яким Сінгапур з’єднаний греблею. Така зосередженість туристичних потоків також зу- мовлена острівним положенням цієї карликової країни, через що вони не розпорошуються між іншими дестинаціями. Три чверті іноземних туристів до Канади прибу- вають зі США. Така ситуація пов’язана з тим, що Канада відрізана океанами від решти світу. Лише на Рис. 2. Частка прибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Рис. 3. Частки прибуттів або вибуттів у структурі міжнародних туристичних потоків і відстані між країнами походження та призначення туристів. Рис. 2. Частка прибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Рис. 2. Частка прибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Рис. 2. теми Ще один випадок, подібний до попереднього, – це туристичний потік із Великої Британії до Австралії, на який доводиться 12% прибуттів. В ос- новному британські туристи відвідують тут родичів і знайомих, хоча останніми роками VFR-потоки ско- рочуються, натомість зростає кількість подорожую- чих задля відпочинку і дозвілля. Таким інтенсивним потокам сприяють ті ж чинники, що й у випадку з Новою Зеландією – мовна спорідненість і спільна історія, адже Австралія так само заселена вихідцями з Британських островів. Якщо розглядати структуру туристичних прибуттів до Ізраїлю, то так само «кидається у вічі» відсутність країн-сусідів, що пов’язане з перманент- ним арабо-ізраїльським конфліктом. Натомість у 5- топ опинилися далекі країни, туристичні потоки з яких зумовлюються ментальним чинником – чисельністю єврейської громади, яка там мешкає. В США проживає 45% усіх євреїв (більше, ніж у само- му Ізраїлі) і саме тому на американських туристів доводиться 24% прибуттів. Якщо розглядати зарубіжний туризм Австралії, то у 2008 р. більше половини туристів із цієї країни (53%) подорожували в межах Азійсько- Тихоокеанського регіону, що підтверджує дію чин- ника відстані. Перше місце очікувано посіла Нова Зеландія, до якої подорожувало 16% австралійських туристів. Однак серед інших топ-дестинацій чільне місце посіли просторово віддалені (більше 10 тис. км) але ментально близькі країни Північної Америки та Європи: на США та Канаду припадало 11% вибуттів; до Європи подорожувало стільки ж австралійців, як і до Нової Зеландії, причому трохи менше половини з них (7%) обирали Велику Британію. Саме потік до «Туманного Альбіону» був зафіксований як екстремальне відхилення на відстані 13680 км. Отже, майже в усіх випадках, які мають найпомітніше відхилення від графіка експоненційної функції, спостерігається спільна риса. Так, більшість із цих країн мають «заокеанське» географічне поло- ження, у т.ч. знаходяться у тій частині світу, де їх оточують ментально далекі країни. Водночас, у кожної із них поруч або далеко є ментально близька велика країна, саме на яку припадає набагато більша частка прибуттів чи вибуттів, ніж пересічно для такої відстані. Туризм по сусідству. Примітним є те, що у структурі міжнародного туризму перші сходинки посідають країни-сусіди, які межують суходолом або кордони котрих розташовані на відстані до 50 км. На них разом може припадати до 80% прибуттів або вибуттів, а в середньому їхня частка сягає 50%. При цьому кількість країн-сусідів не відіграє великого значення, важливішим тут є наявність протяжного кордону суходолом. Очевидно, що в усіх випадках, які пов’язані з туристичними потоками між Великою Британією, Австралією та Новою Зеландією спрацював чинник ментальної близькості, зокрема мовної спорідненості, який був підсилений фактором «історичної батьківщини». теми Частка прибуттів із країни походження до країни призначення туристів і відстань між їхніми кордонами, 2008 р. Рис. 3. Частки прибуттів або вибуттів у структурі міжнародних туристичних потоків і відстані між країнами походження та призначення туристів. Рис. 3. Частки прибуттів або вибуттів у структурі міжнародних туристичних потоків і відстані між країнами походження та призначення туристів. Рис. 3. Частки прибуттів або вибуттів у структурі міжнародних туристичних потоків і відстані між країнами походження та призначення туристів. На більших відстанях виділяються випадки, що представлені туристичними потоками між Австралією, Новою Зеландією та Великою Британією. Так, майже половина новозеландців (48%) у 2008 р. прямували до Австралії. Ці дві країни близькі не лише географічно, але й ментально. Обидві вони перебували під короною Британської імперії та переважно заселенні вихідцями з «Туман- ного Альбіону». Так само в’їзні туристичні потоки до Нової Зеландії переважно складалися з австралійців – майже 40% прибуттів. Відсутність мовного бар’єру та історичне коріння забезпечили Великій Британії чільне місце в структурі туристичних вибуттів із Нової Зеландії. Хоча у 2008 р. на цей туристичний потік припадало 5% подорожуючих, для відстані 18 тис. км – це ано- мально висока частка. Цікавим є той факт, що з- поміж подорожуючих до Великої Британії новозеландців переважують ті, хто відвідують родичів і знайомих – 53%. Водночас, Велика Британія посідає другу сходинку в структурі тури- стичних прибуттів до Нової Зеландії (11,6%). 101 Часопис соціально-економічної географії 2017 випуск 23(2) Британія залишається доволі стабільним постачаль- ником туристів впродовж десятиліть – починаючи від 1950 року, хоча характер подорожей за цей час дещо змінився. Зокрема, все менше британських туристів відвідують у Новій Зеландії родичів і знай- омих і все більше надають перевагу індивідуальним подорожам перед інклюзив-турами. вибуттів із Ізраїлю відсутні дестинації-сусіди. Саме тому ізраїльтяни надають перевагу подорожам до віддалених дестинацій. Через це провідним транс- портним засобом є літак – 81% вибуттів, а більшість вибуттів суходолом припадає на арабів, які їдуть до сусідніх країни, що є рідкістю для євреїв. За таких обставин, не дивлячись на значну віддаленість і візовий режим, США є найпопулярнішою дестинацією в ізраїльських туристів – 30% подоро- жуючих, адже тут доброзичливо ставляться до єврейської культури, а також у США мешкає найбільша іудейська діаспора – 6,5 млн. Британія залишається доволі стабільним постачаль- ником туристів впродовж десятиліть – починаючи від 1950 року, хоча характер подорожей за цей час дещо змінився. Зокрема, все менше британських туристів відвідують у Новій Зеландії родичів і знай- омих і все більше надають перевагу індивідуальним подорожам перед інклюзив-турами. теми Три екстремальні випадки пов’язані із зарубіжним туризмом Японії. У 2008 р. 27% тури- стичних вибуттів з Японії приходилося на США. Очевидно, що ця країна для японських туристів – надзвичайно приваблива дестинація. Це пов’язане з тим, що у 70-ті роки, коли поїздки за кордон вперше стали популярними в Японії, і з другої половини 80-х років до початку 90-х років розвиток туризму в Південній Кореї та Китаї значно відставав від розви- нутих країн, а більшість курортних зон в Азії із ви- соким рівнем обслуговування з’явилися лише на по- чатку 90-х років. Отож, туристичні потоки Японії тоді зорієнтувалися на традиційні розвинуті туристичні регіони світу. Це одна з причин, чому японських мандрівників за звичкою приваблюють далекі напрямки, такі як Північна Америка, зокрема Гавайї та Європа. Серед Європейських країн найбільшою популярністю у японців користувались Італія (8,2%) та Франція (7,1%). На інтенсивність туристичного обміну між країнами-сусідами можуть впливати певні перепони, які пов’язані з перетином державного кордону: візовий режим, митні формальності. Щоб уникнути впливу цих перешкод, для аналізу, насамперед, були обрані країни з «прозорими» кордонами, зокрема континентальної частини Європейського союзу. Роз- гляд цих туристичних потоків розглядався окремо для зарубіжного та іноземного туризму. Зарубіжні (виїзні) туристичні потоки до сусідніх країн. Із 14-ти розглянутих європейських країн у се- редньому 47% зарубіжних туристів прямували до сусідніх дестинацій зі спільним кордоном суходолом [7]. Припускаючи вплив протяжності останнього на обсяги таких потоків, була висунута гіпотеза про те, що частка країни-сусіда в структурі туристичних вибуттів має бути пропорційною частці спільного з нею кордону (суходолом). Для підтвердження або спростування даного припущення був використаний критерій χ2, який називають ще критерієм незалежності, узгодженості та однорідності. Він розраховується за формулою: Потоки між Ізраїлем і США також опинилися з- поміж викидів як для іноземного, так і зарубіжного туризму. Через політичну напругу в стосунках з арабськими країнами, в структурі туристичних 102 Часопис соціально-економічної географії 2017 випуск 23(2) χ2 більше табличного, то розбіжність є не випадко- вою, тобто узгодженість часток не підтверджується [2]. ∑ − = E E O X 2 2 ) ( , де: О – емпіричні частки; Отже, припущення про те, що частка країни- сусіда в структурі туристичних вибуттів має бути пропорційною частці спільного з нею кордону сухо- долом, підтвердилось у 8 із 14 випадках. Для решти 6 країн сумарний χ2, що виділений у таблиці сірим кольором, помітно перевищував табличне значення за рівня істотності 0,05 (див табл. 1). Е – теоретично очікувані частки. теми Відхилення емпіричних часток від теоретичних може бути істотним, якщо спричинене якимось чин- ником, а також незначним, що зумовлене випадко- вими причинами, тоді узгодженість часток підтверджується. Щоби це визначити, розраховане значення χ2 порівнюють з табличним за певної кількості ступенів свободи. Якщо сумарне значення Таблиця 1 тичних потоків до сусідніх країн № Країни χ2 Франція 1 Іспанія 3,498 2 Італія 1,553 3 Німеччина 0,397 4 Бельгія, Люкс. 1,470 5 Швейцарія 3,348 Всього 10,266 Словенія 1 Хорватія 6,903 2 Італія 3,203 3 Австрія 5,906 4 Угорщина 2,178 Всього 18,19 Польща 1 Німеччина 14,837 2 Чехія 1,118 3 Словаччина 0,664 4 Литва 1,346 5 Росія 0,167 6 Україна 4,539 7 Білорусь 2,921 Всього 25,592 Чехія 1 Словаччина 22,045 2 Німеччина 1,308 3 Австрія 0,008 4 Польща 10,649 Всього 34,01 Угорщина 1 Австрія 4,016 2 Хорватія 0,126 3 Румунія 1,583 4 Сербія 0,525 5 Словаччина 0,068 6 Словенія 1,186 7 Україна 0,015 Всього 7,519 Болгарія 1 Греція 2,165 2 Туреччина 5,183 3 Сербія 1,066 4 Румунія 8,745 5 Македонія 0,001 Всього 17,16 Значення χ2 для зарубіжних туристичних потоків до сусідніх країн № Країни χ2 Швеція 1 Фінляндія 1,190 2 Данія 32,884 3 Норвегія 8,215 Всього 42,289 Німеччина 1 Австрія 0,541 2 Франція 2,663 3 Швейцарія 0,205 4 Нідерланди 1,091 5 Польща 0,580 6 Данія 1,207 7 Чехія 3,406 Всього 9,693 Бельгія 1 Франція 2,320 2 Нідерланди 4,716 3 Німеччина 0,002 Всього 7,038 Нідерланди 1 Німеччина 0,060 2 Бельгія 0,080 Всього 0,14 Швейцарія 1 Німеччина 4,996 2 Італія 1,601 3 Франція 0,777 4 Австрія 0,142 Всього 7,516 Австрія 1 Німеччина 0,527 2 Італія 4,692 3 Угорщина 0,834 4 Швейцарія 0,004 5 Чехія 1,799 6 Словенія 2,889 7 Словаччина 0,341 Всього 11,086 Італія 1 Швейцарія 0,579 2 Франція 0,023 3 Австрія 0,374 4 Словенія 0,171 Всього 1,147 Іспанія 1 Франція 3,041 2 Португалія 5,758 № Країни χ2 Франція 1 Іспанія 3,498 2 Італія 1,553 3 Німеччина 0,397 4 Бельгія, Люкс. теми Для підтвердження або Таблиця 2 тичних потоків від сусідніх країн № Країни χ2 Польща 1 Німеччина 6,389 2 Литва 0,234 3 Росія 1,811 4 Україна 1,774 5 Білорусь 3,337 Всього 13,545 Чехія 1 Словаччина 0,041 2 Німеччина 4,261 3 Польща 6,271 Всього 10,574 Угорщина 1 Австрія 0,176 2 Хорватія 1,442 3 Румунія 0,168 4 Сербія і Чорн. 3,279 5 Словаччина 3,622 6 Україна 2,234 Всього 10,921 Україна 1 Росія 0,045 2 Польща 3,308 3 Молдова 0,057 4 Білорусь 0,659 5 Румунія 1,544 6 Угорщина 0,706 7 Словаччина 0,066 Всього 6,385 США 1 Канада 0,432 2 Мексика 0,747 Всього (без Аляски) 1,179 Аргентина 1 Чилі 3,791 2 Бразилія 5,943 3 Уругвай 3,630 4 Парагвай 0,710 Всього 14,074 Значення χ2 для іноземних туристичних потоків від сусідніх країн № Країни χ2 Швеція 1 Фінляндія 0,042 2 Данія 30,737 3 Норвегія 6,968 Всього 37,747 Бельгія 1 Франція 3,395 2 Нідерланди 1,379 3 Німеччина 1,217 Всього 5,991 Нідерланди 1 Німеччина 0,960 2 Бельгія 1,449 Всього 2,409 Німеччина 1 Нідерланди 2,157 2 Швейцарія 0,572 3 Австрія 3,319 4 Франція 0,492 5 Бельгія 0,764 6 Данія 3,218 7 Польща 2,378 Всього 12,902 Швейцарія 1 Німеччина 16,662 2 Італія 6,612 3 Франція 2,473 Всього 25,748 Австрія 1 Німеччина 3,321 2 Італія 7,341 3 Швейцар. і Ліхт. 1,124 Всього 11,786 Італія 1 Швейцарія 0,039 2 Франція 0,274 3 Австрія 0,104 Всього 0,417 Франція 1 Італія 0,038 2 Німеччина 0,971 3 Бельгія 0,603 № Країни χ2 Польща 1 Німеччина 6,389 2 Литва 0,234 3 Росія 1,811 4 Україна 1,774 5 Білорусь 3,337 Всього 13,545 Чехія 1 Словаччина 0,041 2 Німеччина 4,261 3 Польща 6,271 Всього 10,574 Угорщина 1 Австрія 0,176 2 Хорватія 1,442 3 Румунія 0,168 4 Сербія і Чорн. теми 1,470 5 Швейцарія 3,348 Всього 10,266 Словенія 1 Хорватія 6,903 2 Італія 3,203 3 Австрія 5,906 4 Угорщина 2,178 Всього 18,19 Польща 1 Німеччина 14,837 2 Чехія 1,118 3 Словаччина 0,664 4 Литва 1,346 5 Росія 0,167 6 Україна 4,539 7 Білорусь 2,921 Всього 25,592 Чехія 1 Словаччина 22,045 2 Німеччина 1,308 3 Австрія 0,008 4 Польща 10,649 Всього 34,01 Угорщина 1 Австрія 4,016 2 Хорватія 0,126 3 Румунія 1,583 4 Сербія 0,525 5 Словаччина 0,068 6 Словенія 1,186 7 Україна 0,015 Всього 7,519 Болгарія 1 Греція 2,165 2 Туреччина 5,183 3 Сербія 1,066 4 Румунія 8,745 5 Македонія 0,001 Всього 17,16 Значення χ2 для зарубіжних туристичних потоків до сусідніх країн Значення χ для зарубіжни № Країни χ2 Швеція 1 Фінляндія 1,190 2 Данія 32,884 3 Норвегія 8,215 Всього 42,289 Німеччина 1 Австрія 0,541 2 Франція 2,663 3 Швейцарія 0,205 4 Нідерланди 1,091 5 Польща 0,580 6 Данія 1,207 7 Чехія 3,406 Всього 9,693 Бельгія 1 Франція 2,320 2 Нідерланди 4,716 3 Німеччина 0,002 Всього 7,038 Нідерланди 1 Німеччина 0,060 2 Бельгія 0,080 Всього 0,14 Швейцарія 1 Німеччина 4,996 2 Італія 1,601 3 Франція 0,777 4 Австрія 0,142 Всього 7,516 Австрія 1 Німеччина 0,527 2 Італія 4,692 3 Угорщина 0,834 4 Швейцарія 0,004 5 Чехія 1,799 6 Словенія 2,889 7 Словаччина 0,341 Всього 11,086 Італія 1 Швейцарія 0,579 2 Франція 0,023 3 Австрія 0,374 4 Словенія 0,171 Всього 1,147 Іспанія 1 Франція 3,041 2 Португалія 5,758 3 Андорра 6,691 Всього 15,49 103 Часопис соціально-економічної географії 2017 випуск 23(2) спростування даного твердження так само був вико- ристаний критерій χ2. У результаті проведеного аналізу, узгодженість між часткою прибуттів від сусідів із часткою спільного кордону суходолом спостерігалася у 8 із 14 випадках (див. табл. 2). І ли- ше у 6 країнах сумарний χ2, що виділений у таблиці сірим кольором, помітно перевищував табличне зна- чення за рівня істотності 0,05. Іноземні (в’їзні) туристичні потоки від сусідніх країн. У структурі іноземного туризму 14-ти розгля- нутих країн пересічно 53% прибуттів доводилося на сусідів [8]. Як й у випадку із зарубіжними (виїзними) туристичними потоками, зроблене припущення, що частка країни-сусіда в структурі туристичних прибуттів має бути пропорційною частці спільного з нею кордону (суходолом). теми Також велика частка фінських туристів (23%) спостерігається на півночі, тобто там де зі Швецією є спільний кордон суходолом. У Ско- не на півдні Швеції 60% прибуттів припадає на дат- чан, які потрапляють сюди через Ересуннський міст (див. рис. 5). Регіон Сконе з Данією з’єднує лише автомобільно-залізничний міст, що прокладений че- рез протоку Ересунн, тобто, фактично, кордону су- ходолом між Швецією та Данією немає. Проте, до цієї країни виїжджає майже 11% шведів. До Норвегії, наприклад, прямує 9% туристів, хоча на цю країну припадає 72% шведського кордону суходолом, який, до речі, пролягає північними малозаселеними територіями. Такий значний потік шведських туристів у данському напрямку зумовлений ще й тим, що Ересунський міст з’єднує регіон Сконе не просто з Данією, а з її столицею – Копенгагеном – туристичним центром цієї країни. У цьому випадку очевидним є те, що зосередження туристичного потенціалу однієї країни біля кордону з іншою також є тим чинником, що спотворює ізотропність географічного простору та впливає на розподіл міжнародних туристичних потоків. структуру прибуттів із сусідніх країн, прикордонний характер таких подорожей все ж простежується. На- селення країни-походження, яке проживає поближче до кордону, як правило, частіше відвідує прикордонні регіони сусідньої країни. Таку географію туристичних потоків по сусідству можна побачити в розподілі іноземних прибуттів по території Швеції. Зокрема, найбільша частка нор- везьких туристів (55%) спостерігається у Вестра- Геталанд і Халланд, тобто у регіоні, який межує су- ходолом із густозаселеним півднем Норвегії. Найбільше туристів із Фінляндії (60%) нараховується в Стокгольмському регіоні, адже він є найближчим, хоча й через море, до густозаселеного півдня Фінляндії. Також велика частка фінських туристів (23%) спостерігається на півночі, тобто там де зі Швецією є спільний кордон суходолом. У Ско- не на півдні Швеції 60% прибуттів припадає на дат- чан, які потрапляють сюди через Ересуннський міст (див. рис. 5). Якщо розглядати іноземний туризм, то ситуація буде дзеркальною. Так, у структурі прибуттів до Швеції надмірною, порівняно із часткою спільного кордону, виявилася частка туристів із Данії. Цей ви- падок є ще одним прикладом спотворення ізотропності географічного простору через нерівномірне розміщення населення країни поход- ження туристів, зокрема на прикордонних територіях. Як вже зазначалося, міст через протоку Ересунн з’єднує Швецію не просто з Данією, а з її столицею – Копенгагеном, де разом із приміською зоною проживає більше 1 млн. населення. Окрім то- го, подорожам датчан до Швеції сприяє мовна спорідненість. Два інших винятки для іноземного туризму, що представлені Швейцарією та Польщею, пов’язані з туристами із Німеччини. теми 3,279 5 Словаччина 3,622 6 Україна 2,234 Всього 10,921 Україна 1 Росія 0,045 2 Польща 3,308 3 Молдова 0,057 4 Білорусь 0,659 5 Румунія 1,544 6 Угорщина 0,706 7 Словаччина 0,066 Всього 6,385 США 1 Канада 0,432 2 Мексика 0,747 Всього (без Аляски) 1,179 Аргентина 1 Чилі 3,791 2 Бразилія 5,943 3 Уругвай 3,630 4 Парагвай 0,710 Всього 14,074 Значення χ2 для іноземних туристичних потоків від сусідніх країн Значення χ2 для іноземних туристичних потоків від сусідніх країн Значення χ2 для іноземних № Країни χ2 Швеція 1 Фінляндія 0,042 2 Данія 30,737 3 Норвегія 6,968 Всього 37,747 Бельгія 1 Франція 3,395 2 Нідерланди 1,379 3 Німеччина 1,217 Всього 5,991 Нідерланди 1 Німеччина 0,960 2 Бельгія 1,449 Всього 2,409 Німеччина 1 Нідерланди 2,157 2 Швейцарія 0,572 3 Австрія 3,319 4 Франція 0,492 5 Бельгія 0,764 6 Данія 3,218 7 Польща 2,378 Всього 12,902 Швейцарія 1 Німеччина 16,662 2 Італія 6,612 3 Франція 2,473 Всього 25,748 Австрія 1 Німеччина 3,321 2 Італія 7,341 3 Швейцар. і Ліхт. 1,124 Всього 11,786 Італія 1 Швейцарія 0,039 2 Франція 0,274 3 Австрія 0,104 Всього 0,417 Франція 1 Італія 0,038 2 Німеччина 0,971 3 Бельгія 0,603 Всього 1,612 Випадок зі Швецією для зарубіжного туризму є яскравим прикладом впливу нерівномірного розміщення населення по території країни на струк- туру вибуттів до сусідніх дестинацій. 4/5 населення Швеції зосереджено в південній частині країни, а одним із найщільніше заселених регіонів є Сконе (див. рис. 4). Винятки як для іноземного, так і зарубіжного туризму, не суперечать вищевказаним теоретичним викладкам, адже вони ґрунтуються на «ізотропному» геопросторі, який далебі не всюди зустрічається. То- му ці випадки викликають інтерес із точки зору аналізу відхилень задля визначення чинників, які спотворюють ізотропність території та впливають на розподіл туристичних потоків по сусідству [7, 8]. 104 2017 випуск 23(2) Часопис соціально-економічної географії Рис. 4. Розміщення населення Швеції, 2008 р. Рис. 4. Розміщення населення Швеції, 2008 р. структуру прибуттів із сусідніх країн, прикордонний характер таких подорожей все ж простежується. На- селення країни-походження, яке проживає поближче до кордону, як правило, частіше відвідує прикордонні регіони сусідньої країни. Таку географію туристичних потоків по сусідству можна побачити в розподілі іноземних прибуттів по території Швеції. Зокрема, найбільша частка нор- везьких туристів (55%) спостерігається у Вестра- Геталанд і Халланд, тобто у регіоні, який межує су- ходолом із густозаселеним півднем Норвегії. Найбільше туристів із Фінляндії (60%) нараховується в Стокгольмському регіоні, адже він є найближчим, хоча й через море, до густозаселеного півдня Фінляндії. теми В структурі прибуттів до цих двох країн на німців приходиться значно більший відсоток, порівняно з часткою спільного кордону. Окрім положення сусіда, поїздкам німецьких туристів до Швейцарії сприяє ментальний чинник (німецька мова є найпоширенішою в Швейцарії – нею говорять 64% населення країни). Німців у Швейцарії приваблює, насамперед, відпочинок у горах, саме тому близько 40% їхніх ночівель припадає на зимовий сезон (з листопада до квітня). У цьому випадку також спрацьовує мотив контрасту. Хоча у випадку з іноземним туризмом Швеції протяжність спільного кордону не вплинула на 105 Часопис соціально-економічної географії 2017 випуск 23(2) Рис. 5. Розподіл іноземних прибуттів по окремих регіонах Швеції та країнах походження туристів, 2008 р. Рис. 5. Розподіл іноземних прибуттів по окремих регіонах Швеції та країнах походження туристів, 2008 р. табл. 3). табл. 3). табл. 3). Однак чинник ментальної близькості та мотив контрасту не пояснюють надмірних, порівняно з ча- сткою спільного кордону, туристичних потоків із Німеччини до Польщі. Тобто, на перший погляд не зрозуміло, чому, наприклад, майже за тієї ж протяжності кордону до Польщі з України прибуває втричі менше туристів, аніж із Німеччини (див. Це можна пояснити лише тим, що Німеччина є світовим лідером на ринку зарубіжного туризму. В 2008 р. мешканці цієї країни здійснили 73 млн. вибуттів, тоді, як українців за кордон подорожувало значно менше – 15,5 млн. (див. табл. 4). Таблиця 3 Структура туристичних прибуттів із країн-сусідів, 2008 р. Частка, % № Країни походження Прибуттів Кордону Польща 1 Німеччина 36,9 15,3 2 Україна 12,0 17,3 Всього 48,9 32,6 Структура туристичних прибуттів із країн-сусідів, 2008 р. Таблиця 4 Зарубіжні туристичні вибуття, 2008 р. Зарубіжні туристичні вибуття, 2008 р. № Країни Вибуття % 1 Німеччина 73000000 82,5 2 Україна 15499000 17,5 Всього 88499000 100,0 спільного з нею кордону становить 15%. Водночас, частка кордону з ментально ближчою Чехією стано- вить 26%, а подорожують до неї лише 8,5% поляків. Ймовірно, тут провідним чинником виступає соціально-економічний розрив, що засвідчується істотною різницею в питомих (Per capita) споживчих витратах домогосподарств, які в 2008 р. для Польщі становили 8454 US$, а для Німеччини – 25212 US$. спільного з нею кордону становить 15%. Водночас, частка кордону з ментально ближчою Чехією стано- вить 26%, а подорожують до неї лише 8,5% поляків. Ймовірно, тут провідним чинником виступає соціально-економічний розрив, що засвідчується істотною різницею в питомих (Per capita) споживчих витратах домогосподарств, які в 2008 р. для Польщі становили 8454 US$, а для Німеччини – 25212 US$. Виходячи з цього, на частку прибуттів від сусідів впливає не лише протяжність спільного кор- дону суходолом, але й обсяг зарубіжного туризму в країні походження туристів, який вимірюється кількістю вибуттів. Польща також була винятком для зарубіжного туризму. Зокрема 25% польських туристів виїжджають до сусідньої Німеччини, тоді як частка 106 2017 Часопис соціально-економічної географії випуск 23(2) Потоки у цьому випадку формуються від менш розвинутої до розвинутішої країни. Це пов’язане, головне, з двома причинами: по-перше люди відвідують країну, яка знаходиться за своїм розвит- ком на щабель вище не лише з цікавості, але й задля походів магазинами (шопінг-тури прикордонними територіями); по-друге, до багатої країни здійснюються трудові міграції або виїжджають на постійне місце проживання, тому саме родичі та знайомі, які відвідують цих мігрантів, можуть фор- мувати такі туристичні потоки. Висновок. Протяжність географічного простору впливає на розподіл міжнародних туристичні потоків, насамперед, через відстані між країнами походження і призначення туристів, а також через довжину спільного кодону суходолом, яка у багатьох випадках визначає обсяги туристичного обміну між країнами-сусідами. Проведений аналіз не підтвердив припущення про те, що чим ближче знаходиться країна призначення тим частіше вона відвідується. Водночас була виявлена закономірність, за якою ча- стка прибуттів або вибуттів у структурі міжнародного туризму певної країни може бути будь-якою, але, зазвичай, не перевищує значення, яке із відстанню зменшується за експонентою. Інакше кажучи, відстань не визначає, а обмежує інтенсивність туристичного обміну між країнами, тобто є обмежуючим чинником. Повертаючись до інших винятків для зарубіж- ного туризму, у яких частки вибуттів до сусідніх кра- їн не узгоджуються з протяжністю спільних держав- них кордонів, відзначимо, що зі Словенії 55% пря- мують до сусідньої Хорватії. Таблиця 4 Крім спільного кордо- ну, цьому, насамперед, сприяє мовна та історична близькість, бо словенська та хорватська входять до західної підгрупи південнослов’янських мов, а обид- ва народи до 1991 року «проживали» в одній країні – Югославії. Така ж ситуація склалася з Чехією, адже 20% туристів подорожували до Словаччини, хоча частка кордону з нею становила лише 11%. У цьому випадку також спрацьовує чинник ментальної близькості: обидві країни не лише «розмовляють» мовами однієї підгрупи західнослов’янських мов, але й мають спільну історію під «назвою Чехословаччи- на». Також було встановлено, що близько 50% міжнародного туристичного обімну відбувається між країнами, які межують суходолом або кордони кот- рих розташовані на відстані до 50 км, а такі потоки мають характер «дифузії». Тому, за інших рівних умов, частка країни-сусіда в структурі туристичних прибуттів або вибуттів часто узгоджується з протяжністю спільного кордону суходолом. Винятки з цього правила, зазвичай, пов’язані зі спотворенням ізотропності географічного простору, насамперед, на прикордонних територіях. Список використаних джерел: 1. Александрова А.Ю. Международный туризм [Текст]: учебник для студентов высших учеб. заведений / Анна Юрьевна Александрова. – М.: Аспект-Пресс, 2002. – 470 с. 1. Александрова А.Ю. Международный туризм [Текст]: учебник для студентов высших учеб. заведений / Анна Юрьевна Александрова. – М.: Аспект-Пресс, 2002. – 470 с. 2. Єріна А.М. Статистичне моделювання та прогнозування [Текст]: навч. посiбник / Антоніна Михайлівна Єріна. – К. : КНЕУ, 2001. – 170 с. 2. Єріна А.М. Статистичне моделювання та прогнозування [Текст]: навч. посiбник / Антоніна Михайлівна Єріна. – К. : КНЕУ, 2001. – 170 с. 3. Король О.Д. Міжнародний туризм: методика і матеріали статистичних досліджень [Текст] / О.Д. Король, Т.Д. Скутар. – Чернівці : Рута, 2008. – 64 с. 3. Король О.Д. Міжнародний туризм: методика і матеріали статистичних досліджень [Текст] / О.Д. Король, Т.Д. Скутар. – Чернівці : Рута, 2008. – 64 с. 4. Любіцева О.О. Ринок туристичних послуг [Текст]: навч. посiбник / Ольга Олександрівна Любіцева. – 2-е вид., перероб. та доп. – К.: Альтерпрес, 2003. – 436 с. р р р р 5. Hagerstrand T. (1970). What about people in regional science? Lund: Springer – Verlag. g p p g p g g 6. International tourism [Електроний ресурс] : UNWTO Tourism Highlights, 2001-2016 Editions. – Режим доступу до щорічника: http://mkt.unwto.org/publications 7. OECD Tourism Trends and Policies 2010 [Електроний ресурс]: OECD Publishing, 2010. – Режим доступу до електронної книги : https://books.google.com.ua/books?id=UC7WAgAAQBAJ 8. Tourism flows outbound (by countries) [Електроний ресурс]: Euromonitor International, 2012. – Режим доступу до електронного видання : http://www.euromonitor.com/travel References: 1. Aleksandrova, A. u. (2002). Mezhdunarodnyi turizm: uchebnik dlya studentov vysshyx ucheb. zavedeniy [Interna- tional tourism: Textbook for university students]. Moscow: Aspekt-Press, 470. y p 2. Erina, A.M. (2001). Statystychne modelyuvannya ta prognozuvannya [Statistical modeling and forecasting]. Kyiv: KNEU, 170. 3. Korol, O.D., Skutar T.D. (2008). Mizhnarodnyi turyzm: metodyka i materialy statystychnykh doslidzhen [Interna- tional tourism: methods and materials of statistical research]. Chernivtsi: Ruta, 64. 4. Lyubitseva, O.O. (2003). Rynok turystychnykh posluh [The market of tourist services]. Kyiv: Alterp 5. Hagerstrand, T. (1970). What about people in regional science? Lund: Springer – Verlag. T. (1970). What about people in regional science? Lund: Springer – Verlag. tourism: UNWTO Tourism Highlights, 2001-2016 Editions. Available at: nwto.org/publications g , ( ) p p g p g g 6. International tourism: UNWTO Tourism Highlights, 2001-2016 Editions. A http://mkt.unwto.org/publications g p ourism Trends and Policies 2010: OECD Publishing (2010). Available at: .google.com.ua/books?id=UC7WAgAAQBAJ p g p 7. OECD Tourism Trends and Policies 2010: OECD Publishing (2010). https://books.google.com.ua/books?id=UC7WAgAAQBAJ 8. Tourism flows outbound (by countries): Euromonitor International (2012). Available at: http://www.euromonitor.com/travel 8. Tourism flows outbound (by countries): Euromonitor International (2012). Available at: http://www.euromonitor.com/travel Надійшла до редколегії 09.09.2017 р. 107
https://openalex.org/W4387342791	https://www.ijfmr.com/papers/2023/5/6978.pdf	English	null	Unraveling the Academic Odyssey: Challenges, Discipline Dynamics, and Remedial Measures for Head Teachers in Public Secondary Schools of Micheweni District	International Journal For Multidisciplinary Research	2,023	cc-by-sa	3,374	International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com Abstract: The study focused on challenges head teachers face in enhancing students’ academic performance in Zanzibar secondary schools specifically in Micheweni District. Thus, the study aimed to address the challenges that head teachers face in improving academic performance. To determine the possible measures to address the challenges head teachers face in improving academic performance in public secondary schools in Micheweni district. Total sampled respondents were 160 people from 10 public secondary schools. This study involved 10 head teachers, 49 class teachers, 100 students and 1 District Educational Officer was selected as a sample. The study employed qualitative design by using interview, focused group discussions, observation and document review as methods of data collection. The overall findings show that head teachers worked in hard environments as compared to when the free education policy was not in operation from the teachers and the parents. The study recommended establishing extra curriculum for examination classes to improve student performance and to provide chance for teachers’ professional development through in-service training in order to achieve quality education. Keywords: Administrative, Human resources, Financial, and Personal challenges, High Academic performance Unraveling the Academic Odyssey: Challenges, Discipline Dynamics, and Remedial Measures for Head Teachers in Public Secondary Schools of Micheweni District Ahmed Rashid Hamad Jordan University collage Department of Education, P. O. Box 1878, Morogoro, Tanzan Background to the study. Globally, head teachers face challenges in achieving student performance in Public secondary school’s challenges such as, English teacher shortages, lacking materials, and unqualified teachers. Untrained, under-qualified teachers omit syllabus topics and hinder language skills. Challenges in public secondary schools include large classes, teacher workload, poor facilities, limited home support, and poverty Mosha, (2014). In Latin America hasn't fully addressed disparities relating on the Educational expansion with half of completing students lacking proficiency (Brunner 2009). Developing countries aim to balance enrollment and quality. Head teachers' roles focus on creating conducive learning environments (UNESCO, 2009). Eastern African countries, have undergone multiple reforms in its education system to align with global standards. In a study conducted by Mkumbo (2012) posited a low level of teachers' commitment, linked to poor performance among schools and students. In Tanzania, the government initiated Development Vision 2025 aims to achieve four main objectives: providing well-qualified human resources, establishing a highly standardized and relevant education system, addressing developmental disparities among regions, and improving transition rates from primary to secondary and tertiary levels. In relation to this study particularly in Zanzibar whereas the Micheweni district is the case study. At Ministerial level of education and Vocational Training, Performance in secondary education management in services delivery and development in Zanzibar has been raised as in the contemporary years on the development of digital system in the universe. In fact, the country has introduced free education system after Revolution in1964. The education services have been improved year after year. MoEVT, (2006). ational Journal for Multidisciplinary Research (IJFMR) Head teacher: means secondary disciplinary authority at school level where he over sees and regulate teachers conduct at school Public schools: are schools that owned by community and supported by government in terms of salaries of the teachers and supplying teaching and learning facilities like books (URT, 2010). High academic performance: means the matter what employees can achieve and how they achieve it. Similarly, performance means effect use of needed skills, knowledge and experience. 2. General Objective To determine the challenges that head teachers face in improving and maintaining high academic performance in public secondary schools in Micheweni District. 1. Introduction Education is an essential tool for personal and societal development and it is often considered as the key for eradication of poverty especially in developing countries like Tanzania. However, in Micheweni District, the academic performance of students in secondary schools is often poor with low pass rates. According to the Tanzania National Examination council (NECTA), academic performance is the measurement of student’s achievement across various academic subjects and the attainment of their short term or long-term educational goals. While there are many challenges that contribute to poor academic performance, this study focused on the challenges head teaches face that contributing to poor academic performance in Micheweni District. Challenges: which mean difficulties or problems come from all domain at the same time and requires great effort. Volume 5, Issue 5, September-October 2023 1 IJFMR23056978 1 Research design This study employed qualitative approach whereby the researcher used interview focus grouped discussion and documentary review as tools of data collection. The sample of 160 people comprising students 100 (80%) as main respondents district educational officer and head teachers 10 and class teachers 49 (20%) secondary schools in Micheweni District Volume 5, Issue 5, September-October 2023 2 IJFMR23056978 4.3Emergency of different classes among teachers based on political ideology 4.3Emergency of different classes among teachers based on political ideology In the interview with Head Teacher (HT3), it was revealed that there is a presence of different political ideologies among teachers in the school, which can lead to challenges in maintaining a harmonious educational environment. This situation requires strong administrative skills to navigate. While diverse political perspectives among teachers can have positive outcomes, such as promoting critical thinking skills, open dialogue, and exposure to various viewpoints, it also raises concerns. These concerns include the potential for teachers' biases to influence classroom discussions, the risk of polarization within the school community, and the need to ensure that hiring and promotion decisions are based on merit rather than political affiliations. 4.1 Administrative Challenges Faced by Head Teachers g y In this section the researcher involved several head teachers in the study. The question asked relating on the administration aspect. The following interviews for the first Head Teacher (HT1) the first question was ‘What are the administrative challenges that you have been facing as a Head teacher in achieving student’s high academic performance in your office?’ The followings are the results from the respondents 4.2 Shortage of some administrative skills and tools such as policy, Acts, and Regulations. 4.2 Shortage of some administrative skills and tools such as policy, Acts, and Regulations. [HT1: As head teachers in this school have found the following challenges that threat high academic performance to public secondary school in Micheweni. The challenges include shortage of some administrative skills and tools such as policy, acts, and regulations. Normally the Ministry of education appoints head teachers with only their point of views but even myself I don’t know the proper criteria that they considered while appointing. The roles outlined by the ministry does include the administrative skills and tools that may help to develop skills and capability on effective and management control system at the school. It is different to find that skills have been tough to make head teachers have been built their capable in the area mentioned. [HT1: As head teachers in this school have found the following challenges that threat high academic performance to public secondary school in Micheweni. The challenges include shortage of some administrative skills and tools such as policy, acts, and regulations. Normally the Ministry of education appoints head teachers with only their point of views but even myself I don’t know the proper criteria that they considered while appointing. The roles outlined by the ministry does include the administrative skills and tools that may help to develop skills and capability on effective and management control system at the school. It is different to find that skills have been tough to make head teachers have been built their capable in the area mentioned. 4.0 STUDY FINDINGS The first specific objective was to identify the challenges that head teachers face in improving and maintaining high academic performance in their public schools in Micheweni District. In this objective, the researcher wanted to identify the different challenges that have been facing head teachers in the public secondary school in Micheweni District. In this aspect the interview questions as the adopted tool by the researcher were used to ask selected head teachers in Micheweni. There were four different questions used by the researcher to ask the head teachers. Total of 10 different head teachers from 10 public secondary school in Micheweni were involved in providing information concerning the challenges facing head teachers in achieving student’s high academic performance in public secondary school in Micheweni. The Researcher also asked head teachers (HT) from among the 10 secondary schools in Micheweni. The main question was based on the Challenges facing head teachers. 5. Summary of the study The study sought to determine the challenges that head teachers face in improving and maintaining high academic performance in public secondary schools in Micheweni District. It further explored the process that head teachers use in maintaining the levels of school discipline among teachers and students and to determine possible measures to identified challenges of head teachers. Possible Measures to Overcome Head Teachers’ Challenges Possible Measures to Overcome Head Teachers’ Challenges The findings on the third objective which was to determine the possible solutions to the head teachers’ challenges are outlines as follows. The MoEST ought to supply necessary laws, policies, acts and regulations to Head teachers. The school administration and parents of students have to show commitment towards time management so that they can achieve the schools’ desired goal which is students’ high academic performance. There is also need for proper arrangement of in-service training for teachers in order to improve their skills and capacity which are inextricably connected with students’ academic performance. 4.5 Management of the Teachers, Students and Surrounded School Environment Researcher also developed to ask the respondents 5 head teacher (HT5) who was the five head teacher involved in this study. Challenges that Head Teachers Face The findings on the first objective of the current research study identified that, there are three main challenges that head teachers face in improving and maintaining student’s academic performance in their public secondary schools. Firstly, administrative challenges which include lack of administrative skills, policies, acts and regulations. Secondly, human resource challenges and thirdly, financial challenges which bring about shortage of funds for teaching and learning materials as well as ICT equipment. E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com Political interaction by different political interest. This normally happens when political leaders find that when the general election in nearby, they could be closest to the teachers and students to participate in general election. It could be all parties not only ruling part but also other political parties. Several times I have faced with this challenge. (HT4) Head Teachers’ Strategy for Maintaining School Discipline Findings on the second objective of the study identified the strategies that head teachers use to maintain discipline among teachers and students. These were developing close relationship between teachers, students and parents. Head teachers’ ability to encourage and boost the morale of teachers so as to be patriotic through participating in school’s local income generating activities that secure funds for school’s administrative work. They also help to negotiate with Government through MoEST to be given funds for the smooth running of their schools. Enhance the students’ parents or guardians to contribute petty cash to cater for the teaching and learning resources. 4.4 Political Interaction by Different Political Interest The researcher asked respondent four head teacher 4 (HT4) concerning the challenges that facing on administration, he answered that, Volume 5, Issue 5, September-October 2023 IJFMR23056978 3 International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com 7. Recommendations 1. The Ministry of Education should prioritize and invest in continuous professional development for head teachers and teachers alike. This should include training in administrative skills, subject specialization, and effective teaching techniques. 2. The Ministry should provide clear administrative guidelines and tools to head teachers to facilitate effective school management. This can help address challenges related to policy, time management, and political interactions. 2. The Ministry should provide clear administrative guidelines and tools to head teachers to facilitate effective school management. This can help address challenges related to policy, time management, and political interactions. 3. The Ministry should establish programs to enhance the skills and capacity of teachers, especially those handling examination classes or subjects outside their specialization. Regular training and support can improve their effectiveness in the classroom. 4. Schools should develop and implement effective motivational strategies for teachers to boost their morale and job satisfaction. Adequate financial incentives, timely bus fare, and recognition for outstanding performance can help motivate educators. 5. Schools should collaborate with parents and guardians to address issues of student discipline. Clear disciplinary policies and mechanisms should be in place to ensure a conducive learning environment. 6. The Ministry of Education should review its funding mechanisms to ensure that schools receive adequate financial support. This includes addressing issues related to the availability of funds for school development and infrastructure improvement. 7. Investments should be made to improve ICT infrastructure in schools. Access to information and technology is crucial for modern education, and schools should have the necessary resources to provide students with up-to-date information. 7. Investments should be made to improve ICT infrastructure in schools. Access to information and technology is crucial for modern education, and schools should have the necessary resources to provide students with up-to-date information. 8. Encourage greater involvement of parents, guardians, and the community in school affairs. This can help bridge financial gaps and enhance the overall school environment. International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com On the human resource front, head teachers grapple with challenges related to skills mismatch among 6. Conclusions The findings of this research highlight the various challenges faced by head teachers in enhancing students' high academic performance in public secondary schools in Micheweni District. These Volume 5, Issue 5, September-October 2023 IJFMR23056978 4 International Journal for Multidisciplinary Research (IJFMR) E-ISSN: 2582-2160 ● Website: www.ijfmr.com ● Email: editor@ijfmr.com challenges span administrative, human resource, and financial domains, and they significantly impact the overall effectiveness of school management and academic outcomes. Administratively, head teachers struggle with issues such as the shortage of administrative skills and tools, time management, political interactions, and maintaining discipline among students and staff. These challenges underscore the need for continuous professional development and clear administrative guidelines to address them effectively. On the human resource front, head teachers grapple with challenges related to skills mismatch among teachers, inadequate skills and capacity in some educators, poor motivational strategies, and issues related to student discipline. Addressing these challenges necessitates targeted training and support for teachers, as well as effective strategies for fostering a positive school environment. Financial constraints are also a significant hurdle, with inadequate funds affecting important aspects of school development, such as ICT infrastructure, educational materials, and overall school progress. The need for sustainable sources of income at the school level and improved financial support from the Ministry of Education is evident. challenges span administrative, human resource, and financial domains, and they significantly impact the overall effectiveness of school management and academic outcomes. Administratively, head teachers struggle with issues such as the shortage of administrative skills and tools, time management, political interactions, and maintaining discipline among students and staff. These challenges underscore the need for continuous professional development and clear administrative guidelines to address them effectively. On the human resource front, head teachers grapple with challenges related to skills mismatch among teachers, inadequate skills and capacity in some educators, poor motivational strategies, and issues related to student discipline. Addressing these challenges necessitates targeted training and support for teachers, as well as effective strategies for fostering a positive school environment. Financial constraints are also a significant hurdle, with inadequate funds affecting important aspects of school development, such as ICT infrastructure, educational materials, and overall school progress. The need for sustainable sources of income at the school level and improved financial support from the Ministry of Education is evident. challenges span administrative, human resource, and financial domains, and they significantly impact the overall effectiveness of school management and academic outcomes. Administratively, head teachers struggle with issues such as the shortage of administrative skills and tools, time management, political interactions, and maintaining discipline among students and staff. These challenges underscore the need for continuous professional development and clear administrative guidelines to address them effectively. REFERENCES 1. Aboud, F. (2017). Challenges Facing Students in Secondary Schools: A Case Study of Mkoa wa Mjini Magharibi, Zanzibar. Journal of Educational Policies and Implementation Research, 3(2), 61- 68. 2. Brunner, J. J. (2009). Educational expansion and inequalities in Latin America: A background paper for the International Institute for Educational Planning. UNESCO International Institute for Educational Planning. 3. Charles, A. & Mkulu, D. G. (2020). Management challenges facing school administrators and pupils’ academic performance in public primary schools in Sengerema district, Mwanza, Tanzania. Journal of Humanities and Education Development (JHED), 2(3), 191-207. 3. Charles, A. & Mkulu, D. G. (2020). Management challenges facing school administrators and pupils’ academic performance in public primary schools in Sengerema district, Mwanza, Tanzania. Journal of Humanities and Education Development (JHED), 2(3), 191-207. 4. Juma, A., Smith, J., & Hassan, H. (2018). Challenges Faced by Secondary School Students in Zanzibar: A Case Study of Unguja North Region. International Journal of Education and Research, 6(5), 89-100. 4. Juma, A., Smith, J., & Hassan, H. (2018). Challenges Faced by Secondary School Students in Zanzibar: A Case Study of Unguja North Region. International Journal of Education and Research, 6(5), 89-100. 5. Khamis, J. M. (2017). Challenges Facing Head Teachers in Implemeting Free Primary Education Policy: A Case Study of Zanzibar West District Unguja . Dodoma: University of Dodoma . 6. Kuluchumila R..C (2009). The implementation of secondary Education Plan in Tanzania: a case study of community secondary school heads in Shinyanga: Journal of Education and practice. ISSN 2222-1735 vol 4 No 12 7. Mohammed, S. (2016). The Principals’ Supervisory Roles for Quality Education and Effective School Administration of Basic Education Schools in Nigeria. Northwest University Journal of Humanities and Education Development (JHED) 7. Mohammed, S. (2016). The Principals’ Supervisory Roles for Quality Education and Effective School Administration of Basic Education Schools in Nigeria. Northwest University Journal of Humanities and Education Development (JHED) 8. Mosha, H. (2014). New directions in teacher education for quality improvement in Africa. Paper in Education Development, 24, 25-28. 8. Mosha, H. (2014). New directions in teacher education for quality improvement in Africa. Paper in Education Development, 24, 25-28. 9. Msafiri, S. J. (2019). The Challenges Faced by Secondary School Teachers in the Integration of Information and Communication Technology in Teaching and Learning: A Case of Zanzibar Secondary Schools. 8. Suggestions for further study • The role of parents in enhancing students’ academic performance. • The impact of drug and substance abuse on the discipline of students in public secondary schools. • The effect of teachers and parents’ partnership on the students’ academic performance in public secondary schools. Volume 5, Issue 5, September-October 2023 IJFMR23056978 5 p f y gy ational examination council NECTA (2022). evelopment vision (2025). REFERENCES Journal of Education, Society and Behavioural Science, 29(4), UNESCO, 2009 () The new roles of secondary school head teachers Published in 2006 by the United Nations Educational, Scientific and Cultural Organization 7, Place de Fontenoy, 75352 Paris 07 SP. 10. Onyango (2001) Conduct a study in Nairob and Kaka mega to determine the competencies needed by head teachers for effective management. 1. Orodha, et, al (2014). Researchers and policy makers alike suspect lack of equal acces 11. Orodha, et, al (2014). Researchers and policy makers alike suspect lack of equal access to education 12. URT (2010) Education sector development program: primary education initiation plan, BEDS Dar es Salaam: Tanzania. 12. URT (2010) Education sector development program: primary education initiation plan, BEDS Dar es Salaam: Tanzania. 13. Patton, M. (1990). Qualitative Evaluation and research methods, Newbury Park CA: Sage.1-15. 13. Patton, M. (1990). Qualitative Evaluation and research methods, Newbury Park CA: Sage.1-15. 14. Pintrich, P. R., & De Groot, E. V. (1990). Motivational and self-regulated learning components of classroom academic performance. Journal of Educational Psychology, 82(1), 33–40. 15. Tanzania National examination council NECTA (2022). 16. Tanzania Development vision (2025). 17. Zanzibar Education Policy MoEVT (2006). 17. Zanzibar Education Policy MoEVT (2006). Volume 5, Issue 5, September-October 2023 IJFMR23056978 6
https://openalex.org/W2022457765	https://asp-eurasipjournals.springeropen.com/counter/pdf/10.1186/1687-6180-2012-117	English	null	A methodology for time-frequency image processing applied to the classification of non-stationary multichannel signals using instantaneous frequency descriptors with application to newborn EEG signals	EURASIP Journal on Advances in Signal Processing	2,012	cc-by	18,585	Abstract This article presents a general methodology for processing non-stationary signals for the purpose of classification and localization. The methodology combines methods adapted from three complementary areas: time-frequency signal analysis, multichannel signal analysis and image processing. The latter three combine in a new methodology referred to as multichannel time-frequency image processing which is applied to the problem of classifying electroencephalogram (EEG) abnormalities in both adults and newborns. A combination of signal related features and image related features are used by merging key instantaneous frequency descriptors which characterize the signal non-stationarities. The results obtained show that, firstly, the features based on time-frequency image processing techniques such as image segmentation, improve the performance of EEG abnormalities detection in the classification systems based on multi-SVM and neural network classifiers. Secondly, these discriminating features are able to better detect the correlation between newborn EEG signals in a multichannel-based newborn EEG seizure detection for the purpose of localizing EEG abnormalities on the scalp. Keywords: time-frequency signal analysis, multichannel signal analysis, instantaneous frequency, time-frequency image processing, image segmentation, time-frequency feature extraction, seizure detection, EEG classification, loca- lization, newborn EEG A methodology for time-frequency image processing applied to the classification of non- stationary multichannel signals using instantaneous frequency descriptors with application to newborn EEG signals Boualem Boashash1,2, Larbi Boubchir1* and Ghasem Azemi2,3 Boualem Boashash1,2, Larbi Boubchir1* and Ghasem Azemi2,3 Boualem Boashash1,2, Larbi Boubchir1* and Ghasem Azemi2,3 © 2012 Boashash et al; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. RESEARCH Open Access A methodology for time-frequency image processing applied to the classification of non- stationary multichannel signals using instantaneous frequency descriptors with application to newborn EEG signals Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 * Correspondence: larbi@qu.edu.qa 1Electrical Engineering Department, Qatar University College of Engineering, Doha, Qatar Full list of author information is available at the end of the article 1 Introduction to non-stationary EEG signal processing for diagnostic of abnormalities This is important because these affect both adults and infants in a significant way. For example, newborn con- genital anomalies, often referred to as birth defects, have a variety of causes ranging from pregnancy or birth complications to genetic malformations to viral infec- tions in utero [1]. Epileptic seizures are among the most observed abnormalities, and result from problems such as lack of oxygen, haemorrhage, meningitis, infection and strokes [2]. The best tool for diagnosing the differ- ent abnormalities is the EEG which uses special sensors (electrodes) placed on the surface of the scalp to mea- sure the electrical activity of the brain [3]. An illustra- tion of EEG multichannel measurements using EEG Scalp electrodes is shown in Figure 1. This article is a novel contribution in three aspects. The first is a review of time-frequency (T-F) image proces- sing techniques, i.e. image processing techniques that process the T-F representation of a signal considered as an image. The second aspect is the extension of tradi- tional mono-channel T-F analysis to multichannel analy- sis. The third is to combine the previous two aspects and propose a more complete multichannel T-F image processing approach to the problem of Electroencepha- logram (EEG) abnormality diagnostic and localization. * Correspondence: larbi@qu.edu.qa 1Electrical Engineering Department, Qatar University College of Engineering, Doha, Qatar Full list of author information is available at the end of the article * Correspondence: larbi@qu.edu.qa 1Electrical Engineering Department, Qatar University College of Engineering, Doha, Qatar Full list of author information is available at the end of the article The EEG is a representative signal containing informa- tion about the electrical activity generated by the Page 2 of 21 Page 2 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Figure 1 EEG scalp sensor electrodes (from [3]): (a) Lateral view of the electrode positions in a realistic display. (b) Electrode position and labels for various systems. Black circles indicate positions of the original 10-20 system, gray circles indicate additional positions in the 10-10 extension and small dots indicate additional positions in the 10-5 extension [3]. Figure 1 EEG scalp sensor electrodes (from [3]): (a) Lateral view of the electrode positions in a realistic display. (b) Electrode position and labels for various systems. 1.1 EEG abnormalities detection using DSP The focus of this study is to analyze abnormal newborn EEG signals in order to detect the different abnormal- ities and to locate their spatial distribution on the scalp. This abnormality detection and localization problem can be solved by analyzing newborn EEG signals and extracting features, which are then classified. Step 2, feature extraction, then extracts this informa- tion in order to assign it to one abnormality type using the classification method used in Step 3. A typical scheme for EEG abnormalities detection using DSP includes the following three steps: (1) signal analysis in either time, frequency, time-scale or joint time frequency domain, (2) features extraction which characterize the different abnormalities and (3) classifi- cation of these features in order to assign the signal to one abnormality class with a relevant degree of severity: mild, moderate or severe abnormality. Research on EEG abnormalities detection is wide and varied, but almost in all previous studies, a single channel signal analysis approach is used (see Section 1.3). Although, the EEG abnormalities detection and classification system receives an input of more than one EEG channel, the ana- lysis is made for a single channel only. Hence, one main goal of this study is to not only examine the EEG abnorm- ality detection problem but also consider the localization of this abnormality in the brain and on the scalp, using a mul- tichannel analysis of the multichannel EEG signal. Step 1, signal analysis, can be either a single channel analysis or a multichannel analysis, depending on the nat- ure of the data and the application considered. The use of single channel analysis for complex systems description can produce incorrect system model and get false results. So, it is often more desirable to use instead a multichannel analysis which offer a much more accurate model although it may complicate the system model and lead to an increase in computation load. In the multichannel case, 1 Introduction to non-stationary EEG signal processing for diagnostic of abnormalities Black circles indicate positions of the original 10-20 system, gray circles indicate additional positions in the 10-10 extension and small dots indicate additional positions in the 10-5 extension [3]. non-seizure waveforms. The manual detection of these EEG abnormalities includes visual scanning of EEG recordings which is very time consuming especially in the case of long recordings [5]. It also requires skilled cerebral cortex nerve cells; it is the most utilized signal to clinically assess brain activities, and is widely used to detect abnormalities such as epilepsy and seizure [4]. Figure 2 shows an example of newborn EEG seizure and Figure 2 Newborn EEG seizure and non-seizure patterns (1st row) with their power spectral densities (2nd row). Figure 2 Newborn EEG seizure and non-seizure patterns (1st row) with their power spectral densities (2nd row). Page 3 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 EEG data can be represented by a d dimensional vector [s1 (t)s2(t) ... sd(t)], where d represents the number of chan- nels. Figure 3 shows an example of array EEG for newborn in the normal and seizure case, and their corresponding electrode placement on the scalp. interpreters; i.e. a neurophysiologist, who could be prone to subjective judgment and error. So, instead, it is desired to get the EEG signal parameters extracted and analyzed using computer based digital signal processing (DSP) techniques. Such an approach is highly useful in diagnostics and more suitable for automatic EEG abnormalities detection and classification [6,7]. For a single channel, EEG data would be represented by only the symbolic notation s1(t). Then, the signal analysis permits to exploit all information necessary to characterize different abnormalities in the newborn EEG signal. However, this is in general only an approxima- tion aimed at simplifying the problem and this leads to inaccuracies that are detrimental to the application, and it is therefore preferable to consider the multichannel case when possible. 1.2 EEG abnormality localization using multichannel analysis The EEG abnormality localization can be achieved by: (1) extracting features, which are able to detect and Figure 3 An example of array EEG for newborn. (a) Normal and seizure newborn EEGs. (b) Electrode placement for measuring neonatal EEG: the measurement of brain activity via the EEG is performed by attaching several electrodes (or antennae) to the head of the newborn. These electrodes are placed according to the International 10-20 system. This example shows a montage which is built from recordings of F3, F4, Cz, C3, C4, T3, T4, O1, O2. Figure 3 An example of array EEG for newborn. (a) Normal and seizure newborn EEGs. (b) Electrode placement for measuring neonatal EEG: the measurement of brain activity via the EEG is performed by attaching several electrodes (or antennae) to the head of the newborn. These electrodes are placed according to the International 10-20 system. This example shows a montage which is built from recordings of F3, F4, Cz, C3, C4, T3, T4, O1, O2. Figure 3 An example of array EEG for newborn. (a) Normal and seizure newborn EEGs. (b) Electrode placement for measuring neonatal EEG: the measurement of brain activity via the EEG is performed by attaching several electrodes (or antennae) to the head of the newborn. These electrodes are placed according to the International 10-20 system. This example shows a montage which is built from recordings of F3, F4, Cz, C3, C4, T3, T4, O1, O2. Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 4 of 21 transform (DWT) of EEG signals [11], combination of DWT coefficients and chaotic measures [23], energy dis- tribution of EEG signals in the T-F representation [24,25], and T-F matched filtering methods [26,27]. A comparison of the performance of different feature sets in classifying EEG signals can be found in [5]. In [12,13], a combination of features in time, frequency, and time-scale domains and chaotic measures are used for neonatal seizure detection. characterize the abnormalities, from the multichannel EEG signals; and (2) using these features to then identify the channels or electrodes in which the abnormalities are present. This procedure locates the sources of the abnormality on the scalp. Figure 4 presents the relevant block diagrams of the two discussed signal analysis schemes in the joint T-F domain. 1.2 EEG abnormality localization using multichannel analysis Figure 4-(a) illustrates the single signal analysis scheme and Figure 4-(b) exhi- bits the multichannel analysis scheme. Due to the non-stationarity of EEG signals, T-F and instantaneous frequency (IF) based methods seem naturally more suitable for seizure detection and clas- sification [7,26,28-32] and [[6], Section 15.5]. Recently, a study exploited the additional information provided by signal variations in terms of non-stationarity observed with the T-F approach, and then, developed novel features extracted from the T-F representation of EEG signals for the purpose of classifying them [14-16]. 1.3 The T-F approach to non-stationary EEG analysis Several automated techniques have been proposed for detecting EEG seizures [5,8-16]. These techniques are based on time, frequency, non-stationary and nonlinear characteristics of EEG signals. Most of these methods use single channel EEG as their input and are based on a methodology similar to the one previously defined and illustrated in Figure 4-(a)[14-16]. Among these methods, we note those which use time domain statistics [8], spectral features [9], autoregressive modeling [17], and linear prediction error energy [18]. Other methods use a combination of time and fre- quency features [19], chaotic features [20-22], fast Four- ier transform (FFT) coefficients [10]; more advanced methods use the coefficients of the discrete wavelet A variety of methods have already been applied for automatic seizure detection using multichannel EEG. This includes, for example, the use of T-F matched fil- ters [7,33,34], multichannel matching pursuit [35], spa- tial and temporal contextual information obtained from multichannel EEGs [36]. Figure 4 EEG-based automated abnormalities detection and classification scheme: (a) Block diagram of traditional single analysis. (b) Block diagram of multichannel analysis. Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 5 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 newborn EEG signals given in Figure 2, using the modi- fied-B distribution (MBD). One of the aims of this article is to extend the pro- posed approach in [14-16] using the full multichannel EEG signal in order to define a new multichannel abnormalities detection and localization approach for the newborn EEG. 2.1.1 EEG time-frequency based feature selection approach A T-F based feature selection approach can improve the classification of EEG signals (illustrated in Figure 4-(a) acquired from healthy and epileptic subjects [14-16]. This is achieved by exploiting the link between the non- stationary features of the signal and considering their impact on the performance of the classification scheme. This new classification method is based on a feature vector composed of ten T-F features extracted from the TFD of EEG segments. These include: the IF and singu- lar value decomposition (SVD) based features, the T-F complexity, and the energy concentration measures (see detail in Section 4). The article is organized as follows. Section 2 presents a brief review of T-F signal analysis and IF estimation using quadratic time-frequency distributions (QTFDs). 1.3 The T-F approach to non-stationary EEG analysis Section 3 describes a review of T-F image processing techniques applied to the time-frequency distribution (TFD) considered as an image. The methodology employed to extract new T-F features using T-F image processing tools for classifying EEG signals is described in Section 4. Experimental results with discussion are reported in Section 5. Finally, the article concludes with Section 6. Statistical analysis and experimental results given in [14-16], have shown that the proposed T-F features extracted from all TFDs perform well and achieve a highly significant discrimination between the EEG sig- nals collected from healthy and those from epileptic patients. Also, the use of reduced interference TFDs (RIDs)-such as MBD, Smoothed Wigner-Ville distribu- tion (SWVD), Gaussian Kernel distribution (GKD), Spectrogram (SPEC), Separable Kernel based RID (SEPK) [[6], Chapter 2 & Section 5.7] slightly improves the performance of the proposed method. This is because RIDs have the ability to reduce the effects of the cross-terms while still providing a good resolution [6]. 2.1 Representing EEGs with TFDs and their IFs p g Time-frequency signal analysis can be seen as the analy- sis of non-stationary signals with time-varying frequency content. Such signals are best represented by a TFD, which is intended to describe how the energy of the sig- nal is distributed over the two-dimensional (2D) T-F space. The TFD not only shows the start and stop times of signal components and their frequency range, but also shows the component variation in frequency with time (which is called the IF) [37,38]. The IF can be esti- mated using a peak detector in the T-F domain that selects the frequency with the maximum value in the T- F representation as a function of time. The use of a TFD for EEG abnormalities detection and classification is therefore the most natural approach, given their prop- erties [6]. Figure 5 shows an example of TFDs of 2.1.2 Need for image processing techniques to extract new features 2.1.2 Need for image processing techniques to extract new features 2 T-F analysis of multichannel non-stationary EEG signals g This section introduces the concept of T-F signal analy- sis applied to both mono-channel and multichannel non-stationary EEG Signals using quadratic TFDs. 2.1 Representing EEGs with TFDs and their IFs 2.2.1 Formulation of quadratic TFDs Quadratic time-frequency distributions were shown to be the most appropriate methods for the analysis and processing of non-stationary signals in many practical applications. They can be formulated as [6,39]: As all the T-F features discussed in [14-16] were extracted from TFDs of the EEG segments, there is still room for improving the performance of the proposed approach by exploiting the information provided by the full T-F representation viewed as an image. ρ(t, f) = Wz(t, f) ∗∗ (t, f) γ (t, f) (1) (1) The use of T-F image techniques can then improve the method performance by selecting and extracting new T-F features. Section 3.1 reviews the proposed approaches based on T-F image techniques in order to select a new methodology to define new features. where r(t, f) denotes the TFD, Wz(t, f) the Winger- Ville distribution (WVD), g(t, f) the T-F kernel of the distribution, and ∗∗ (t, f) the 2D convolution operation in time and frequency. According to Equation 1, QTFDs can be considered as smoothed versions of the WVD, with g(t, f) being a 2D smoothing filter. Different kernels in Equation 1 define different distributions in the class, that are most specifically adapted to particular classes of signals [[6], pp. 71-76]. The above formulation can also be given in any of the other three joint domains (time- lag (t, τ), doppler-frequency (ν, f) and doppler-lag (ν, τ)) linked to the T-F domain via Fourier transform [6]. where r(t, f) denotes the TFD, Wz(t, f) the Winger- Ville distribution (WVD), g(t, f) the T-F kernel of the distribution, and ∗∗ (t, f) the 2D convolution operation in time and frequency. According to Equation 1, QTFDs can be considered as smoothed versions of the WVD, with g(t, f) being a 2D smoothing filter. Different kernels in Equation 1 define different distributions in the class, that are most specifically adapted to particular classes of signals [[6], pp. 71-76]. The above formulation can also be given in any of the other three joint domains (time- lag (t, τ), doppler-frequency (ν, f) and doppler-lag (ν, τ)) linked to the T-F domain via Fourier transform [6]. 2.1.3 Array processing and multichannel analysis The abnormality localization problem needs to consider the multiple EEG channels simultaneously since EEG data often come as correlated time series from multiple electrodes on the scalp. 2.1.2 Need for image processing techniques to extract new features The improvements provided by the method described in [14-16] are attributed to the T-F approach as this allows the use of extra information related to the non-stationarity of the signals, which is directly available in the T-F domain. Such information includes the distribution of the signal energy over T-F plane and the IF characterizing the variation of its frequency content over time. The impact of this new method is that such variations are the key critical Figure 5 The TFD of newborn EEG signals with seizure (left) and non-seizure (right) patterns given in Figure 2, using the MBD. The TFD in the left illustrates clearly the non-stationary nature and multi-component characteristics of the EEG signal. Figure 5 The TFD of newborn EEG signals with seizure (left) and non-seizure (right) patterns given in Figure 2, using the MBD. The TFD in the left illustrates clearly the non-stationary nature and multi-component characteristics of the EEG signal. Figure 5 The TFD of newborn EEG signals with seizure (left) and non-seizure (right) patterns given in Figure TFD in the left illustrates clearly the non-stationary nature and multi-component characteristics of the EEG signal. Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 6 of 21 Page 6 of 21 information that describes the non-stationarity of the signals. 2.1.4 Multichannel T-F approach for localization The design of a novel approach for automatic abnormal- ities detection and localization using a multichannel EEG signal for newborn is necessary to help the specialists to detect the presence of different abnormalities by interpret- ing the EEG signal in terms of abnormality type and its source localization, and to decide on a suitable diagnosis. The selected approach is to exploit the additional information provided by the signal variation in terms of non-stationarity observed in the TFD; and then, develop novel features extracted from the TFD of the EEG sig- nals using the T-F image processing techniques, for the purpose of detecting the different neonatal abnormalities and classifying the EEG signals. Also, the localization of the abnormality sources on the scalp is dealt with by analyzing the multichannel EEG signal. Using Equation 2, r(t, f) in Equation 1 can be rewrit- ten as [[6], pp. 67]: ρ(t, f) = F τ→f G(t, τ) ∗ t z t + τ 2 z∗ t −τ 2 = ∞ −∞ ∞ −∞ G(u, τ)z t −u + τ 2 z∗ t −u −τ 2 due−j2πfτdτ 2.2.1 Formulation of quadratic TFDs In practice, each channel of the 1D time-domain signal is first transformed to reveal its 2D T-F representation. By stacking all the 2D T-F arrays from all the channels, we form a 3D array in the space “T-F domain”. Then, this 3D array is processed in order to detect and locate with accuracy the abnormality source on the scalp, and also their propagation in the head. To do this, the techni- ques borrowed from radar and communications statisti- cal signal processing can be applied on this 3D data by considering the correlation between 2D T-F arrays. This processing can be made in two ways: (1) correlation between the 2D T-F arrays corresponding to the TFDs of multiple EEG signals, or (2) correlation between the “actual features” extracted from these 2D T-F arrays. At the core of the class formulated by Equation 1 is the WVD which is characterized by a T-F kernel defined as g(t, f) = δ(t)δ(f), where δ is the Dirac delta function. For a real-valued signal s(t), the WVD is defined as [6,39]: Wz(t, f) = +∞ −∞ z t + τ 2 z∗ t −τ 2 e−j2πfτdτ (2) (2) where z (t) = s (t) + H (s (t)) is the analytic associate of s(t), H stands for the Hilbert transform, and z(t) its complex conjugate [[6], pp. 13-15]. The WVD provides the best joint T-F concentration among all QTFDs for linear frequency modulation (LFM) signals [6]. As EEG seizures were shown to be mostly Piece-wise LFM (PWLFM), the WVD seems to be the ideal tool. How- ever, being quadratic in nature, the WVD introduces artifacts, or cross-terms, in the case of multi-component signals and/or non-linear FM signals. The presence of these artifacts can mask the true signal components and make the interpretation of the TFD a difficult task [40]. For those signals, reduced interference TFDs and poly- nomial WVDs may be used [6]. 2.2.2 Instantaneous frequency The IF is a key parameter of a non-stationary signal which describes how its frequency content changes with time [37]. For a mono-component analytical signal z(t) = a(t)ej(t) the IF is defined as [37]: There is a special class of kernels, referred to as separ- able kernel [6], whose members are expressed as the product of two single-variable functions (i.e., g(t, f) = g1 (t)G2(f)). This formulation allows the smoothing in time direction and frequency direction to be quasi-indepen- dently performed. The length and shape of the smooth- ing window g1(t) control outer cross-term reduction. And, G2(f) is the Fourier transform of the analysis win- dow g2(τ) whose length and shape control inner-artifacts reduction [[6], Chapter 3]. The MBD is one such method that was shown to be the best to represent new- born EEG signals using an objective measure for com- parison [6,29]. Figure 5 shows an example of a joint T-F representation of newborn EEG signals using MBD which illustrates the non-stationary and multi-compo- nent characteristics of the EEG signal. Compared to time domain and frequency domain (given in Figure 2), the TFD provides a more informative description of the EEG seizure by revealing the temporal variation of its frequency content through the IF. There is a special class of kernels, referred to as separ- able kernel [6], whose members are expressed as the product of two single-variable functions (i.e., g(t, f) = g1 (t)G2(f)). This formulation allows the smoothing in time direction and frequency direction to be quasi-indepen- dently performed. The length and shape of the smooth- ing window g1(t) control outer cross-term reduction. fi (t) = 1 2π dϕ (t) dt (5) (5) where a(t) and (t) are respectively the instantaneous amplitude and instantaneous phase of the signal. The concept of IF has been successfully used in many tech- nical fields such as radar, sonar, seismic, biomedical applications and communications (e.g. [41]). Many IF estimating techniques have been developed and a detailed review can be found in [38]. where a(t) and (t) are respectively the instantaneous amplitude and instantaneous phase of the signal. The concept of IF has been successfully used in many tech- nical fields such as radar, sonar, seismic, biomedical applications and communications (e.g. [41]). Many IF estimating techniques have been developed and a detailed review can be found in [38]. 2.2 T-F formulation Quadratic time-frequency distributions form a class of TFDs which distribute the signal energy in a joint T-F domain from which we can generate features for detection and classification. Related backgrounds regarding QTFDs and IF estimation are presented in the next sections. (3) Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 7 of 21 Page 7 of 21 where G(t, τ) = F−1 f→τ {γ(t, f)} is the time-lag kernel of the TFD. 2.2.2 Instantaneous frequency For mono-component signals, the peaks of the con- stant-time cross-sections of the TFD as well as its first order moment give estimates of the IF law of the signal. The notion of a single IF becomes inappropriate for a multi-component signal. To characterize this type of sig- nals, an IF law is assigned to each signal component. Various IF estimation approaches for multi-component signals have been proposed (see e.g., [[6], Chapter 10] and [42,43]). Such methods first localize and extract sig- nal components from the signal T-F representation and then apply an IF estimation procedure. The methods have the advantage of not requiring prior knowledge about the signal under analysis except that its compo- nents are separated in the T-F domain, a condition often satisfied by newborn EEG signals [6]. For a given analytic signal z[n] associated with the real discrete time signal x[n], n = 0, 1, ..., N - 1, the discrete version of Equation 3 is given by [[6], pp. 237]: ρ[n, k] = 2 DFT n→k {G[n, m] ∗ n (z[n + m]z∗[n −m])}.(4) For an N-point real signal x[n], r[n, k] is represented by an N × M matrix r where M (M ≥N) is the number of FFT points used in calculating the TFD. Note that n = t.fs and k = 2M fs f where t and f are, respectively, the continuous time and frequency variables, and fs is the sampling frequency of the signal [[6], pp. 236]. Table 1 lists the discrete time-lag kernels of the TFDs used in this study. A typical implementation of a basic multi-component IF estimation method such as that proposed in [42] includes the following two steps: (a) T-F transformation: The given signal is first mapped to the T-F domain using a suitable TFD which is chosen based on the characteristics of the signal under analysis. For newborn EEG signals, due to their non-stationary and multi-component nature, the selected TFD should provide high spectral reso- lution and have good cross-term reduction-capabil- ity. Previous studies have shown that separable kernel TFDs, i.e. g(t, f) = g1(t) G2(f) are the most sui- table ones [[6], Section 5.7]. We can use typical spectral analysis windows such as a Hanning window and a Gaussian window for g1(t) and G2(f) respec- tively, or the windows used in MBD which was shown to be best for representing EEG signals. 2.2.3 T-F analysis of multichannel EEG using spatial T-F distributions (d) Reduced interference: The TFD attenuates the unwanted cross-terms in the T-F domain relative to the signal terms. The concept of spatial T-F distribution (STFD) was introduced to analyze an array of non-stationary signals and has been successfully used in the area of array sig- nal processing for source separation, for example (see [[6], Chapter 8]). STFDs are a generalization of TFDs to a vector of multi-sensor signals. (e) High resolution: The reduced interference property is achieved while preserving a good T-F resolution. Several measures that take into account the above properties were proposed in [29,44,45]. Comparative studies of these measures indicated that the selection of TFDs for a particular application is data-dependent [29]. Several measures that take into account the above properties were proposed in [29,44,45]. Comparative studies of these measures indicated that the selection of TFDs for a particular application is data-dependent [29]. Let us denote the EEG signal recorded at the ith chan- nel at the discrete time n as xi[n]; i = 1, 2, ..., P; n = 1, 2, ..., N where P is the total number of EEG channels, and the vector Z[n] as the analytic associate of the vec- tor X[n] as a set of recorded signals. The general form of the STFD of the signal vector Z[n] is defined as [[6], pp. 349-350]: Table 1 lists the expressions for the time-lag kernels G [n, m] of some of these QTFDs known as RIDs which are used in this study. These RIDs were designed to suppress or attenuate the cross-terms in the WVD [[6], Section 5.2]. The WVD is also considered in this study as it is a reference, for performance comparison. The exact formula for each method can be found in [[6], Section 6.1]. Dxx n, k = M0 m=−M0 M0 p=−M0 G n −p, m ⊙Z p + m Z∗ p −m e−j4π mk n (6) 2.2.4 Selecting a data adapted TFD 2.2.4 Selecting a data adapted TFD The selection of a suitable TFD for representing the EEG signals is the first step in forming the expressions (6) and (7) for any detection and classification scheme in the T-F domain. A suitable TFD is the one which is capable of highlighting the signal non-stationary features that best discriminate between different classes under consideration. For the analysis of multi-component non- stationary EEG signals [6,32] and to generate suitable features for detection, a TFD needs to have the follow- ing properties [6]: A component linking algorithm then detects linked components in B(t, f) by examining the pixel con- nectivity and the number of connected pixels. This is based on the fact that ideally, the IF of a signal component where signal energy concentrates will be observed in the TFD as a ridge describing the IF. Among the sets of connected pixels, only those with a number of pixels exceeding a preset threshold are identified as true linked components of seizure IFs. To remove the IFs of non-seizure components, this threshold is set to the minimum time duration of a seizure component in samples (equivalent to 10 s). (a) Realness: r[n, k] must be real. (a) Realness: r[n, k] must be real. (b) IF estimation: The IF of the signal can be esti- mated using the peak of the TFD of the signal. (c) Local energy: The energy of the signal in a certain region in the T-F plane can be found by the integral of the TFD over that region. 2.2.2 Instantaneous frequency (b) l k d l k Table 1 The time-Lag kernels of TFDs used in this article [6] QTFD G[n, m] Kernel type WVD δ[n] Lag independent/Doppler independent SWVD δ[n]w[m] Doppler independent SPEC w[n + m] w [n - m] Non-separable kernel GKD √πσ 2\|m\| exp −π2σn2 4m2 Product kernel MBD cosh−2β n cosh−2βn Lag independent SEPK HammM[n] HannL[n] Separable kernel The parameters b and s are positive real and w[n] represents the window function. HammM[n] and HannL[n] represent respectively M-point Hamming and L-point Hanning functions Table 1 The time-Lag kernels of TFDs used in this article (b) TFD Local peaks detection & component linking: The IF estimation method proposed in [42], consid- ers the resulted TFD (r(t, f)) as a 2D image with time and frequency as its row and column coordi- nates and identifies the Local maxima (with respect The parameters b and s are positive real and w[n] represents the window function. HammM[n] and HannL[n] represent respectively M-point Hamming and L-point Hanning functions The parameters b and s are positive real and w[n] represents the window function. HammM[n] and HannL[n] represent respectively M-point Hamming and L-point Hanning functions Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 8 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 to frequency) using both the first and second deriva- tive tests. Only those local maxima which are higher than a threshold (i.e. r(t, f) >0.01 × max(r(t, f))) are considered as valid peaks. By assigning value 1 to the locations of valid peaks and value 0 to all others, the T-F image can then be transformed to a binary image B(t, f) [42]. with the spatial diversity provided by the multiple sen- sors. This property makes them suitable for analyzing multichannel EEGs. 3 T-F image processing methodology for newborn EEG analysis where [G[n, m]]ij = Gij [n, m] is the time-lag kernel associated with the signals zi[n] and zj [n], ⊙represents the Hadamard product, Z is the conjugate transpose (Hermitian transpose) of the vector Z, and M0 = N−1 2 . Each element of the STFD is given by: y As mentioned earlier, T-F image processing techniques represent a set of image processing methods that pro- cess the TFD of a given signal considered as an image. Several studies applied image processing techniques to the signal TFD in order to select a methodology suitable for the problem considered, leading to an improvement in terms of precision, resolution, performance or robust- ness. The applications of these techniques include: seg- mentation [46-55], denoising and signal-to-noise ratio (SNR) enhancement [56-59], classification [60-63], cod- ing and compression [64,65], watermarking [66-70], and IF estimation and cross-term reduction [42,52,71,72]. A review of proposed approaches is given in the following section. Dxx n, k ij = ij M0 m=−M0 M0 p=−M0 Gij n −p, m zi p + m z∗ j p −m e−j4π mk n (7) The off-diagonal elements of the STFD are the cross- TFDs and the diagonal entries are the classical auto- TFDs of the EEG signals. STFD-based methods exploit the non-stationary characteristics of the signals together Page 9 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 assuming a prior knowledge of the interference geometry. 3.1 T-F image processing techniques: overview A comprehensive review of proposed T-F image proces- sing approaches shows that they can be grouped in six classes (corresponding to the numbering of the next subsections). The intent is to evaluate these methods to determine their suitability in the main application con- sidered in this article, i.e. feature selection and extrac- tion, and classification. The use of the above mentioned methods not only improves the performance of the extracted features in discriminating between different classes, but also facili- tates the estimation of the IF laws of multi-component signals. The statistics of the estimated IF laws of the sig- nal using image processing techniques can be used as T-F features to characterize the non-stationarity of the EEG signal. 3.1.3 Segmentation of T-F images Image segmentation and edge detection methods can be applied to TFDs to localize significant structures in them and extract T-F signatures of different compo- nents of the signal under analysis [74]. The goal of a T-F image segmentation is to create homogeneous regions that correspond to separate components of the signal. This is performed by a complete partitioning of image C (the original TFD) into n disjoint sub-regions, C1, C2, ..., Cn. Morphological filtering was used in the analysis of TFDs in several applications, for example, musical transcription of audio recording [47] and ana- lysis of submarine and seismic signals [48,49]. Similar to the method presented in [49], authors in [46] apply the water-shed segmentation procedure on T-F images of non-stationary signals consisting of multiple closely- spaced components in the presence of noise. In the context of segmentation, it is not only important to extract desirable patterns from noisy background but also to separate the T-F image subsets corresponding to different signal components. This can be achieved by computing an over-segmented watershed image and re-merging the adjacent regions into pertinent struc- tures, based on their statistical and geometry features [46]. An extension of the segmentation procedure based on morphological T-F image processing techni- ques is also proposed in [50] for speech spectrogram segmentation. There are several denoising T-F methods that use T-F image processing techniques for the removal of noise from a signal [56]. An adaptive block thresholding denoising algorithm based on Stein Unbiased Risk esti- mator is proposed to remove noise from audio signals in the T-F representation in [73]. Another approach uses thresholding denoising algorithms in conjunction with the S-transform [57]. A Bayesian image denoising method is also proposed in [58] which removes Gaus- sian noise from the T-F representation of EEG signals [59]. These methods can be adapted following the approach based on LPA-RICI.a The idea consists of transforming a 2D image into 1D array using the row- wise or diagonal-wise transformation, and its denoising using the LPA-RICI method. The above indicates that there is a choice of several algorithms for denoising methods which reduce noise and artifacts in the EEG signals from their TFDs consid- ered as images. The power of noise and artifact in the signal resulting from reconstructing the enhanced TFDs are expected to be much less than those in the original signals. 3 T-F image processing methodology for newborn EEG analysis 3.1.1 Denoising or SNR enhancement of T-F images The presence of noise and/or artifact (e.g. muscle activ- ity and eye blinks) in the recorded EEG signals degrades the performance of any EEG based signal processing method. Image processing techniques can be used to reduce noise and artifact from the TFD of EEG signals while retaining key details such as edges and texture of the TFD. T-F image techniques can then be adapted and used to enhance the quality of T-F representations in order to better represent the T-F features which char- acterize different abnormalities. 3.1.3 Segmentation of T-F images This in turn can significantly improve the per- formance of any newborn EEG classification/source localization method. In addition, a new approach for blind separation of nonstationary sources using their TFDs, was proposed in [75]. This approach is based on the observation that a monocomponent FM signal is represented by a linear feature corresponding to the energy concentration points in the T-F image. The idea is based on a line detection algorithm, which is an adaptation of the road network extraction method [76], to extract separately all the components using a spatially averaged T-F image of their mixtures. The line detection is done at the pixel level by determining whether a pixel belongs to a line crossing it along a particular direction. 3.1.2 T-F image processing for IF estimation and cross-term reduction Image processing techniques can be used to estimate the signal IF and attenuate cross-terms which appear in QTFDs due to their bilinear nature. One such image processing techniques estimates the IF laws of a multi- component signal based on component linking [42]. Another approach uses morphological image processing operators to reduce cross-terms in the WVD [71]. Alter- native approaches are also proposed in [52,72], all These techniques are especially useful for analyzing EEG signals as they have the ability to provide us with new features which capture the non-stationarity and Page 10 of 21 Page 10 of 21 Page 10 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 multi-component nature of the EEG signals. These fea- tures can be grouped into the following four classes: SVMs in terms of classification accuracy, and can be useful in the context of EEG signal classification. SVMs in terms of classification accuracy, and can be useful in the context of EEG signal classification. 3.1.5 Coding and compression of T-F images multi-component nature of the EEG signals. These fea- tures can be grouped into the following four classes: SVMs in terms of classification accuracy, and can be useful in the context of EEG signal classification. 3.1.5 Coding and compression of T-F images • Topological features such as the number of com- ponents of the signals, Time-frequency image compression aims at reducing irrelevant and redundant information in the T-F data in order to store or transmit non-stationary data in an effi- cient form. 3.1.3 Segmentation of T-F images So, T-F image compression and coding tech- niques can be applied to TFD records as a solution to the transmission and storage of complex non-stationary sig- nals. For example, one can use a T-F image coding tech- nique by dividing the T-F image into a combination of time and frequency components and then code them separately [64]. The key of this algorithm is to locate the sharp edges in both horizontal and vertical directions, and to select a threshold size for which the sharp edges are to be coded. Another approach is based on the esti- mation of the sequence duration using the autocorrela- tion of the T-F image; it was used to detect and characterize a frequency hopping (FH) signal in the con- text of spectrum surveillance [65]. Since the obtained T- F image is often very large, a compression step reduces the image before processing it. Following the above approach, an audio data compression algorithm based on the T-F image of music signals could reproduce an audio signal which keeps only the main sense of music with smaller size [78]. The method can be used for tasks of identification of a signal, search of pieces of music, etc. • Morphometric features based on the geometric shapes of the T-F components of the signals such as: size, position, and orientation of the components, • Intensity features based on the histograms of the TFDs over relevant zones (typically components) such as moments, median, and mode; and • Texture features such as statistics of spatial grays- cale variations of the TFDs such as contrast, entropy, and energy. The use of the above-mentioned features is expected to improve the performance of any EEG abnormality detector. 3.1.4 Classification of T-F images Electroencephalogram signals can be classified based on their TFDs using image processing techniques which allocate a particular T-F image belongs to a certain class. Several approaches have been proposed for differ- ent applications. In the context of ECG, three methods were proposed to classify the signals using the images obtained from their WVDs [60]. The first method uses features that are problem specific and highly sensitive to noise such as the frequency range and the multi-compo- nent characteristics of the signal. The second one uses correlation in the T-F domain. The third method is based on a decision theoretic pattern recognition which uses the SVD and the ambiguity function as features for the classification [60]. The above indicates that the T-F representation or T- F image can be used in general coding and compression algorithms of EEG signals. This implies that these T-F methods can be adapted and used in the case of large and complex biomedical data, including EEG, as a possi- ble innovative solution to the transmission and storage of key T-F features describing specific abnormalities. Another approach determines the modulation mode of the signal through a T-F method with autoregressive modeling to detect phase shifts, frequency shifts and amplitude shifts [61]. Then the classification is per- formed through a simple decision tree. Another approach is to use the regional correlation between TFDs in order to achieve a better performance such as automatically selecting regions of interest before per- forming the correlation [62]. For audio applications, an approach considers the T-F spectrograms related to music samples as textured images and extracts texture- based features to classify these music samples according to the predominant musical instrument [63]. Another method directly classifies each signal according to the best match of its TFD (such as Spectrogram, QTFDs or Gabor) with a set of ideal masks. The best fit is deter- mined through Frobenius inner products as in [77]. 3.1.6 T-F watermarking The main goal of T-F watermarking is to hide a message m (watermark) in a non-stationary signal using a T-F representation of data w to obtain new data w’ practi- cally indistinguishable from w, by people, in such a way that an eavesdropper cannot remove or replace m in w’. This allows to develop tools for data hiding and copy- right protection, using the characteristics of non-statio- narity signals. Time-frequency image watermarking methods can be classified in two traditional approaches: spatial and spec- tral domain techniques. In the spatial domain, the watermark is embedded in selected regions chosen based on the texture of the given T-F image. In the spectral domain, the watermark is embedded in the transform domain using methods such as discrete cosine transform (DCT) and DWT, in the mid-frequency range to ensure transparency and robustness of the watermark. A joint TFD of images can give a more comprehensive representation of the image compared to looking at Most of the classification methods proposed in the image processing literature can be adapted and applied to EEG classification. These classifiers may outperform conventional classifiers such as Neural Networks and Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 11 of 21 Table 2 T-F image processing techniques and corresponding quality criteria T-F image techniques Resolution Performance Precision Robustness Segmentation √ Denoising/SNR enhancement √ Classification √ √ Coding and compression √ √ Watermarking √ √ T-F signal estimation √ √ Table 2 T-F image processing techniques and corresponding quality criteria each domain individually [67,70]. Image watermarking in the T-F domain can be done using the basic WVD [68]. In this method, the watermarked cells are selected based on their impact on image quality as well as resili- ence to JPEG. Another watermarking algorithm based on the WVD [69] does not assume any particular structure for the embedded watermark and offers a theoretical framework for designing the minimum probability of error detector. The embedding algorithm in the joint T-F domain is based on two aspects: the first one embeds the water- mark directly into the WVD of the image, while the sec- ond one embeds the WVD of the watermark. This result reduces the computational complexity of embedding and detecting the watermark. This method is shown to be transparent and robust under attack. 4 EEG abnormality detection and localization methodology The proposed method for EEG abnormalities detection and classification, illustrated in Figure 4, includes three stages: monochannel or multichannel T-F analysis, fea- ture selection and extraction, and final classification. The above indicates that image watermarking methods can be adapted and used for biomedical data. For exam- ple, watermarking methods are needed for EEG data if information such as acquisition measures and/or patient information, is confidential or registered. The idea is that this information can be embedded as a hidden mes- sage in the T-F image of the EEG signals, using the methods proposed in [66,67,69,70]. As mentioned earlier, spectral analysis of non-station- ary signals leads to T-F analysis. It is thus necessary to perform a T-F analysis on the EEG signals in the pre- sence and the absence of abnormality in order to deter- mine features such as the frequency bands in which the energy of the signals is concentrated for each case. The proposed T-F analysis aims at extracting new key fea- tures from the EEG signals which carry information 3.1.6 T-F watermarking A variant of the above method led to a robust and high-capacity image watermarking algorithm using the discrete evolutionary transform (DET) calculated by the Gabor expansion to represent an image in the joint T-F domain [70]. The watermark is then embedded onto selected coefficients of the TFD. Tables 3 and 4 show the applicability of these T-F image processing techniques to the EEG abnormalities detection and classification problem. T-F image proces- sing techniques can therefore improve the analysis of multichannel EEG signals in key areas, namely: pre-pro- cessing, feature extraction, cross-term reduction, and classification. Section 4.1 discusses a number of features that can be extracted from the T-F image and used for classification, as a supplement to signal based features used in previous work [14-16]. Another approach to watermarking of speech signals in the T-F domain [66] is based on the cross-terms free S-transform and the time-varying filtering used for watermark modeling. The main advantages and draw- backs of various TFDs applied to digital watermarking are discussed in [67]. 3.2 Range of applicability of T-F image processing methods to EEG classification The previous section indicates that there is a wide range of T-F image processing methods that can be developed once one has formed the T-F image using a QTFD that describes the characteristics of a non-stationary signal. These methods described above can be assessed in terms of performance against quality criteria for EEG abnormalities detection and classification. Table 2 shows some quality criteria for each T-F image processing category. The performance quality criterion indicates the applicability of a given T-F image processing techni- que in increasing the EEG classification accuracy. Per- formance and precision are the key quality criteria which can be enhanced by relevant T-F image proces- sing techniques. Table 3 T-F image processing techniques and their applicability (or not) to the EEG abnormalities detection and classification problem T-F image processing EEG abnormalities detection and classification system Segmentation √ Denoising/SNR enhancement √ Classification √ Coding and compression Watermarking T-F signal estimation √ Table 3 T-F image processing techniques and their applicability (or not) to the EEG abnormalities detection and classification problem Table 3 T-F image processing techniques and their applicability (or not) to the EEG abnormalities detection and classification problem g p g q applicability (or not) to the EEG abnormalities detection and classification problem T-F image processing EEG abnormalities detection and classification system Segmentation √ Denoising/SNR enhancement √ Classification √ Coding and compression Watermarking T-F signal estimation √ Page 12 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 related in particular to their non-stationary nature. These features include the T-F features proposed recently in [14-16], and also some features based on T-F image processing techniques as detailed in the following subsection. orthogonal alternate subspace of the form given as follows [14-16] ρ = USVH (11) Table 4 Selected T-F image processing techniques and their applicability to the different steps of the EEG abnormalities detection and classification problem T-F image techniques Pre-processing Feature extraction Features classification Segmentation √ Denoising/SNR enhancement √ Classification √ T-F signal estimation √ √ Table 4 Selected T-F image processing techniques and their applicability to the different steps of the EEG abnormalities detection and classification problem T-F image techniques Pre-processing Feature extraction Features classification Segmentation √ Denoising/SNR enhancement √ Classification √ T-F signal estimation √ √ Segmentation Denoising/SNR enhancement Classification T-F signal estimation orthogonal alternate subspace of the form given as follows [14-16] orthogonal alternate subspace of the form given as follows [14-16] related in particular to their non-stationary nature. These features include the T-F features proposed recently in [14-16], and also some features based on T-F image processing techniques as detailed in the following subsection. (11) ρ = USVH 4.1 Feature extraction methodology 4.1.1 Signal related features where U is an N × N matrix, S is an N × M diagonal matrix with non-negative real numbers (Si, i = 1, 2, ..., N) on the diagonal, and VH (the conjugate trans- pose of V) is an M × M real unitary matrix. The diagonal entries of S are known as the singular values of r. The singular values and vectors have proved useful in characterizing EEG abnormalities [13,74]. Here, the maximum and variance of the sin- gular values, denoted as MASV and VASV respec- tively, are selected as characteristics of the singular values of r and chosen as features FS3 and FS4, fol- lowing [14-16]. The discrete-time EEG signal and its TFD are respec- tively denoted as x[n] and r[n, k]. Ten T-F features from the TFD r[n, k] corresponding to each signal are described below. These features are extracted for each EEG segment of length T seconds after calculating r[n, k]. Five other features based on T-F image techniques are then also extracted. Only the features denoted F1 - F6 take into account directly the non-stationary nature of the EEG signals, as discussed in [14-16]. (1) IF-based features: As mentioned earlier, the IF can be estimated either as the peak of the TFD of the signal (Equation 8) g (3) T-F complexity (TFCM): The TFCM feature denoted as FS5 is an extension of the measure pro- posed in [12,79] to the T-F plane as proposed in [14-16]. It uses both SVD and Shannon entropy and is given by fi [n] = fs 2M arg max k ρ n, k (8) (8) FS5 = − N i=1 Si log Si (12) (12) or the first order moment of the TFD (Equation 9). or the first order moment of the TFD (Equation 9). or the first order moment of the TFD (Equation 9). fi [n] = fs 2M M k=1 k ρ n, k M k=1 ρ n, k (9) (9) Where Si , i = 1, 2, ..., N are the normalized singular values,i.e.: Si = Si N i=1 Si Where Si , i = 1, 2, ..., N are the normalized singular values,i.e.: Si = Si N i=1 Si Where Si , i = 1, 2, ..., N are the normalized singular values,i.e.: Si = Si N i=1 Si According to the basic theorem of information the- ory, FS5 ≤log N, and the equality holds only when all the normalized singular values are equal, i.e. Si = 1 N ; i = 1, 2, ..., N (see for example [[80], pp. 88]. TFCM therefore represents the magnitude and the number of the non-zero singular values of the TFD of the EEG signals. TFCM is a useful feature as the number of non-zero singular values of r and their magnitudes have a strong relationship with the information content in the TFD [14-16,81]. According to the basic theorem of information the- ory, FS5 ≤log N, and the equality holds only when all the normalized singular values are equal, i.e. The first selected feature, denoted as FS1, is chosen as the mean of fi[n] (MEIF), i.e. FS1 = 1 N N n=1 fi [n]. The second feature FS2 represents the deviation of the IF (DEIF) of the EEG signals, i.e. Si = 1 N ; i = 1, 2, ..., N (see for example [[80], pp. 88]. TFCM therefore represents the magnitude and the number of the non-zero singular values of the TFD of the EEG signals. TFCM is a useful feature as the number of non-zero singular values of r and their magnitudes have a strong relationship with the information content in the TFD [14-16,81]. FS2 = fi [n] = max fi [n] −min fi [n] (10) FS2 = fi [n] = max fi [n] −min fi [n] (10) (10) (2) SVD-based features: The SVD is performed on the N × M matrix r. This divides the TFD matrix into two subspaces, signal subspace and an (2) SVD-based features: The SVD is performed on the N × M matrix r. 4.1.2 T-F image processing-based features 4.1.2 T-F image processing-based features Further to the features presented earlier that followed [14-16], we are also interested in the use of the T-F image segmentation techniques described in Section 3.1 to extract other features so as to complement the signal related features. We therefore propose new T-F features to describe visually the normal and seizure patterns in the TFD. These features are based on visual descriptors, and are extracted from T-F representations considered as images using image processing techniques. The idea is, for example, to apply a segmentation algorithm on T-F images to detect regions where all EEG information appear (e.g., all components, normal/seizure pattern). The watershed segmentation technique in [46,50,86] is adapted to detect this homogeneous region and then to compute their statistical and geometrical features. In this context, T-F image data is interpreted as a topo- graphic surface where watershed boundaries separate individual catchment basins. The boundaries of the groupings are placed along the crest lines of the gradi- ent image and as a result the watershed transformation can partition the image into meaningful regions. Figure 6 shows an example of the T-F image of a newborn EEG signal obtained using the MBD and its binary seg- mented image, obtained using the watershed segmenta- tion; this makes it possible to detect and select the normal/seizure pattern i.e., the region where the nor- mal/seizure pattern appears. From the binary segmented image, we could get several shape descriptors that include bounding box (a rectangle that circumscribes the segmented object), convex hull (the smallest convex shape that contains the object), centroid and contour. (5) Sub-bands energies: Sub-band energy-based fea- tures represent the energy of the EEG signal x[n] in the sub-band where this energy is mostly concen- trated. The sub-band interval is determined in the T-F representation by analyzing the EEG signal. The energy is computed by proper frequency windowing of r[n, k] and is derived using the formula represent- ing the partial information contained in a certain region of the T-F domain [[6], pp. 46]. In [14-16], the use of four sub-bands energies fea- tures FSi, i = 7, 8, 9, 10 based respectively on 0-4 Hz (δ), 4-8 Hz (θ), 8-12 Hz (a), and 12-30 Hz (b) sub- bands, were proposed. Their sub-bands have been chosen by analyzing the EEG signals taken from the EEG database [84]. or the first order moment of the TFD (Equation 9). This divides the TFD matrix into two subspaces, signal subspace and an (4) Energy concentration measure (ECOME): For a multi-component signal, the energy concentration Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 13 of 21 We note that the features FS7 - FS10 may also be estimated from the spectrum of the signal as they do not contain information about non-stationarity. For the newborn EEG database used in this study (as described in Section 5.2 and also in [15,16]), the energies of the EEG signals in 0-4 Hz (δ) and 4-8 Hz (θ) sub-bands were chosen as features. Only the fea- tures FS7 and FS8 will be used for this database in our experimental simulation in the next section. measure, denoted as ECOME, determines the con- centration of the dominant component at each loca- tion in the T-F domain [82]. We use the definition given in [14-16,83] as it does not discriminate between low concentrated components and highly concentrated ones. It is given by: We note that the features FS7 - FS10 may also be estimated from the spectrum of the signal as they do not contain information about non-stationarity. For the newborn EEG database used in this study (as described in Section 5.2 and also in [15,16]), the energies of the EEG signals in 0-4 Hz (δ) and 4-8 Hz (θ) sub-bands were chosen as features. Only the fea- tures FS7 and FS8 will be used for this database in our experimental simulation in the next section. FS6 = N n=1 M k=1 \|ρ n, k \| 1 2 2 (13) (13) 4.1.2 T-F image processing-based features The white region in the binary-segmented image corresponds to the region where all components appear in the TFD. 4.1.2 T-F image processing-based features This database is described in [85], and will be used in our experimental simulation in the next section. These features are defined as follows [14-16]: FS7 = N n=1 Mδ k=1 ρ n, k (14) (14) FS8 = N n=1 2Mδ k=Mδ ρ n, k (15) FS9 = N n=1 3Mδ k=2Mδ ρ n, k (16) FS10 = N n=1 Mβ k=3Mδ ρ n, k (17) FS8 = N n=1 2Mδ k=Mδ ρ n, k (15) (15) FS9 = N n=1 3Mδ k=2Mδ ρ n, k (16) FS10 = N n=1 Mβ k=3Mδ ρ n, k (17) FS9 = N n=1 3Mδ k=2Mδ ρ n, k (16) (16) We now describe five morphometric features based on the geometric shape of the segmented regions. FS10 = N n=1 Mβ k=3Mδ ρ n, k (17) (17) 1. Image grayscale moments: The moments of the T- F binary-segmented image denoted r’[n, k] can be used to compute shape features, such as height, width, area, perimeter, compactness, centroid, etc. Using these shape descriptors we could further com- pute other ratios of geometric features, such as: where Mδ = [8M/fs], Mb = [60M/fs] and [.] stands for the floor operator. Note that k = M corresponds to the maximum frequency component in the signal (fs/2). It follows that Mδ and Mb correspond to the frequency components at 4 and 30 Hz, respectively. Aspect Ratio = Areaobject Areabounding box Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 14 of 21 where p, q = 0, 1, 2, .... The central moments are expressed as where p, q = 0, 1, 2, .... The central moments are expressed as Figure 6 Illustration of the time-frequency image methodology: An example of the T-F image of newborn EEG signal using the MBD (a) with the corresponding grayscale (b) and binary-segmented image using the methodology presented (c). The white region in the binary-segmented image corresponds to the region where all components appear in the TFD. 4.1.2 T-F image processing-based features μpq = n k (n −n)p k −k q ρ′ n, k (19) (19) where n = m10 m00 and k = m01 m00 are centroids of the T-F image. The μ00 is the area of the segmented region, n and k give the x and y coordinates of the centroid for the segmented region. Higher order moments may give even more detailed shape characteristic, but they are not considered in this study. 2. Convex Hull: The convex Hull H(r’) of a T-F bin- ary-segmented image subset r’[n, k] is defined as the smallest convex set containing a given subset [87]. The only parameter required in our study is the area of the convex Hull (i.e. the number of pixels, d(H (r’)))). This can be calculated using the central moments of order (0, 0) defined in Equation 19. In summary, the following five morphometric features based on the geometric shape of the segmented T-F regions are considered in this study as a supplement to the signal-based features listed in the previous subsec- tion: • Convex Hull/area: FI1 = μ00 • Perimeter: FI2 = (m30 + m12)2 + (m03 + m21)2 • Compactnessb: FI3 = (FI2)2/FI1 • Coordinates of the centroid for the segmented region: FI4 = m10 m00 and FI5 = m01 m00 • Convex Hull/area: FI1 = μ00 • Perimeter: FI2 = (m30 + m12)2 + (m03 + m21)2 b • Compactnessb: FI3 = (FI2)2/FI1 • Compactnessb: FI3 = (FI2)2/FI1 • Coordinates of the centroid for the segmented The total number of features for this study was there- fore 15 for the adult EEG data ({FS1, FS2, ..., FS10, FI1, ..., FI5}) and 13 for the newborn EEG data ({FS1, FS2, ..., FS8, FI1, ..., FI5}). These features include both signal related features and image related features. Figure 6 Illustration of the time-frequency image methodology: An example of the T-F image of newborn EEG signal using the MBD (a) with the corresponding grayscale (b) and binary-segmented image using the methodology presented (c). The white region in the binary-segmented image corresponds to the region where all components appear in the TFD. Figure 6 Illustration of the time-frequency image methodology: An example of the T-F image of newborn EEG signal using the MBD (a) with the corresponding grayscale (b) and binary-segmented image using the methodology presented (c). 4.2 Classification During the classification step, an EEG signal is allocated a certain class based on the location of its feature vec- tor. The T-F features extracted from EEG signals, and described above, are used to train a classifier. We opted for the classification algorithms which were recently used to classify the EEG signals in the T-F domain in [14-16,88]. In particular, two classification algorithms are used in this study: Form Factor = 4π × Areaobject Perimeter2 object Form Factor = 4π × Areaobject Perimeter2 object . The moment of order (p, q) for r’[n, k] is defined as [87] The moment of order (p, q) for r’[n, k] is defined as [ ] mpq = n k npkqρ′ n, k (18) • The Multi-class SVM classifier used in [14-16]. • The Multi class SVM classifier used in [14 16]. • The Neural network-based classifier used in [88]. (18) • The Neural network-based classifier used in [88]. Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 15 of 21 (Figure 4-(a)) is first assessed in the case of real adult EEG data. These methods extract features from the TFD of the EEG signals under analysis and then feed them to the classifier as a vector. These classification methods are then compared by evaluating their performance in clas- sifying the EEG signals using the selected T-F features defined in the previous section, in terms of sensitivity, specificity and classification accuracy. The data set used is described in [85]; it consists of five sets denoted as A, B, C, D and E. Each set contains 100 single channel EEG segments with duration of 23.6 s. The signals have been recorded at fs = 173.6 Hz sam- pling rate and therefore each EEG signal in each set has fs × 23.6 = 4, 096 samples. Sets A and B have been taken from surface EEG recordings of five healthy volunteers with eye open and closed, respectively. EEG signals in sets C and D, respectively, were acquired in seizure-free intervals from five epileptic patients in the epileptogenic zone and from the hippocampal formation of the opposite hemisphere of the brain. Set E contains seizure activity intervals in the EEG signals acquired from the five patients. 4.3 Localization The classification step consists in detecting the presence or absence of the abnormality in the EEG signals and allocating it to a particular class. For the problem of localizing the spatial location of an abnormality on the scalp, it requires the simultaneous analysis of the multi- ple EEG signals which are provided from the different electrodes placed on the scalp. In this study, the selected methodology to detect an abnormality and locate its source on the scalp is as fol- lows: (1) For each electrode, the EEG signal is transformed to reveal its T-F representation, described by its TFD, using the MBD. • Class H includes sets A and B which contain the EEG signals acquired from Healthy volunteers. • Class H includes sets A and B which contain the EEG signals acquired from Healthy volunteers. (2) Construct the 3D array data in the space-T-F domain by stacking all the 2D T-F arrays corre- sponding to the TFDs (by concatenation of all 2D data). The stacking order corresponds to the location of electrodes on the scalp. This assigns the space direction in the space-T-F domain. • Class F includes sets C and D which contain sei- zure-Free intervals acquired from epileptic volun- teers, and • Class S includes set E which contains Seizure- activity intervals of the EEG signals collected from epileptic patients. (3) Calculate the correlation between the 2D T-F arrays in order to detect the presence of an abnorm- ality. The correlation can be calculated between the 2D T-F arrays or the T-F features extracted from these 2D T-F arrays [31]. The T-F feature set {FS1, FS2, ..., FS10, FI1, ..., FI5} was extracted from the TFD of each EEG segment of length T seconds. Only five TFDs are chosen in this simulation: MBD, SPEC, SWVD, GKD and WVD. The parameters of the MBD and GKD were respectively chosen as b = 0.01 and s = 0.9. These values are typical ones for which the MBD and GKD have shown good perfor- mances in analyzing EEG signals [[6], Sections 7.4 and 15.5] and [15,16]. The window w[n] for SWVD and SPEC distributions, was chosen to be a Hanning window of length ⌊N/4⌋samples. Note that we did not optimize the window functions and lengths used for SPEC and SWVD. The simulations were carried out in Matlab. 4.3 Localization (4) Determine the electrodes for which an abnormal- ity is detected in their corresponding 2D T-F arrays. We assume here that the problem of EEG synchroni- zation does not arise or has been already dealt with in a pre-processing stage. Furthermore, fusion approaches can be applied on the 3D array data for detecting and localizing the abnormalities [89-91]. 4.2 Classification More details about the dataset can be found in [84]. Here, for practical reasons, the desired outcome of the classification stage is one of three different classes of the EEG signals in the dataset, namely: classes H, F, and S defined as follows: 5 Experimental results and discussion For performance evaluation, the multi-class SVM and neural network classifiers were trained using the T-F features extracted from the EEG signals. We compared the results of the classification for different TFDs. The original database was split in two parts, 60% of the data were used for training and 40% for testing the classifier. This section assesses the performance of the EEG abnormalities detection and localization schemes using the T-F image processing techniques, defined in the pre- vious section in terms of features selection and extrac- tion, localization and classification. Table 5 shows the confusion matrices for classification results using 15 features extracted from different T-F representations of EEG segments of length T = 11.8 s (N = 2,048 samples). The results between parentheses 5.1 Adult EEG seizure detection and classification 5.1 Adult EEG seizure detection and classification The performance of T-F features, defined in Section 4.1, for EEG abnormalities detection and classification Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 16 of 21 Table 5 Confusion matrices for EEG classification using T-F features set {FS1, FS2, ..., FS10, FI1, ..., FI5} extracted from the TFD with SVM classifier and Neural Network-based classifier. 5.1 Adult EEG seizure detection and classification Classifier outputs Statistical parameters (%) TFD Class Total H F S Sensitivity Specificity Total accuracy Multi-class SVM classifier H 100 98 (98) 1 (2) 1 (0) 98 (98) 92 (90.5) MBD F 100 6 (5) 89 (92) 5 (3) 89 (92) 96.5 (93.5) 94 (93) S 100 1 (2) 4 (9) 95 (89) 95 (89) 93.5 (95) H 100 97 (99) 1 (1) 1 (0) 97 (99) 93 (91) SPEC F 100 5 (5) 89 (92) 6 (3) 89 (92) 97 (94.5) 94.33 (93.67) S 100 1 (2) 2 (7) 97 (90) 97 (90) 93 (95.5) H 100 99 (98) 0 (2) 1 (0) 99 (98) 93.5 (90) SWVD F 100 3 (5) 93 (91) 4 (4) 93 (91) 96.5 (93.5) 95.33 (92.67) S 100 0 (2) 4 (9) 94 (89) 94 (89) 96 (94.5) H 100 99 (98) 1 (2) 0 (0) 99 (98) 90.5 (89.5) GKD F 100 5 (5) 88 (91) 7 (4) 88 (91) 96 (93) 93.33 (92.33) S 100 1 (2) 5 (9) 93 (88) 93 (88) 93.5 (94.5) H 100 98 (98) 1 (2) 1 (0) 98 (98) 91.5 (90) WVD F 100 4 (8) 89 (87) 7 (5) 89 (87) 96.5 (95.5) 93.67 (92.67) S 100 2 (2) 3 (4) 94 (93) 94 (93) 93.5 (92.5) Neural network-based classifier H 100 99 (89) 1 (11) 0 (0) 99 (89) 90.5 (93.5) MBD F 100 5 (8) 89 (88) 6 (0) 89 (88) 95.5 (94) 93.33 (92) S 100 0 (0) 8 (1) 92 (99) 92 (99) 94 (88.5) H 100 99 (91) 1 (7) 0 (0) 99 (91) 94 (95.5) SPEC F 100 2 (4) 93 (92) 5 (4) 93 (92) 97 (95) 95.67 (94) S 100 2 (0) 3 (1) 95 (99) 95 (99) 96 (91.5) H 100 98 (94) 0 (4) 0 (0) 98 (94) 93 (92) SWVD F 100 7 (7) 88 (89) 5 (4) 88 (89) 98 (94.5) 94.67 (92.67) S 100 0 (0) 2 (5) 98 (95) 98 (95) 93 (91.5) H 100 99 (98) 1 (0) 0 (0) 99 (98) 92 (91) GKD F 100 4 (8) 88 (85) 8 (7) 88 (85) 97.5 (97.5) 94.33 (93.33) S 100 1 (0) 3 (3) 96 (97) 96 (97) 93.5 (91.5) H 100 98 (91) 2 (7) 0 (0) 98 (91) 91.5 (92.5) WVD F 100 6 (7) 88 (88) 6 (5) 88 (88) 96.5 (94) 93.67 (92) S 100 0 (0) 4 (3) 95 (97) 95 (97) 93 (89.5) The results, for each TFD, are given using two classifiers: multi-class SVM classifier and neural network-based classifier, with a real EEG database organized in three classes H, F and S. 5.1 Adult EEG seizure detection and classification The T-F features are extracted from EEG segments of length 11.8 s (N = 2,048 samples). The results between parentheses are the classification results using the ten signal-related features {FS1, FS2, ..., FS10}. Sensitivity and specificity of each classifier and for each particular class as well as its total accuracy are also given. Table 5 Confusion matrices for EEG classification using T-F features set {FS1, FS2, ..., FS10, FI1, ..., FI5} extracted from the TFD with SVM classifier and Neural Network-based classifier. The results, for each TFD, are given using two classifiers: multi-class SVM classifier and neural network-based classifier, with a real EEG database organized in three classes H, F and S. The T-F features are extracted from EEG segments of length 11.8 s (N = 2,048 samples). The results between parentheses are the classification results using the ten signal-related features {FS1, FS2, ..., FS10}. Sensitivity and specificity of each classifier and for each particular class as well as its total accuracy are also given. used in order to evaluate the classification performance and are defined as follows: are the classification results using the 10 features {FS1, FS2, ..., FS10} (without the features based on T-F image processing techniques) and the multi-class SVM classi- fier [14-16]. Each row shows, for a particular TFD, the total number of EEG segments correctly classified as well as those misclassified as other classes. The total number of EEG segments in each class is 100 segments. The results indicate that class F signals are most often confused with class H and S signals; likewise class S with class F. are the classification results using the 10 features {FS1, FS2, ..., FS10} (without the features based on T-F image processing techniques) and the multi-class SVM classi- fier [14-16]. Each row shows, for a particular TFD, the total number of EEG segments correctly classified as well as those misclassified as other classes. The total number of EEG segments in each class is 100 segments. The results indicate that class F signals are most often confused with class H and S signals; likewise class S with class F. 5.1 Adult EEG seizure detection and classification Sensitivity = TP TP + FN (20a) Speciﬁcity = TN TN + FP (20b) Sensitivity = TP TP + FN (20a) (20a) Speciﬁcity = TN TN + FP (20b) (20b) Total accuracy = TP + TN (TP + FN) + (TN + FP) (20c) (20c) Table 5 also includes the values of the statistical para- meters of the classifier; namely: its sensitivity, specificity, and total classification accuracy. These parameters are Page 17 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Table 6 Confusion matrices for newborn EEG classification using T-F features set {FS1, FS2, ..., FS8, FI1, ..., FI5} extracted from the TFD Classifier outputs Statistical parameters (%) TFD Class Total N S Sensitivity Specificity Total accuracy N 50 48 (47) 2 (3) 96 (94) 96 (86) 96 (90) MBD S 50 2 (7) 48 (43) 96 (86) 96 (94) N 50 49 (45) 1 (5) 98 (90) 96 (84) 97 (87) SPEC S 50 2 (8) 48 (42) 96 (84) 98 (90) N 50 46 (45) 4 (5) 92 (90) 94 (84) 93 (87) SWVD S 50 3 (8) 47 (42) 94 (84) 92 (90) N 50 43 (44) 7 (6) 86 (88) 72 (77) 79 (79) GKD S 50 14 (15) 36 (35) 72 (70) 86 (88) N 50 46 (33) 4 (17) 92 (66) 92 (86) 90 (76) WVD S 50 6 (7) 44 (43) 88 (86) 88 (66) The results, for each TFD, are given using the multi-class SVM classifier with newborn EEG database organized in two classes N, and S. The result between parentheses is the classification result using only the signal-related features {FS1, FS2, ..., FS8}. Sensitivity and specificity for each particular class as well as its total accuracy are also given. Table 6 Confusion matrices for newborn EEG classification using T-F features set {FS1, FS2, ..., FS8, FI1, ..., FI5} extracted from the TFD where TP, TN, FN, and FP, respectively represent the numbers of true positive, true negative, false negative, and false positive. p We notice that the classification results using the fea- ture vector {FS1, FS2, ..., FS10, FI1, ..., FI5} with the SVM- based classifier, are better compared with the results using the feature vector {FS1, FS2, ..., FS10}. 5.1 Adult EEG seizure detection and classification This is con- firmed by the total classification accuracy calculated for each TFD. Also, the results given by the neural net- work-based classifier are comparable with the SVM clas- sification results. We note that the neural network-based classifier is sensitive to over-training compared to the SVM-based classifier. The total classification accuracy is in the range (93% to 95%) for 100 EEG segments. This can be improved by increasing the number of EEG seg- ments in the training-step. These results indicate that T-F features based on image processing techniques are also able to characterize non-stationary EEG signals, and therefore, detect the different abnormalities with accuracy. The performance of the proposed T-F features in new- born EEG seizure detection and classification (Figure 4- (a)) is now assessed using real newborn EEG data. The results, for each TFD, are given using the multi-class SVM classifier with newborn EEG database organized in two classes N, and S. The result between parentheses is the classification result using only the signal-related features {FS1, FS2, ..., FS8}. Sensitivity and specificity for each particular class as well as its total accuracy are also given. The results, for each TFD, are given using the multi-class SVM classifier with newborn EEG database organized in two classes N, and S. The result between parentheses is the classification result using only the signal-related features {FS1, FS2, ..., FS8}. Sensitivity and specificity for each particular class as well as its total accuracy are also given. The newborn EEG data used here includes 14-channel EEG recordings of nine neonates using mono-polar montages. The recordings have been marked by a neu- rologist for seizures. The EEG signals have been band- pass filtered in [0.7-50] Hz band with a notch filter at 50 Hz and collected at 256 Hz sampling frequency. From the dataset, we extracted two sets of seizure and non-seizure EEG epochs referred to as sets S and N respectively. Each set contains 50 segments. Each seg- ment was band-pass filtered in the range 0.5-10 Hz and down-sampled from 256 to 20 Hz. This was based on findings that neonatal EEG seizures have spectral activ- ities mostly below 12 Hz [92]. The length of each seg- ment is 12.8 s with 256 samples. vector {FS1, FS2, ..., FS8}. This is confirmed by the total classification accuracy calculated for each TFD. 5.1 Adult EEG seizure detection and classification The best total classification accuracy is obtained using the MBD and SPEC; and is in a range (96% to 97%) for 100 segments. This can be improved by increasing the number of EEG segments in the training-step. The results confirm that including T-F image features improves the classification performance. Table 6 shows the confusion matrices representing the classification results using the 15 features extracted from different T-F representations of newborn EEG segments. The result between parentheses is the classi- fication result using only the signal-based features {F1, F2, ..., F10} [15,16]. This result is obtained using the multi-class SVM classifier. Each row shows, for a parti- cular TFD, the total number of newborn EEG seg- ments correctly classified as well as those misclassified as another class, and also, the three statistical para- meters: sensitivity, specificity and total classification accuracy. Table 6 shows that the classification results using the combined T-F signal & image related feature vector {FS1, FS2, ..., FS8, FI1, ..., FI5} are better compared to results using only the T-F signal-related feature 5.3 Seizure localization in newborn EEG signals The performance of T-F features in EEG abnormalities detection and localization (Figure 4-(b)) is now assessed using real multichannel EEG signals recorded from newborns. For the simulation, we have chosen only five newborn EEG segments for five channels (namely channel A, B, C, D and E) from the newborn EEG database described in Section 5.2. Figure 7 shows these newborn EEG seg- ments in the time domain with their corresponding magnitude spectra. For the problem of seizure detection and localization, we assume that the seizure was detected in the newborn EEG signal of channel A and then we want to locate other channels where the EEG signal have the same Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Page 18 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Figure 7 An example of real newborn multichannel EEG. Only five channels are shown. These channels are labeled A, B, C, D and E. The channels A, D and E have newborn EEG seizure patterns. This was detected manually by a neurologist and also confirmed using the methodology described in Section 4.3. Figure 7 An example of real newborn multichannel EEG. Only five channels are shown. These channels are labeled A, B, C, D and E. The channels A, D and E have newborn EEG seizure patterns. This was detected manually by a neurologist and also confirmed using the methodology described in Section 4.3. The results in the time domain does not provide enough significant information for determining the existence of significant correlations between EEG signals from all channels. seizure pattern. We applied the methodology described in Section 4.3 which consists of calculating the correla- tion between signals in the T-F domain. Table 7 shows the correlation resulting between chan- nel A and channels {B, C, D, E} in the spatial domain and T-F domain. We have used the correlation coeffi- cient to measure the correlation between sample vectors between newborn EEG signals in the time domain, and also between their corresponding T-F representations using the MBD. The results indicate that a strong corre- lation exists in the T-F domain between channels A ↦ D and A ↦E. This implies that the seizure pattern observed in the newborn EEG signal from channel A is also detected in the EEG signals from channels D and E. References 1. JS Hahn, BR Tharp, in Neonatal and Pediatric Electroencephalography, Electrodiagnosis in Clinical Neurology, ed. by M Aminoff (Churchill Livingstone, New York, 1992) 1. JS Hahn, BR Tharp, in Neonatal and Pediatric Electroencephalography, Electrodiagnosis in Clinical Neurology, ed. by M Aminoff (Churchill Livingstone, New York, 1992) 2. RS Fisher, W Boas, W Blume, C Elger, P Genton, P Lee, J Engel, Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). Epilepsia. 46(4), 470–472 (2005). doi:10.1111/j.0013-9580.2005.66104.x 3. R Oostenveld, P Praamstra, The five percent electrode system for high- resolution EEG and ERP measurements. Clin Neurophysiol. 112(4), 713–719 (2001). doi:10.1016/S1388-2457(00)00527-7 4. BA Khalil, KE Misulis, Atlas of EEG & Seizure Semiology, (Elsevier, Butterworth Heinehmann Edition, Philadelphia, 2006) 5. HR Mohseni, A Maghsoudi, MB Shamsollahi, Seizure detection in EEG signals: a comparison of different approaches, in International Conference of IEEE Engineering in Medicine and Biology Society (EMBS’2006), New York, USA, 6724–6727 (30 Aug–3 Sept 2006) 6. B Boashash, Time-Frequency Signal Analysis and Processing: A Comprehensive Reference, (Elsevier, Oxford, UK, 2003) 7. B Boashash, M Mesbah, Using DSP to detect seizures in newborns. IEEE Electron Syst Softw. 1(3), 34–37 (2003) 8. J Gotman, Automatic seizure detection: improvements and evaluation. Electroencephalogr Clin Neurophysiol. 76(4), 317–324 (1990). doi:10.1016/ 0013-4694(90)90032-F 9. G Widman, T Schreiber, B Rehberg, A Hoeft, CE Elger, Quantification of depth of anesthesia by nonlinear time series analysis of brain electrical activity. Phys Rev E. 62(4), 4898–4903 (2000). doi:10.1103/PhysRevE.62.4898 10. K Polat, S Gunes, Classification of epileptiform EEG using a hybrid system based on decision tree classifier and fast fourier transform. Appl Math Comput. 187(2), 1017–1026 (2007). doi:10.1016/j.amc.2006.09.022 11. A Subasi, EEG signal classification using wavelet feature extraction and a mixture of expert model. Exp Syst Appl. 32(4), 1084–1093 (2007). doi:10.1016/j.eswa.2006.02.005 2. RS Fisher, W Boas, W Blume, C Elger, P Genton, P Lee, J Engel, Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). Epilepsia. 46(4), 470–472 (2005). doi:10.1111/j.0013-9580.2005.66104.x In particular, this article has adapted and applied a segmentation algorithm on the T-F image in order to detect the region where all EEG information appear. Then, new morphometric features were defined based on the geometric shape of the segmented region. Two T-F image classification methods were selected to evalu- ate the performance of the EEG classification using the selected T-F features. 5.3 Seizure localization in newborn EEG signals Page 19 of 21 Page 19 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 d ⃗x, ⃗y = ⃗x −⃗y T S−1 ⃗x −⃗y (21) based on a complete multichannel T-F image processing approach. In addition, this work can be further refined by considering the TFD which is most appropriate for this T-F image processing approach following the algorithmic perspective developed in [6,94-96] and the selection cri- teria presented in [29]. (21) where ⃗x and ⃗y are two feature vectors and S is the covariance matrix. The results show also that the correlation is better detected using the complete 15 T-F features vector. This suggests that features based on T-F image proces- sing techniques are able to help characterize better the abnormality patterns in newborn EEG signals, and also to detect and locate them on the scalp. Competing interests The authors declare that they have no competing interests. Received: 24 November 2011 Accepted: 29 May 2012 Published: 29 May 2012 Received: 24 November 2011 Accepted: 29 May 2012 Published: 29 May 2012 Author details 1 1Electrical Engineering Department, Qatar University College of Engineering, Doha, Qatar 2Centre for Clinical Research and Perinatal Research Centre, Royal Brisbane & Women’s Hospital, University of Queensland, Herston, QLD 4029, Australia 3Department of Electrical Engineering, Razi University, Kermanshah, Iran Endnotes aLocal polynomial approximation-relative intersection of confidence intervals (for more detail see [97]). bThe compactness is the most popular shape feature, and is obtained by perimeter2/area. Acknowledgements Th h ld lik This article concludes that it is possible to improve the analysis, classification and localization of non-stationary signals by using a general methodology that combines techniques from three complementary fields: T-F signal analysis, image processing and multichannel analysis. The approach is illustrated on the key problem of new- born EEG abnormality diagnostic and localization for which a solution would help improve health outcomes for newborns. In addition, the same problem was also considered for the case of adults. More precisely, this article reports the results of a study on EEG abnormal- ities detection and classification in the case of mono- channel T-F analysis first, and then, the EEG abnormal- ities sources localization in the case of multichannel T-F analysis. The novelties of this article include: (1) a com- prehensive review of T-F image processing techniques, and (2) a review of methods to process the T-F repre- sentation of EEG signal considered as an image, as well as (3) a novel approach that combines the above with T-F features to characterize the non-stationary newborn EEG signal, following [14-16]. g The authors would like to thank Prof. Paul Colditz and his colleagues at the University of Queensland, Centre for Clinical Research, Brisbane, Australia for providing the Newborn data used in this article. This publication was made possible by a grant from the Qatar National Research Fund under its National Priorities Research Program award number NPRP 09-465-2-174. 5.3 Seizure localization in newborn EEG signals In addition, we notice that only the T-F features vec- tor is able to detect and locate the seizure pattern in dif- ferent channels. This is confirmed by the Mahalanobis distance [93] for which results are given in Table 8. This distance calculates the correlation/similarity between feature vectors, and is defined as follows: Table 8 Mahalanobis distance between channel A and channels B, C, D and E in the T-F domain using the feature vector Channel A 8 T-F features 13 T-F features Channel B 0.5084 0.3883 Channel C 0.5049 0.3856 Channel D 0.1832 0.1399 Channel E 0.1733 0.1324 The MBD distribution is used to represent the newborn EEG signal in the T-F domain. The columns in the table correspond respectively to the correlation between: (a) 8 T-F features vector ({FS1, FS2, ..., FS8}), and (b) 13 T-F features vector ({FS1, FS2, ..., FS8, FI1, ..., FI5}). The results indicate that a strong correlation exists between the channel A ↦D and A ↦E. Also, this correlation is better detected using 13 T-F features vector. Table 8 Mahalanobis distance between channel A and channels B, C, D and E in the T-F domain using the feature vector Table 7 Correlation between channel A and channels B, C, D and E in the spatial domain and T-F domain Channel A Time domain T-F domain Channel B -0.0562 0.2166 Channel C 0.0105 0.1157 Channel D -0.2527 0.5945 Channel E -0.0502 0.6113 The MBD is used to represent the signal in the T-F domain. The columns in the table correspond respectively to the correlation between: (1) newborn EEG signals in the time domain, and (2) T-F representations in the T-F domain. The results indicate that a strong correlation exists between channel A ↦D and A ↦E in the T-F domain. Table 7 Correlation between channel A and channels B, C, D and E in the spatial domain and T-F domain The MBD distribution is used to represent the newborn EEG signal in the T-F domain. The columns in the table correspond respectively to the correlation between: (a) 8 T-F features vector ({FS1, FS2, ..., FS8}), and (b) 13 T-F features vector ({FS1, FS2, ..., FS8, FI1, ..., FI5}). The results indicate that a strong correlation exists between the channel A ↦D and A ↦E. Also, this correlation is better detected using 13 T-F features vector. 9. G Widman, T Schreiber, B Rehberg, A Hoeft, CE Elger, Quantification of depth of anesthesia by nonlinear time series analysis of brain electrical activity. Phys Rev E. 62(4), 4898–4903 (2000). doi:10.1103/PhysRevE.62.4898 y 8. J Gotman, Automatic seizure detection: improvements and evaluation. Electroencephalogr Clin Neurophysiol. 76(4), 317–324 (1990). doi:10.1016/ 0013-4694(90)90032-F 7. B Boashash, M Mesbah, Using DSP to detect seizures in newborns. IEEE Electron Syst Softw. 1(3), 34–37 (2003) 10. K Polat, S Gunes, Classification of epileptiform EEG using a hybrid system based on decision tree classifier and fast fourier transform. Appl Math Comput. 187(2), 1017–1026 (2007). doi:10.1016/j.amc.2006.09.022 References These T-F features were also able to evaluate the correlation between newborn EEG sig- nals in order to locate the abnormality source on the scalp. The experimental results show that the features based on T-F image processing techniques are able to improve the performance of both EEG classification and EEG abnormality source localization. 3. R Oostenveld, P Praamstra, The five percent electrode system for high- resolution EEG and ERP measurements. Clin Neurophysiol. 112(4), 713–719 (2001). doi:10.1016/S1388-2457(00)00527-7 4. BA Khalil, KE Misulis, Atlas of EEG & Seizure Semiology, (Elsevier, Butterworth Heinehmann Edition, Philadelphia, 2006) 5. HR Mohseni, A Maghsoudi, MB Shamsollahi, Seizure detection in EEG signals: a comparison of different approaches, in International Conference of IEEE Engineering in Medicine and Biology Society (EMBS’2006), New York, USA, 6724–6727 (30 Aug–3 Sept 2006) 6. B Boashash, Time-Frequency Signal Analysis and Processing: A Comprehensive Reference, (Elsevier, Oxford, UK, 2003) 7. B Boashash, M Mesbah, Using DSP to detect seizures in newborns. IEEE Electron Syst Softw. 1(3), 34–37 (2003) 8. J Gotman, Automatic seizure detection: improvements and evaluation. Electroencephalogr Clin Neurophysiol. 76(4), 317–324 (1990). doi:10.1016/ 0013-4694(90)90032-F 8. J Gotman, Automatic seizure detection: improvements and evaluation. Electroencephalogr Clin Neurophysiol. 76(4), 317–324 (1990). doi:10.1016/ 0013-4694(90)90032-F 9. G Widman, T Schreiber, B Rehberg, A Hoeft, CE Elger, Quantification of depth of anesthesia by nonlinear time series analysis of brain electrical activity. Phys Rev E. 62(4), 4898–4903 (2000). doi:10.1103/PhysRevE.62.4898 Finally, the above results allow further progress with a new focus on the adaptation of T-F image processing tech- niques in order to better use them in the problem consid- ered. This should result in a more advanced design for automatic abnormality diagnosis and localization systems 10. K Polat, S Gunes, Classification of epileptiform EEG using a hybrid system based on decision tree classifier and fast fourier transform. Appl Math Comput. 187(2), 1017–1026 (2007). doi:10.1016/j.amc.2006.09.022 11. A Subasi, EEG signal classification using wavelet feature extraction and a mixture of expert model. Exp Syst Appl. 32(4), 1084–1093 (2007). doi:10.1016/j.eswa.2006.02.005 Page 20 of 21 Page 20 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 12. BR Greene, S Faul, WP Marnane, G Lightbody, I Korotchikova, GB Boylan, A comparison of quantitative EEG features for neonatal seizure detection. Clin Neurophysiol. 119(6), 1248–1261 (2008). doi:10.1016/j.clinph.2008.02.001 34. References V Pierson, M Martin, Watershed segmentation of time-frequency images, in Proceeding of IEEE NSIP’95, Haldikiki, Greecem, 1000–1003 (1995) 25. AT Tzallas, MG Tsipouras, D Fotiadis, Epileptic seizure detection in EEGs using timefrequency analysis. IEEE Trans Inf Technol Biomed. 13(5), 703–710 (2009) 49. B Leprette, N Martin, Extraction of pertinent subsets from time-frequency representations for detection and recognition purposes. Signal Process. 82, 229–238 (2002). doi:10.1016/S0165-1684(01)00181-5 26. B Boashash, M Mesbah, A time-frequency approach for newborn seizure detection. IEEE Eng Med Biol Mag. 20(5), 54–64 (2001) 50. R Steinberg, D O’Shaughnessy, Segmentation of a speech spectrogram using mathematical morphology, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP’2008), Las Vegas, Nevada, USA, 1637–1640 (30 Mar–4 Apr 2008) 27. B Boashash, M Mesbah, Time-Frequency Methodology for Newborn Electroencephalographic Seizure Detection (chapter 9), Applications in Time- Frequency Signal Processing, ed. by A Papandreou-Suppappola (CRC Press, 2003), pp. 339–369 28. B Boashash, AP Reilly, Algorithms for Time-Frequency Signal Analysis (chapter 7), Time Frequency Signal Analysis–Methods and Applications, ed. by B Boashash Melbourne (Longman Cheshire Edition 1992), pp. 163–181 51. C Meyer, M Spiertz, Audio segmentation using different time-frequency representations, in Proc of the 12th International Student Conference on Electrical Engineering (POSTER’08), Prague, Czech Republic, (15 May 2008) 29. B Boashash, V Sucic, Resolution measure criteria for the objective assessment of the performance of quadratic time-frequency distributions. IEEE Trans Signal Process. 51(5), 1253–1263 (2003). doi:10.1109/TSP.2003.810300 52. L Tao, Y Shao-Quan, Improvement on joint time-frequency representation with application of image processing technique, in Proceedings of the 5th International Conference on Computational Intelligence and Multimedia Applications (ICCIMA’2003), Xi’an, China, 243–248 (27–30 Sept 2003) http:// portal.acm.org/citation.cfm?id=942797.943372 30. B Boashash, PJ O’Shea, Use of the cross Wigner-Ville distribution for estimation of instantaneous frequency. IEEE Trans Signal Process. 41(3), 1439–1445 (1993). doi:10.1109/78.205752 53. C Tantibundhit, F Pernkopf, G Kubin, Joint time-frequency segmentation algorithm for transient speech decomposition and speech enhancement. IEEE Trans Audio Speech Lang Process. 18(6), 1417–1428 (2010) 31. B Boashash, PJ O’Shea, A methodology for detection and classification of some underwate acoustic signals using time-frequency analysis techniques. IEEE Trans Acoust Speech Signal Process. 38(11), 1829–1841 (1990). doi:10.1109/29.103085 54. F Millioz, N Martin, Reassignment vector field for time-frequency segmentation, in Proceedings of International Congress on Sound and Vibration (ICSV14), Cairns, Australia, (9–12 July 2007) 32. L Rankine, N Stevenson, M Mesbah, B Boashash, A nonstationary model of newborn EEG. IEEE Trans Biomed Eng. 54, 19–28 (2007) 55. References MS Khlif, M Mesbah, B Boashash, P Colditz, Neonatal seizure detection and localization using time-frequency analysis of multichannel EEG, in IEEE International Conference on Signal Processing and Communications (ICSPC’2007), Dubai, UAE, 1567–1570 (24–27 Nov 2007) y 13. A Temko, E Thomas, W Marnane, G Lightbody, G Boylan, EEG-based neonatal seizure detection with support vector machines. Clin Neurophysiol. 122(3), 464–473 (2011). doi:10.1016/j.clinph.2010.06.034 (ICSPC’2007), Dubai, UAE, 1567–1570 (24–27 Nov 2007) 35. MS Khlif, M Mesbah, B Boashash, P Colditz, Detection of neonatal EEG seizure using multichannel matching pursuit. Int Conf IEEE Eng Med Biol Soc. 2008, 907–910 (2008) 14. M Fani, G Azemi, B Boashash, EEG-based automatic epilepsy diagnosis using the instantaneous frequency with sub-band energies, in The 17th International Workshop on Systems, Signal Processing and their Applications (WoSSPA’2011), Algeria, 187–190 (9–11 MAy 2011) 36. A Aarabi, R Grebeb, F Walloisa, A multistage knowledge-based system for EEG seizure detection in newborn infants. Clin Neurophysiol. 118(12), 2781–2797 (2007). doi:10.1016/j.clinph.2007.08.012 2781–2797 (2007). doi:10.1016/j.clinph.2007.08.012 37. B Boashash, Estimation and interpreting the instantaneous frequency of a signal-I. Fundamentals Proc IEEE. 80(4), 520–538 (1992) 38. B Boashash, Estimation and interpreting the instantaneous frequency of a signal-II. Algorithms ans applications Proc IEEE. 80(4), 540–568 (1992) 37. B Boashash, Estimation and interpreting the instantaneous frequency of a signal-I. Fundamentals Proc IEEE. 80(4), 520–538 (1992) 15. B Boashash, L Boubchir, G Azemi, Time-frequency signal and image processing of non-stationary signals with application to the classification of newborn EEG abnormalities, in The 11th IEEE International Symposium on Signal Processing and Information Technology (IEEE ISSPIT’2011), Bilbao, Spain, 120–129 (14–17 Dec 2011) 38. B Boashash, Estimation and interpreting the instantaneous frequency of a signal-II. Algorithms ans applications Proc IEEE. 80(4), 540–568 (1992) 39. B Boashash, Digital signal processing in the time-frequency domain using the Wigner-Ville distribution, in Conference on Computers and Engineering 1985: Real-time Applications of Computers in Engineering, (Institution of Engineers Australia Barton, ACT, 1985), pp. 46–50. National conference publication; No. 85/8, ISBN:0858252686 16. B Boashash, L Boubchir, G Azemi, Improving the classification of newborn EEG time-frequency representations using a combined time-frequency signal and image approach, in The 11th International Conference on Information Sciences, Signal Processing and their Applications (ISSPA’2012), Montreal, Quebec, Canada, (3–5 July 2012) 40. B Boashash, Note on the use of the Wigner distribution for time-frequency signal analysis. IEEE Trans Acoust Speech Signal Process. 36(9), 1518–1521 (1988). doi:10.1109/29.90380 17. References A Subasi, E Erelebi, A Alkan, E Koklukaya, Comparison of subspace-based methods with AR parametric methods in epileptic seizure detection. Comput Biol Med. 36(2), 195–208 (2006). doi:10.1016/j.compbiomed.2004.11.001 41. G Azemi, B Senadji, B Boashash, Mobile unit velocity estimation based on the instantaneous frequency of the received signal. IEEE Trans Veh Technol. 53(3), 716–724 (2004). doi:10.1109/TVT.2004.827157 18. S Altunay, Z Telatar, O Erogul, Epileptic EEG detection using the linear prediction error energy. Exp Syst Appl. 37(8), 5661–5665 (2010). doi:10.1016/ j.eswa.2010.02.045 42. L Rankine, M Mesbah, B Boashash, IF estimation for multicomponent signals using image processing techniques in the time-frequency domain. Signal Process. 87(6), 1234–1250 (2007). doi:10.1016/j.sigpro.2006.10.013 19. V Srinivasan, C Eswaran, N Sriraam, Artificial neural network based epileptic detection using time-domain and frequency-domain features. J Med Syst. 29, 647–660 (2005). doi:10.1007/s10916-005-6133-1 43. J Lerga, V Sucic, B Boashash, An efficient algorithm for instantaneous frequency estimation of nonstationary multicomponent signals in low SNR. EURASIP J Adv Signal Process. 2011, 16 (2011). (Article ID 725189) doi:10.1155/2011/725189. doi:10.1186/1687-6180-2011-16 20. NF Guler, ED Ubeyli, I Guler, Recurrent neural networks employing lyapunov exponents for EEG signals classification. Expert Syst Appl. 29(3), 506–514 (2005). doi:10.1016/j.eswa.2005.04.011 44. TH Sang, W Williams, Renyi information and signal-dependent optimal kernel design, in International Conference on Acoustics, Speech and Signal Processing (ICASSP’95), Detroit, Michigan, USA, 2, 997–1000 (8–12 May 1995) 21. N Kannathal, ML Choo, UR Acharya, PK Sadasivan, Entropies for detection of epilepsy in EEG. Comput Methods Program Biomed. 80(3), 187–194 (2005). doi:10.1016/j.cmpb.2005.06.012 g g y 45. PM Oliveira, V Barroso, Uncertainty in the time-frequency plane, in The 10th IEEE Workshop on Statistical Signal and Array Processing (SSAP’2000), Pennsylvania, USA, 607–611 (14-16 Aug 2000) 22. N Kannathal, UR Acharya, CM Lim, PK Sadasivan, Characterization of EEG: a comparative study. Comput Methods Program Biomed. 80, 17–23 (2005). doi:10.1016/j.cmpb.2005.06.005 j 23. I Guler, ED Ubeyli, Multiclass support vector machines for EEG signals classification. IEEE Trans Inf Technol Biomed. 11(2), 117–126 (2007) 46. A Loza, N Canagarajah, D Bull, Region feature-based segmentation of time- frequency images, in The 11th International Workshop on Systems, Signals and Image Processing (IWSSIP’2004), Poznan, Poland, 375–378 (13–15 Sept 2004) 24. AT Tzallas, MG Tsipouras, DI Fotiadis, Automatic seizure detection based on time-frequency analysis and artificial neural networks. Intell Neurosci. 2007, 18:1–18:13 (2007) 47. A Sterian, Model-based segmentation of time-frequency imagews for musical transcription, PhD thesis, (University of Michigan, 1999) 48. Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 References F Millioz, J Huillery, N Martin, Short time Fourier transform probability distribution for time-frequency segmentation, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP’2006), Toulouse, France, 3, 448–451 (14–19 May 2006) 33. MS Khlif, M Mesbah, B Boashash, P Colditz, Multichannel-based newborn EEG seizure detection using time-frequency matched filter. Int Conf IEEE Eng Med Biol Soc. 2007, 1265–1268 (2007) Page 21 of 21 Page 21 of 21 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 Boashash et al. EURASIP Journal on Advances in Signal Processing 2012, 2012:117 http://asp.eurasipjournals.com/content/2012/1/117 56. D Nelson, G Cristobal, V Kober, F Cakrak, PL Loughlin, L Cohen, Denoising using time-frequency and image processing methods. Adv Signal Process Algor Architec Implement IX. 3807, 564–581 (1999) distribution. Appl Math Comput Sci. 20, 135–147 (2010). doi:10.2478/ v10006-010-0010-x distribution. Appl Math Comput Sci. 20, 135–147 (2010). doi:10.2478/ v10006-010-0010-x 78. AV Shekhirev, EV Rabinovich, Music files compression based on time- frequency representation of audio signal, in Third International Forum on Strategic Technologies (IFOST’2008), Novosibirsk-Tomsk, Russia, 340–342 (23–29 June 2008) g 57. S Parolai, Denoising of seismograms using the S transform. Bull Seismolog Soc Am. 99, 226–234 (2009). doi:10.1785/0120080001 58. L Boubchir, J Fadili, Bayesian denoising based on The MAP estimation in wavelet-domain using Bessel K form prior. IEEE Int Conf Image Process. I, 113–116 (2005) 79. S Roberts, W Penny, I Rezek, Temporal and spatial complexity measures for electroencephalogram based brain-computer interfacing. Med Biol Eng Comput. 37, 93–98 (1999). doi:10.1007/BF02513272 59. L Boubchir, B Boashash, Bayesian wavelet denoising based on the MAP estimation using the BKF prior. IEEE Trans Signal Process (Submitted February 2012). (in review) 80. JG Proakis, Digital Communications, 4th edn. (McGraw-Hill, New York, 200 80. JG Proakis, Digital Communications, 4th edn. (McGra 81. B Boashash, B Lovell, L White, Time-frequency analysis and pattern recognition using singular value decomposition of the Wigner-Ville distribution, in SPIE Conference on Advanced Algorithms and Architectures for Signal Processing, San Diego, CA, USA, 18, 104–114 (18–19 Aug 1987) f f 60. S Abeysekera, B Boashash, Methods of signal classification using the images produced by the Wigner-Ville distribution. Pattern Recogn Lett. 12, 717–729 (1991). doi:10.1016/0167-8655(91)90010-J 61. H Ketterer, F Jondral, AH Costa, Classification of modulation modes using time-frequency methods, in IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP’99), Phoenix, Arizona, USA, 5, 2471–2474 (15–19 Mar 1999) 82. References E Sejdic, I Djurovic, J Jiang, Time-frequency feature representation using energy concentration: An overview of recent advances. Digital Signal Process. 19, 153–183 (2009). doi:10.1016/j.dsp.2007.12.004 j 83. L Stankovic, A measure of some time-frequency distributions concentration. Signal Process. 81(3), 621–631 (2001). doi:10.1016/S0165-1684(00)00236-X 62. E Sejdić, J Jiang, Pattern Recognition in Time-Frequency Domain: Selective Regional Correlation and Its Applications, Pattern recognition in time- frequency domain: selective regional correlation and its applications, ed. by Peng-Yeng Yin (I-Tech Education and Publishing, Vienna, Austria, 2008), pp. 613–626 g 84. EEG Database, http://epileptologie-bonn.de/cms/front_content.php? idcat=193&lang=3&changelang=3 84. EEG Database, http://epileptologie- idcat=193&lang=3&changelang=3 85. R Andrzejak, K Lehnertz, F Mormann, C Rieke, P David, C Elger, Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: dependence on recording region and brain state. Phys Rev E. 64, 061907-1–061907-8 (2001). doi:10.1103/PhysRevE.64.061907 63. G Yu, Jj Slotine, Audio classification from time-frequency texture, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP’2009), Taipei, Taiwan, 1677–1680 (19–24 April 2009) 86. P Soille, Morpological Image Analysis, Principles and Applications, (Springer- Verlag, Berlin, 1999). isbn 978-3540656715 64. H Cheung, CM Wong, SW Leung, A time-frequency image coding technique, in 7th IEE International Conference on Image Processing And Its Applications, Manchester, UK, 1, 166–168 (July 1999) 87. RC Gonzalez, R Woods, Digital Image Processing, (Addison-Wesley, New York, 1993) 65. O Berder, C Bourder, G Burel, in Identification of Frequency Hopping Communications, Problems in Modern Applied Mathematics, ed. by N Mastorakis (WSES, 2000), pp. 259–264 88. C Gope, N Kehtarnavaz, D Nair, Neural network classification of EEG signals using time-frequency representation, in Proceedings of IEEE International Joint Conference on Neural Networks (IJCNN’2005), Montréal, Québec, Canada, 4, 2502–2507 (31 July–4 Aug 2005) 66. S Stankovic, I Orovic, N Zaric, Robust speech watermarking procedure in the time-frequency domain. EURASIP J Adv Signal Process. 2008, 9 (2008). (Article ID 519206) doi:10.1155/2008/519206 89. L Gupta, B Chung, MD Srinath, DL Molfese, H Kook, Multichannel fusion models for the parametric classification of differential brain activity. IEEE Trans Biomed Eng. 52(11), 1869–1881 (2005). doi:10.1109/TBME.2005.856272 67. S Stankovic, Time-frequency analysis and its application in digital watermarking. EURASIP J Adv Signal Process. 2010, 20 (2010). (Article ID 579295) doi:10.1155/2010/579295 90. MB Krishnan, M Mesbah, B Boashash, Newborn seizure detection based on fusion of multichannel EEG, in 5th International Workshop on Signal Processing and its Applications (WoSPA’2008), Sharjah, UAE, (18–20 Mar 2008) 68. References BG Mobasseri, Digital watermarking in joint time-frequency domain, in Proceedings of the International Conference on Image Processing (ICIP’2002), Rochester, NY, USA, 3, 481–484 (22–25 Sept 2002) 91. L Huaying, A Jeremic, Neonatal seizure detection using blind multichannel information fusion, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP’2011), Prague, Czech Republic, 649–652 (22–27 May 2011) 69. M Al-khassaweneh, S Aviyente, A time-frequency inspired robust image watermarking, in Conference Record of the Thirty-Eighth Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, USA, 1, 392–396 (7–10 Nov. 2004) 92. RR Clancy, The contribution of EEG to the understanding of neonatal seizures. Epilepsia. 37, S52–S59 (1996). doi:10.1111/j.1528-1157.1996.tb06022.x y, g seizures. Epilepsia. 37, S52–S59 (1996). doi:10.1111/j.1528-1157.1996.tb06022.x 93. RD Maesschalck, D Jouan-Rimbaud, D Massart, The Mahalanobis distance. Chemom Intell Lab Syst. 50, 1–18 (2000). doi:10.1016/S0169-7439(99)00047-7 93. RD Maesschalck, D Jouan-Rimbaud, D Massart, The Mahalanobis distance. Chemom Intell Lab Syst. 50, 1–18 (2000). doi:10.1016/S0169-7439(99)00047-7 70. M Öztürk, A Akan, Y Çekiç, A robust image watermarking in the joint time- frequency domain. EURASIP J Adv Signal Process. 2010, 9 (2010). (Article ID 509757) doi:10.1155/2010/509757 94. B Boashash, Advances in Spectral Estimation and Array Processing, Time- Frequency Signal Analysis, ed. by S Haykin (Prentice Hall, Englewood Cliffs, 1991), pp. 418–517 71. S Gomez, V Naranjo, R Miralles, Removing interference components in time- frequency representations using morphological operators. J Vis Commun Image Represent. 22(5), 401–410 (2011). doi:10.1016/j.jvcir.2011.03.007 95. M Abed, A Belouchrani, M Cheriet, B Boashash, Compact support kernels based time-frequency distributions: performance evaluation, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP’2011), Prague, Czech Republic, 4180–4183 (227–27 May 2011) 72. M Sun, CC Li, LN Sekhar, RJ Sclabassi, Elimination of crosscomponents of the discrete pseudo Wigner distribution via image processing, in IEEE International Conference on Acoustics, Speech and Signal Processing, Glasgow, Scotland, 717–720 (22–25 May 1989) 96. M Abed, A Belouchrani, M Cheriet, B Boashash, Time-frequency distributions based on compact support kernels: properties and performance evaluation. IEEE Trans Signal Process. 60(6), 2814–2827 (2012) 73. G Yu, S Mallat, E Bacry, Audio denoising by time-frequency block thresholding. IEEE Trans Signal Process. 56(5), 1830–1839 (2008) 74. H Hassanpour, M Mesbah, B Boashash, Time-frequency feature extraction of newborn EEG seizure using SVD-based techniques. EURASIP J Adv Signal Process. 2004, 2544–2554 (2004). doi:10.1155/S1110865704406167 97. References J Lerga, V Sucic, D Sersic, An image processing approach for underdetermined blind separation of nonstationary sources, in 51st International Symposium ELMAR’2009, Zadar, Croatia, 57–60 (30 Sept 2009) 75. K Abed-Meraim, N Linh-Trung, V Sucic, F Tupin, B Boashash, An image processing approach for underdetermined blind separation of nonstationary sources, in Proceedings of the 3rd International Symposium on Image and Signal Processing and Analysis (ISPA’2003), Rome, Italy, 347–352 (187–20 Sept 2003) doi:10.1186/1687-6180-2012-117 Cite this article as: Boashash et al.: A methodology for time-frequency image processing applied to the classification of non-stationary multichannel signals using instantaneous frequency descriptors with application to newborn EEG signals. EURASIP Journal on Advances in Signal Processing 2012 2012:117. doi:10.1186/1687-6180-2012-117 doi:10.1186/1687-6180-2012-117 Cite this article as: Boashash et al.: A methodology for time-frequency image processing applied to the classification of non-stationary multichannel signals using instantaneous frequency descriptors with application to newborn EEG signals. EURASIP Journal on Advances in Signal Processing 2012 2012:117. p 76. F Tupin, H Maitre, JE Mangin, JM Nicolas, E Pechersky, Detection of linear features in SAR images: Application to road network extraction. IEEE Trans Geosci Remote Sens. 36(2), 434–453 (1998). doi:10.1109/36.662728 77. E Swiercz, Classification in the Gabor time-frequency domain of non- stationary signals embedded in heavy noise with unknown statistical 76. F Tupin, H Maitre, JE Mangin, JM Nicolas, E Pechersky, Detection of linear features in SAR images: Application to road network extraction. IEEE Trans Geosci Remote Sens. 36(2), 434–453 (1998). doi:10.1109/36.662728 77. E Swiercz, Classification in the Gabor time-frequency domain of non- stationary signals embedded in heavy noise with unknown statistical 77. E Swiercz, Classification in the Gabor time-frequency domain of non- stationary signals embedded in heavy noise with unknown statistical
https://openalex.org/W2002532370	https://repositori.udl.cat/bitstream/10459.1/69056/1/plosone_a2011v6n12e28652.PDF	English	null	How Spatial Heterogeneity of Cover Affects Patterns of Shrub Encroachment into Mesic Grasslands	PloS one	2,011	cc-by	8,898	Abstract This is an open-access article distributed under the terms of the Creative Comm unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by a grant for an Academic Mission at the University of Northern British Columbia (2008 BE-1 00493) awarded to FM, a PhD fellowship (2005FI 00801) awarded by DURSI-Generalitat de Catalunya, and by the European Social Fund. Additional support was receive from the Spanish Ministerio de Educacio´n y Ciencia (VULCA, CGL2005-08133-CO3) and Ministerio de Medio Ambiente y Medio Rural y Marino (Parques Nacionales 69/2005). PC was supported by a Ramo´n y Cajal contract (Ministerio de Ciencia e Innovacio´n, Spain). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: francesc.montane@gmail.com * E-mail: francesc.montane@gmail.com ¤ Current address: Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, Ontario, Canada How Spatial Heterogeneity of Cover Affects Patterns of Shrub Encroachment into Mesic Grasslands Francesc Montane´ 1¤, Pere Casals1, Mark R. T. Dale2 1 Forest Sciences Center of Catalonia (CTFC), Sant Llorenc¸ de Morunys, Solsona, Spain, 2 University of Northern British Columbia, Prince George, Canada ane´ 1¤, Pere Casals1, Mark R. T. Dale2 r of Catalonia (CTFC), Sant Llorenc¸ de Morunys, Solsona, Spain, 2 University of Northern British Columbia, Prince George, Canada PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 Abstract We used a multi-method approach to analyze the spatial patterns of shrubs and cover types (plant species, litter or bare soil) in grassland-shrubland ecotones. This approach allows us to assess how fine-scale spatial heterogeneity of cover types affects the patterns of Cytisus balansae shrub encroachment into mesic mountain grasslands (Catalan Pyrenees, Spain). Spatial patterns and the spatial associations between juvenile shrubs and different cover types were assessed in mesic grasslands dominated by species with different palatabilities (palatable grass Festuca nigrescens and unpalatable grass Festuca eskia). A new index, called RISES (‘‘Relative Index of Shrub Encroachment Susceptibility’’), was proposed to calculate the chances of shrub encroachment into a given grassland, combining the magnitude of the spatial associations and the surface area for each cover type. Overall, juveniles showed positive associations with palatable F. nigrescens and negative associations with unpalatable F. eskia, although these associations shifted with shrub development stage. In F. eskia grasslands, bare soil showed a low scale of pattern and positive associations with juveniles. Although the highest RISES values were found in F. nigrescens plots, the number of juvenile Cytisus was similar in both types of grasslands. However, F. nigrescens grasslands showed the greatest number of juveniles in early development stage (i.e. height,10 cm) whereas F. eskia grasslands showed the greatest number of juveniles in late development stages (i.e. height.30 cm). We concluded that in F. eskia grasslands, where establishment may be constrained by the dominant cover type, the low scale of pattern on bare soil may result in higher chances of shrub establishment and survival. In contrast, although grasslands dominated by the palatable F. nigrescens may be more susceptible to shrub establishment; current grazing rates may reduce juvenile survival. Citation: Montane´ F, Casals P, Dale MRT (2011) How Spatial Heterogeneity of Cover Affects Patterns of Shrub Encroachment into Mesic Grasslands. PLoS ONE 6(12): e28652. doi:10.1371/journal.pone.0028652 Editor: Jeff Ollerton, University of Northampton, United Kingdom Editor: Jeff Ollerton, University of Northampton, United Kingdom Received June 2, 2011; Accepted November 12, 2011; Published December 8, 2011 Received June 2, 2011; Accepted November 12, 2011; Published December 8, 2011 Copyright: 2011 Montane´ et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: 2011 Montane´ et al. Spatial Heterogeneity and Shrub Encroachment Spatial Heterogeneity and Shrub Encroachment Pyrenees located in the Alt Pirineu Natural Park (Catalonia, Spain). Mean annual and winter temperature are 2.5uC and 23uC, respectively, and mean annual precipitation is 1397 mm based on the closest meteorological station to the study site (Boı´, 42u279580N, 0u529220E; 2540 m. a.s.l.). The area is usually under snow from December to April. Soils at the site develop over slates, and the soil profile is approximately 60–80 cm deep. most conspicuous features of grasslands [17,18], and this spatial heterogeneity is a major driver of woody plant encroachment [19,20,21]. Positive interactions are common in plant communities [22,23], where protection from herbivores is one of the primary mechanisms by which plants facilitate their neighbours [24]. Facilitation by unpalatable plants, also termed ‘associational resistance’, is therefore considered one of the processes driving woody plant establishment in grasslands [25]. The study site is a gentle south-facing slope. Vegetation is a mosaic of grassland and shrubland. The mesic grasslands are dominated by patches of Festuca nigrescens Lam. or Festuca eskia Ramond ex DC., although other species are present, including Calluna vulgaris (L.) Hull, Carex sp. and Juniperus communis L. F. nigrescens grows mostly in mesic grasslands while F. eskia grows in a wide range of subalpine and alpine grassland types, from steep slopes to mesic grasslands. In our site, both these grasses coexist without apparent differences in soils under F. nigrescens or F. eskia patches. Grasslands in the Pyrenees offer a dramatic example of shrub encroachment [26,27]. Pyrenean grasslands are usually dominated by patches of grasses with different palatability, which provides an exceptional opportunity to understand how fine-scale spatial heterogeneity of cover types drives the patterns of juvenile shrub encroachment into grasslands. If the ‘associational resistance’ mechanism prevails in these grasslands, we can expect that juvenile shrubs would more frequently appear close to unpalatable grass species (positive spatial associations), and that grasslands dominated by unpalatable species would show more abundant high-density clusters (spatial hot spots) of juveniles. However, in the Pyrenees, shrub encroachment rate seems to be slower in grasslands dominated by unpalatable grass species than in other grassland communities, suggesting that other negative interactions between unpalatable grasses and shrubs may outweigh the potential effects of ‘associational resistance’ [14,26]. Once juvenile shrubs have established in a site, the effects of grazing and cover types may change with shrub development stage. Spatial Heterogeneity and Shrub Encroachment A detailed spatial analysis of juvenile shrubs at different development stages (e.g. juvenile shrubs of different sizes) and spatial heterogeneity in grasslands dominated by different grass species in grassland- shrubland ecotones may provide valuable insight into the process of shrub encroachment into grasslands. Studies documenting shrub expansion in Pyrenean grasslands [27] have identified the legume Cytisus balansae ssp europaeus (G. Lo´pez & Jarvis) Mun˜oz Garmendia as one of the most abun- dant shrubs encroaching on this area’s montane and subalpine grasslands. Historical aerial photographs confirm that shrubland surface has increased at the site and that shrub encroachment has been taking place since the mid-20th century. Both clonal growth from roots and seed dispersal play important roles in the spread of Cytisus into grasslands [29]. Despite Cytisus being a widespread shrub in southern-European mountain grasslands, little is known about the mechanisms and consequences under- pinning Cytisus encroachment in different grassland communities. Over the last few decades, the site has been grazed by cattle, horse and sheep, which usually graze in F. nigrescens patches, at a grazing rate of approximately 80–90 livestock unit grazing days (LUGD) ha21 (Taull M., pers. comm. 2009). Domestic herbivores graze strongly on F. nigrescens but ignore F. eskia and adult Cytisus due to their unpalatable characteristics [29,30]. Cytisus is a leafless legume shrub and a shade-intolerant species that spreads in sites with low overstory cover [31]. Its lifespan is approximately 30–40 years. Although shepherds in the Pyrenees have traditionally used burning as a management tool for transforming encroached land to grassland, the site has not been burned for at least 30 years (Forest Office Pallars Sobira`, pers. comm. 2007). To address how spatial heterogeneity of cover affects patterns of shrub encroachment, we focused on two mesic grasslands presenting different palatability (i.e. palatable grass-dominated vs. unpalatable grass-dominated grasslands) in the Pyrenees. If the ‘associational resistance’ mechanism [25,28] is important for shrub expansion, we would expect to find a higher number of juvenile shrubs and spatial hot spots in unpalatable-dominated grasslands than in palatable grass-dominated grasslands. In addition, due to grazing pressure, we would also expect that juveniles and unpalatable grasses would frequently appear together (positive spatial associations) while juveniles and palatable grasses would frequently appear less close together (negative spatial associations). Ethics statement All necessary permits were obtained for the described field studies. The authority responsible for the Alt Pirineu Natural Park issued the permission for our field studies. Spatial Heterogeneity and Shrub Encroachment As direct assessment of the variables potentially driving shrub encroachment requires long-term experimental manipulations or observations, we use space as a surrogate in an attempt to capture how fine-scale spatial heterogeneity of grasslands affects shrub establishment. Our aim is to understand how spatial heterogeneity of cover affects patterns of shrub encroachment in different grassland communities. This study shows that the use of space as surrogate can be an effective way to understand the processes taking place in spatially heterogeneous environments, such as shrub encroachment into spatially heterogeneous grasslands. Plot selection and field sampling In summer 2008, we randomly picked eight plots (20610 m) in grassland-shrubland ecotones from 368 ha. Half of the plots were located in F. eskia-dominated grasslands while the other half were located in F. nigrescens-dominated grasslands. Although both F. eskia and F. nigrescens coexisted in most of the plots, only the species that showed the greatest percentage cover in the plot was considered dominant (approximately 50%; Table 1). For each grassland type, two plots were placed in areas where shrub encroachment was occurring in an uphill direction, and two plots were placed in areas where shrub encroachment was occurring in a downhill direction. We elected to sample in different encroachment directions due to the fact that shrub seed dispersal by gravity is limited upslope but favoured downslope, thus modifying encroachment rates. However, since juveniles en- croaching upslope and downslope both showed similar counts and spatial patterns, the plots with different encroachment directions in the same grassland community were combined for statistical analysis. Thus, we contrasted two experimental conditions (F. nigrescens and F. eskia grasslands) with four plots per experimental condition. In each rectangular plot, the 10-m Introduction across different spatial scales [6,11]. For instance, whereas factors such as climate or fire [7,12,13] may drive shrub encroachment at large scale (regional or landscape scale), interactions between grass species and shrubs [14] may drive shrub encroachment at fine scale. Whether one specific factor driving shrub encroachment acts at large (e.g. climate) or fine scale (e.g. biotic interactions) represents the opposite ends of a continuum of possibilities, with some specific factors (e.g. grazing) acting simultaneously at both large and fine scale. Thus, when selecting forage, herbivores are attracted by grass dominance at the large scale, while they preferentially select grass patches dominated by more palatable species at the fine scale [15]. All ecological systems exhibit heterogeneity and patchiness on a broad range of scales, and this patchiness is fundamental to popu- lation dynamics, community organization and element cycling [1]. Since the seminal work of Watt [2], ecologists have been trying to understand the crucial relationship between vegetation pattern and the processes that generate it. Given that plant-plant interactions are local and that plants respond to the so-called ‘‘plant’s-eye view’’ of the community [3,4], the use of space as a surrogate may be an effective way to infer processes from spatial patterns [5]. Overall, all the factors affecting shrub encroachment into grasslands most likely act in a hierarchical manner [16], and if conditions for shrub encroachment are satisfied at large scale (e.g. suitable climate), factors acting at fine scale (e.g. suitable cover types in grasslands) will be essential for precise predictions of shrub encroachment. Thus, it is crucial to understand how fine-scale spatial heterogeneity of cover types in grasslands affects patterns of shrub encroachment. Fine-scale spatial heterogeneity is one of the Woody proliferation into grasslands is a worldwide phenome- non [6] that can cause dramatic changes in community structure and function, such as species diversity and carbon storage [7,8]. The expansion of woody species in grasslands has been attributed to a number of individual factors, such as climate change, elevated CO2 levels, changes in fire frequency, grazing regime, changes in grass competitive ability, and combinations of these factors [9,10]. These different factors may act as drivers of shrub encroachment 1 December 2011 \| Volume 6 \| Issue 12 \| e28652 PLoS ONE \| www.plosone.org Study site This research was conducted in Collada de Montalto, Campirme (42u379470N, 1u119150E; 2100 m a.s.l.), a mountain mesic site in the PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 2 Spatial Heterogeneity and Shrub Encroachment Table 1. Total number of juvenile Cytisus shrubs in each plot and total cover (%) for each cover type (species, bare soil and litter). Table 1. Total number of juvenile Cytisus shrubs in each plot and total cover (%) for each cover type (species, bare soil and litter). Festuca nigrescens grassland Festuca eskia grassland Uphill Downhill Uphill Downhill NigUp1 NigUp2 NigDown1 NigDown2 EskUp1 EskUp2 EskDown1 EskDown2 Number of juvenile Cytisus shrubs 130 64 76 120 42 122 48 70 Cover (%) Adult Cytisus shrubs 15.5 18.7 21.8 18.8 16.3 18.9 16.9 21.2 Festuca nigrescens 61.1 55.5 53.8 50.9 10.1 3.2 4.5 15.4 Festuca eskia - 12.3 - 15.4 54.5 55.8 50.9 47.1 Calluna vulgaris 1.2 - - 4.3 0.5 3.2 1.0 0.1 Carex sp. 10.8 11.8 3.3 3.6 0.1 0.8 2.0 2.8 Juniperus communis 0.3 - 0.9 - - - - - Litter 6.4 1.1 6.7 3.8 14.0 16.6 20.6 10.6 Bare soil 4.7 0.6 13.5 3.2 4.5 1.5 4.1 2.8 Plot codes: NigUp: plot dominated by F. nigrescens with shrubs encroaching uphill; NigDown: plot dominated by F. nigrescens with shrubs encroaching downhill; EskUp: plot dominated by F. eskia with shrubs encroaching uphill; EskDown: plot dominated by F. eskia with shrubs encroaching downhill. doi:10.1371/journal.pone.0028652.t001 Plot codes: NigUp: plot dominated by F. nigrescens with shrubs encroaching uphill; NigDown: plot dominated by F. nigrescens with shrubs encroaching downhill; EskUp: plot dominated by F. eskia with shrubs encroaching uphill; EskDown: plot dominated by F. eskia with shrubs encroaching downhill. doi:10.1371/journal.pone.0028652.t001 Plot codes: NigUp: plot dominated by F. nigrescens with shrubs encroaching uphill; NigDown: plot dominated by F. nigrescens with shrubs encroaching downhill; EskUp: plot dominated by F. eskia with shrubs encroaching uphill; EskDown: plot dominated by F. eskia with shrubs encroaching downhill. doi:10.1371/journal.pone.0028652.t001 equation: sides were located following maximum slope direction (slope = 10–25%), whereas the 20-m sides of the plots were located following contour lines (slope = 0). Thus, one of the 20-m sides bordered onto shrubland while the other side bordered onto open grassland. Study site We chose 20610 m plots because in a preliminary survey, neither adult nor juvenile shrubs were found at more than 10 m distance from shrubland boundary, and also to ensure a minimum number of 30 juveniles per plot. Adult shrubs covered less than 25% of the plot surface area. equation: G i (d)~ P n j~1 wij(d)xj P n j~1 xj where wij(d) expresses the binary connections based on distance (1.5 units = 37.5 cm) between quadrats i and j in a plot, and xj denotes the counts of juvenile shrubs in quadrat j. A significant Gi * value (p,0.05) reveals hot spots of juveniles. An ANOVA was run to test for differences in the number of hot spots between different grassland communities. The spatial hot spots were calculated using PASSaGE v2 software [35]. To assess fine-scale effects on shrub establishment, each plot (20610 m) was gridded into 3200 25625-cm quadrats. A 1-m vegetation sampling frame divided into sixteen 25625-cm squares was used to sample all the quadrats in each plot. In each quadrat, we measured the cover of F. nigrescens, F. eskia, Calluna vulgaris, Carex sp., Juniperus communis, litter and bare soil in different cover classes (0%, 1%–25%, 26–50%, 51%–75% or 76%–100%). Cover types were only considered if they accounted for at least 10% cover within a quadrat. In addition to grid sampling, we also measured the coordinates of all juvenile shrubs. As the study focus is grassland encroachment, only juvenile shrubs that had a minimum distance of 20 cm to the closest adult shrub were measured and mapped. Height and diameter of each juvenile shrub were measured using a ruler and a digital caliper, respectively. According to diameter-age relationships for Cytisus, the ages of juveniles in the plots ranged between 0 and 6 years. Scale of pattern of cover types and spatial associations between juvenile shrubs and covers The scales of heterogeneity in a landscape detected by pattern analysis are important characteristics of the landscape, as they affect the responses of organisms to their environment and other ecological processes [32]. The scale of pattern is defined as half the average distance between patch centres [36]. The scale of pattern of each cover type (plant species, litter and bare soil) in the plots was analyzed using four-term local quadrat variance, 4TLQV [36], an extension of the two-term local quadrat variance (TTLQV) method used to analyze scale of pattern in one- dimensional transects. The local variance in 4TLQV is calculated based on the total cover in each of four mutually contiguous square blocks, each consisting of 1, 4, 9, 16… of the original sample quadrats. The original quadrats are combined into square blocks, which also form a square of four blocks for 4TLQV. The variance is essentially the squared difference between the total cover in one block and the average total cover in the adjacent three blocks. This calculation is then performed for a range of block sizes. Peaks in 4TLQV indicate the scale of pattern in the data. To confirm that these peaks did not occur by chance, randomization tests were performed to construct a null variance model and the associated 95% confidence intervals. December 2011 \| Volume 6 \| Issue 12 \| e28652 Spatial Heterogeneity and Shrub Encroachment total cover for one cover type), the difference in the pair of variables (e.g. abundance of juveniles and total cover for one cover type) is multiplied. This calculation is then performed for a range of block sizes. Peaks in this covariance are indicative of scale of spatial association between juveniles and a particular cover type (being positive when juveniles and a particular cover type tend to be found together and negative in the opposite case). To confirm that these peaks did not occur by chance, randomization tests were performed to construct a null covariance model and the associated 95% confidence intervals. Peaks with either positive or negative values greater in absolute magnitude than the confidence limits were interpreted as indicative of positive or negative associations. In addition, to test whether the spatial associations between juveniles and different cover types changed at different shrub development stages, we assessed the spatial associations (4TLQC) between the different cover types and all short (height,10 cm) and tall (height.30 cm) juveniles in a particular plot. Both scale of pattern and the spatial associations between the different cover types and juveniles were calculated using PASSaGE v2 soft- ware [35]. definition’’ significant, because the term si ensures that only significant spatial associations (based on the null model and 95% confidence intervals) are included. There are many predictive habitat distribution models in ecology [37], but most of them do not incorporate biotic interactions terms (e.g. competition, facilitation) into spatial modelling [38]. The better predictions for biotic interactions can be integrated using indirect variables, such as cover type, for modelling at small spatial scales [37]. Therefore, the new RISES index, which integrates biotic interactions based on covariation of spatial patterns, may help fill this gap. As the RISES index is based on results of 4TLQC spatial analysis, it allows us to integrate the sign of spatial associations and the spatial scales at which these spatial associations occur, which is something that could not be achieved with an index based, for example, on contingency tables using presence/absence data. Spatial patterns of juvenile shrubs The total number of juvenile shrubs found in the plots was similar between different grassland communities (p = 0.400), with values ranging from 42 to 130 juvenile shrubs per plot (Table 1). However, the number of short (height,10 cm) juvenile shrubs per plot was greater in F. nigrescens than F. eskia grasslands (45.0 vs. 17.3 juveniles per plot; p = 0.049). In contrast, total number of tall (height.30 cm) juvenile shrubs per plot was lower in F. nigrescens than F. eskia grasslands (1.2 vs. 9.2 juveniles per plot; p = 0.023). Number of spatial hot spots was not significantly different between grassland communities (p = 0.421; Figure 1). However, F. nigrescens plots always had over 100 hot spots while two F. eskia plots had a hot spot count of less than 80. RISES~ P si\|(Bmax{Bi)\|ci (Bmax{1)\|cgrass RISES~ P si\|(Bmax{Bi)\|ci (Bmax{1)\|cgrass For each cover type (i) in a particular plot, si takes the value +1 or 21 depending on whether cover type i was positively or negatively associated with Cytisus. When no association was found between Cytisus and cover type i, si takes the value 0. Bmax is the maximum block size (expressed as number of quadrats) used in the 4TLQC calculations, whereas Bi is the smallest block size (expressed as number of quadrats) in which 4TLQC found a significant association. ci is the cover (%) for the cover type i in the plot, while cgrass is the cover in % for the grassland surface in the plot, and cgrass~ P ci. RISES summarizes the results of the covariance analyses by creating a weighted sum of the scales at which the first significant (positive or negative) peaks are found in plots of covariance as a function of block size (scale). The index takes values close to +1 when there are only strong positive associations with all cover types, and values close to 21 when there are only strong negative associations found. Values close to 0 can indicate either a mixture of positive and negative associations, or a neutral susceptibility to shrub encroachment. RISES values are ‘‘by Spatial Heterogeneity and Shrub Encroachment Although this index does have its drawbacks (for instance, it assumes that juvenile shrubs have the same chance of establishing in a particular cover type regardless of proximity to adult shrubs), RISES makes it possible to assess potential vulnerability of grassland communities to shrub encroachment based on the surface area of different cover types and the spatial associations (strength and sign) between the different cover types (plant species, litter and bare soil) and juvenile Cytisus. Thus, grasslands with the same proportion of cover types may have different RISES values if there are distinct spatial associations between cover types and juveniles. RISES can be computed on any other plot of contiguous quadrats for which the cover or abundance of each cover type is known. RISES can be used not only to assess shrub encroachment into grasslands but also to assess invasibility in spatially heter- ogeneous habitats by a non-native plant species or in any other setting where there are signicant patterns of spatial covariance between species or cover types. In addition, to test whether the spatial associations between juveniles and different cover types changed at different shrub development stages, we assessed the spatial associations (4TLQC) between the different cover types and all short (height,10 cm) and tall (height.30 cm) juveniles in a particular plot. Both scale of pattern and the spatial associations between the different cover types and juveniles were calculated using PASSaGE v2 soft- ware [35]. Relative Index of Shrub Encroachment Susceptibility (RISES) The associations between juveniles and each cover type were summarized to evaluate the chances of shrub encroachment for a given plot, using a new index, ‘‘Relative Index of Shrub Encroachment Susceptibility’’ (RISES). Although a given 4TLQC analysis can find significant peaks at different block sizes, RISES is based on the first significant peak (i.e. the one at the smallest block size). RISES was calculated for each plot using the following equation: Spatial Heterogeneity and Shrub Encroachment Spatial Heterogeneity and Shrub Encroachment 4TLQC [36], a method similar to 4TLQV but which examines covariation between variables of interest in quadrats. The local covariance in 4TLQC is calculated based on the abundance of juveniles and the total cover of a particular cover type in each of four mutually contiguous square blocks, each consisting of 1, 4, 9, 16… of the original sample quadrats. The original quadrats are combined into square blocks, which also form a square of four blocks for 4TLQC. The covariance calculation is similar to the variance calculation described above for 4TLQV, but instead of squaring the difference of one variable (e.g. total cover for one cover type), the difference in the pair of variables (e.g. abundance of juveniles and total cover for one cover type) is multiplied. This calculation is then performed for a range of block sizes. Peaks in this covariance are indicative of scale of spatial association between juveniles and a particular cover type (being positive when juveniles and a particular cover type tend to be found together and negative in the opposite case). To confirm that these peaks did not occur by chance, randomization tests were performed to construct a null covariance model and the associated 95% confidence intervals. Peaks with either positive or negative values greater in absolute magnitude than the confidence limits were interpreted as indicative of positive or negative associations. In addition, to test whether the spatial associations between juveniles and different cover types changed at different shrub development stages, we assessed the spatial associations (4TLQC) between the different cover types and all short (height,10 cm) and tall (height.30 cm) juveniles in a particular plot. Both scale of pattern and the spatial associations between the different cover types and juveniles were calculated using PASSaGE v2 soft- ware [35]. 4TLQC [36], a method similar to 4TLQV but which examines covariation between variables of interest in quadrats. The local covariance in 4TLQC is calculated based on the abundance of juveniles and the total cover of a particular cover type in each of four mutually contiguous square blocks, each consisting of 1, 4, 9, 16… of the original sample quadrats. The original quadrats are combined into square blocks, which also form a square of four blocks for 4TLQC. The covariance calculation is similar to the variance calculation described above for 4TLQV, but instead of squaring the difference of one variable (e.g. Spatial patterns of juveniles p p j We considered the total number of juveniles in each plot. In addition, two size classes of juvenile Cytisus, i.e. short (height,10 cm) and tall (height.30 cm), were used as indicators of different shrub development stages. For each plot, the total number of juveniles and the number of juveniles in different development stages (short and tall) were compared between different grassland communities using ANOVA. Spatial hot spots of juvenile shrubs (high-density clusters of juvenile shrubs in a particular plot) were detected using local spatial statistics [32,33]. In order to detect spatial hot spots of juveniles, the local Gi * statistic [34] was calculated for juvenile shrub counts in each 25625-cm quadrat in a given plot, using the Associations between juvenile shrubs and the different cover types were assessed using four-term local quadrat covariance, PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 December 2011 \| Volume 6 \| Issue 12 \| e28652 3 December 2011 \| Volume 6 \| Issue 12 \| e28652 PLoS ONE \| www.plosone.org Scale of pattern of cover types and spatial associations between juvenile shrubs and cover types In each grassland type, the dominant grass species cover was approximately 50%. The highest litter cover values were found in F. eskia plots. Bare soil cover was similar between grassland types (Table 1). The most important difference was found in the scale of pattern of bare soil depending on grassland community (Figure 2): whereas bare soil showed a scale of pattern at 0.5 m in most of the F. eskia plots, its scale of pattern in the F. nigrescens plots was far more variable among plots (from 1.25 to 4.75 m). The remaining cover types did not show clear trends (Figure 2). Regardless of grassland community, F. nigrescens and F. eskia grass patches showed opposite associations with juvenile Cytisus (Figure 3; App. 1). Positive associations were found with F. PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 December 2011 \| Volume 6 \| Issue 12 \| e28652 4 Spatial Heterogeneity and Shrub Encroachment nigrescens (scales between 0.75 and 2.5 m) while negative associa- tions were found with F. eskia (from 0.5 to 3.25 m). For the remaining cover types we found contrasting associations with grasslands, most of the non-dominant cover types showed associations, except Calluna vulgaris which showed both po negative associations and Carex sp which showed a Figure 1. Juvenile Cytisus hot spots detected with local Gi * statistic in F. nigrescens or F. eskia grasslands. Number of hot spots brackets. doi:10.1371/journal.pone.0028652.g001 Figure 1. Juvenile Cytisus hot spots detected with local Gi * statistic in F. nigrescens or F. eskia grasslands. Number of hot spots is given in brackets. doi:10.1371/journal.pone.0028652.g001 nigrescens (scales between 0.75 and 2.5 m) while negative associa- tions were found with F. eskia (from 0.5 to 3.25 m). For the remaining cover types, we found contrasting associations with juvenile Cytisus depending on grassland type (Figure 3). In F. nigrescens grasslands, most of the non-dominant cover types showed negative associations, except litter and bare soil which did not show clear association trends (Figure 3). In contrast, in F. eskia grasslands, most of the non-dominant cover types showed positive associations, except Calluna vulgaris which showed both positive and negative associations, and Carex sp. which showed a positive association in just two out of four plots. In F. December 2011 \| Volume 6 \| Issue 12 \| e28652 Susceptibility of grassland communities to shrub establishment Total number of juvenile shrubs at early development stage (i.e. height,10 cm) was greater in grasslands dominated by the palatable F. nigrescens than in grasslands dominated by the non- palatable F. eskia. F. eskia grass patches and juveniles were negatively associated (see Appendix S1), suggesting that ‘associa- tional resistance’ plays a less relevant role than other negative mechanisms driving Cytisus shrub establishment. In contrast, F. nigrescens was positively associated with juvenile Cytisus (see Appendix S1). These associations cannot be due to differences in edaphic factors, since both F. eskia and F. nigrescens species appear in similar soils that greatly differ from the soils occupied by other grassland communities in the Pyrenees, such as Festuca paniculata grasslands [14]. Therefore, the associations found between both Festuca species and juvenile Cytisus still require experimental evidence and identification of the mechanisms driving them. Scale of pattern of cover types and spatial associations between juvenile shrubs and cover types eskia grasslands, litter and bare soil always showed positive fine-scale associations (from 0.75 to 1.25 m for litter and from 1.0 to 1.5 m for bare soil; Figure 3). nigrescens (scales between 0.75 and 2.5 m) while negative associa- tions were found with F. eskia (from 0.5 to 3.25 m). For the remaining cover types, we found contrasting associations with juvenile Cytisus depending on grassland type (Figure 3). In F. nigrescens grasslands, most of the non-dominant cover types showed negative associations, except litter and bare soil which did not show clear association trends (Figure 3). In contrast, in F. eskia grasslands, most of the non-dominant cover types showed positive associations, except Calluna vulgaris which showed both positive and negative associations, and Carex sp. which showed a positive association in just two out of four plots. In F. eskia grasslands, litter and bare soil always showed positive fine-scale associations (from 0.75 to 1.25 m for litter and from 1.0 to 1.5 m for bare soil; Figure 3). December 2011 \| Volume 6 \| Issue 12 \| e28652 PLoS ONE \| www.plosone.org PLoS ONE \| www.plosone.org 5 Spatial Heterogeneity and Shrub Encroachment Figure 2. Scale of pattern for each cover type in F. nigrescens or F. eskia grasslands. Cover types: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. (A = cover type absent in the plot; N = scale of pattern is non-significant below 5 m). doi:10.1371/journal.pone.0028652.g002 Figure 2. Scale of pattern for each cover type in F. nigrescens or F. eskia grasslands. Cover types: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. (A = cover type absent in the plot; N = scale of pattern is non-significant below 5 m). doi:10.1371/journal.pone.0028652.g002 Short juvenile shrubs showed opposite associations with F. nigrescens and F. eskia grasses, regardless of grassland community (Figure 4). Positive associations were found with F. nigrescens (scales between 0.5 and 2.75 m) while negative associations were found with F. eskia (from 1.5 to 4.75 m). In contrast with the associations found for short juvenile shrubs, tall juvenile Cytisus showed fairly positive associations with F. eskia (except in one plot) and unclear associations with F. nigrescens (negative associations in three out of five plots; Figure 4). In addition, only very few plots in F. nigrescens- dominated grasslands contained tall (height.30 cm) juveniles (Figure 4). Relative Index of Shrub Encroachment Susceptibility (RISES) (A = cover type absent in the plot; J = juvenile Cytisus with height.30 cm absent in the plot; N = association is non-significant below 5 m). All the scales of association except those indicated (N) are significant (p,0.05) based on randomization tests used to build a null covariance model and the associated 95% confidence intervals. doi:10.1371/journal.pone.0028652.g004 negative associations in F. nigrescens-dominated grasslands to mostly positive associations in F. eskia-dominated grasslands. For instance, associations between Calluna vulgaris or Carex sp. and juvenile Cytisus were mainly negative in F. nigrescens grasslands yet mostly positive or neutral in F. eskia grasslands, suggesting the existence of indirect interactions in these communities [24]. These association drifts might be related to alterations of competitive effects between species, and may thus determine species assemblages. Litter and bare soil also showed positive and strong associations with juveniles in F. eskia grasslands (see Appendix S1) but unclear effects in F. nigrescens grasslands. Grass litter plays a complex role in plant establishment [43], and our results may also be partially explained by the different physical and chemical properties of litter [44]. Bare soil has significant and diverse effects on the establishment of woody plants [45,46]. Either due to intrinsic community structure or to factors such as grazing and/or trampling, grassland communities show different bare soil patterns which may, in turn, affect shrub establishment. Based on our results, the lower scale of pattern on bare soil in F. eskia than in F. nigrescens grasslands may offer major chances of shrub establishment and survival, with juveniles on bare soil being protected by surrounding tall and unpalatable F. eskia tussocks. F. eskia may exert stronger belowground competition with shrubs than other grass species [39], since being ungrazed it can allocate resources mainly to belowground biomass and develop a deeper rooting system than other grass species [40]. In contrast, the fact that F. nigrescens invests in resprouting after grazing may result in slower belowground biomass growth [41], which may favour shrub establishment. Furthermore, aboveground constraints (e.g. light) may also explain the associations found [42]. Since grazing pressure is greater in F. nigrescens patches than in F. eskia patches, our results run opposite to what would be predicted by classic indirect plant-herbivore studies [24,25]. Looking at the RISES values, F. nigrescens-dominated grass- lands appear more susceptible to shrub encroachment than F. eskia-dominated grasslands. Thus, shrub establishment into F. Relative Index of Shrub Encroachment Susceptibility (RISES) The RISES index indicated opposite susceptibility to shrub encroachment between F. nigrescens and F. eskia grasslands (p = 0.002). Mean RISES values were higher in F. nigrescens grasslands than F. eskia grasslands (+0.43 vs. 20.18, respectively). Figure 3. Scales of positive and negative associations between juvenile Cytisus and each cover type in F. nigrescens or F. eskia grasslands. Cover type: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. (A = cover type absent in the plot; N = association is non-significant below 5 m). All the scales of association except those indicated (N) are significant (p,0.05) based on randomization tests used to build a null covariance model and the associated 95% confidence intervals. doi:10.1371/journal.pone.0028652.g003 Figure 3. Scales of positive and negative associations between juvenile Cytisus and each cover type in F. nigrescens or F. eskia grasslands. Cover type: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. (A = cover type absent in the plot; N = association is non-significant below 5 m). All the scales of association except those indicated (N) are significant (p,0.05) based on randomization tests used to build a null covariance model and the associated 95% confidence intervals. doi:10.1371/journal.pone.0028652.g003 December 2011 \| Volume 6 \| Issue 12 \| e28652 PLoS ONE \| www.plosone.org 6 6 Spatial Heterogeneity and Shrub Encroachment Figure 4. Scales of positive and negative associations between each cover type and a) short juvenile Cytisus (height,10 cm) and b) tall juvenile Cytisus (height.30 cm) in F. nigrescens or F. eskia grasslands. Cover types: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. (A = cover type absent in the plot; J = juvenile Cytisus with height.30 cm absent in the plot; N = association is non-significant below 5 m). All the scales of association except those indicated (N) are significant (p,0.05) based on randomization tests used to build a null covariance model and the associated 95% confidence intervals. doi:10.1371/journal.pone.0028652.g004 Figure 4. Scales of positive and negative associations between each cover type and a) short juvenile Cytisus (height,10 cm) and b) tall juvenile Cytisus (height.30 cm) in F. nigrescens or F. eskia grasslands. Cover types: FesNig = Festuca nigrescens; FesEsk = Festuca eskia; CalVul = Calluna vulgaris; CarSp = Carex sp. PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 Acknowledgments Our study suggests that fine-scale spatial heterogeneity of cover types is a crucial factor in shrub encroachment into grasslands. Cytisus shrub encroachment into mesic grasslands is constrained first by the chances of new individuals being able to establish, and second by the success of their expansion. Shrub establishment is shaped by spatial patterns of suitable cover types. Grasslands dominated by F. nigrescens are more susceptible to shrub establishment than grasslands dominated by F. eskia. In F. eskia grasslands, the strong interspecific competition between F. eskia and Cytisus may reduce the number of The authors thank the Alt Pirineu Natural Park technicians for allowing them to work in the Campirme study site. Differential shrub encroachment into grasslands ifferential shrub encroachment into grasslands g Although RISES values together with the number of juveniles at early development stage (height,10 cm) showed that F. nigrescens grassland was a more susceptible habitat for shrub establishment than F. eskia grasslands, the number of juvenile shrubs at late stage of development (height.30 cm) was lower in F. nigrescens than F. eskia grasslands. This may be explained by the shifts from positive to negative associations between these two dominant grass species (F. nigrescens and F. eskia) and juvenile shrubs depending on development stage [47]. Assuming that the timeframe for juvenile establishment into grasslands (6 years) is short enough to avoid major changes of cover types in time, our results suggest that unpalatable F. eskia patches seem to limit Cytisus establishment (e.g. plant emergence) but improve the survival of established shrub juveniles. In contrast, palatable F. nigrescens patches seem to favour Cytisus establishment but diminish the survival of established juvenile shrubs. Given that one of the most apparent differences between these two grasslands is the palatability of the dominant grass, we suggest that the current grazing regime slows down Cytisus encroachment into grasslands dominated by the palatable grass F. nigrescens. Differences in juvenile survival between grassland communities may also partly explain the similar number of spatial hot spots found in both grassland types, with the highest survival of juveniles in F. eskia grasslands counterbalancing the lower susceptibility to shrub encroachment in these grasslands. There are several reports of mortality rates of various plant species decreasing significantly as abundance of neighbouring unpalatable species increases [48,49]. Thus, associa- tional resistance in F. eskia patches may not appear in the first development stages of Cytisus but instead emerge later on, thus increasing the survival of juvenile Cytisus. There is previous evidence of ontogenetic effects on grass-woody interactions [50], and these ontogenetic constraints are often critical in understanding and managing plant populations where particular life stages require specific conditions not shared by other life stages [47]. Supporting Information Appendix S1 Plots of the main associations between juvenile shrubs and different cover types in grasslands. Plots of the main associations between juvenile Cytisus (empty circles) and different cover types in F. nigrescens- and F. eskia-dominated grasslands. Cover in the plots was measured in 25625 cm quadrats. The sign of the association (positive or negative) is given in brackets. (PDF) Appendix S2 Changes in RISES values in different grasslands. Changes in RISES values in a) F. nigrescens-dominated grasslands and b) F. eskia-dominated grasslands with dominant species cover ranging from 50 to 90% and an additional represented non-dominant cover type ranging from 0 to 40%. Grassland cover (Cgrass) was assumed to be 100% in all the examples. For each case, an equally-distributed total surface cover of 10% was assumed for the rest of the non-represented cover types (e.g. Carex sp, litter, etc.) in all the situations, and the spatial association values used were derived from our plots. According to the results obtained, changes in surface cover of the dominant grass species are not enough to predict changes in RISES, as RISES values are also dependent on surface cover of non-dominant cover types in both F. nigrescens- and F. eskia-dominated grasslands. (PDF) Author Contributions Conceived and designed the experiments: FM PC MRTD. Performed the experiments: FM PC. Analyzed the data: FM MRTD. Contributed reagents/materials/analysis tools: FM MRTD. Wrote the paper: FM PC MRTD. Conceived and designed the experiments: FM PC MRTD. Performed the experiments: FM PC. Analyzed the data: FM MRTD. Contributed reagents/materials/analysis tools: FM MRTD. Wrote the paper: FM PC MRTD. References 1. Levin SA (1992) The problem of pattern and scale in ecology. Ecology 73: 1943–1967. 1. Levin SA (1992) The problem of pattern and scale in ecology. Ecology 73: 1943–1967. 10. Briggs JM, Hoch GA, Johnson LC (2002) Assessing the rate, mechanisms, and consequences of the conversion of tallgrass prairie to Juniperus virginiana forest. Ecosystems 5: 578–586. 2. Watt AS (1947) Pattern and process in the plant community. Journal of Ecology 35: 1–22. 2. Watt AS (1947) Pattern and process in the plant community. Journal of Ecology 35: 1–22. 11. Browning DM, Archer SR (2011) Protection from livestock fails to deter shrub proliferation in a desert landscape with history of heavy grazing. Ecological Applications;doi: 10.1890/10-0542.1. 3. Turkington R, Harper JL (1979) The growth, distribution and neighbour relationships of Trifolium repens in a permanent pasture. I Ordination, pattern and contact. Journal of Ecology 67: 201–218. 3. Turkington R, Harper JL (1979) The growth, distribution and neighbour relationships of Trifolium repens in a permanent pasture. I Ordination, pattern and contact. Journal of Ecology 67: 201–218. 12. Liu F, Wu XB, Bai E, Boutton T, Archer SR (2010) Quantifying soil organic carbon in complex landscapes: an example of grassland undergoing encroach- ment of woody plants. Global Change Biology 17: 1119–1129. 13. Montane´ F, Casals P, Taull M, Lambert B, Dale MRT (2009) Spatial patterns of shrub cover after different fire disturbances in the Pyrenees. Annals of Forest Science 66: 612. 5. McIntire EJB, Fajardo A (2009) Beyond description: the active and effective way to infer processes from spatial patterns. Ecology 90: 46–56. 6. Archer S, Schimel DS, Holland EA (1995) Mechanisms of shrubland expansion: Land use, climate or CO2? Climatic Change 29: 91–99. 14. Montane´ F, Casals P, Taull M, Lambert B, Dale MRT (2010) Spatial patterns of shrub encroachment in neighbouring grassland communities in the Pyrenees: floristic com- position heterogeneity drives shrub proliferation rates. Plant Ecology 211: 267–278. 7. Briggs JM, Knapp AK, Blair JM, Heisler JL, Hoch GA, et al. (2005) An ecosystem in transition: causes and consequences of the conversion of mesic grassland to shrubland. BioScience 55: 243–254. 15. Palmer SCF, Hester AJ, Elston DA, Gordon IJ, Hartley SE (2003) The perils of having tasty neighbours: grazing impacts of large herbivores at vegetation boundaries. Ecology 84: 2877–2890. 8. McKinley DC, Blair JM (2008) Woody plant encroachment by Juniperus virginiana in a mesic native grassland promotes rapid carbon and nitrogen accrual. Relative Index of Shrub Encroachment Susceptibility (RISES) eskia-dominated grasslands may primarily be constrained by the abundance of secondary cover types having positive associations with juveniles, such as F. nigrescens patches, litter or bare soil. The role of these secondary cover types must not be overlooked in either grassland community (see Appendix S2 for examples of changes in RISES values with secondary cover types). Interestingly, although F. nigrescens and F. eskia patches associations with Cytisus were similar in both grassland types, there was a general drift for the rest of the cover types, from mainly PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 7 6. Archer S, Schimel DS, Holland EA (1995) Mechanisms of shrubland expansion: Land use, climate or CO2? Climatic Change 29: 91–99. 1. Levin SA (1992) The problem of pattern and scale in ecology. Ecology 73: 1943–1967. 2. Watt AS (1947) Pattern and process in the plant community. Journal of Ecology 35: 1–22. 3. Turkington R, Harper JL (1979) The growth, distribution and neighbour relationships of Trifolium repens in a permanent pasture. I Ordination, pattern and contact. Journal of Ecology 67: 201–218. 4. Purves DW, Law RW (2002) Fine-scale spatial structure in a grassland community: quantifying the plant’s-eye view. Journal of Ecology 90: 121–129. 5. McIntire EJB, Fajardo A (2009) Beyond description: the active and effective way to infer processes from spatial patterns. Ecology 90: 46–56. 6. Archer S, Schimel DS, Holland EA (1995) Mechanisms of shrubland expansion: Land use, climate or CO2? Climatic Change 29: 91–99. 7. Briggs JM, Knapp AK, Blair JM, Heisler JL, Hoch GA, et al. (2005) An ecosystem in transition: causes and consequences of the conversion of mesic grassland to shrubland. BioScience 55: 243–254. 8. McKinley DC, Blair JM (2008) Woody plant encroachment by Juniperus virginiana in a mesic native grassland promotes rapid carbon and nitrogen accrual. Ecosystems 11: 454–468. 9. Van Auken OW (2000) Shrub invasions of North American semiarid grasslands. Annual Review of Ecology, Evolution and Systematics 31: 197–215. Spatial Heterogeneity and Shrub Encroachment Spatial Heterogeneity and Shrub Encroachment Differential shrub encroachment into grasslands juvenile shrubs at early development stage compared with F. nigrescens grasslands. However, the low scale of pattern of bare soil in F. eskia grasslands may offer a major opening for shrub establishment and survival in these grasslands. The associational resistance mechanism may act not at the establishment phase but later on, increasing the survival and proliferation of juvenile shrubs into unpalatable F. eskia grasslands, while current grazing rates may reduce the success of proliferation in the more palatable F. nigrescens grasslands, despite them being more susceptible to shrub establishment. Spatial Heterogeneity and Shrub Encroachment (1999) Shifting mosaics in grazed woodlands driven by the alternation of plant facilitation and competition. Plant Biology 1: 127–137. 42. Burke IC, Lauenroth WK, Vinton MA, Hook PB, Kelly RH, et al. (1998) Plant- soil interactions in temperate grasslands. Biogeochemistry 42: 121–143. soil interactions in temperate grasslands. Biogeochemistry 42: 121– 43. Xiong SJ, Nilsson C (1999) The effects of plant litter on vegetation: a meta- analysis. Journal of Ecology 87: 984–994. p gy 26. Pasche F, Armand M, Gouaux P, Lamaze T, Pornon A (2004) Are meadows with high ecological and patrimonial value endangered by heathland invasion in the French central Pyrenees? Biological Conservation 118: 101–108. 44. Hovstad KA, Ohlson M (2008) Physical and chemical effects of litter on plant establishment in semi-natural grasslands. Plant Ecology 196: 251–260. the French central Pyrenees? Biological Conservation 118: 101–1 y g 27. Roura-Pascual N, Pons P, Etienne M, Lambert B (2005) Transformation of a rural landscape in the Eastern Pyrennes between 1953 and 2000. Mountain Research and Development 25: 252–261. 45. Castro J, Zamora R, Hodar JA, Gomez JM (2004) Seedling establishment of a boreal tree species at its southernmost distribution limit: consequences of being in a marginal Mediterranean habitat. Journal of Ecology 92: 266–277. p 28. Smit C, Beguin D, Buttler A, Muller-Scharer H (2005) Safe sites for tree regeneration in wooded pastures: a case of associational resistance? Journal of Vegetation Science 16: 209–214. 46. Baraza E, Zamora R, Hodar JA (2006) Conditional outcomes in plant-herbivore interactions: neighbours matter. Oikos 113: 148–156. 47. Miriti MN (2006) Ontogenetic shift from facilitation to competition in a desert shrub. Journal of Ecology 94: 973–979. g 29. Debussche M, Escarre J, Lepart J (1980) Changes in mediterranean shub communities with Cytisus purgans and Genista scorpius. Vegetatio 43: 73–82. 48. Smit C, den Ouden JAN, Muller-Scharer H (2006) Unpalatable plants facilitate tree sapling survival in wooded pastures. Journal of Applied Ecology 43: 305–312. 30. Daget P, Poissonet J (1972) Une me´tode d’analyse phytologique des prairies. Criteres d’aplication. Annales Agronomiques 22: 5–41. 31. Gracia M, Montane´ F, Pique´ J, Retana J (2007) Overstory structure and topographic gradients determining diversity and abundance of understory shrub species in temperate forests in central Pyrenees (NE Spain). Forest Ecology and Management 242: 391–397. 49. Spatial Heterogeneity and Shrub Encroachment 17. Lavorel S, Oneill RV, Gardner RH (1994) Spatio-temporal dispersal strategies and annual plant species coexistence in a structured landscape. Oikos 71: 75–88. 34. Getis A, Ord JK (1992) Tha analysis of spatial association by use of distance statistics. Geographical Analysis 24: 189–206. 18. Rusch G, Ferna´ndez-Palacios JM (1995) The influence of spatial heterogeneity on regeneration by seed in a limestone grassland. Journal of Vegetation Science 6: 417–426. 35. Rosenberg MS (2009) PASSaGE. Pattern Analysis, Spatial Statistics, and Geographic Exegesis. Version 2. http://www.passagesoftware.net. 36. Dale MRT (1999) Spatial pattern analysis in plant ecology Cambridge University Press. 19. Skarpe C (1991) Spatial patterns and dynamics of woody vegetation in an arid savanna. Journal of Vegetation Science 2: 565–572. 37. Guisan A, Zimmermann NE (2000) Predictive habitat distribution models in ecology. Ecological Modelling 135: 147–186. 20. Maestre FT, Bautista S, Cortina J, Bellot J (2001) Potential for using facilitation by grasses to establish shrubs in semiarid degraded steppe. Ecological Applications 11: 1641–1655. 20. Maestre FT, Bautista S, Cortina J, Bellot J (2001) Potential for using facilitation by grasses to establish shrubs in semiarid degraded steppe. Ecological Applications 11: 1641–1655. 38. Austin M (2007) Species distribution models and ecological theory: A critical assessment and some possible new approaches. Ecological Modelling 200: 1–19. pp 21. Jurena PN, Archer S (2003) Woody plant establishment and spatial heterogeneity in grasslands. Ecology 84: 907–919. 39. Aerts R, Boot RGA, Vanderaart PJM (1991) The relation between above- and belowground biomass allocation patterns and competitive ability. Oecologia 87: 551–559. 22. Brooker RW, Maestre FT, Callaway RM, Lortie CL, Cavieres LA, et al. (2008) Facilitation in plant communities: the past, the present, and the future. Journal of Ecology 96: 18–34. 40. Pornon A, Escaravage N, Lamaze T (2007) Complementary in mineral nitrogen use among dominant plant species in a subalpine community. American Journal of Botany 94: 1778–1785. gy 23. Bertness MD, Callaway R (1994) Positive interactions in communities. Trends in Ecology & Evolution 9: 191–193. 41. Cornelissen JHC, Lavorel S, Garnier E, Diaz S, Buchmann N, et al. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51: 335–380. gy 24. Callaway RM (2007) Positive interactions and interdependence in plant communities Springer. p g 25. Olff H, Vera FWM, Bokdam J, Bakker ES, Gleichman JM, et al. References Ecosystems 11: 454–468. 16. Milbau A, Stout JC, Graae BJ, Nijs I (2009) A hierarchical framework for integrating invisibility experiments incorporating different factors and spatial scales. Biological Invasions 11: 941–950. 9. Van Auken OW (2000) Shrub invasions of North American semiarid grasslands. Annual Review of Ecology, Evolution and Systematics 31: 197–215. PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 December 2011 \| Volume 6 \| Issue 12 \| e28652 8 December 2011 \| Volume 6 \| Issue 12 \| e28652 PLoS ONE \| www.plosone.org Spatial Heterogeneity and Shrub Encroachment Van Uytvanck J, Maes D, Vandenhaute D, Hoffmann M (2008) Restoration of woodpasture on former agricultural land: the importance of safe sites and time gaps before grazing for tree seedlings. Biological Conservation 141: 78–88. 32. Fortin M-J, Dale MRT (2005) Spatial analysis. A guide for ecologists. Cambridge, . 50. Soliveres S, DeSoto L, Maestre FT, Olano JM (2010) Spatio-temporal heterogeneity in abiotic factors modulate multiple ontogenetic shifts between competition and facilitation. Perspectives in Plant Ecology, Evolution and Systematics 12: 227–234. 33. Nelson TA, Boots B (2008) Detecting spatial hot spots in landscape ecology. Ecography 31: 556–566. PLoS ONE \| www.plosone.org December 2011 \| Volume 6 \| Issue 12 \| e28652 9
https://openalex.org/W3038341888	https://innfoodsecr.elpub.ru/jour/article/download/558/412	Russian	null	PRODUCTIVITY OF SOME LAKES OF THE KOSH-AGACH DISTRICT OF THE REPUBLIC OF ALTAI	Innovacii i prodovolʹstvennaâ bezopasnostʹ	2,020	cc-by	5,367	Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management УДК 597.2/.5 Л. В. Веснина, доктор биологических наук, профессор Т. О. Ронжина, кандидат биологических наук Н. В. Зеленцов, научный сотрудник Г. А. Романенко, младший научный сотрудник И. Ю. Теряева, младший научный сотрудник Д. Г. Елизарьев, старший лаборант Алтайский филиал федерального государственного бюджетного научного учреждения «Государственный научно-производственный центр рыбног E-mail: artemiaalt@mail.ru Алтайский филиал федерального государственного бюджетного научного учреждения «Государственный научно-производственный центр рыбного хозяйства» E-mail: artemiaalt@mail.ru PRODUCTIVITY OF SOME LAKES OF THE KOSH-AGACH DISTRICT OF THE REPUBLIC OF ALTAI L. V. Vesnina, Doctor of Biological Sciences, Professor T. O. Ronzhina, Candidate of Biological Sciences N. V. Zelentsov, Researcher G. A. Romanenko, Junior Researcher I. Yu. Teryaeva, Junior Researcher D. G. Elizariev, Senior Laboratory Assistant L. V. Vesnina, Doctor of Biological Sciences, Professor T. O. Ronzhina, Candidate of Biological Sciences N. V. Zelentsov, Researcher G. A. Romanenko, Junior Researcher I. Yu. Teryaeva, Junior Researcher D. G. Elizariev, Senior Laboratory Assistant Key words: phytoplankton, zooplankton, fish fauna, monitoring, ecosystem. Abstract. The Republic of Altai has a large length of watercourses and significant areas of lakes of fish- ery importance. The collection of materials was carried out in the field season of 2018 on the lakes: Uch-Kol (Three Lakes), Kamyshevoye and Lake without a name, located within the borders of Kosh-Agach district of the Altai Republic. Research has covered the entire water area. In the course of the research it was noted that Lake Uch-Kol consists of three successive basins with a total area of 7.9 hectares. The upper lake basin has an area of 5.5 hectares. The average lake basin has an area of 1.8 hectares. The lower lake basin has an area of 0.6 ha. Higher aquatic vegetation in the pond is missing. Lake zooplankton is represented by one species of Cladocera and two species of Copepoda. Lake Zoobenthos Uch-Kol is represented by a representative of the r. Gammarus. The area of Kamyshevoye Lake is 7.4 ha. The pond has a slightly elongated shape from west to east. In the zooplankton of the reservoir of mass development, branchy and rotifers reach. Chironomid larvae and caddisfruits dominate in the structure of zoobenthos. The area of the lake without a name is 29.4 hectares. The reservoir has a form elongated from the northeast to the southwest. In terms of numbers and biomass, Copepoda copepads are the dominant group in zooplankton. The dominant position in the zoobenthos compo- sition is occupied by caddis flies and gammarids. All studied water bodies have low biological productivity. Phytoplankton is represented mainly by diatoms and green algae. The ichthyological fauna of Lake Uch-Kol and Lake Kamyshevoye is represented by an oligoid type of ichthyocenosis. The ichthyological fauna in the Lake without a name is missing. Республика Алтай располагает большой протяжённостью водотоков и значительными площадями озёр, имеющих рыбохозяйственное значение. Ключевые слова: фитопланктон, зоопланктон, ихтиофауна, мониторинг, экосистема. Реферат. Республика Алтай располагает большой протяжённостью водотоков и значительными площадями озёр, имеющих рыбохозяйственное значение. Сбор материалов осуществлялся в полевой сезон 2018 г. на озерах Уч-Кол (Три Озера), Камышевое и озере без названия, расположенных в грани- цах Кош-Агачского района Республики Алтай. Исследованиями была охвачена вся акватория водоемов. В ходе проведенных исследований отмечено, что оз. Уч-Кол состоит из трех последовательно рас- положенных котловин общей площадью 7,9 га. Верхняя котловина озера имеет площадь 5,5 га, сред- няя – 1,8, нижняя – 0,6 га. Высшая водная растительность в водоеме отсутствует. Зоопланктон озе- ра представлен одним видом Cladocera и двумя видами Copepoda. Зообентос оз. Уч-Кол представлен представителем рода Gammarus. Площадь оз. Камышевое составляет 7,4 га. Водоем имеет немного вытянутую с запада на восток форму. В зоопланктоне водоема массового развития достигают вет- вистоусые и коловратки. В структуре зообентоса доминирующее положение занимают личинки хиро- номид и ручейники. Площадь озера без названия составляет 29,4 га. Водоем имеет вытянутую с севе- ро-востока на юго-запад форму. По численным показателям и биомассе доминирующей группой в зоо- планктоне являются веслоногие – Copepoda. Доминирующее положение в составе зообентоса занима- ют ручейники и гаммариды. Все изученные водные объекты имеют низкую биологическую продуктив- ность. Фитопланктон представлен в основном диатомовыми и зелеными водорослями. Ихтиофауна оз. Уч-Кол и оз. Камышевое представлена олиговидовым типом ихтиоценоза. Ихтиофауна в озере без названия отсутствует. 94 «Инновации и продовольственная безопасность» № 2 (24)/2019 Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management тованного Испытательного лабораторного центра (Алтайский край, г. Славгород). Гидрохимический анализ воды выполнен по общепринятой методике [3]. тованного Испытательного лабораторного центра (Алтайский край, г. Славгород). Гидрохимический анализ воды выполнен по общепринятой методике [3]. Сбор и обработку гидробиологических проб проводили по общепринятым методикам [4—9]. Гидробиологический материал был собран в дневное время на заранее намеченных станциях в раз- личных местах водоема (в зависимости от развития береговой линии, глубин и степени зарастания макрофитами) сетью Апштейна с ситом № 72. Пробы фиксировали в 4 %-м растворе формалина и ана- лизировали в лабораторных условиях в камере Богорова под бинокулярным микроскопом МБС-10 по качественным (видовой состав) и количественным характеристикам (численность и биомасса). Для изучения зообентоса пробы грунта отбирали дночерпателем Петерсена с площадью захвата 0,025 м 2, промывали в мешке из газа № 32, организмы фиксировали 4 %-м раствором формалина, в ла- бораторных условиях разбирали по группам и взвешивали на торсионных весах с точностью до 0,001 г. В дальнейшем проводили пересчет биомассы на единицу площади и на площадь того или иного био- топа водоема с учетом изменения массы организмов после фиксации формалином. Камеральная обра- ботка проб проведена по общепринятым методикам [10]. Видовой состав личинок насекомых, олигохет и пиявок приведен в соответствии с систематикой, принятой в «Определителе пресноводных беспозво- ночных России и сопредельных территорий» [11—14]. Видовой состав и пространственное распределение рыб изучали методом контрольных обловов набором ставных жаберных сетей с ячеей 22,0—85,0 мм, длина каждой сети – 25,0 м, общая длина на- бора сетей – 350,0 м. Длительность экспозиции – 12 ч. При сборе, обработке и анализе ихтиологического материала применяли стандартные, общеприня- тые методики [15—17]. Обработку ихтиологического материала проводили в соответствии с методикой полевых ихтиологических исследований [18]. Статистическая обработка, подготовка таблиц и графических изображений данных проведены с использованием программ Microsoft Excel, Microsoft Word. Место проведения исследований – Юго-Восточная Алтайская провинция – отличается значитель- ным своеобразием ландшафтов и имеет больше сходных черт с соседними территориями Монголии, чем с другими провинциями Алтая. Специфические особенности природы обусловлены здесь вы- соким положением поверхности территории (средняя высота 2300—2700 м), изолированностью от Центрального Алтая высокими хребтами, большой суровостью и континентальностью климата, су- щественным влиянием соседних областей Центральной Азии, что выражается в формировании цен- трально-азиатского типа высотной поясности. Орографическую основу провинции образуют хребты Чихачева, Сайлюгем, Курайский, средняя высота которых 3200—3400 м; отроги Южно-Чуйского и ча- стично Северо-Чуйского хребта, а также плоскогорье Укок, юго-восточная оконечность Чулышманского плоскогорья с Джулукульской впадиной и межгорные котловины: Чуйская и Курайская. PRODUCTIVITY OF SOME LAKES OF THE KOSH-AGACH DISTRICT OF THE REPUBLIC OF ALTAI Большинство водоёмов республики относятся к первой и высшей рыбохозяйственной категории благодаря обитанию и размножению в них ценных видов рыб: лососёвых (обыкновенный таймень) и сиговых (обыкновенный сиг, пелядь). Однако рыбохозяйственное использование водоёмов республики и их ихтиофауны находится на низком уровне в связи с труднодоступностью высокогорных водоёмов и крайне низким уровнем орга- низации спортивно-любительского рыболовства. Целью данной работы было проведение мониторинговых исследований на водоемах Кош-Агачского района Республики Алтай в вегетационный период 2018 г. для выявления современного состояния раз- вития биоты водных объектов. Сбор материалов осуществлялся в полевой сезон 2018 г. на озерах Уч-Кол (Три Озера), Камышевое и озере без названия, расположенных в границах Кош-Агачского района Республики Алтай. Сбор материалов осуществлялся в полевой сезон 2018 г. на озерах Уч-Кол (Три Озера), Камышевое и озере без названия, расположенных в границах Кош-Агачского района Республики Алтай. Исследованиями была охвачена вся акватория водоемов. Для выявления гидрологического режима водных объектов использованы данные гидролого-гидрохимической службы, сверенные с показателями сайта http://www.gismeteo.ru/, и собственные наблюдения за температурой воды, воздуха и количеством осадков. Исследование морфометрических показателей выполнено согласно методикам практической гидрометрии [1, 2]. Из них изучены и приведены основные: площадь водной поверхности и глубина. Гидрохимические пробы объемом 1,0—1,5 л отбирали одновременно с отбором проб зоопланктона. Анализ гидрохимических проб проводили по общепринятой методике с определением состава основ- ных ионов. Гидрохимические пробы обрабатывались на договорных условиях сотрудниками аккреди- «Инновации и продовольственная безопасность» № 2 (24)/2019 95 Рациональное природопользование Rational nature management Питание в основном происходит за счет поверхностного стока. Озеро проточное – ручьи с тающего ледника впадают в верхнюю часть озера. Уровень воды постоянный, с незначительными колебаниями в течение года. Вода гидрокарбонатно-кальциевая, мягкая, близкая к нейтральной, с рН 7,3±0,2. Сумма ионов составляет 43,25 мг/л, Na+ + K+ – 19,0, Ca++ – 7,3±1,1, Mg++ – 3,89+0,08, Cl– – 0,64±0,10, HCO3 - – 1,5±0,2 мг/л. Прозрачность по диску Секки 4,5 м. Берег каменистый, береговая линия слабо изрезана. Растительные сообщества береговой зоны представлены в основном горно-степными ландшафтами. Растительный покров преимущественно со- стоит из мхов и лишайников. Древесная и кустарниковая растительность полностью отсутствует. Озеро Камышевое расположено в 4 км от с. Кош-Агач Кош-Агачского района Республики Алтай, в 500 м от трассы 84К-27 Кош-Агач – Теленгит – Сорторой. Озеро расположено в пойме р. Чуя на вы- соте 1758 м над уровнем моря. Координаты 50°0ʹ57,2ʹʹ СШ, 88°43ʹ40,1ʹʹВД. Озеро Камышевое расположено в 4 км от с. Кош-Агач Кош-Агачского района Республики Алтай, в 500 м от трассы 84К-27 Кош-Агач – Теленгит – Сорторой. Озеро расположено в пойме р. Чуя на вы- соте 1758 м над уровнем моря. Координаты 50°0ʹ57,2ʹʹ СШ, 88°43ʹ40,1ʹʹВД. Площадь водоема составляет 7,4 га. Средние отметки глубин около 3,5 м. Максимальные глубины наблюдаются в центральной части озера и составляют 5,4 м. Озеро имеет немного вытянутую с запа- да на восток форму. Максимальная длина водоема 386,0, ширина – 238,0 м. Длина береговой линии 1022,0 м, коэффициент изрезанности – 0,6 (слабоизрезанная). По происхождению оз. Камышевое относится к типу моренных озер. Юго-восточная часть озера представляет собой мелководную литоральную часть с глубинами не более 1,5–2,0 м. На остальной части озера глубины нарастают быстро, обрывисто. Донные грунты озера повсеместно представлены валунно-каменисто-галечниковыми фракциями, с глубин 0,4—0,6 м покрытыми легкими илистыми от- ложениями. Питание в основном происходит за счет поверхностного стока в весенний период, отмечено также родниковое питание. Из юго-восточной части озера вытекает ручей. Уровень воды постоянный, с не- значительными колебаниями в течение года. Вода гидрокарбонатно-кальциевая, мягкая, близкая к ней- тральной с pH 7,2. Содержание Ca++ – 17,3 мг/л, Mg++ – 10,6, Na+ + K+ – 3,7, HCO3 - – 20,7, SO4 - – 22,1, Cl– – 5,8, сумма ионов – 80,2 мг/л. Прозрачность по диску Секки 4,5 м. Берег болотистый, местами топкий, покрыт дерновинами злаковых и осоковых трав. Зарастание бордюрное (тростник, рогоз широколистный, сусак зонтичный). Растительные сообщества берего- вой зоны представлены в основном горно-степными ландшафтами. Растительный покров состоит из дернин ковыльника, осоки твердой. «Инновации и продовольственная безопасность» № 2 (24)/2019 Рациональное природопользование Rational nature management Озеро Уч-Кол (Три Озера) является истоком р. Себыстей в Кош-Агачском районе Республики Алтай и находится на высоте 2550 м над уровнем моря. Координаты 49°45′08″ СШ, 88°14′46″ ВД. у Озеро Уч-Кол состоит из трех последовательно расположенных котловин общей пло Верхняя котловина озера имеет площадь 5,5 га. Средние отметки глубин около 5,0 м. Максимальные глубины наблюдаются в центральной части озера и составляют 19,5 м. Озеро имеет вытянутую с юго-востока на северо-запад форму. Максимальная его длина 349,0, ширина – 244,0 м. Длина берего- вой линии 0,91 км, коэффициент изрезанности – 0,62. Средняя котловина озера имеет площадь 1,8 га. Средние отметки глубин около 4,0 м. Максимальные глубины наблюдаются в центральной части озера и составляют 17,5 м. Озеро имеет вытянутую с юго-востока на северо-запад форму. Максимальная его длина 251,0, ширина – 86,0 м. Длина берего- вой линии 0,61 км, коэффициент изрезанности – 0,73. Нижняя котловина озера имеет площадь 0,6 га. Средние отметки глубин около 2,5 м. Максимальные глубины наблюдаются в центральной части озера и составляют 6,5 м. Озеро имеет овальную форму, немного вытянуто с юга на север. Максимальная его длина 136,0, ширина – 64,0 м. Длина береговой линии 0,32 км, коэффициент изрезанности – 0,66. По происхождению оз. Уч-Кол относится к типу моренных озер. Глубины нарастают быстро, об- рывисто. Донные грунты озера повсеместно представлены валунно-каменисто-галечниковыми фрак- циями. По происхождению оз. Уч-Кол относится к типу моренных озер. Глубины нарастают быстро, об- рывисто. Донные грунты озера повсеместно представлены валунно-каменисто-галечниковыми фрак- циями. «Инновации и продовольственная безопасность» № 2 (24)/2019 96 Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management Древесная растительность отсутствует полностью, кустарниковая представлена малочисленными скоплениями акации на западном и северном берегах озера. Озеро без названия расположено в низовьях р. Тангыт, в 3,5 км от ее впадения в р. Джазатор (Жасатер) в Кош-Агачском районе Республики Алтай на высоте 2143 м над уровнем моря. Координаты 49,674 °CШ, 87,716° ВД. Площадь водоема составляет 29,4 га. Средние отметки глубин около 1,2 м. Максимальные глуби- ны наблюдаются в западной части озера и составляют 6,5 м. Озеро имеет вытянутую с северо-востока на юго-запад форму. Максимальная длина водоема 1026,0, ширина – 451,0 м. Длина береговой линии 2,53 км, коэффициент изрезанности – 1,3. По происхождению озеро без названия относится к типу моренных озер. Вся восточная часть озера представляет собой мелководную литоральную часть с глу- бинами не более 1,0—1,2 м. Яма в восточной части озера имеет площадь 6,0 га, глубины нарастают быстро, обрывисто. Донные грунты озера повсеместно представлены валунно-каменисто-галечнико- выми фракциями, с глубин 0,4—0,6 м покрытыми легкими илистыми отложениями, глубина которых достигает 0,8—1,0 м. Питание в основном происходит за счет поверхностного стока. Озеро проточное – ручей впадает в северо-западную часть озера, вытекает из северо-восточной. Уровень воды постоянный, с незначи- тельными колебаниями в течение года. Вода гидрокарбонатно-кальциевая, мягкая, близкая к нейтраль- ной, с pH 6,9—7,2. Прозрачность по диску Секки 2,5 м. Берег каменистый, покрыт сплошным ковром мхов и лишайников, береговая линия слабо изрезана. Растительные сообщества береговой зоны представлены в основном горно-степными ландшафтами. Растительный покров состоит из дернин ковыльника, осоки твердой. Древесная и кустарниковая рас- тительность представлена одиночно стоящими лиственницами и сплошными зарослями карликовой березы на северном и восточном берегах озера. 97 «Инновации и продовольственная безопасность» № 2 (24)/2019 Рациональное природопользование Rational nature management экз/м 3) и биомасса (г/м 3) зоопланктона оз. Камышевое Rotifera Cladocera Cop Следует отметить достаточно высокие показатели численности и биомассы зоопланктона в водое- ме применительно к сезону года, что может являться индикатором высокого потенциала этого компо- нента кормовой базы. Рациональное природопользование Rational nature management Биологическая продуктивность водоемов. В связи с расположением в высокогорье озеро Уч-Кол отличается низким температурным режимом, имеет невысокие показатели биологической продуктив- ности и является олиготрофным водоемом. Высшая водная растительность в водоеме отсутствует. Фитопланктон озера представлен преимуще- ственно зелеными и диатомовыми водорослями. По численности и биомассе преобладает Spirogyra sp. Зоопланктон озера представлен одним видом Cladocera (Daphnia longispina O. F. Müller) и двумя видами Copepoda (Mesocyclops (s.str.) leuckarti Claus, Diaptomus graciloides Lill), коловратки в кон- трольных пробах отсутствовали. По численным показателям Cladocera и Copepoda имеют равное значение (табл. 1). Доминирующая группа по численности Diaptomus graciloides, по биомассе – Daphnia longispina. Средняя биомасса зо- опланктона составляет 0,36 г/м 3. Зообентос озера Уч-Кол представлен представителем рода Gammarus. Наибольшая продуктив- ность зообентоса отмечена на мелководных участках. Средняя биомасса зообентоса озера составляет 1,1 г/м 2. Таблица 1 Таблица 1 Средние значения численности (тыс. экз/м 3) и биомассы (г/м 3) основных групп зоопланктона оз. Уч-Кол Станция Cladocera Copepoda численность биомасса численность биомасса 1 4,050 0,289 3,700 0,145 2 2,000 0,143 2,350 0,115 3 3,030 0,216 3,030 0,130 Среднее 3,030 0,216 3,026 0,130 Средние значения численности (тыс. экз/м 3) и биомассы (г/м 3) основных групп зоопланктона оз. Уч-Кол В связи с расположением в среднегорье оз. Камышевое обладает низким температурным режимом, имеет невысокие показатели биологической продуктивности. В связи с расположением в среднегорье оз. Камышевое обладает низким температурным режимом, имеет невысокие показатели биологической продуктивности. Видовой состав водной растительности представлен роголистником темно-зеленым (Ceratophillum demersum), харой мелкошиповой (Chara aculeolata), харой алтайской (C. altaica), хвостником обыкно- венным (Hippuris vulgaris), гидриллой мутовчатой (Hydrilla verticillata), пузырчаткой обыкновенной (Utricularia vulgaris), рдестом гребенчатым (Potamogeton pectinatus), рдестом плавающим (P. natans), урутью колосистой (Myriophyllum spicatum), урутью сибирской (M. sibiricum), рогозом широколистным (Typha latifolia), сусаком зонтичным (Butomus umbellatus). В составе фитопланктона отмечены преимущественно зеленые и диатомовые водоросли. В зоопланктоне водоема массового развития достигают ветвистоусые и коловратки. Доминирующее положение занимают ветвистоусые. Средняя биомасса зоопланктона составляет 1,1 г/м 3 (табл. 2). В составе фитопланктона отмечены преимущественно зеленые и диатомовые водоро В зоопланктоне водоема массового развития достигают ветвистоусые и коловратки. Доминирующее положение занимают ветвистоусые. Средняя биомасса зоопланктона составляет 1,1 г/м 3 (табл. 2). Т б 2 Таблица 2 ц Численность (тыс. экз/м 3) и биомасса (г/м 3) зоопланктона оз. Камышевое Станции Rotifera Cladocera Copepoda численность биомасса численность биомасса численность биомасса 1 0,3999 0,0056 5,8666 0,2551 10,5667 0,7028 2 0,0 0,0 16,3 0,53 5,2 0,31 3 0,4 0,0056 7,9 0,3362 15,0 1,003 Среднее 0,26 0,003 10,02 0,38 10,26 0,67 Численность (тыс. Рациональное природопользование Rational nature management Рациональное природопользование Rational nature management Озеро без названия в связи с расположением в высокогорье, низким температурным режимом име- ет невысокие показатели биологической продуктивности и является олиготрофным водоемом. Видовой состав водной растительности представлен роголистником темно-зеленым (Ceratophillum demersum), рдестом гребенчатым (Potamogeton pectinatus), рдестом длиннейшим (Potamogeton praelongus), болотником (Callitriche sp.); наиболее распространен горец земноводный (Persicaria amphibia). Для прибрежных валунов (открытой литорали) характерны обрастания, в основном в виде бурого налета диатомовых водорослей (родов Diatoma, Synedra); реже зеленовато-бурые пятна зеленых водорослей (родов Ulothrix, Spirogyra, Cladophora). Фитопланктон озера представлен преимущественно зелеными и диатомовыми водорослями. По численности и биомассе преобладает Dinobryon divergens. Зоопланктон озера характеризуется бедностью как по видовому, так и по количественному со- ставу. В структуре зоопланктонного сообщества отмечено 7 видов коловраток (Rotifera: Poliarthra trigla vulgaris Carlin, Asplanchna priodonta Gosse, Euchlanis deflexa Gosse, Brachionus quadridentatus hyphalmyros Tschugunoff, Keratella quadrata O. F. Müller, Kellicottia longispina Kellicott, Filinia longiseta Ehrenb); 5 видов ветвистоусых (Cladocera: Daphnia pulex De Geer, Daphnia longispina O. F. Müller, Simocephalus vetulus O. F. Müller, Bosmina coregoni Baird, Leptodora kindtii Focke) и 2 вида веслоногих ракообразных (Copepoda: Mesocyclops (s.str.) leuckarti Claus, Diaptomus graciloides Lill). По численным показателям и биомассе доминирующей группой являются веслоногие ракообраз- ные (Copepoda). Среди них наибольшую численность и биомассу имеет Mesocyclops (s.str.) leuckarti (табл. 3). Rotifera и Cladocera по численности равнозначны, по биомассе среди них ветвистоусые рачки превосходят коловраток. Средняя биомасса зоопланктона составляет 0,63 г/м 3. Таблица 3 Таблица 3 Средние значения численности (тыс. экз./м 3) и биомассы (г/м 3) основных групп зоопланктона озера без названия Станция Rotifera Cladocera Copepoda численность биомасса численность биомасса численность биомасса Центр 2,61 0,007 3,18 0,222 10,26 0,448 Прибрежная зона 3,29 0,005 2,41 0,177 11,22 0,395 Среднее 2,95 0,006 2,80 0,199 10,74 0,422 Табл редние значения численности (тыс. экз./м 3) и биомассы (г/м 3) основных групп зоопланктона озера без названия ния численности (тыс. экз./м 3) и биомассы (г/м 3) основных групп зоопланктона озера без названия Развитию донных организмов благоприятствует высокая степень зарастания озера макрофитами. В структуре зообентоса отмечены личинки хирономид, моллюски, гаммариды, ручейники, олигохеты и т. д. Доминирующее положение в составе зообентоса занимают ручейники и гаммариды. Наибольшая продуктивность зообентоса отмечена на мелководных участках. Средняя биомасса зообентоса озера составляет 1,4 г/м 2. В настоящее время в озере без названия отсутствует коренная ихтиофауна. Состав ихтиофауны. В настоящее время в оз. Уч-Кол коренная ихтиофауна представлена исклю- чительно сибирским хариусом (Thumallus arcticus (Pallas, 1776)). Необходимо отметить, что данный вид присутствует только в среднем и нижнем озерах. Верхнее озеро безрыбно. Следует отметить достаточно высокие показатели численности и биомассы зоопланктона в водое- ме применительно к сезону года, что может являться индикатором высокого потенциала этого компо- нента кормовой базы. Следует отметить достаточно высокие показатели численности и биомассы зоопланктона в водое- ме применительно к сезону года, что может являться индикатором высокого потенциала этого компо- нента кормовой базы. Развитию донных организмов благоприятствует высокая степень зарастания озера макрофитами. В структуре зообентоса отмечены личинки хирономид и стрекоз, моллюски, гаммариды, ручейники, олигохеты и т. д. Доминирующее положение в составе зообентоса занимают личинки хирономид и ру- чейники. Наибольшая продуктивность зообентоса отмечена на юго-восточном мелководном участке озера. Средняя биомасса зообентоса озера составляет 1,8 г/м 2. «Инновации и продовольственная безопасность» № 2 (24)/2019 98 «Инновации и продовольственная безопасность» № 2 (24)/2019 «Инновации и продовольственная безопасность» № 2 (24)/2019 Рациональное природопользование Rational nature management Стадо сибирского хари- уса в контрольных уловах 2018 г. в озере Уч-Кол сформировано особями 1+ … 4+ лет. Длина тела рыб варьирует от 80,0 до 190,0 мм, а масса рыб от 7,0 до 119,0 г. Основу численности сибирского хариуса составляли особи в возрасте четырех лет со средней массой 64,5 г при средней длине 157,3 мм соот- ветственно. Соотношение самцов и самок в контрольных уловах составило 1: 2. Половозрелым сибирский ха- риус становится в трехлетнем возрасте. Абсолютная плодовитость хариуса составляла от 1678 (2+) до 3076 (4+) икринок и закономерно увеличивалась с возрастом. Оз. Камышевое характеризуется олиговидовым типом ихтиоценоза. Среди представителей ихтио- фауны доминирующее положение занимает алтайский осман Потанина (Oreoleuciscus potanini (Kessler, 1879)), серебряный карась (Carassius auratus (Linnaeus, 1758)) имеет второстепенное значение. На водоеме осуществляется исключительно спортивно-любительское рыболовство. Основной про- мысловый вид водоема – алтайский осман Потанина. В уловах представлены особи от 8+ до 27+ лет, преобладают особи 12+ и 13+ лет. Средняя промысловая длина особей в уловах составляет 192,9 мм, масса – 111,7 г. На водоеме осуществляется исключительно спортивно-любительское рыболовство. Основной про- мысловый вид водоема – алтайский осман Потанина. В уловах представлены особи от 8+ до 27+ лет, преобладают особи 12+ и 13+ лет. Средняя промысловая длина особей в уловах составляет 192,9 мм, масса – 111,7 г. 99 «Инновации и продовольственная безопасность» № 2 (24)/2019 БИБЛИОГРАФИЧЕСКИЙ СПИСОК ова Т. А., Демидова А. Г. Гидрология и гидрохимия. – М.: Пищ. пром-сть, 1977. – 312 с. 1. Берникова Т. А., Демидова А. Г. Гидрология и гидрохимия. – М.: Пищ. пром-сть, 1977. – 312 с. 2 Лучшева А А Практическая гидрометрия – Л : Гидрометеоиздат 1983 – 423 с 1. Берникова Т. А., Демидова А. Г. Гидрология и гидрохимия. – М.: Пищ. пром-сть, 1977. – 312 с. р Д р р р 2. Лучшева А. А. Практическая гидрометрия. – Л.: Гидрометеоиздат, 1983. – 423 с. 3. Алекин О. А. Основы гидрохимии. – Л.: Гидрометеоиздат, 1970. – 443 с. 3. Алекин О. А. Основы гидрохимии. – Л.: Гидрометеоиздат, 1970. – 443 с. 4. Методические указания по оценке численности рыб в пресноводных водоемах. – М.: ВНИИПРХ, 1990. –51 с. 4. Методические указания по оценке численности рыб в пресноводных водоемах. – М.: ВНИИПРХ 1990. –51 с. 5. Справочник по гидрохимии. – Л., 1989. – 391 с. р р 6. Гидрохимия: курс лекций / И. В. Морузи, Е. В. Пищенко, Л. В. Веснина, П. В. Белоусов. – Новосибирск: Новосиб. гос. аграр. ун-т, 2008. – 44 с. 6. Гидрохимия: курс лекций / И. В. Морузи, Е. В. Пищенко, Л. В. Веснина, П. В. Белоусов. – Новосибирск: Новосиб. гос. аграр. ун-т, 2008. – 44 с. 7. Никольский Г. В. Теория динамики стада рыб как биологическая основа рациональной эксплуа- тации и воспроизводства рыбных ресурсов. – М., 1974. – 375 с. 7. Никольский Г. В. Теория динамики стада рыб как биологическая основа рациональной эксплуа- тации и воспроизводства рыбных ресурсов. – М., 1974. – 375 с. 8. Методики изучения биогеоценозов внутренних водоёмов / под ред. Ф. Д. Мордухай- Болтовского. – М.: Наука, 1975. – 240 с. 8. Методики изучения биогеоценозов внутренних водоёмов / под ред. Ф. Д. Мордухай- Болтовского. – М.: Наука, 1975. – 240 с. 9. Методические рекомендации по сбору и обработке материалов при гидробиологических иссле- дованиях на пресноводных водоемах. Зоопланктон и его продукция. – Л.: ГосНИОРХ, 1983. – 24 с. 9. Методические рекомендации по сбору и обработке материалов при гидробиологических иссле- дованиях на пресноводных водоемах. Зоопланктон и его продукция. – Л.: ГосНИОРХ, 1983. – 24 с. 10. Методические рекомендации по сбору и обработке материалов при гидробиологических исследованиях на пресноводных водоемах. Зообентос и его продукция / под ред. Г. Г. Винберга, Г. М. Лаврентьевой. – Л.: Изд-во ГосНИОРХ, 1984. – 52 с. 10. Методические рекомендации по сбору и обработке материалов при гидробиологических исследованиях на пресноводных водоемах. Рациональное природопользование Rational nature management Вторым по численности является серебряный карась. В уловах отмечены особи 3 возрастных групп, причем преобладают особи 4+ и 5+ лет. Для вида характерна концентрация в юго-восточной мелководной части озера в зарослях высшей водной растительности. Таким образом, по степени развития кормовой базы оз. Уч-Кол можно отнести к водоемам с низкой кормностью по зоопланктону и зообентосу. Ихтиофауна водоема представлена сибирским хариусом. Озеро Камышевое по степени развития кормовой базы также можно отнести к водоемам с низ- кой кормностью по зоопланктону и зообентосу. Ихтиофауна водоема представлена алтайским османом Потанина и серебряный карасем. Промышленный лов рыбы на озере не ведется, рыбные запасы осва- иваются не в полной мере спортивно-любительским рыболовством. Озеро без названия, как и другие два, является водоемом с низкой кормностью по зоопланктону и зообентосу. БИБЛИОГРАФИЧЕСКИЙ СПИСОК – М.: Наука, 1974. – 209 с. 17. Методическое пособие по изучению питания и пищевых отношений рыб в естественных усло- виях. – М.: Наука, 1974. – 209 с. у 18. Чугунова Н. И. Руководство по изучению возраста и роста рыб. – М.: Изд-во АН СССР, 1959. – 164 с. 18. Чугунова Н. И. Руководство по изучению возраста и роста рыб. – М.: Изд-во АН СССР, 1959. – 164 с. БИБЛИОГРАФИЧЕСКИЙ СПИСОК Зообентос и его продукция / под ред. Г. Г. Винберга, Г. М. Лаврентьевой. – Л.: Изд-во ГосНИОРХ, 1984. – 52 с. р 11. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1994. – Т. 1. – 395 с. 11. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1994. – Т. 1. – 395 с. 12. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С Я Ц СПб ЗИН РАН 1997 Т 3 440 12. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1997. – Т. 3. – 440 с. 12. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1997. – Т. 3. – 440 с. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1997. – Т. 3. – 440 с. 13. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. 13. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1999. – Т. 4. – 999 с. 13. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1999. – Т. 4. – 999 с. С. Я. Цалолихина. – СПб.: ЗИН РАН, 1999. – Т. 4. – 999 с. 14. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С Я Цалолихина – СПб : Наука 2001 – Т 5 – 837 с 14. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: Наука, 2001. – Т. 5. – 837 с. 14. Определитель пресноводных беспозвоночных России и сопредельных территорий / под ред. С. Я. Цалолихина. – СПб.: Наука, 2001. – Т. 5. – 837 с. С. Я. Цалолихина. – СПб.: Наука, 2001. – Т. 5. – 837 с. 15. Правдин И. Ф. Руководство по изучению рыб (преимущественно пресноводных). – М.: Пищ. пром-сть, 1966. – 376 с. 15. Правдин И. Ф. Руководство по изучению рыб (преимущественно пресноводных). – М.: Пищ. пром-сть, 1966. – 376 с. р 16. Плохинский А. Н. Биометрия. – Новосибирск: СО АН СССР, 1961. – 364 с. 16. Плохинский А. Н. Биометрия. – Новосибирск: СО АН СССР, 1961. – 364 с. 17. Методическое пособие по изучению питания и пищевых отношений рыб в естественных усло- виях. REFERENCES 1. Bernikova T. A. Gidrologiya i gidrokhimiya (Hydrology and hydrochemistry), T. A. Bernikova, A. G. Demidova., Moscow: Pishchevaya promyshlennost», 1977, 312 p. 1. Bernikova T. A. Gidrologiya i gidrokhimiya (Hydrology and hydrochemistry), T. A. Bernikova, A. G. Demidova., Moscow: Pishchevaya promyshlennost», 1977, 312 p. 2. Luchsheva A. A. Prakticheskaya gidrometriya (Practical hydrometry), Leningrad: Gidrometeoizdat, 1983, 423 р. 2. Luchsheva A. A. Prakticheskaya gidrometriya (Practical hydrometry), Leningrad: Gidrometeoizdat, 1983, 423 р. «Инновации и продовольственная безопасность» № 2 (24)/2019 100 18. Chugunova N. I. Rukovodstvo po izucheniyu vozrasta i rosta ryb (Guide to the study of age and growth of fish), Moscow: AN SSSR, 1959, 164 p. y y ( ) р 17. Metodicheskoye posobiye po izucheniyu pitaniya i pishchevykh otnosheniy ryb v yestestvennykh usloviyakh (Methodical manual on the study of nutrition and nutritional relationships of fish in natural conditions), Moscow: Nauka, 1974, 209 p. y ( y )) y p y р 16. Plokhinskiy A. N. Biometriya (Biometrics), Novosibirsk: SO AN SSSR, 1961, 364 р. Рациональное природопользование Rational nature management 3. Alekin O. A. Osnovy gidrokhimii (Basics of hydrochemistry), Leningrad: Gidrometeoizdat, 1970, 443 р. 4. Metodicheskiye ukazaniya po otsenke chislennosti ryb v presnovodnykh vodoyemakh (Guidelines for estimating the number of fish in freshwater), Moscow: VNIIPRKH, 1990, 51 p. 3. Alekin O. A. Osnovy gidrokhimii (Basics of hydrochemistry), Leningrad: Gidrometeoizdat, 1970, 443 р. 4. Metodicheskiye ukazaniya po otsenke chislennosti ryb v presnovodnykh vodoyemakh (Guidelines for estimating the number of fish in freshwater), Moscow: VNIIPRKH, 1990, 51 p. 3. Alekin O. A. Osnovy gidrokhimii (Basics of hydrochemistry), Leningrad: Gidrometeoizdat, 1970, 443 р. 4. Metodicheskiye ukazaniya po otsenke chislennosti ryb v presnovodnykh vodoyemakh (Guidelines for estimating the number of fish in freshwater), Moscow: VNIIPRKH, 1990, 51 p. 5. Spravochnik po gidrokhimii (Handbook of hydrochemistry), Leningrad, 1989, 391 р. 6. Moruzi I. V., Pishchenko Ye.V., Vesnina L. V., Belousov P. V., Gidrokhimiya: kurs lektsiy, Novosibirsk: Novosibirskiy gosudarstvennyy agrarnyy universitet, 2008, 44 р. 6. Moruzi I. V., Pishchenko Ye.V., Vesnina L. V., Belousov P. V., Gidrokhimiya: kurs lektsiy, Novosibirsk: Novosibirskiy gosudarstvennyy agrarnyy universitet, 2008, 44 р. y g yy g yy р 7. Nikol’skiy G. V. Teoriya dinamiki stada ryb kak biologicheskaya osnova ratsional’noy ekspluatatsii i vosproizvodstva rybnykh resursov (The theory of fish stock dynamics as a biological basis for rational exploitation and reproduction of fish resources), Moscow, 1974, 375 p. 7. Nikol’skiy G. V. Teoriya dinamiki stada ryb kak biologicheskaya osnova ratsional’noy ekspluatatsii i vosproizvodstva rybnykh resursov (The theory of fish stock dynamics as a biological basis for rational exploitation and reproduction of fish resources), Moscow, 1974, 375 p. p p ) p 8. Mordukhay-Boltovskogo F.D Metodiki izucheniya biogeotsenozov vnutrennikh vodoyomov (Methods of studying the biogeocenoses of inland waters), Moscow: Nauka, 1975, 240 p. p p ) p 8. Mordukhay-Boltovskogo F.D Metodiki izucheniya biogeotsenozov vnutrennikh vodoyomov (Methods of studying the biogeocenoses of inland waters), Moscow: Nauka, 1975, 240 p. 9. Metodicheskiye rekomendatsii po sboru i obrabotke materialov pri gidrobiologicheskikh issledovaniyakh na presnovodnykh vodoyemakh. Zooplankton i yego produktsiya (Guidelines for the collection and processing of materials in hydrobiological studies on freshwater bodies. Zooplankton and its products), Leningrad: GosNIORKH, 1983, 24 p. 9. Metodicheskiye rekomendatsii po sboru i obrabotke materialov pri gidrobiologicheskikh issledovaniyakh na presnovodnykh vodoyemakh. Zooplankton i yego produktsiya (Guidelines for the collection and processing of materials in hydrobiological studies on freshwater bodies. Zooplankton and its products), Leningrad: GosNIORKH, 1983, 24 p. 10. Vinberg G. G., Lavrent’yeva G. M. Рациональное природопользование Rational nature management Metodicheskiye rekomendatsii po sboru i obrabotke materialov pri gidrobiologicheskikh issledovaniyakh na presnovodnykh vodoyemakh. Zoobentos i yego produktsiya (Guidelines for the collection and processing of materials in hydrobiological studies on freshwater bodies. Zoobenthos and its products), Leningrad: GosNIORKH, 1984, 52 p. 10. Vinberg G. G., Lavrent’yeva G. M. Metodicheskiye rekomendatsii po sboru i obrabotke materialov pri gidrobiologicheskikh issledovaniyakh na presnovodnykh vodoyemakh. Zoobentos i yego produktsiya (Guidelines for the collection and processing of materials in hydrobiological studies on freshwater bodies. Zoobenthos and its products), Leningrad: GosNIORKH, 1984, 52 p. p ) g p 11. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: ZIN RAN, 1994, T. 1, 395 р. 11. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: ZIN RAN, 1994, T. 1, 395 р. 12. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: ZIN RAN, 1997, T. 3, 440 р. 12. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: ZIN RAN, 1997, T. 3, 440 р. 13. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: ZIN RAN, 1999, T. 4, 999 р. 14. Tsalolikhin S.YA. Opredelitel» presnovodnykh bespozvonochnykh Rossii i sopredel’nykh territoriy (Key to freshwater invertebrates in Russia and adjacent territories), St. Petersburg: Nauka, 2001, T. 5, 837 р. 15. Pravdin I. F. Rukovodstvo po izucheniyu ryb (preimushchestvenno presnovodnykh) (Guidelines for the study of fish (mostly freshwater)) Moscow: Pishchevaya promyshlennost», 1966, 376 р. y ( y )) y p y р Plokhinskiy A. N. Biometriya (Biometrics), Novosibirsk: SO AN SSSR, 1961, 364 р. y y ( ) р 17. Metodicheskoye posobiye po izucheniyu pitaniya i pishchevykh otnosheniy ryb v yestestvennykh usloviyakh (Methodical manual on the study of nutrition and nutritional relationships of fish in natural conditions), Moscow: Nauka, 1974, 209 p. 101 «Инновации и продовольственная безопасность» № 2 (24)/2019
https://openalex.org/W2148512680	https://europepmc.org/articles/pmc3661395?pdf=render	English	null	Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye	BioMedical engineering online	2,013	cc-by	10,057	Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye Robert Koprowski1, Zygmunt Wróbel1, Sławomir Wilczyński2, Anna Nowińska3 and Edward Wy Correspondence: robert. koprowski@us.edu.pl 1Department of Biomedical Computer Systems, Institute of Computer Science, University of Silesia, Będzińska 39 Str, Sosnowiec 41-200, Poland Full list of author information is available at the end of the article * Correspondence: robert. koprowski@us.edu.pl 1Department of Biomedical Computer Systems, Institute of Computer Science, University of Silesia, Będzińska 39 Str, Sosnowiec 41-200, Poland Full list of author information is available at the end of the article RESEARCH Open Access © 2013 Koprowski et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Keywords: Eye, Image processing, Iridocorneal angle, OCT Keywords: Eye, Image processing, Iridocorneal angle, OCT Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye Koprowski et al. Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Abstract Introduction: This paper presents the problem of automatic measurement of the iridocorneal angle in tomographic images of the anterior segment of the eye. It includes the results of the comparison of well-known methods for measuring the iridocorneal angle with new methods, proposed in this paper. All these methods concern tomographic image analysis and processing. Full list of author information is available at the end of the article Material and method: In total, approximately 100’000 tomographic images (from about 6’000 patients) were analysed. They were obtained using two devices: SOCT Copernicus (Optopol Tech. SA, Zawiercie, Poland) and Visante OCT (Carl Zeiss Meditec, Inc, Dublin, California, USA). The patients, aged 12 to 78 years with varying degrees of the iridocorneal angle pathology, were from the region of Silesia, Poland. The images were in DICOM or RAW formats and analysed in the software developed by the authors for the purposes of this study. Results: The results indicate that the measurement method proposed by the authors, which is based on the calculation of the minimum distance between the iris and the cornea in the adopted area, is the most accurate. For this method sensitivity was 0.88, specificity 0.89 and the area under the Receiver Operating Characteristic curve (AUC) was 0.88. The other known methods for measuring the iridocorneal angle gave worse results, that is, for example, for the measurement of the distance between the iris and the cornea AUC = 0.87, sensitivity = 0.86 and specificity = 0.71. For another well-known method of measuring the iridocorneal angle AUC = 0.77, sensitivity = 0.82 and specificity = 0.61. Conclusions: The study proved that the proposed method of measuring the minimum distance between the iris and the cornea within the adopted area is the most effective in the classification of the iridocorneal angle in patients with a high degree of pathology of all the compared measurement methods based on tomographic images. However, it requires fully automated measurement. Introduction This paper compares well-known methods for the iridocorneal angle analysis in tomo- graphic images of the anterior segment of the eye with the new ones, suggested by the authors. The iridocorneal angle is the structure responsible for the outflow of aqueous humor from the anterior chamber of the eye. Normal intraocular pressure is determined by the production of aqueous humor by the ciliary epithelium and the rate of humor Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 2 of 16 Page 2 of 16 outflow via two pathways – the trabecular meshwork and the uveoscleral pathway [1,2]. Anatomical anomalies such as the angle narrowing or closure result in impeded outflow and increased intraocular pressure. The iridocorneal angle is situated on the circumference of the anterior chamber be- tween the sides of the cornea and sclera and the base of the iris and the anterior sur- face of the ciliary body. Currently, a primary diagnostic tool and technique that enables the analysis of the angle structures is gonioscopy which uses contact gonio lenses [3]. The angle is also studied with OCT devices and the technique is called automatic gonioscopy. The main advantage of this method is its non-invasiveness, whereas the advantage of classical gonioscopy is the ability to visualize pathological structures, such as neovascularization or hyperpigmentation of the weave of the trabecular meshwork [4]. The assessment of the iridocorneal angle using OCT Visante involves morpho- logical assessment and a morphometric analysis of the angle parameters. Morphometric measurements can be carried out using the measuring tool "caliper" included in Visante OCT commercial software. The tool is intended to assist manual operation – Figure 1. An operator typically indicates a point on the scleral spur in a OCT image (marked with a white circle in Figure 1) whereas a device adjusts the position of the other points (marked with red circles – in Figure 1). In order to perform these morphometric mea- surements, it is necessary to manually locate the scleral spur, which is a landmark for determining morphometric measurements, in the scan "ASS". In practice, it often hap- pens that the scleral spur is not visible. According to the results given in paper [5], such a situation occurs in approximately 20% of cases. Introduction c) TISA 500 (Trabecular-Iris Space Area) involves measuring an area covering 500 μm located in the area bounded by the cornea and the iris. Figure 2 Methods of measuring the iridocorneal angle. a) AOD 500 (Angel Opening Distance) involves measuring a distance between a point of the cornea which is 500 μm away from the scleral spur and the opposite point of the iris. b) TIA (Trabecular-Iris Angle) involves a direct measurement of the angle. c) TISA 500 (Trabecular-Iris Space Area) involves measuring an area covering 500 μm located in the area bounded by the cornea and the iris. Figure 2 Methods of measuring the iridocorneal angle. a) AOD 500 (Angel Opening Distance) involves measuring a distance between a point of the cornea which is 500 μm away from the scleral spur and the opposite point of the iris. b) TIA (Trabecular-Iris Angle) involves a direct measurement of the angle. c) TISA 500 (Trabecular-Iris Space Area) involves measuring an area covering 500 μm located in the area bounded by the cornea and the iris. with a suitable corneal contour point. A division of contours is performed based on the point of the greatest curvature (point marked in black, the starting point of the coord- inate system – Figure 3). In this way, a sequence of measurements at various distances from the apex of the measured angle is obtained – the chart shown in Figure 3. This method, referred to as AOS (Angle Opening Sequence), ensures obtaining much more information on the iridocorneal angle when compared to AOD, TIA and TISA. The varying degree of the iridocorneal angle pathology visible in Figure 4 (especially narrow or closed iridocorneal angle), which is difficult to measure with conventional methods such as TIA, TISA and AOD, is successfully and reliably evaluated using AOS [8]. It is apparent from the examples shown in Figure 4 that a difficulty in a reliable assessment of the iridocorneal angle with the AOD, TIA and TISA methods lies primarily in a large extent of pathology – distorted sides of the angle, especially of the iris angle. Therefore, in pathological conditions, measurement results of TISA and AOD are strictly dependent on measurement locations. Slightly less sensitive to this type of pathology is the TISA method. Introduction The known methods TIA, AOD 500, AOD 750, TISA 500, TISA 750 can be defined on the basis of these morphometric mea- surements – Figure 2 [6] and Table 1. These methods are described in detail in papers [1-3,7]. The values "500" and "750" refer to the distance, expressed in microns, from the scleral spur. As mentioned above, on the basis of the scleral spur location point indicated by an operator, the other char- acteristic points are drawn automatically. They are necessary to perform calculations for the presented methods, namely TIA, TISA and AOD. Undoubtedly, this semi- automatic method facilitates operator's work but still it is not done fully automatically. Fully automatic measurement was suggested by the authors in 2011 [7]. It involves the use of information about iridocorneal contours. Each iris contour point is combined Figure 1 Tomographic image of the anterior segment of the eye and examples of commercial software operation provided with the Visante OCT device. In the illustrated case, an operator indicates one point for the scleral spur position (highlighted with a white circle) and the machine adjusts its position and draws the location of consecutive points marked with red circles. The results for the analysed case are AOD = 393 μm, TISA = 0.152 mm2, TIA = 38.2°. Figure 1 Tomographic image of the anterior segment of the eye and examples of commercial software operation provided with the Visante OCT device. In the illustrated case, an operator indicates one point for the scleral spur position (highlighted with a white circle) and the machine adjusts its position and draws the location of consecutive points marked with red circles. The results for the analysed case are AOD = 393 μm, TISA = 0.152 mm2, TIA = 38.2°. Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Page 3 of 16 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Page 3 of 16 Page 3 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 p g g http://www.biomedical-engineering-online.com/content/12/1/40 Figure 2 Methods of measuring the iridocorneal angle. a) AOD 500 (Angel Opening Distance) involves measuring a distance between a point of the cornea which is 500 μm away from the scleral spur and the opposite point of the iris. b) TIA (Trabecular-Iris Angle) involves a direct measurement of the angle. Introduction However, its practical application is limited due to the area unit (μm2) which is less intuitive and difficult to quickly compare with the other methods, TIA and AOD. In practice, the described methods for measuring the iridocorneal angle, namely TIA, TISA and AOD, have a number of inconsistencies and irregularities in the interpretation of results. As a consequence, results are not repeatable, reliable and diffi- cult to verify and compare with the model and other doctors’ results. The situation be- comes critical when the progress of treatment or the disease progression of a patient diagnosed by different doctors in different medical centres equipped with different types of OCT devices needs to be assessed. Archiving results is an important issue in practical measurements. In the case of TIA, TISA and AOD, each measurement is connected with archiving one scalar value, e.g. the angle value measured with TIA with a typical accuracy of one decimal place. In the Table 1 Known methods for measuring the iridocorneal angle and their definitions Method symbol Method name Definition AOD Angle Opening Distance (Figure 2a) involves measuring a distance between a point of the cornea which is 500 μm away from the scleral spur and the opposite point of the iris TIA Trabecular-Iris Angle (Figure 2b) involves a direct measurement of the angle TISA Trabecular-Iris Space Area (Figure 2c) involves measuring an area covering 500 μm located in the area bounded by the cornea and the iris Table 1 Known methods for measuring the iridocorneal angle and their definitions Method symbol Method name Definition AOD Angle Opening Distance (Figure 2a) involves measuring a distance between a point of the cornea which is 500 μm away from the scleral spur and the opposite point of the iris TIA Trabecular-Iris Angle (Figure 2b) involves a direct measurement of the angle TISA Trabecular-Iris Space Area (Figure 2c) involves measuring an area covering 500 μm located in the area bounded by the cornea and the iris Table 1 Known methods for measuring the iridocorneal angle and their definitions Method symbol Method name Definition Page 4 of 16 Page 4 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 http://www.biomedical-engineering-online.com/content/12/1/40 Figure 3 Measuring principle of the iridocorneal angle with AOS (Angel Opening Sequence) suggested by the authors in 2011. Material In the study, about 100’000 tomographic images (from about 6’000 patients) were ex- amined. The images were acquired using the following devices: SOCT Copernicus (Optopol Tech. SA, Zawiercie, Poland) and Visante OCT (Carl Zeiss Meditec, Inc, Dublin, California, USA). The patients, aged 12 to 78 years with varying degrees of the iridocorneal angle pathology, were from the region of Silesia, Poland. The obtained im- ages were in DICOM or RAW formats with a resolution of 256 × 1024 pixels, within a measuring range of 8 mm × 16 mm, which gives 31.3 μm/pixel. The image analysis was carried out in a Matlab software package with Image and Signal Processing toolboxes and the code was optimized in the C programming language. Introduction BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Table 2 Alphabet for the iridocorneal angle description proposed by the authors in 2011 Symbols Function / increasing distance for successive values on the axis 0x (Figure 3) ^ local minimum v local maximum _ constant distance value for increasing values of 0x Numerical parameters angular value, maximum, minimum or fixed distance for specific 0x range of values on the axis 0x in which a given situation occurs Table 2 Alphabet for the iridocorneal angle description proposed by the authors in 2011 Symbols Function Table 2 Alphabet for the iridocorneal angle description proposed by the authors in 2011 Symbols Function / increasing distance for successive values on the axis 0x (Figure 3) ^ local minimum v local maximum _ constant distance value for increasing values of 0x Numerical parameters angular value, maximum, minimum or fixed distance for specific 0x range of values on the axis 0x in which a given situation occurs conventional, previously known, methods (TIA, TISA, AOD). As a result, medical errors are more likely due to improper recording. Therefore, the authors suggested a new method, modified in relation to AOS, for measuring the iridocorneal angle, namely AOM (Angel Opening Minimum) which is described below. Introduction Each iris contour point (in red) is connected to a suitable corneal contour point (in yellow). A division into the appropriate contour is performed based on the point in the largest curvature (shown in black). In this way, a sequence of measurements at various distances from the apex of the measured angle is obtained. Figure 3 Measuring principle of the iridocorneal angle with AOS (Angel Opening Sequence) suggested by the authors in 2011. Each iris contour point (in red) is connected to a suitable corneal contour point (in yellow). A division into the appropriate contour is performed based on the point in the largest curvature (shown in black). In this way, a sequence of measurements at various distances from the apex of the measured angle is obtained. Figure 3 Measuring principle of the iridocorneal angle with AOS (Angel Opening Sequence) suggested by the authors in 2011. Each iris contour point (in red) is connected to a suitable corneal contour point (in yellow). A division into the appropriate contour is performed based on the point in the largest curvature (shown in black). In this way, a sequence of measurements at various distances from the apex of the measured angle is obtained. case of AOS, a data vector of successive measurements of iridocorneal contour distances is archived. In paper [7] in 2011, the authors presented a new way of recording the results obtained with the AOS method, which included the alphabet shown in Table 2. The alphabet has been adopted in clinical practice. However, its practical application is somewhat troublesome due to more complicated recording in comparison to Figure 4 Various degrees of the iridocorneal angle pathology with marked measurement points for AOS. The presented cases a), b) and c) are not correctly measured with TIA, TISA and AOD. This is due to both a visible pathology (a variable iris shape) as well as the definition of TIA, TISA and AOD. Figure 4 Various degrees of the iridocorneal angle pathology with marked measurement points for AOS. The presented cases a), b) and c) are not correctly measured with TIA, TISA and AOD. This is due to both a visible pathology (a variable iris shape) as well as the definition of TIA, TISA and AOD. Page 5 of 16 Page 5 of 16 Koprowski et al. AOM method The AOM (Angle Opening Minimum) method involves determining a minimum dis- tance, which is one scalar value, between the contours of the iris and cornea. To be more exact, it is the distance between different points of the cornea and the nearest point of the iris. The measured area of the cornea covers a range of 500 μm or 750 μm starting from the scleral spur. Details on the principle of the AOM method of measure- ment are shown in Figure 5a. Figure 5b, on the other hand, shows an opposite situ- ation, where a minimum distance between various points of the iris and the nearest point of the cornea is calculated – the AOM2 method. The AOM method is based on a sequence of calculations of the minimum angle values for individual pixels of the cornea edge from all the pixels of the iris edge (or the other way round in AOM2 – Figure 5). Measurements can be carried out in the entire range for each of the contour points of the iris or cornea (as in Figure 5c and 5d) but only the contour points which are realized in the range of 500 μm of the corneal contour are relevant. This narrowing of the measure- ment range usually takes place in the last stage of calculations. The obtained results differ depending on the measurement method (AOS, AOM or AOM2 – Table 3). Examples of a sequence of distance measurements are shown in Figure 6. The graph shows changes in values of the angle opening for successive pixels of Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Page 6 of 16 http://www.biomedical-engineering-online.com/content/12/1/40 Page 6 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Figure 5 The conception of AOM (Angle Opening Minimum) and AOM2 methods and the obtained results. a) shows how the shortest distance between all the points of the cornea and one of the sample points of the iris is chosen in the AOM method. The shortest distance for both variants for one analysed pixel is highlighted in bold arrow. b) shows how the shortest distance between all the points of the iris and one of the sample points of the cornea is chosen in the AOM2 method. AOM method c) and d) show practically obtained results for all pixels in the implementation of the two variants: the AOM method shown in a) and the AOM2 method shown in b). In order to better visualize the results, the analysis was not limited to the appropriate range of 500 μm. he conception of AOM (Angle Opening Minimum) and AOM2 methods and the obtained Figure 5 The conception of AOM (Angle Opening Minimum) and AOM2 methods and the obtained results. a) shows how the shortest distance between all the points of the cornea and one of the sample points of the iris is chosen in the AOM method. The shortest distance for both variants for one analysed pixel is highlighted in bold arrow. b) shows how the shortest distance between all the points of the iris and one of the sample points of the cornea is chosen in the AOM2 method. c) and d) show practically obtained results for all pixels in the implementation of the two variants: the AOM method shown in a) and the AOM2 method shown in b). In order to better visualize the results, the analysis was not limited to the appropriate range of 500 μm. the contour of the cornea or iris, respectively, for the images in Figure 5c and 5d. The lar- gest visible difference is between the AOM method and the other methods, namely AOM2 and AOS. The difference is due to different measurement methods and amounts to approximately 10 pixels (10 pixels * 31.3 μm/pixel = 313 μm) in the range of 500 μm which is marked in yellow (Figure 6). In practice, due to a greatly variable shape of the contour of the iris in comparison with the contour of the cornea, the AOM method pro- duces underestimated measurement values. However, they are more reliable when com- pared to the AOS or AOM2 methods. They enable to take into account the iridocorneal angle pathology to a greater extent. Such situations concern the narrowing of the iridocorneal angle in the area outside the range of 500 μm when accurate (consistent with the definition) calculations with the methods TISA, TIA and AOD do not produce satisfac- tory results. AOM method The area highlighted in yellow includes the range of 500 μm measured from the scleral spur (16 pixels ⋅31.3 μm/pixel ≅500 μm). Figure 6 Graph of changes in values of the angle opening for successive pixels of the contour of the cornea or iris, respectively (depending on the measurement method) shown in Figure 5 c) and d). The graph shows the biggest difference between the AOM method and the other methods (AOM2 and AOS). Visible differences are due to different measurement methods described in the paper. The area highlighted in yellow includes the range of 500 μm measured from the scleral spur (16 pixels ⋅31.3 μm/pixel ≅500 μm). 500 μm starting from the scleral spur. The above mentioned range of 500 μm of the corneal contour (16 pixels fall into this range for the analysed image resolution) and all the pixels in the contour of the iris are taken into account in the measurements. AOM method The result of the measurement of the iridocorneal angle with the AOM method is therefore a minimum distance between the iris and the cornea measured in the area of Table 3 Methods for measuring the iridocorneal angle proposed by the authors and their definitions Method symbol Method name Definition AOS Angle Opening Sequence (Figure 3) each iris contour point is combined with a suitable corneal contour point AOM Angle Opening Minimum (Figure 5a) shortest distance between all the points of the cornea and selected range of the iris AOM2 Angle Opening Minimum 2 (Figure 5b) shortest distance between all the points of the iris and selected range of the cornea Table 3 Methods for measuring the iridocorneal angle proposed by the authors and their definitions Method symbol Method name Definition AOS Angle Opening Sequence (Figure 3) each iris contour point is combined with a suitable corneal contour point AOM Angle Opening Minimum (Figure 5a) shortest distance between all the points of the cornea and selected range of the iris AOM2 Angle Opening Minimum 2 (Figure 5b) shortest distance between all the points of the iris and selected range of the cornea Table 3 Methods for measuring the iridocorneal angle proposed by the authors and their definitions Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Page 7 of 16 Page 7 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 http://www.biomedical-engineering-online.com/content/12/1/40 Figure 6 Graph of changes in values of the angle opening for successive pixels of the contour of the cornea or iris, respectively (depending on the measurement method) shown in Figure 5 c) and d). The graph shows the biggest difference between the AOM method and the other methods (AOM2 and AOS). Visible differences are due to different measurement methods described in the paper. The area highlighted in yellow includes the range of 500 μm measured from the scleral spur (16 pixels ⋅31.3 μm/pixel ≅500 μm). Figure 6 Graph of changes in values of the angle opening for successive pixels of the contour of the cornea or iris, respectively (depending on the measurement method) shown in Figure 5 c) and d). The graph shows the biggest difference between the AOM method and the other methods (AOM2 and AOS). Visible differences are due to different measurement methods described in the paper. Implementation of the AOM method The AOM method requires full automation of image analysis. This automation enables to detect the cornea and iris edges and then perform adequate calculations of the minima (according to the methodology of AOM described above). The image analysis algorithm, in particular, should include the following elements [9-12] shown in Figure 7. The need to use a profiled algorithm in this case is connected with inadequate results obtained with other known algorithms for detecting lines and/or areas in the image. Hough’s transform [7] enables to detect lines in images of a pre-selected shape. However, the results in the case of large inter-individual variability are not satisfactory. Hough’s transform [7] enables to detect lines in images of a pre-selected shape. Houghs transform [7] enables to detect lines in images of a pre-selected shape. However, the results in the case of large inter-individual variability are not satisfactory. However, the results in the case of large inter-individual variability are not satisfactory. wavelet analysis method [13] gives incorrect results when objects are hard to see and the lines overlap – such situations are quite common in the case of the analysed images, analysis methods of elongated objects cannot be applied here due to a possibility of large changes in the size of both the object itself and its thickness and a possibility of its division into multiple parts (e.g. the iris or cornea). object recognition methods in cases of large pathology can give unpredictable results. object recognition methods in cases of large pathology can give unpredictable results. other known methods, for example, texture analysis [14], also do not produce satisfactory results. other known methods, for example, texture analysis [14], also do not produce satisfactory results. Based on this and the literature review [8,15-18] and given the medical evidence presented below, a profiled algorithm for the analysis and processing of images of the front part of the eye was suggested. Page 8 of 16 Page 8 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 http://www.biomedical-engineering-online.com/content/12/1/40 Figure 7 Block diagram of the tomographic image analysis algorithm in the AOM method. The presented algorithm enables fully automatic measurement of the iridocorneal angle. The obtained result is in the form of one scalar value which is the minimum distance between the selected ranges of the iris and cornea. Implementation of the AOM method The presented algorithm is versatile and provides correct results for any tomographic images of the eye. Figure 7 Block diagram of the tomographic image analysis algorithm in the AOM method. The presented algorithm enables fully automatic measurement of the iridocorneal angle. The obtained result is in the form of one scalar value which is the minimum distance between the selected ranges of the iris and cornea. The presented algorithm is versatile and provides correct results for any tomographic images of the eye In the implemented algorithm, an input image with the resolution of 256 × 1024 pixels, mentioned in the introduction, is entered into the developed software in DICOM format. Then filtration with a median filter (with a 3 × 3 pixel mask) is carried out followed by the analysis of each column. As a result of this analysis, a binarization threshold is calculated for each column (Otsu method [19]). The created binary image is shown in Figure 8b). In the next stage, the method of filling the holes is applied in Figure 8 Some stages of processing in the calculation of the iridocorneal angle with the AOM method: a) the sclera and cornea boundaries (in green and red) set automatically, b) approximation of the automatically set sclera boundary – in blue and yellow, c) the iridocorneal angle set automatically with the AOM method and its enlargement d). According to the definition, the iridocorneal angle calculated with the proposed AOM method is a minimum distance calculated between the cornea pixels (the area of 500 μm) and all the pixels of the iris boundary. It is marked with a bold yellow line – d). Figure 8 Some stages of processing in the calculation of the iridocorneal angle with the AOM method: a) the sclera and cornea boundaries (in green and red) set automatically, b) approximation of the automatically set sclera boundary – in blue and yellow, c) the iridocorneal angle set automatically with the AOM method and its enlargement d). According to the definition, the iridocorneal angle calculated with the proposed AOM method is a minimum distance calculated between the cornea pixels (the area of 500 μm) and all the pixels of the iris boundary. It is marked with a bold yellow line – d). Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. Implementation of the AOM method BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 9 of 16 Page 9 of 16 order to eliminate minor inclusions or detachment. In this pre-prepared image, the sclera boundaries are determined followed by approximation of the boundaries with a polynomial of degree 4 (Figure 8b). After the analyses of the iris, ciliary appendages and scleral spur, there follows the analysis of iris termination points which uses informa- tion from the inside of the sclera boundary (Figure 8c). At this stage, the iridocorneal angle can be determined using the TIA, TISA and AOD methods. The value of the iridocorneal angle measured with the AOM method (or AOM2) requires a calculation of Euclidean distance for each possible pair of points (iris-cornea). The result is a minimum distance between them calculated in the range of 500 μm, highlighted in Figure 8d in bold yellow. This algorithm correctly identifies and calculates the iridocorneal angle not only with the AOM method but also with the other methods, namely AOS, AOM2 or TIA, TISA and AOD at 500 and 750 μm from the scleral spur. All the algorithm parameters are calculated automatically. The algorithm automatically adjusts to the type and brightness of a derived image (pre-processing, filtering with a median filter with a 5 × 5 pixel mask or normalization). During the programme installation, an operator only gives a distance attributable to the pixel (strictly dependent on the type of the used tomographic cam- era) if it is not specified in the DICOM header. This information is necessary to cali- brate the measured values of the iridocorneal angle (in the analysed images obtained with Visante OCT it is 31.3 μm/pixel). Accuracy of the measurements and the comparison of the results obtained with the above methods (AOD, TIA, TISA, AOS, AOM and AOM2) are presented below. Comparison of AOD, TIA, TISA, AOS, AOM and AOM2 methods in practice Comparison of AOD, TIA, TISA, AOS, AOM and AOM2 methods in practice A practical application of the measurement methods, namely AOD, TIA, TISA, AOS, AOM and AOM2, requires designation of the contours or a single point of the cornea and iris. In addition, it is necessary to determine the scleral spur location. The currently available software for OCT devices does not enable the mentioned fully automatic measurement. Calculations are carried out either manually or semi-automatically. For this reason, the results are not repeatable – they depend on an individual choice of the scleral spur (iridocorneal angle) location by an operator. The presented algorithm en- ables fully automatic measurement with the above mentioned methods. Therefore, their comparison is possible. Of all the mentioned methods (AOD, TIA, TISA, AOS, AOM and AOM2), only AOD, TIA, AOM2 and AOM are further taken into account. The TISA method is rarely used in practice due to the unit of the iridocorneal angle, that is, μm2 (mm2). The AOS method is also difficult to compare with other results because of a difficulty in com- parison of an AOS data sequence (or the mentioned alphabet) with scalar values obtained as results from other methods. The other methods, AOD, TIA, AOM2 and AOM, can be compared using one of two ways. One of them uses a model (an artificial image) with a known and measured iridocorneal angle and the other one compares results with those obtained by ophthal- mology experts. The first way, which will not be used here, is to use an artificial image containing known values of the iridocorneal angle realized with resolution error preci- sion at the time of creating the image. The other way is to use the results obtained from an assessment performed by an ophthalmology expert or from other more Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 10 of 16 Page 10 of 16 accurate measurement methods (if they exist). Both ways have their advantages and dis- advantages. The disadvantage of the first method is a difficulty in close conjunction of artificial images with real images which reflect a full range of variability. Its advantage is an ability to quickly compare the results and no need for the presence of an ophthal- mologist. Comparison of AOD, TIA, TISA, AOS, AOM and AOM2 methods in practice In the other method, results are obtained from a real image, but it requires a tedious procedure that involves manual marking (by an operator) of individual values of AOD, TIA, AOM2 and AOM. This method is also highly dependent on the observer’s subjective judgment which is mainly related to the determination of significance of de- tails in an image. As a result, the comparison of the methods AOD, TIA, AOM2 and AOM was carried out in two stages. Comparison with the results obtained by an expert In the first stage, an ophthalmologist manually marked the iridocorneal angle for the first 100 images with the AOD, TIA, AOM2 and AOM methods. In this case, a mea- surement error δq was calculated. It was defined as: δq ¼ wM−wP wP :100% ð1Þ δq ¼ wM−wP wP :100% ð1Þ where: where: wM – the measured value, wM – the measured value, wM – the measured value, t l d ith t th d ( g ll ) wM – the measured value, wP – a correct value measured with more accurate methods (e.g. manually) or an wP – a correct value measured with more accurate methods (e.g. manually) or an average value of a series of measurements, q – an index indicating the measurement method: AOD, TIA, AOM2 or AOM. The obtained values of δq are not greater than 5% for the AOD and TIA methods (δAOD = 4.5%, δTIA = 4.8%) and for AOM and AOM2, they are δAOM = 7.2% and δAOM2 = 9%, respectively. Therefore, it seems that the presented AOM and AOM2 methods are worse than the known methods. The reason for larger error values for these methods is the principle of measurement. It is much harder to manually select iris contour points located closest to the analysed corneal contour pixel (Figure 5) and then determine, of all the points, which of them is the smallest. Thus, it is necessary to compare the methods in terms of reliability of the iridocorneal angle calculation and detection of pathological cases. This methodology is described below. Comparison with the results obtained with other known methods In the second stage, the comparison of the AOD, TIA, AOM and AOM2 methods for all the obtained images (approximately 100’000) was performed by standardization of the measurement unit to relative values. This comparison was related to the effective- ness of detection of cases with normal and narrow iridocorneal angle, including patho- logical conditions. In order to compare the results, the following typical measures were used: accuracy ACC=(TP+TN)/(TP+TN+FP+FN), specificity SPC=TN/(FP+TN) and sensitivity TPR=TP/(TP+FN) where: TP – true positive, TN – true negative, FN – false negative and FP – false positive. These measures were used to evaluate the AOD, TIA, AOM2 and AOM methods. All cases where the iris or cornea was not visible were pre-eliminated. Such situations occurred when OCT imaging was not focused on calculating the iridocorneal angle. Page 11 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 After removing such cases, 83'574 images remained for further analysis. It was carried out in the steps presented below. Removal of all atypical cases for which the algorithm designated too short or too long iris edge Atypical cases are those for which the iris edge designation contains less than 16 pixels (500 μm, Figure 9a – in blue) or more than 70 to 80 pixels, which is equivalent to 2.5 mm (Figure 9a – in yellow). These limitations are natural and result from the presented methodology of measurement of AOD, TIA, AOM2 and AOM. The distance from the scleral spur boundary used in the measurement is 500 μm, which corresponds to a minimum of 16 pixels. This limitation (16 pixels) arises only from the need to compare these methods, and smaller values (<16) do not indicate errors in the algo- rithm operation. There are 48'022 images of this type whose length of the iris contour is below 16 pixels. This represents 57.5% of the analysed images. Values above approxi- mately 70 to 80 pixels of the iris edge length do not occur in medicine; therefore, they indicate an error in the algorithm operation. As is apparent from the graph showing the iris contour length for individual images, a number of such cases (exceeding 70 pixels) is relatively low (3'458, which is 4.1% of all cases). In total, 32’094 images remained for further analysis after applying these two limitations. Comparison of the results obtained for the proposed AOM method and the AOD method Comparison of the results obtained for the proposed AOM method and the AOD method For this purpose, ΔAOD,AOM was calculated as a difference between AOD and AOM. The results are shown in Figure 9b. The graph (Figure 9b) was divided in two parts. The changes in the value of ΔAOD,AOM highlighted in green when compared to the AOM and AOD methods indicate undervaluation of the iridocorneal angle measure- ment for the AOM method. This situation is consistent with the definition according to which in the case of the AOM method, it is the smallest distance between the corre- sponding points in the iris and cornea that is sought. The part of the graph marked in yellow (Figure 9b) covers a range of pathological cases for which a direct comparison Figure 9 Graph of changes in the iris length a) and the graph of changes in the value of ΔAOD,AOM b) for the sorted (from the smallest value) images. The area marked in blue in the graph a) covers the properly set iris contours but due to their length (<16 pixels, 500 μm) they cannot be used for further analysis. The acceptable values of the iris length are highlighted in green and the incorrectly designated iris contours in yellow. On the graph b), properly set values of the difference ΔAOD,AOM are marked in green and the algorithm errors are in yellow. Figure 9 Graph of changes in the iris length a) and the graph of changes in the value of ΔAOD,AOM b) for the sorted (from the smallest value) images. The area marked in blue in the graph a) covers the properly set iris contours but due to their length (<16 pixels, 500 μm) they cannot be used for further analysis. The acceptable values of the iris length are highlighted in green and the incorrectly designated iris contours in yellow. On the graph b), properly set values of the difference ΔAOD,AOM are marked in green and the algorithm errors are in yellow. Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 12 of 16 Page 12 of 16 of AOM and AOD was not possible (incorrect location of the area of analysis for the AOD method or incorrect indication of a minimum in the AOM method). Comparison of the results obtained for the proposed AOM method and the AOD method The graph of ΔAOD,AOM as a function of a sequence of images was divided in two parts (Figure 9b): with differences between AOD and AOM arising from their definition (in practice to about 22 pixels) and, to about 22 pixels) and, with differences between AOD and AOM caused by the algorithm errors. with differences between AOD and AOM caused by the algorithm errors. The division between the two areas (the threshold value of 22 pixels) was verified in practice and included a manual analysis of borderline cases. This value also means that in carrying out a calculation for the afore mentioned group of images, the difference in the results for AOD and AOM methods will not be greater than 22 pixels for the most obtuse iridocorneal angle. As it turns out, in practice, the AOM method enables a proper evaluation of specific cases of the iridocorneal angle. These examples are shown in Figure 10. Thus it is pos- sible to obtain correct classification of images with a proper, open iridocorneal angle from pathological cases with the risk of iridocorneal angle closure. Figure 10 Sample image fragments of the iridocorneal angle for which the known AOD, TIA and TISA methods give incorrect results. For these examples, the iridocorneal angle is determined properly with AOM and AOD methods. The results obtained with the AOM method are marked with a bold yellow line, and the results obtained for the AOD method are in red. The other coloured lines refer to the cornea and iris contours (yellow, magenta), and sides of the angle of the TIA method (sides are shown as a linear interpolation of the iris and cornea contours) are marked in white. The results obtained for the examples are: a) AOD = 12.9, AOM = 7.8, TIA = 29.7 b) AOD = 15.6, AOM = 9.2, TIA = 27.1 c) AOD = 22.6, AOM = 6.5, TIA = 29 d) AOD = 26, AOM = 16, TIA = 24.5 e) AOD = 21.6, AOM = 15.9, TIA = 22.6 f) AOD = 30.2, AOM = 15.2, TIA = 29.8 (AOM and AOD values are given in pixels, TIA values in degrees). Figure 10 Sample image fragments of the iridocorneal angle for which the known AOD, TIA and TISA methods give incorrect results. For these examples, the iridocorneal angle is determined properly with AOM and AOD methods. Comparison of the results obtained for the proposed AOM method and the AOD method The results obtained with the AOM method are marked with a bold yellow line, and the results obtained for the AOD method are in red. The other coloured lines refer to the cornea and iris contours (yellow, magenta), and sides of the angle of the TIA method (sides are shown as a linear interpolation of the iris and cornea contours) are marked in white. The results obtained for the examples are: a) AOD = 12.9, AOM = 7.8, TIA = 29.7 b) AOD = 15.6, AOM = 9.2, TIA = 27.1 c) AOD = 22.6, AOM = 6.5, TIA = 29 d) AOD = 26, AOM = 16, TIA = 24.5 e) AOD = 21.6, AOM = 15.9, TIA = 22.6 f) AOD = 30.2, AOM = 15.2, TIA = 29.8 (AOM and AOD values are given in pixels, TIA values in degrees). Figure 10 Sample image fragments of the iridocorneal angle for which the known AOD, TIA and TISA methods give incorrect results. For these examples, the iridocorneal angle is determined properly Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 13 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Comparison of quality of classification for the methods AOM, TIA and AOD Comparison of quality of classification for the methods AOM, TIA and AOD This comparison was made for a classification of healthy subjects and patients with the risk of iridocorneal angle closure (TIA < 15°, AOD < 190μm, TISA < 0.11 mm2 [5,20]). Model values of the classification (due to the size of the group) were determined as the average of measurements performed with the AOS and TISA methods; additionally, in some cases, as the average of the results corrected by an expert. The division gave 20'536 cases with a correct iridocorneal angle and 2'454 cases with an incorrect narrowed angle. The AOM, TIA and AOD methods were compared for which the ROC (Receiver Operating Characteristic) curves shown in Figure 11 were obtained. For each method, the cut-off threshold was changed in the range from 0 to 180 by every 1, cov- ering both the range of angular values (in the case of TIA) and the range of distance changes given in pixels (AOM and AOD). Conclusions The paper presents the comparison of known methods and proposes a new method of measuring the iridocorneal angle on the basis of tomographic images. This comparison shows that the best method is the one presented by the authors, namely AOM for which AUC = 0.88. This result is by 0.01 better than that obtained for the well-known AOD method (Tables 1, 2, 3). In terms of specificity, the difference between AOD and AOM is even greater (Figure 11) and amounts to 0.18. The differences between AOD and AOM result from specificity of measurement. For example, measurements for the AOD method (as well as for another method, i.e. TIA) are carried out pointwise [27-29]. The whole area of the iridocorneal angle, i.e. in the range of 500 μm starting from the scleral spur [5], is not taken into account. Moreover, minimum distance values are not calculated unlike in the AOM method. AOM enables to obtain reliable results in the form of a single scalar value in comparison with AOS (for which a sequence of values is obtained). Although AOS enables visualization of more complex cases (of the iridocorneal angle calculated for large pathology) in the form of a graph, it requires re- cording a series of numbers (which refer to distances between the retina and the cor- nea). One of the disadvantages of the AOM method is the need to perform calculations (of minimum distance) on a computer. Manual calculations of the iridocorneal angle with the AOM method may lead to worse results than in the case of the TIA and AOD methods [7,30,31]. In summary, this paper shows the advantage of automatic calculations of the mini- mum distance between the iris and the cornea within a specified range over the other previously known methods. The disadvantages of these known methods are: point measurement, lack of full automation or obtaining a series of information instead of a single number. The new proposed method is free of these defects. It enables to success- fully diagnose pathological cases of iridocorneal angle narrowing, which is not possible with other available methods. It also enables to obtain very high repeatability of mea- surements and record the results in the form of single numbers. Owing to this method, implemented to any OCT device, ophthalmologists will receive more accurate results fully automatically – without any manual intervention. Comparison of the results obtained for the proposed AOM method and the AOD method For the TIA method, optimal values were obtained for the threshold TIA = 18°, i.e.: TPR = 0.82 and SPC = 0.61. These are the worst results obtained for the compared methods. In this case, AUC is 0.77. The other compared method is AOD. In the case of this method, an optimal cut-off threshold is AOD = 8 pixels (about 240 μm). With this threshold value, TPR = 0.86 and SPC = 0.71 and AUC is 0.87. The last compared method is the new AOM method. It gave the best results for a threshold cut-off AOM = 6 pixels (about 180 μm) where TPR = 0.88, SPC = 0.89 and AUC = 0.88. The adopted measurement definitions result from differences between the results. Distance point measurements are the main reason for errors in AOD. This method is very sensitive to noise and artefacts in images which can be observed especially in cases of high degrees of pathology. For the TIA method the situation is similar. A fixed point position of the angle measurement is very sensitive to noise. Another difficulty lies in finding an appropriate point of the angle apex [5,21,22]. Perhaps arranging the angle Figure 11 ROC (Receiver Operating Characteristic) graph of the evaluation of classification for TIA, AOD and AOM. The area under the curve for the best classifier of AOM, i.e. AUC = 0.88 is marked in blue. The other methods produce worse results i.e.: for AOD, AUC = 0.87 and for TIA, AUC = 0.77. Differences in the obtained classification quality result from the conception of measurement with TIA and AOD methods. Figure 11 ROC (Receiver Operating Characteristic) graph of the evaluation of classification for TIA, AOD and AOM. The area under the curve for the best classifier of AOM, i.e. AUC = 0.88 is marked in blue. The other methods produce worse results i.e.: for AOD, AUC = 0.87 and for TIA, AUC = 0.77. Differences in the obtained classification quality result from the conception of measurement with TIA and AOD methods. Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Page 14 of 16 sides as a line approximating the iris and cornea contours would produce better results. The TISA method seems to be the most appropriate here. Comparison of the results obtained for the proposed AOM method and the AOD method However, local narrowing of the iridocorneal angle enables to obtain the same results as in the case of slight angle narrowing, but in a larger area (500 μm) [23-26]. Conclusions Not only will it increase comfort but it will also give rapid and reproducible results. In addition, it will enable to perform an automatic statistical analysis of a selected population of patients and monitor the progress of treatment, which will directly influence the costs of treatment, prevention and diagnosis of patients. Abbreviations AOD: Angle opening distance; TIA: Trabecular-iris angle; TISA: Trabecular-iris space area; AOS: Angle Opening Sequence; AOM: Angle Opening Minimum; AOM2: Angle Opening Minimum 2; ROC: Receiver Operating Characteristic; ACC: Accuracy; SPC: Specificity; TP: True positive; TN: True negative; FN: False negative; FP: False positive. Competing interests The authors declare that they have no competing interests. g p g ; g ; p ; g p g Sequence; AOM: Angle Opening Minimum; AOM2: Angle Opening Minimum 2; ROC: Receiver Operating Characteristic; ACC: Accuracy; SPC: Specificity; TP: True positive; TN: True negative; FN: False negative; FP: False positive. Abbreviations AOD: Angle opening distance; TIA: Trabecular-iris angle; TISA: Trabecular-iris space area; AOS: Angle Opening Sequence; AOM: Angle Opening Minimum; AOM2: Angle Opening Minimum 2; ROC: Receiver Operating Characteristic; Competing interests Th h d l h Page 15 of 16 Page 15 of 16 Page 15 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 Author contributions RK and ZW suggested the algorithm for image analysis and processing, implemented it and analysed the images. SW, AN, EW performed the acquisition of the OCT images and consulted the obtained results. All authors have read and approved the final manuscript. References Katowice, Poland: University of Silesia; 2011 [http://www.ncbi.nlm.nih.gov/books/NBK97169/]. y 8. Sakata LM, Lavanya R, Friedman DS, Aung HT, Seah SK, Foster PJ, Aung T: Assessment of the scleral spur in anterior segment optical coherence tomography images. Arch Ophthalmol 2008, 126:18–185. 8. Sakata LM, Lavanya R, Friedman DS, Aung HT, Seah SK, Foster PJ, Aung T: Assessment of the scleral spur anterior segment optical coherence tomography images. Arch Ophthalmol 2008, 126:18–185. 9. Koprowski R, Wróbel Z: Hierarchic approach in the analysis of tomographic eye image. Comput Recognit Syst 3 Book Ser Adv Intell Soft Comput 2009, 57:463–470. 9. Koprowski R, Wróbel Z: Hierarchic approach in the analysis of tomographic eye image. Comput Recognit Syst 3 Book Ser Adv Intell Soft Comput 2009, 57:463–470. 10. Koprowski R, Wróbel Z: Layers recognition in tomographic eye image based on random contour analysis. Comput Recognit Syst 3 Adv Intell Soft Comput 2009, 57:471–478. Comput Recognit Syst 3 Adv Intell Soft Comput 2009, 57:471–478. 11. Korzynska A, Iwanowski M: Multistage morphological segmentation of bright-field and fluorescent microscopy images. Opto-Electron Rev 2012, 20(2):174–186. 12. Porwik P, Para T: Some handwritten signature parameters in biometric recognition process. In Proceedings of the ITI 2007 29th International Conference on Information Technology Interfaces Book Series. Dubrovnik Croatia: ITI; 2007:185–190. 13. Sonka M, Michael Fitzpatrick J: Medical Image Processing and Analysis. In Handbook of Medical Imaging. Belligham: SPIE; 2000. 14. Korzynska A, Hoppe A, Strojny W, et al: Investigation of a combined texture and contour method for segmentation of light microscopy cell images. In Biomedical Engineering: Proceedings of the Second IASTED International Conference. 2004:234–239. 15. Yoo C, Hyun OJ, Yeon KY, Hai RJ: Peripheral Anterior Synechiae and Ultrasound Biomicroscopic Parameters in Angle-Closure Glaucoma Suspects. Kor J Ophthalmol 2007, 21(2):106–110. g 16. Kim M, Park KH, Kim TW, Kim DM: Changes in anterior chamber configuration after cataract surgery as measured by anterior segment optical coherence tomography. Korean J Ophthalmol 2011, 25(2):77–83. 17. Furuya T, Mabuchi F, Chiba T, Kogure S, Tsukahara S, Kashiwagi K: Comparison of the anterior ocular segment measurements using swept-source optical coherent tomography and a scanning peripheral anterior chamber depth analyzer. Jpn J Ophthalmol 2011, 55(5):472–479. p y p p 18. Liu S, Yu M, Ye C, Lam DS, Leung CK: Anterior chamber angle imaging with swept-source optical coherence tomography: an investigation on variability of angle measurement. Acknowledgements No outside funding was received for this study. References 1. Ishikawa H, Schuman JS: Anterior segment imaging: ultrasound biomicroscopy. Ophthalmol Clin North Am 2004, 7(1):7–20. 1. Ishikawa H, Schuman JS: Anterior segment imaging: ultrasound biomicroscopy. Ophthalmol Clin North Am 2004, 7(1):7–20. 2. Kai-shun LC, Li H, Neal WR, Liu J, Yim LC, Yiu KLR, Pui PC, Shun CD: Anterior Chamber Angle Measurement with Anterior Segment Optical Coherence Tomography. Comparison Between Slit Lamp OCT Visante OCT IOVS 2008, 49(8):3469–3474. 2. Kai-shun LC, Li H, Neal WR, Liu J, Yim LC, Yiu KLR, Pui PC, Shun CD: Anterior Chamber Angle Measurement with Anterior Segment Optical Coherence Tomography. Comparison Between Slit Lamp OCT Visante OCT IOVS 2008, 49(8):3469–3474. 3. Wirbelauer C, Karandish A, Häberle H, Thoai D: Pham Noncontact Goniometry With Optical Coherence Tomography. Arch Ophthalmol 2005, 123:179–185. g p y p 4. Radhakrishnan S, See J, Smith SD, Nolan WP, Ce Z, Friedman DS, Huang D, Li Y, Aung T, Chew PTK: 4. Radhakrishnan S, See J, Smith SD, Nolan WP, Ce Z, Friedman DS, Huang D, Li Y, Aung T, Chew PTK: Reproducibility of Anterior Chamber Angle Measurements Obtained with Anterior Segment Optical Coherence Tomography. IOVS 2007, 48(8):3683–3688. 4. Radhakrishnan S, See J, Smith SD, Nolan WP, Ce Z, Friedman DS, Huang D, Li Y, Aung T, Chew PTK: Reproducibility of Anterior Chamber Angle Measurements Obtained with Anterior Segment Optical Coherence Tomography. IOVS 2007, 48(8):3683–3688. Reproducibility of Anterior Chamber Angle Measurements Obtained with Anterior Segment Optical Coherence Tomography. IOVS 2007, 48(8):3683–3688. g p y 5. Liu S, Li H, Dorairaj S, Cheung CY, Rousso J, Liebmann J, Ritch R, Lam DS, Leung CK: Assessment of scleral spur visibility with anterior segment optical coherence tomography. J Glaucoma 2010, 19(2):132–135. 6. Koprowski R, Wrobel Z, Nowinska A, et al: Automatic Measuring of the Iridocorneal Angle in the Optical Coherence Tomographic Image of the Anterior Segment of the Eye. Inf Technol In Biomed Vol 2 Book Ser Adv Intell Soft Comput 2010, 69:175–182. 6. Koprowski R, Wrobel Z, Nowinska A, et al: Automatic Measuring of the Iridocorneal Angle in the Optical Coherence Tomographic Image of the Anterior Segment of the Eye. Inf Technol In Biomed Vol 2 Book Ser Adv Intell Soft Comput 2010, 69:175–182. p 7. Koprowski R, Wróbel Z: Image Processing in Optical Coherence Tomography: Using Matlab. Katowice, Poland: University of Silesia; 2011 [http://www.ncbi.nlm.nih.gov/books/NBK97169/]. 7. Koprowski R, Wróbel Z: Image Processing in Optical Coherence Tomography: Using Matlab. Author details 1 1Department of Biomedical Computer Systems, Institute of Computer Science, University of Silesia, Będzińska 39 Str, Sosnowiec 41-200, Poland. 2Department of Biophysics, School of Pharmacy, Medical University of Silesia, Jednosci 8 Str, Sosnowiec 41-200, Poland. 3Oddział Okulistyki Okręgowego Szpitala Kolejowego w Katowicach, Okręgowy Szpital Kolejowy w, Katowicach, Poland. Received: 10 February 2013 Accepted: 6 May 2013 Published: 10 May 2013 Received: 10 February 2013 Accepted: 6 May 2013 Published: 10 May 2013 References Invest Ophthalmol Vis Sci 2011, 52(12):8598–8603 19 Ot N A th h ld l ti th d f l l hi t IEEE T S M C b 1979 9(1) 62 66 tomography: an investigation on variability of angle measurement. Invest Ophthalmol Vis Sci 2011, 52(12):8598–8603. 19. Otsu N: A threshold selection method from gray-level histograms. IEEE Trans Sys Man Cyber 1979, 9(1):62–66. 19. Otsu N: A threshold selection method from gray-level histograms. IEEE Trans Sys Man Cyber 1979, 9(1):62–66 20. Christopher Wirbelauer MD, Amir Karandish MD, Heike Häberle MD, Duy Thoai Pham MD: Noncontact Goniometry With Optical Coherence Tomography Arch Ophthalmol 2005 123(2):179–185 20. Christopher Wirbelauer MD, Amir Karandish MD, Heike Häberle MD, Duy Thoai Pham MD: N Goniometry With Optical Coherence Tomography. Arch Ophthalmol 2005, 123(2):179–185. 21. Leung CK, Li H, Weinreb RN, Liu J, Cheung CY, Lai RY, Pang CP, Lam DS: Anterior chamber angle measurement with anterior segment optical coherence tomography: a comparison between slit lamp OCT and Visante OCT. Invest Ophthalmol Vis Sci 2008, 49(8):3469–3474. 22. Ang GS, Wells AP: Changes in Caucasian eyes after laser peripheral iridotomy: an anterior segment optical coherence tomography study. Clin Experiment Ophthalmol 2010, 38(8):778–785. 23. Dada T, Sihota R, Gadia R, Aggarwal A, Mandal S, Gupta V: Comparison of anterior segment optical coherence tomography and ultrasound biomicroscopy for assessment of the anterior segment. J Cataract Refract Surg 2007, 33:837–840. Page 16 of 16 Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 24. Kalev-Landoy M, Day AC, Cordeiro MF, Migdal C: Optical coherence tomography in anterior segment imaging. Acta Ophthalmol Scand 2007, 85(4):427–430. 25. Sakata LM, Lavanya R, Friedman DS, Aung HT, Seah SK, Foster PJ, Aung T: Assessment of the scleral spur in anterior segment optical coherence tomography images. Arch Ophthalmol 2008, 126(2):181–185. 26. Xu L, Cao WF, Wang YX, Chen CX, Jonas JB: Anterior Chamber Depth and Chamber Angle and Their Association with Ocular and General Parameters. The Beijing Eye Study. Am J Ophthalmol 2008, 145(5):929–936. y y 27. Baikoff G, Lutun E, Ferraz C, Wei J: Static and dynamic of the Anterior Segment with Optical Coherence Tomography. J Cataract Refract Surg 2004, 30(9):1843–1850. Tomography. J Cataract Refract Surg 2004, 30(9):1843–1850. 28. Phillips CI: Closed-iridocorneal angle: significance of sectorial variations in angle depth. Br J Ophthalmol 1956, 40:136–143. 29. Koprowski et al. BioMedical Engineering OnLine 2013, 12:40 http://www.biomedical-engineering-online.com/content/12/1/40 doi:10.1186/1475-925X-12-40 Cite this article as: Koprowski et al.: Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye. BioMedical Engineering OnLine 2013 12:40. References Dorairaj SK, Tello C, Liebmann JM, et al: Narrow angles and angle closure: anatomic reasons for earlier closure of the superior portion of the iridocorneal angle. Arch Ophthalmol 2007, 125:734–739. 30. Muftuoglu O, Hosal BM, Zilelioglu G: Ciliary body thickness in unilateral high axial myopia. Eye (Lond) 2009, 23:1176–1181. 31. Oliveira C, Tello C, Liebmann JM, et al: Ciliary body thickness increases with increasing axial myopia. Am J Ophthalmol 2005, 140:324–325. doi:10.1186/1475-925X-12-40 Cite this article as: Koprowski et al.: Methods of measuring the iridocorneal angle in tomographic images of the anterior segment of the eye. BioMedical Engineering OnLine 2013 12:40. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W3012347440	https://www.frontiersin.org/articles/10.3389/fams.2020.00008/pdf	English	null	Editorial: Risk Management Models and Theories	Frontiers in applied mathematics and statistics	2,020	cc-by	1,187	Risk Management Models and Theories The aim of this Research Topic—“Risk Management Models and Theories” is to create a platform for authors to explore, analyze and discuss current and innovative ﬁnancial models and theories that ﬁrms use/prescribe to determine, measure, monitor, forecast, and manage risk in the face of disruptors, such as the increased use of artiﬁcial intelligence and technology, change in regulations, climate change, etc. Since the topic of Risk Management is quite vast, contributions in this ﬁrst issue are related to various areas. One of these areas explored ﬁnancial risks and derivative use by non-ﬁnancial companies in Turkey, wherein the author studied the ﬁnancial risks arising from the activities of non-ﬁnancial companies in BIST 100 index operating in Turkey, and the derivatives used in the management of these risks. He found that the risks that the companies were exposed to included credit risk, liquidity risk, interest risk, currency risk, and other risks and only half of the companies, on average, used derivatives in the management of these risks. In addition, he observed that they hedged themselves against currency risks through futures contracts and used more derivatives in the transportation and energy sectors. Besides, it was noted that the companies using derivative products are big and low liquid companies. Simon Grima 1, Jonathan V. Spiteri 1 and Eleftherios Thalassinos 2 1 Department of Insurance, Faculty of Economics, Management and Accountancy, University of Malta, Msida, Malta, 2 Department of Maritime Studies, University of Piraeus, Piraeus, Greece Keywords: risk models, risk management, risk measurement, risk forecasting, risk theories EDITORIAL EDITORIAL published: 18 March 2020 doi: 10.3389/fams.2020.00008 Edited and reviewed by: Young Shin Aaron Kim, Stony Brook University, United States Correspondence: Simon Grima simon.grima@um.edu.mt Specialty section: This article was submitted to Mathematical Finance, a section of the journal Frontiers in Applied Mathematics and Statistics Received: 11 January 2020 Accepted: 25 February 2020 Published: 18 March 2020 Citation: Grima S, Spiteri JV and Thalassinos E (2020) Editorial: Risk Management Models and Theories. Front. Appl. Math. Stat. 6:8. doi: 10.3389/fams.2020.00008 Edited and reviewed by: Young Shin Aaron Kim, Stony Brook University, United States One author’s contribution related to the evaluation of credit counterparty risk of American options via the Monte Carlo methods. He established a comparison between the Tilley Bundling and Longstaﬀ-Schwartz LSM. Here, the author explained the steps involved in evaluating the value of an American option and how these can be extended to evaluate risk metrics. *Correspondence: Simon Grima simon.grima@um.edu.mt Another contribution was provided on risk measures and inequality, wherein the authors’ objective was to present the inequality evolution through the evolution of the top 1%’s income share, while simultaneously attempting to connect and compare the performance of alternative ARCH/GARCH univariate models for the estimation of 95% Value-at-Risk (VaR) and 95% Expected Shortfall (ES) measures for three equity indices. They noticed that business cycles are present in risk measures as in the inequality measures and highlighted that at the beginning of 2014 there are low values of VaR, and at the same time, the top 1% share increases in 3 out of the 4 countries studied. However, the opposite held for the period just before 2014. They note that this could be an indication that risk measures are not closely related to income inequality. We also had a contribution on the exchange rate pass-through investigation for the Turkish economy, wherein the author investigated the eﬀect of exchange rate changes on the producer and consumer prices in Turkey by using the VAR model. He found that the degree of transition with eﬀect-response does function and that the prices were aﬀected by variance decomposition by using the 2005-2019 monthly data. He found that the CPI’s response to the change in nominal exchange rate was found to be greater than the PPI, and that the end of the impact is shorter than CPI’s impact. Citation: March 2020 \| Volume 6 \| Article 8 Frontiers in Applied Mathematics and Statistics \| www.frontiersin.org Editorial: Risk Management Models and Theories Grima et al. Moreover, we ﬁnd a contribution relating to the ways in which developed and developing stock markets reacted to the Dow Jones during the 2008 crisis. Here, the authors ﬁnd that the developed and emerging stock markets react diﬀerently to the DJIA. Another author contributed to the volatility spillover between stock prices and trading volume during the pre-, in- and post global ﬁnancial crisis periods. Herein, she found the existence of bidirectional volatility spillovers between stock price and trading volume in the pre- and post-crisis periods, and noted that in the crisis period, there is a unidirectional volatility spillover from stock prices to trading volume. This indicated that while the volatility of stock price aﬀects the trading volume with lags in the crisis period, the volatility of stock price and that of trading volume in the non-crisis periods aﬀect each other. As can be noted from the above paragraphs, the contributions and ﬁndings of the ﬁrst issue of this topic are quite focused around the area of ﬁnance and related to studies of current and during-crisis practices, tools, techniques and ﬁnancial instruments used by the Risk Manager to manage identiﬁed exposures to risks. Therefore, the issue is aimed at the interest of academics and practitioners in this ﬁeld. p We also have a contribution on the volatility and shock transmission patterns between the BIST sustainability and BIST 100 indices. The aim of the authors in this case was to empirically investigate the volatility and shock transmission patterns between the BIST 100 index and the relatively new BIST sustainability index, which is a platform for companies with high performance on the international sustainability criteria. They utilized 678 daily data from 05/11/2014, the day the XSURD Index was launched, to 31/08/2017. The analysis employed a bivariate BEKK-GARCH (1, 1) model and ﬁndings indicate the existence of bidirectional volatility spillovers between the two indices. Additionally, they show that current volatility is aﬀected by its past volatility for each index. As for the shock transmission, they show that the BIST Sustainability Index is responsive to both its shocks and shocks arriving from the BIST 100 Index, notwithstanding that the BIST 100 Index responds only to its shocks without any signiﬁcant shock transmission from the BIST Sustainability Index. Frontiers in Applied Mathematics and Statistics \| www.frontiersin.org AUTHOR CONTRIBUTIONS All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Copyright © 2020 Grima, Spiteri and Thalassinos. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. March 2020 \| Volume 6 \| Article 8 Frontiers in Applied Mathematics and Statistics \| www.frontiersin.org 2
https://openalex.org/W4200468868	http://scielo.pt/pdf/riis/v4n2/2184-3791-riis-4-02-52.pdf	Portuguese	null	Tutoria por pares na educação em enfermagem: a voz dos tutores	Revista de investigação & inovação em saúde	2,021	cc-by	5,676	Augusto, C., Mendes, G. Ermelinda Macedo, E., Machado, M., Candeias, A., Encarnação, P. Freire, T., & Araújo, O. (2021). Tutoria por pares na educação em enfermagem: a voz dos tutores. Revista de Investigação & Inovação em Saúde, 4(2), 53-61. doi.org/10.37914/riis.v4i2.166 ABSTRACT * PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0001-5450-7307 - Contribuição no artigo: Study conception and design, Data analysis and interpretation, Drafting of the article PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0003-4035-3743 - Contribuição no artigo: Study conception and design, Data analysis and interpretation, Drafting of the article * PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0003-4053-2864 - Contribuição no artigo: Data collection, ** PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0003-2867-435X - Contribuição no artigo: Study conception and design, Critical revision of the article * MsC, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0001- 9620-163X - Contribuição no artigo: Data collection ** PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0002- 7458-8105 - Contribuição no artigo: Data collection *** PhD, School of Psychology, University of Minho; Psychology Research Centre, university of Minho - https://orcid.org/0000-0001-5773-381X - Contribuição no artigo: Study conception and design, Critical revision of the article ****** PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0001- 9016-9528 - Contribuição no artigo: Data analysis and interpretation, Drafting of the article Background: pedagogical experiences involving peer tutoring have been increasingly diversified. In the case of nursing education, tutoring has registered a significant increase This study is integrated in Project for Innovation and Development of Teaching and Learning - Peer tutoring in the academic context of nursing (TutorParE). It aims to promote innovative pedagogical strategies in order to facilitate both the development of academic and transversal skills in nursing students. Objectives: analyze the experience of tutors on academic tutoring in nursing education. Methodology: this study, qualitative, exploratory and descriptive design enrolled eleven students from the 4th year of the nursing degree in one discipline of the 1st year. Results: two categories emerged from the Results: (1) personal development and; (2) collaborative learning. TUTORIA POR PARES NA EDUCAÇÃO EM ENFERMAGEM: A VOZ DOS TUTORES Peer tutoring in nursing education: the voice of tutors Tutoría entre pares en educación en enfermería: la voz de los tutores Peer tutoring in nursing education: the voice of tutors Tutoría entre pares en educación en enfermería: la voz de los tutores Cláudia Augusto, Goreti Mendes, Ermelinda Macedo, Manuela Machado, Analisa Candeias*, Paula Encarnação**, Teresa Freire***, Odete Araújo**** RESUMO Enquadramento: as experiências pedagógicas que envolvem tutoria por pares têm sido cada vez mais diversificadas. No caso do ensino de enfermagem, a tutoria tem registado um aumento significativo. O presente estudo insere-se num Projeto de Inovação e Desenvolvimento do Ensino e da Aprendizagem - Tutorias por pares em contexto académico de Enfermagem (TutorParE). Tem como finalidade a promoção de estratégias pedagógicas inovadoras e promotoras do desenvolvimento de competências académicas e transversais nos estudantes do ensino superior. Objetivos: analisar a experiência dos tutores sobre a tutoria académica no ensino de enfermagem. Metodologia: o estudo desenvolvido, qualitativo, exploratório e descritivo, envolveu onze estudantes tutores do 4.º ano do curso de licenciatura em enfermagem, numa Unidade Curricular do 1.º ano do curso. Resultados: emergiram duas categorias: (1) desenvolvimento pessoal e; (2) aprendizagem colaborativa. A estratégia de tutoria por pares foi considerada útil na aquisição de conhecimentos e de competências nos estudantes, facilitando o desenvolvimento pessoal e a aprendizagem colaborativa. Conclusão: os achados reforçam a importância da adoção de práticas pedagógicas ativas e inovadoras, baseadas na tutoria por pares. Palavras chave: tutoria; enfermagem; estudantes; aprendizagem ABSTRACT The strategy adopted in peer tutoring in nursing education was considered useful to enhance the acquisition of students' knowledge and skills, allowing personal development and collaborative learning. Conclusion: the findings achieved reinforce the importance of adopting active and innovative pedagogical practices, based on peer tutoring. * MsC, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0001- 9620-163X - Contribuição no artigo: Data collection Contribuição no artigo: Data collection ** PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0002- 7458-8105 Keywords: tutoring; nursing; students; learning - Contribuição no artigo: Data collection *** PhD, School of Psychology, University of Minho; Psychology Research Centre, university of Minho - https://orcid.org/0000-0001-5773-381X RIIS \| vol.4(2), 53 - 61 RIIS \| vol.4(2), 53 - 61 RIIS \| vol.4(2), 53 - 61 ARTIGO DE INVESTIGAÇÃO Recebido para publicação:23/07/2021 Aceite para publicação:15/11/2021 INTRODUÇÃO A tutoria por pares, frequentemente utilizada no ensino de enfermagem, compreende uma relação entre dois estudantes em estadios de aprendizagem diferentes (Hawkins & Fontenot, 2010). Assenta na ideia de que um estudante num estadio mais inicial da sua formação sentir-se-á mais confortável e seguro ao abordar um colega (tutor), com idade e status aproximados, em contexto de aprendizagem, em detrimento do professor, cujo papel está frequentemente associado à figura de avaliador do processo de ensino aprendizagem (Bulut, Hisar, & Demir, 2010). Em contextos laboratoriais e de simulação o estudante tem oportunidade de executar procedimentos técnicos, sem receio de falhar, num ambiente de aprendizagem colaborativo, que o ajudará a adquirir progressivamente maior confiança nas suas capacidades, facilitando o desenvolvimento de competências práticas (Kim, Jillapali & Boyd, 2021). Promove uma aprendizagem mais significativa e a transferência de competências comunicacionais para as experiências clínicas (Svellingen, Røssland, & Røykenes, 2021). Os programas de tutoria têm sido descritos como uma estratégia, centrada no estudante, que permite melhorar a sua aprendizagem e o seu desempenho académico, potencia o desenvolvimento de competências transversais, como o trabalho em equipa e as relações interpessoais, promove a motivação e o crescimento pessoal de todos os envolvidos no processo, tutores e tutorandos (Kim, Jillapali & Boyd, 2021). Têm sido amplamente adotados no ensino superior e configuram uma estratégia importante de facilitação da integração dos estudantes do primeiro ano, ajudando-os nessa transição. Constituem uma oportunidade de construção identitária, podendo os tutores ser um “modelo” para os seus tutorandos que, por sua vez, desenvolvem competências intelectuais e emocionais (Freire & Beiramar, 2017; Hall & Jaugietis, 2011). São facilitadoras de realização, em contextos gerais de vida, ao permitirem o contacto com experiências desenvolvimentais positivas, o desenvolvimento da auto complexidade, e que conferem um elevado sentido de autoeficácia (Freire & Beiramar, 2017). Kim e colaboradores (2021) reportam a satisfação do tutor e do tutorando com as sessões de tutoria, a criação de um ambiente promotor da aprendizagem, o fornecimento de feedback útil, a comunicação efetiva tutor/tutorando e a melhoria da capacidade na resolução de problemas. A implementação de programas de tutoria por pares no ensino de enfermagem, focadas em abordagens centradas no estudante, são muito relevantes para o desenvolvimento da autonomia, do pensamento crítico e da responsabilidade profissional do futuro enfermeiro. RESUMEN Marco contextual: las experiencias pedagógicas que involucran la tutoría entre pares se han diversificado cada vez más. En el caso de la formación en enfermería, la tutoría ha registrado un aumento significativo. Este estudio es parte de un Proyecto de Innovación y Desarrollo para la Enseñanza y el Aprendizaje - Tutoría entre pares en el contexto académico de enfermería (TutorParE). Su propósito es promover estrategias pedagógicas innovadoras que promuevan el desarrollo de habilidades académicas y transversales en los estudiantes de Educación Superior. Objetivo: analizar la experiencia de los tutores en la tutoría académica en la educación en enfermería. Metodologia: este estudio, cualitativo, exploratorio y descriptivo involucró a once estudiantes de cuarto año de la carrera de Enfermería en una matéria de primero ano. Resultados: surgieron dos categorías: (1) desarrollo personal y; (2) aprendizaje colaborativo. La estrategia de tutoría entre pares se consideró útil para mejorar la adquisición de conocimientos y habilidades de los estudiantes al permitir su desarrollo personal al tiempo que mejora el aprendizaje colaborativo. Conclusión: los hallazgos son pistas válidas que refuerzan la importancia de adoptar prácticas pedagógicas activas e innovadoras, basadas en la tutoría entre pares. - Contribuição no artigo: Study conception and design, Critical revision of the article - Contribuição no artigo: Study conception and design, Critical revision of the article ****** PhD, School of Nursing of the University of Minho; Health Sciences Research Unit: Nursing (UICISA: E), Nursing School of Coimbra (ESEnfC) - https://orcid.org/0000-0001- 9016-9528 - Contribuição no artigo: Data analysis and interpretation, Drafting of the article Recebido para publicação:23/07/2021 Aceite para publicação:15/11/2021 Palabras clave: tutoría; enfermería; estudantes; aprendizaje 53 Disponível em: https://doi.org/10.37914/riis.v4i2.166 Disponível em: https://doi.org/10.37914/riis.v4i2.166 RIIS Revista de Investigação & Inovação em Saúde Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores À data da realização da tutoria, os tutores, encontravam-se a realizar uma UC de estágio a qual, no projeto individual do estudante, valoriza o desenvolvimento de competências académicas e transversais (nomeadamente o trabalho em equipa, a relação interpessoal, capacidade de liderança, entre outras) decorrentes de várias experiências. Sendo este estágio a última UC do CLE, estas competências revestem-se de particular importância, uma vez que correspondem em parte às competências requeridas pelo órgão regulador da profissão de enfermagem em Portugal (Ordem dos Enfermeiros) no Perfil de Competências do Enfermeiro de Cuidados Gerais (Ordem dos Enfermeiros, 2011) INTRODUÇÃO Tendo em conta as vantagens enunciadas, surge o projeto “Tutorias por pares em contexto académico de enfermagem” (TutorParE) desenvolvido na Escola Superior de Enfermagem da Universidade do Minho. Com o propósito de analisar a experiência dos tutores que realizaram tutorias em contexto académico do ensino de enfermagem, foi desenvolvido o presente estudo de natureza qualitativa. 54 RIIS RIIS Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores ria por pares na educação em enfermagem: a voz dos tutores antes o que experiência pedagógica decorreu entre fevereiro e maio de 2019. ECTS e média de curso ≥ 14 valores. Estes estudantes frequentaram 6h de formação de preparação que versou, por um lado, a caraterização da UC onde decorreriam as sessões de tutoria e, por outro, orientações para o desempenho do papel de tutor, designadamente preparação teórica para as sessões, comunicação com o tutorando, responsabilidade com o planeamento, entre outros. Para melhor se enquadrarem nos conteúdos teóricos que estavam a ser lecionados na UC do 1.º ano, os estudantes tutores foram inscritos na mesma, na plataforma digital, com o perfil de estudante convidado. experiência pedagógica decorreu entre fevereiro e maio de 2019. experiência pedagógica decorreu entre fevereiro e maio de 2019. experiência pedagógica decorreu entre fevereiro e maio de 2019. No final das sessões de tutoria, foi solicitado que docentes, tutores e tutorandos efetuassem uma apreciação global qualitativa da participação na experiência. Os resultados deste estudo dizem respeito aos dados obtidos da avaliação qualitativa dos estudantes tutores. Os tutores foram convidados a responder através da ferramenta formulário da plataforma Google Forms, à questão “De um modo geral, que apreciação faz da sua experiência como tutor?” Os tutores foram sensibilizados para a importância de integrar o estudo de forma voluntária, tendo sido assegurado o anonimato e a confidencialidade dos dados, bem como todos os princípios éticos previstos em estudos de investigação científica. A função de cada tutor foi mediada por um docente da equipa pedagógica da UC do 1.º ano. Aquando da abordagem dos conteúdos teóricos pelo docente, foi permitido ao tutor assistir à lecionação destes conteúdos para que, de uma forma concertada, pudesse posteriormente sustentar as aulas de prática laboratorial. Cada tutor realizou sessões de tutoria a grupos de 10 a 12 tutorandos, ao longo de 20 horas letivas da UC do 1.º ano. Nas aulas de prática- laboratorial, o tutor participou nas atividades de simulação, adequação e treino de procedimentos técnicos, bem como, esclarecimento de dúvidas. A participação dos tutores incluiu ainda, complementarmente, testemunhos de experiências de aprendizagem, de prática clínica e vivências académicas. A informação foi analisada tendo por base os pressupostos da análise de conteúdo de Bardin (2018). Tutoria por pares na educação em enfermagem: a voz dos tutores Recorreu-se à utilização de dois códigos: (1) o código (...) significa um corte na narrativa sem significado relevante para a interpretação dos dados, ou seja, não compromete a interpretação e a compreensão dos mesmos; e, (2) recorreu-se à utilização de um código de dois dígitos composto por uma letra, letra E (de estudante) e um número decimal que corresponde à ordem pela qual foi obtida a resposta. Assim obteve-se E1, E2, E3, …E11, que corresponde ao número total de estudantes participantes. Os docentes envolvidos no projeto participaram também em sessões de formação sobre tutoria por pares e qual o papel do docente na promoção da aprendizagem de tutores e tutorandos. Cada tutor acompanhou um docente vinculado às aulas práticas laboratoriais dos estudantes do 1.º ano. Esta METODOLOGIA O presente estudo, de carácter exploratório e descritivo, insere-se no paradigma de investigação qualitativa. Envolveu a participação de 11 estudantes do 4.º ano, finalistas do Curso de Licenciatura em Enfermagem (CLE), que assumiram o papel de tutores dos seus pares, estudantes do 1.º ano do mesmo curso, no âmbito do projeto TutorParE. O projeto TutorParE, inovador no ensino de enfermagem, foi financiado por um Programa de Apoio a Projetos de Inovação e Desenvolvimento do Ensino e da Aprendizagem. Foi implementado no ano letivo 2018/2019. A tutoria decorreu ao longo de 20 horas de aulas de prática laboratorial no âmbito da Unidade Curricular (UC) Fundamentos de Enfermagem II, do 1º ano, 2º semestre do CLE e de uma UC do último semestre do mesmo curso (4º ano, 2º semestre), Estágio de Integração à Vida Profissional. Integraram este projeto 85 estudantes do 1.º ano (tutorandos); 11 estudantes do 4.º ano, finalistas (tutores) e 4 docentes da equipa pedagógica da UC do 1.º ano. Aquando da realização do estágio, os estudantes tutores encontravam-se dispersos por diferentes instituições de saúde, algumas das quais distantes geograficamente do campus universitário onde decorreram as práticas laboratoriais do 1.º ano. O exercício do papel de tutor implica a deslocação destes estudantes ao campus, variável que influenciou a captação de tutores na experiência piloto desenvolvida. Previamente a este projeto, havia já sido desenvolvida uma experiência piloto no ano letivo 2017/2018, experiência que motivou a continuidade de implementação da estratégia pedagógica e da qual emergiram contributos para o aprimoramento da mesma para o ano letivo subsequente. A avaliação feita dessa experiência, permitiu chegar a algumas conclusões qualitativas interessantes tendo sido, o número reduzido de tutores, um aspeto apreciado de forma menos favorável pelos tutorandos, o que desafiou a refletir sobre novas estratégias de captação de tutores, de forma a ampliar a iniciativa. O financiamento do projeto veio permitir colmatar o constrangimento identificado, ao possibilitar, pela via da gratificação, ou seja, ao suportar os custos acrescidos decorrentes das deslocações dos estudantes tutores ao campus, ampliar os benefícios da experiência, com maior expressão do número de tutores. O exercício do papel de tutores obedeceu a critérios previamente estabelecidos pela equipa pedagógica e de investigação. Puderam candidatar-se de forma livre, os estudantes com manifesto interesse em ser tutor e a participar no projeto, com aproveitamento em 180 55 RIIS Revista de Investigação & Inovação em Saúde RIIS RESULTADOS Recolheram-se respostas de 8 tutores participantes. A maioria é do género feminino (87,4%), com uma idade média de 22 anos. 56 RIIS RIIS Revista de Investigação & Inovação em Saúde ares na educação em enfermagem: a voz dos tutores resultados, a inferência e a interpretação. Nesta análise, o critério de categorização adotado foi o semântico, originando categorias temáticas que foram criadas à posteriori, ou seja, tomaram forma no curso da própria análise, respeitando as características específicas definidas por Bardin (2018): exclusão mútua; homogeneidade; objetividade, fidelidade e produtividade. De referir, ainda, que o processo de categorização teve por base a análise indutiva e pormenorizada levada a cabo por dois investigadores independentes. A evidência dos resultados é visível na Tabela 1. Da análise dos dados, relativamente à questão aberta “De um modo geral, que apreciação faz da sua experiência como tutor?” resultou um conjunto de linhas discursivas que foram agrupadas em duas categorias: (i) Desenvolvimento Pessoal e (ii) Aprendizagem colaborativa. Toda a informação foi analisada tendo por base os pressupostos da análise de conteúdo de Bardin (2018) o que permite a explicitação e sistematização do conteúdo das respostas, a categorização dos dados, com base nas seguintes fases: i) a pré-análise; ii) a exploração do material; iii) o tratamento dos Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores A experiência de tutoria é compreendida pelos tutores como uma oportunidade de desenvolverem capacidades como a confiança, o sentido de responsabilidade e a liderança. Os estudos desenvolvidos por Dennison (2010); Ford (2015); Hall e Jaugietis (2011) e Rosenau e colaboradores (2015), suportam esta ideia ao afirmarem que o processo de tutoria traduz maior autoconfiança, sentido de responsabilidade e desenvolvimento intelectual nos tutores que participam em experiências desta natureza. Os resultados mostraram, que os estudantes tutores envolvidos nas atividades de tutoria experimentaram sentimentos de gratificação e realização, mas também de responsabilidade e autonomia, resultados que refletem o contributo no seu próprio desenvolvimento pessoal, mas também dos estudantes que acompanharam. Os estudantes tutores reconheceram, na sua maioria, que da interação entre tutores e tutorandos resulta uma aprendizagem colaborativa sendo por meio da participação de todos os intervenientes que se realiza o processo de construção coletiva do conhecimento. Tratando-se de tutoria académica, os tutores reportaram que esta prática pedagógica constituiu uma exigência pessoal, que gerou simultaneamente aprendizagem para o próprio tutor, nomeadamente ao nível da comunicação e da mobilização de conhecimentos técnicos. Alguma investigação desenvolvida sobre as experiências de tutoria por pares no ensino superior, demonstra o impacto das experiências de tutoria na melhoria generalizada do desempenho académico do tutor (Aponte et al., 2015; Rosenau et al., 2015; Tabloski, 2016). Tabela 1 Categorias emergentes (“desenvolvimento pessoal” e “aprendizagem colaborativa”) e respetivas unidades de análise Unidades de Análise Desenvolvimento Pessoal Aprendizagem Colaborativa “proporcionou crescimento profissional e pessoal (...) um momento relevante de atualização e de estudo” (E1); “(…) permitiu-me aprimorar competências transversais (…) comunicação e cooperação (…) relações humanas, na própria mobilização de conhecimentos técnicos (…)” (E4; E5); “(…) ajudou bastante no desenvolvimento de competências comunicacionais” (...) (E1); “(…) colocarmo-nos no lugar de um aluno recém-chegado ao ensino superior (…) com tantas dúvidas para esclarecer, com muitos medos do que pode ser o seu percurso académico, é muito gratificante e um grande desafio pessoal para nós (…)” (E8); “(…) responsabilidade na transmissão de conhecimentos e estabelecimento interpessoal (…) uma experiência gratificante, promotora de desenvolvimento a todos os níveis (…)” (E10); “(…) uma verdadeira gratificação para os tutores nas relações humanas (...)” (E5); “(…) uma mais valia para ambos os estudantes (…) oferece oportunidades de uma aprendizagem colaborativa (...) aprendemos todos juntos, tutores e tutorandos” (E1); “(...) há uma entrega de todos” (...)” (E7); “(…) há um envolvimento de todos os alunos (…)” (E10); “(...) tanto os tutores como os tutorandos beneficiam da experiência (...)” (E8); “(…) gostava que isto tivesse acontecido enquanto fui aluna no meu primeiro ano” (…) (E1); (…) fico muito feliz pelos nosso colegas (tutorandos) que tiveram a oportunidade que nunca tivemos (…)” (E2); (…) uma aprendizagem mais focada e interessada (...)” E3. ergentes (“desenvolvimento pessoal” e “aprendizagem colaborativa”) e respetivas unidades de “(…) colocarmo-nos no lugar de um aluno recém-chegado ao ensino superior (…) com tantas dúvidas para esclarecer, com muitos medos do que pode ser o seu percurso académico, é muito gratificante e um grande desafio pessoal para nós (…)” (E8); 57 RIIS Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores RIIS Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores consideram o treino de competências comunicacionais como um dos principais benefícios deste tipo de programas, na medida em que envolve habilidades básicas de comunicação destacando o feedback, a capacidade de auscultar ou mesmo a capacidade de ouvir uma questão sem influenciar a resposta. programas de tutoria por pares na educação em enfermagem traz benefícios e desafios para os estudantes tutores e tutorandos. Os estudantes tutores, participantes neste estudo, reconheceram que todos os envolvidos são agentes ativos que contribuem para a construção do conhecimento mútuo. Também na perspetiva de Klein & Vosgerau (2018), a prática de aprendizagem colaborativa permite que docentes e estudantes construam em conjunto a aprendizagem e compartilhem conhecimentos e experiências. O professor assume também um papel fundamental em todo o processo, na medida em que deve proporcionar aos tutores o acesso ao maior conhecimento possível, é esperado que os estudantes compreendam bem o que estão a estudar, com o máximo de profundidade possível (Pereira, 2017). q p No que concerne à aprendizagem colaborativa, metodologia de ensino pautada pela interação, colaboração e participação ativa dos estudantes, é reconhecido o seu potencial de promover uma aprendizagem mais ativa por meio do estímulo, particularmente, ao nível do pensamento crítico (Pereira, 2017; Rosenau et al., 2015; Torres & Irala, 2014) e do desenvolvimento de competências de interação, de resolução de problemas e do desenvolvimento da capacidade de autorregulação do processo de ensino-aprendizagem (Torres & Irala, 2014). Numa visão mais ampla do que significa aprendizagem colaborativa, estes autores advogam que é esperado que a aprendizagem ocorra como efeito paralelo a uma interação entre os pares que trabalham num processo de interdependência na resolução de problemas ou na realização de tarefas propostas pelo tutor, conduzindo deste modo a uma aprendizagem mais eficiente, em vez de competitiva e isolada (Torres & Irala, 2014). A aprendizagem colaborativa é um processo partilhado entre os estudantes e mediado pelo docente que, além de fomentar a interação e a colaboração promove o desenvolvimento do grupo, o desenvolvimento da autonomia, da aprendizagem pelo processo de interação e o desenvolvimento da responsabilidade sobre a sua própria aprendizagem (Klein & Vosgerau, 2018). DISCUSSÃO Em consonância com outras investigações, os estudantes tutores participantes deste estudo reconheceram que a tutoria por pares traz genericamente contributos para o seu desenvolvimento de competências (Bulut et al., 2010; Sim-Sim, Marques, Frade, & Chora, 2013). Relativamente ao desenvolvimento pessoal, este pressupõe o crescimento de habilidades pessoais que potenciam o conhecimento humano. A par da gratificação monetária, a participação voluntária dos tutores determinou a motivação e o envolvimento na experiência pedagógica. Alguma da investigação desenvolvida sobre as experiências de tutoria sugere que os estudantes mais motivados e mais envolvidos nestas atividades, experimentam sentimentos de gratificação e realização, ajudando, por um lado, os tutorandos no processo de aprendizagem e, por outro, na atribuição de um maior sentido à sua vida profissional (Dennison, 2010). Ao assumir o papel de tutores, os estudantes de enfermagem experimentam auto crescimento e desenvolvem competências de liderança, representando uma aproximação ao exercício do seu papel enquanto futuros enfermeiros (Ford, 2015; Rosenau et al., 2015). Estes dados são corroborados por James e colaboradores (2014) ao revelarem que os tutores expressam, frequentemente, sentido de crescimento e de maturidade pessoal, constituindo a vivência de tutoria em enfermagem uma verdadeira gratificação para os tutores, tal como os participantes deste estudo. James e colaboradores (2014) 58 RIIS RIIS Revista de Investigação & Inovação em Saúde RIIS Revista de Investigação & Inovação em Saúde Tutoria por pares na educação em enfermagem: a voz dos tutores unicacionais te tipo de habilidades programas de tutoria por pares na educação em enfermagem traz benefícios e desafios para os estudantes tutores e tutorandos. Tutoria por pares na educação em enfermagem: a voz dos tutores A aprendizagem colaborativa possibilita a construção da aprendizagem e do conhecimento de forma compartilhada entre os estudantes, permanecendo como grandes desafios a compreensão do conceito pelos próprios docentes e a aceitação por parte dos estudantes (Klein & Vosgerau, 2018). Esta prática pedagógica consegue envolver os estudantes em um nível de problematização que as aulas expositivas tradicionais dificilmente conseguiriam. A capacidade de ouvir outros colegas sobre determinado assunto torna-se chave para aprendizagens mais profundas (Pereira, 2017). Os contributos deste estudo constituem, assim, pistas válidas que reforçam a importância da adoção de práticas pedagógicas ativas e inovadoras, baseadas na tutoria por pares em contexto académico. A participação em atividades que desenvolvem as competências transversais no contexto da formação em enfermagem conduz a uma melhor performance académica, ao estabelecimento de relações interpessoais positivas, pela aprendizagem colaborativa que proporciona. Ao discutir-se os benefícios e desafios da aprendizagem colaborativa no ensino superior, propõe-se contribuir para o fortalecimento e promoção do uso destas práticas pedagógicas. Sugere- se que estudos futuros enquadrem as experiências de tutoria por pares como estratégia pedagógica centradas nos estudantes potenciadora da qualidade ao nível dos diferentes ciclos de estudo. A implementação do projeto TutorParE apela à continuidade, pelo que se sugere que estudos futuros possam contemplar a perspetiva dos estudantes tutorandos e, ainda, o desenvolvimento de competências monitorizadas por instrumentos de avaliação das práticas laboratoriais e de simulação. Tutoria por pares na educação em enfermagem: a voz dos tutores Tutoria por pares na educação em enfermagem: a voz dos tutores uentemente derada uma os benefícios erida pelos estudantes, nomeadamente competências relacionais e comunicacionais promovendo o seu desenvolvimento pessoal, ao mesmo tempo que potencia a aprendizagem colaborativa. A aprendizagem colaborativa, frequentemente defendida no ensino superior e considerada uma forma de ensinar e aprender com inúmeros benefícios (Klein & Yosgerau, 2018), foi referida pelos participantes neste estudo como uma estratégia potenciadora da aprendizagem, uma vez que coloca todos os elementos envolvidos como agentes do seu próprio desenvolvimento. O estudo desenvolvido por Rosenau e colaboradores (2015), evidencia o impacto que estas estratégias de ensino e aprendizagem têm no desenvolvimento dos tutores, ao promoverem uma consciência aumentada das suas próprias crenças e valores e enquanto pessoas e profissionais. Além de promover a responsabilidade profissional, esta estratégia de ensino aprendizagem estimula os processos de pensamento reflexivo do estudante tutor, melhora a sua confiança e competência (Rosenau et al., 2015). A aprendizagem colaborativa possibilita a construção da aprendizagem e do conhecimento de forma compartilhada entre os estudantes, permanecendo como grandes desafios a compreensão do conceito pelos próprios docentes e a aceitação por parte dos estudantes (Klein & Vosgerau, 2018). Esta prática pedagógica consegue envolver os estudantes em um nível de problematização que as aulas expositivas tradicionais dificilmente conseguiriam. A capacidade de ouvir outros colegas sobre determinado assunto torna-se chave para aprendizagens mais profundas (Pereira, 2017). estudantes, nomeadamente competências relacionais e comunicacionais promovendo o seu desenvolvimento pessoal, ao mesmo tempo que potencia a aprendizagem colaborativa. A aprendizagem colaborativa, frequentemente defendida no ensino superior e considerada uma forma de ensinar e aprender com inúmeros benefícios (Klein & Yosgerau, 2018), foi referida pelos participantes neste estudo como uma estratégia potenciadora da aprendizagem, uma vez que coloca todos os elementos envolvidos como agentes do seu próprio desenvolvimento. O estudo desenvolvido por Rosenau e colaboradores (2015), evidencia o impacto que estas estratégias de ensino e aprendizagem têm no desenvolvimento dos tutores, ao promoverem uma consciência aumentada das suas próprias crenças e valores e enquanto pessoas e profissionais. Além de promover a responsabilidade profissional, esta estratégia de ensino aprendizagem estimula os processos de pensamento reflexivo do estudante tutor, melhora a sua confiança e competência (Rosenau et al., 2015). Tutoria por pares na educação em enfermagem: a voz dos tutores A aprendizagem colaborativa tem vindo a ser usada, há já algum tempo no ensino superior e os achados deste estudo revelaram que participação em No que concerne à aprendizagem colaborativa, metodologia de ensino pautada pela interação, colaboração e participação ativa dos estudantes, é reconhecido o seu potencial de promover uma aprendizagem mais ativa por meio do estímulo, particularmente, ao nível do pensamento crítico (Pereira, 2017; Rosenau et al., 2015; Torres & Irala, 2014) e do desenvolvimento de competências de interação, de resolução de problemas e do desenvolvimento da capacidade de autorregulação do processo de ensino-aprendizagem (Torres & Irala, 2014). Numa visão mais ampla do que significa aprendizagem colaborativa, estes autores advogam que é esperado que a aprendizagem ocorra como efeito paralelo a uma interação entre os pares que trabalham num processo de interdependência na resolução de problemas ou na realização de tarefas propostas pelo tutor, conduzindo deste modo a uma aprendizagem mais eficiente, em vez de competitiva e isolada (Torres & Irala, 2014). A aprendizagem colaborativa é um processo partilhado entre os estudantes e mediado pelo docente que, além de fomentar a interação e a colaboração promove o desenvolvimento do grupo, o desenvolvimento da autonomia, da aprendizagem pelo processo de interação e o desenvolvimento da responsabilidade sobre a sua própria aprendizagem (Klein & Vosgerau, 2018). A aprendizagem colaborativa tem vindo a ser usada, há já algum tempo no ensino superior e os Um dos tutores classificou esta experiência como focada, despertando o interesse por aprender com alguém que já enfrentou desafios semelhantes. O envolvimento dos participantes e a troca de experiências, na perspetiva de Brody e colaboradores (2016) será tão melhor conseguida se realizada por quem já passou pela experiência, ou seja, por quem está familiarizado com as exigências do curso e os desafios académicos do mesmo. Da interação entre tutores e tutorandos resulta uma construção coletiva, sendo por meio da participação de todos os que interagem entre si que se realiza o processo de construção do conhecimento (Torres & Irala, 2014). Como referem Rosenau e colaboradores (2015), a tutoria por pares promove a maturidade e a responsabilidade dos tutores e, por conseguinte, competências de liderança, essenciais ao perfil profissional do enfermeiro. 59 RIIS RIIS Revista de Investigação & Inovação em Saúde RIIS Revista de Investigação & Inovação em Saúde RIIS Revista de Investigação & Inovação em Saúde Bardin, L. (2018). Análise de Conteúdo. Edições 70. Brody, A. A., Edelman, L., Siegel, E. O., Foster, V., Bailey, D. E., Jr., Bryant, A. L., & Bond, S. M. (2016). Evaluation of a peer mentoring program for early career gerontological nursing faculty and its potential for application to other fields in nursing and health sciences. Nursing Outlook, 64(4), 332-338. doi:10.1016/j.outlook.2016.03.004 Klein, E., & Vosgerau, D. (2018). Possibilidades e desafios da prática de aprendizagem colaborativa no ensino superior. Educação (UFSM), 4(43), 667-698. Ordem dos Enfermeiros (2011). Regulamento do Perfil de Competências do Enfermeiro de Cuidados Gerais. Ordem dos Enfermeiros. Bulut, H., Hisar, F., & Demir, S. G. (2010). Evaluation of mentorship programme in nursing education: A pilot study in Turkey. Nurse Education Today, 30(8), 756- 762. doi:10.1016/j.nedt.2010.01.019 Pereira, F. (2017). Aprendizagem por pares e os desafios da educação para o senso crítico. Int. J. Activ. Learn., 2(1), 6-12. De Backer, L., Van Keer, H., & Valcke, M. (2012). Exploring the potential impact of reciprocal peer tutoring on higher education students' metacognitive knowledge and regulation. Instructional Science, 40(3), 559-588. doi:10.1007/s11251-011-9190-5 Rosenau, P. A., Lisella, R. F., Clancy, T. L., & Nowell, L. S. (2015). Developing future nurse educators through peer mentoring. Nursing-Research and Reviews, 5, 13- 21. doi:10.2147/nrr.s73432 Sim-Sim, M.-M., Marques, M.-d.-C., Frade, M.-d.-A., & Chora, M.-A. (2013). Tutoria: perspetiva de estudantes e professores de enfermagem. Revista Iberoamericana de Educación Superior, 4(11), 45-59. doi:https://doi.org/10.1016/S2007-2872(13)71932-2 Dennison, S. (2010). Peer Mentoring: Untapped Potential. Journal of Nursing Education, 49(6), 340- 342. doi:10.3928/01484834-20100217-04 Ford, Y. (2015). Development of Nurse Self-Concept in Nursing Students: The Effects of a Peer-Mentoring Experience. Journal of Nursing Education, 54(9), S107- S111. doi:10.3928/01484834-20150814-20 Dennison, S. (2010). Peer Mentoring: Untapped Potential. Journal of Nursing Education, 49(6), 340- 342. doi:10.3928/01484834-20100217-04 Ford, Y. (2015). Development of Nurse Self-Concept in Nursing Students: The Effects of a Peer-Mentoring Experience. Journal of Nursing Education, 54(9), S107- S111. doi:10.3928/01484834-20150814-20 Svellingen, A., Røssland, A. & Røykenes, K. (2021). Students as Facilitators: Experiences of Reciprocal Peer Tutoring in Simulation-Based Learning. Clinical Simulation in Nursing, 54, 10-16. https://doi.org/10.1016/j.ecns.2021.01.008. Freire, T., & Beiramar, A. (2017). Tutorias por pares: acolher, promover e potenciar os estudantes do ensino superior. In L. S. Almeida & R. V. Castro (Eds.), Ser estudante do ensino superior: as respostas institucionais à diversidade de públicos. Centro de Investigação em Educação. Tabloski, P. A. (2016). Setting the stage for success: mentoring and leadership development. Journal of Professional Nursing, 32(5), S54-S58. doi:10.1016/j.profnurs.2016.03.003 Hall, R., & Jaugietis, Z. (2011). REFERÊNCIAS BIBLIOGRÁFICAS Hawkins, J. W., & Fontenot, H. B. (2010). Mentorship: the heart and soul of health care leadership. Journal of Healthcare Leadership, 2, 31-34. Aponte, J., Figueroa, B. N., Madera, M., Campos- Dominguez, G., Panora, E., & Jaramillo, D. (2015). Mentoring Hispanic Undergraduate and Graduate Research Assistants: Building Research Capacity in Nursing. Journal of Nursing Education, 54(6), 328-334. doi:10.3928/01484834-20150515-03 James, A., Smith, P., & Radford, L. (2014). Becoming grown-ups: a qualitative study of the experiences of peer mentors. Pastoral Care in Education, 32(2), 104- 115. doi:10.1080/02643944.2014.893008 Kim, S., Jillapali, R. & Boyd, S. (2021). Impacts of peer tutoring on academic performance of first-year baccalaureate nursing students: A quasi-experimental study. Nursing Education Today, 96, 1046-58. Bardin, L. (2018). Análise de Conteúdo. Edições 70. Tutoria por pares na educação em enfermagem: a voz dos tutores Tutoria por pares na educação em enfermagem: a voz dos tutores CONCLUSÃO Agradecemos ao Centro IDEA UMinho, designadamente ao Programa de Apoio a Projetos de Inovação e Desenvolvimento do Ensino e da Aprendizagem, o financiamento do projeto TutorParE onde este estudo se insere. A análise dos resultados demonstrou a importância da participação de estudantes de anos mais avançados na tutoria académica dos seus pares, no sentido em que melhora o conhecimento e as competências dos 60 RIIS RIIS Revista de Investigação & Inovação em Saúde RIIS Revista de Investigação & Inovação em Saúde Bardin, L. (2018). Análise de Conteúdo. Edições 70. Developing Peer Mentoring through Evaluation. Innovative Higher Education, 36(1), 41-52. doi:10.1007/s10755-010- 9156-6 Torres, P., & Irala, E. (2014). Aprendizagem colaborativa: teoria e prática. In P. Torres (Ed.), Complexidade: Redes e Conexões na Produção do Conhecimento (pp. 61-93). Curitiba: SENARP 61 RIIS Revista de Investigação & Inovação em Saúde
https://openalex.org/W1491840665	https://dspace.mit.edu/bitstream/1721.1/96232/2/3148.1996.603430.pdf	English	null	Rigid-Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading	Shock and vibration	1,996	cc-by	7,998	MIT Open Access Articles Rigid-Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation: Hoo Fatt, Michelle S., Tomasz Wierzbicki, Minos Moussouros, and John Koenig. “Rigid- Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading.” Shock and Vibration 3, no. 3 (1996): 169–181. © 1996 by John Wiley & Sons, Inc. Rigid-Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation: Hoo Fatt, Michelle S., Tomasz Wierzbicki, Minos Moussouros, and John Koenig. “Rigid- Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading.” Shock and Vibration 3, no. 3 (1996): 169–181. © 1996 by John Wiley & Sons, Inc. Citation: Hoo Fatt, Michelle S., Tomasz Wierzbicki, Minos Moussouros, and John Koenig. “Rigid- Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading.” Shock and Vibration 3, no. 3 (1996): 169–181. © 1996 by John Wiley & Sons, Inc. As Published: http://dx.doi.org/10.3233/SAV-1996-3303 Publisher: Hindawi Publishing Corporation Persistent URL: http://hdl.handle.net/1721.1/96232 Version: Final published version: final published article, as it appeared in a journal, conference proceedings, or other formally published context Version: Final published version: final published article, as it appeared in a journal, conference proceedings, or other formally published context Terms of use: Creative Commons Attribution Terms of use: Creative Commons Attribution Terms of use: Creative Commons Attribution Michelle S. Hoo Fatt T omasz Wierzbicki Department of Ocean Engineering Massachusetts Institute of Technology Cambridge, MA 02139 Rigid-Plastic Approximations for Predicting Plastic Deformation of Cylindrical Shells Subject to Dynamic Loading John Koenig Naval Surface Warefare Center Indian Head Division Silver Spring, MD A theoretical approach was developed for predicting the plastic deformation of a cylindrical shell subject to asymmetric dynamic loads. The plastic deformation of the leading generator of the shell is found by solving for the transverse deflections of a rigid-plastic beam/ string-on-foundation. The axial bending moment and tensile force in the beam/string are equivalent to the longitudinal bending moments and membrane forces of the shell, while the plastic foundation force is equivalent to the shell circumfer- ential bending moment and membrane resistances. Closed-form solutions for the transient and final deformation profile of an impulsive loaded shell when it is in a "string" state were derived using the eigenfunction expansion method. These results were compared to D YNA 3D predictions. The analytical predictions of the transient shell and final centerline deflections were within 25% of the DYNA 3D results. © 1996 John Wiley & Sons, Inc. Received July 18, 1995; Accepted December 5, 1995. Shock and Vibration, Vol. 3, No.3, pp. 169-181 (1996) © 1996 by John Wiley & Sons, Inc. THEORETICAL FORMULATION Consider a long cylindrical shell of thickness n and radius R as shown in Fig. 1. The cylinder is subjected to an inward radial pressure pulse p(x, 8, t) only on the upper half of the cylinder. The load is asymmetric, but it has two planes of symmetry at x = 0 and 8 = O. The load intensity is such that the shell experiences localized plastic deformation, i.e., the maximum extent in the hoop direction is of the order of the shell radius or smaller, and the axial distribution of the shell deformation spans over a few shell radii. In addi- tion to the inward radial pressure are radial shear Material Idealization In the theoretical formulation of the problem, the material is assumed to be isotropic, time-indepen- dent, and rigid-plastic. Neglect of elasticity and strain rate tends to increase the value of the flow stress U o of most ductile materials, and the rigid- plastic approximation provides an upper bound for shell deformations. Material strain hardening may be taken into account by using the concept of an average flow stress U o ' which lies somewhere between the yield and ultimate strength. (The flow strength is calculated by requiring equal areas under the actual material stress-strain curve and the ap- proximate rigid-plastic stress-strain curve.) When the material undergoes considerable work hardening, the flow stress represents a constant, elevated stress corresponding to an average strain eav during the loading process, U o = u(eav). An average strain during the loading process is then evaluated from the solution for the deformation based on the assumed flow stress. If the calcu- lated average strain corresponds to the assumed flow stress, then the rigid-plastic solution is con- sistent with the assumed flow stress. If both are not consistent with each other, the flow stress that corresponds to the average strain is evaluated and used to derive a new solution for shell defor- mation. The iterative process is repeated until the average strain and flow stress are in agreement with the material stress-strain relation. In developing the theoretical model for the shell, we assume a material that is isotropic, rate independent, and rigid-plastic. A general method- ology for reducing the 2-D shell problem sub- jected to dynamic pressure loads into a I-D prob- lem involving a rigid-plastic beam/string resting on a rigid-plastic foundation and subjected to equivalent line load is first derived in this article. This is then followed by the solution for a special case of an impulsively loaded shell using the eigenfunction expansion method. Finally, the an- alytical solutions for the transient and final defor- mation of the impulsively loaded shell are com- pared to DYNA 3D predictions. INTRODUCTION The beam-on-foundation model for the shell has been used to model shells undergoing large plastic deformation and gave useful results in applications such as the pinching of tubes (Reid, 1978), the denting of tubes (Wierzbicki and Suh, 1988), and projectile impact into thin cylin- drical shells (Yu and Stronge, 1990). anism of a ring in plane strain. The circumferential distribution of the ring deformation is assumed to be the same for each cross section of the shell, and the function used to describe it yields "equiv- alent functions," i.e., average values of shell quantities found by integration with respect to the circumferential coordinate. It will be shown subsequently that by deriving equivalent shell functions, one may represent the longitudinal bending moment and membrane force of the shell as the axial bending moment and tensile force in a beam/string, and the circumferential bending moment and membrane compression as a plastic foundation that supports the beam/string. The equivalent functions are functions only of axial direction, and thus enable us to solve the two- dimensional (2-D) shell as a leD beam/string-on- foundation. The beam-on-foundation model for the shell has been used to model shells undergoing large plastic deformation and gave useful results in applications such as the pinching of tubes (Reid, 1978), the denting of tubes (Wierzbicki and Suh, 1988), and projectile impact into thin cylin- drical shells (Yu and Stronge, 1990). As a result of these loads, the shell undergoes deformation u(x, 8, t), where x, 8 denotes the axial and circumferential coordinates and t de- notes time, and rigid body displacement in the radial direction w lend and axial direction u lend' and rotation w I lend at the ends. INTRODUCTION for finding the transient deformation of the shell, and the analytical model will be compared with DYNA 3D numerical predictions produced by Moussouros and Koenig (1994). The particular example chosen for comparing the analytical so- lution with DYNA 3D predictions is an aluminum shell subjected to asymmetric impulsive loading. The objective of this study is to develop a gen- eral approach for predicting the plastic deforma- tion of a cylindrical shell subject to dynamic load- ing. The shell is subject to a "side-on" pressure load (i.e., one in which the loading is asymmetric in the circumferential direction). Previous analyt- ical attempts by Witmer et al. (1960) and Green- spon (1970) to solve this problem resulted in closed-form expressions for the final deformation and response time of the shell, but these solutions do not give the transient deformation of the shell. Several commercially available numerical codes (see for example, Underwood, 1972; Stricklin et aI., 1974; Jiang and Olson, 1991) are also available It is well known that the coupled nonlinear partial differential equations that govern shell de- formation are mathematically intractable if the loading to the shell is asymmetric because deriva- tives with respect to the circumferential direction must be retained. In the following theory, the distribution of the shell deformation in the cir- cumferential direction is specified by considering a kinematically admissible plastic collapse mech- CCC 1070-9622/96/030169-13 169 Hoo Fat! et al. 170 force F,., axial force Fr , and bending moment T at the ends of the shell. anism of a ring in plane strain. The circumferential distribution of the ring deformation is assumed to be the same for each cross section of the shell, and the function used to describe it yields "equiv- alent functions," i.e., average values of shell quantities found by integration with respect to the circumferential coordinate. It will be shown subsequently that by deriving equivalent shell functions, one may represent the longitudinal bending moment and membrane force of the shell as the axial bending moment and tensile force in a beam/string, and the circumferential bending moment and membrane compression as a plastic foundation that supports the beam/string. The equivalent functions are functions only of axial direction, and thus enable us to solve the two- dimensional (2-D) shell as a leD beam/string-on- foundation. Strain-Displacement Relations If we confine our analysis to moderately large deflection and small strains, the Lagrangian de- scription of the axial strain and curvature rates are (Brush and Almroth, 1975) 8xx = u' + w'w', Kxx = -w", (7) (8) (7) (7) (8) The bending moments M,,(3 and membrane forces N,,(3 are coupled through a yield condition, (8) (3) (3) (3) where ( )' denotes afax. where ( )' denotes afax. which is assumed to be a plastic potential for the generalized strain rates Dynamic Equilibrium The overall shell equilibrium is expressed via the principle of virtual velocities D == f uUe;jdV = f pw dS - J pww dV, (1) v s v (1) where ( . ) denotes a/at, D is the rate of plastic work dissipated, uij and sij denote stress and Plastic Deformation of Cylindrical Shells 171 ~ P (X,e, t) Fr F x ~ ~...L-l----L---L-.J'--L---L....O>---, I}/ x Fr x FIGURE 1 Geometry of and loading on the cylindrical shell. FIGURE 1 Geometry of and loading on the cylindrical shell. strain rate, w is the displacement vector, and p is a generalized surface traction. In shell coordinates for which a differential shell element is dS = dx R de, the above equation is expressed as Previous analyses by Wierzbicki and Suh (1988) and Moussouros and Hoo Fatt (1995) showed that the contribution of plastic work asso- ciated with shear deformation is 10-15% of the total plastic work dissipated in tubes subject to transverse "knife" loading. As a first-order ap- proximation, assume that MxoKxo = N¥08xO = 0, so that (he rate of plastic dissipation simplifies to 2 r 2R r (M,,(3K,,(3 + N,,(38,,(3) de dx = 2 r 2R r Pow de dx + 2· 2R r Fxulend de + 2· 2R r Frwlend de + 2· 2R f7T Tw' lend de o 0 - 2 fog 2R r m(iiu + vv + ww') de dx, (2) (6) (6) (6) where g is the extent of plastic deformation, [ex, m = [x, e]; M,,(3 and N,,(3 are the corresponding tensors of the bending moment and membrane force; K"(3 and 8"(3 are curvature and strain rates; the velocity vector w[u, v, w] corresponds to the x, e, r axis; p is a vector of surface tractions with components p[O, 0, Po] in the x, e, r direction; and m = ph is the mass per unit shell area. Equivalent Functions Substituting Eqs. (7) and (8) into Eq. (2) gives the following statement of dynamic equilibrium: . \ af K,,(3="-M ' a a(3 . af Sa(3=A- N ' a a(3 (4) (4) -2 fg 2R r Mxxw" de dx o 0 -2 fg 2R r Mxxw" de dx o 0 where A is a proportionality constant. An expression of the rate of plastic work dissi- pated in the shell b is Hoo Fatt et al. 172 = 2 f g 2R fIT pow dO dx + 2· 2R r F,ulend dO o 0 0 + 2· 2R fIT Frwlend dO + 2· 2R r Tw'lend dO o 0 - 2 f 2R r m(iiu + vi; + ww) dO dx. (9) o 0 (16) (16) an equivalent applied bending moment, T· w ~ lend = 2R r Tw'lend dO. o (17) (17) (9) All equivalent functions depend on variables x and t; they are a consequence of the dynamic response of the shell. Even the equivalent mass varies with position and time because of varying inertia forces induced by the shell motion. Assume: tangential deformations are negligi- ble, i.e., v = 0; there is no warping, i.e., u is independent of 0; and cross-sections of the de- forming shell are similar and correspond to a plas- tically deforming ring in plane strain whose defor- mation field may be described in terms of its centerline deflection w(x, 0 = 0, t) = wjx, t). For a given deformation field, each term in Eq. (9) may be integrated with respect to the circumfer- ential coordinate to give the following expression for deriving equivalent functions: an equivalent line load, Introducing equivalent functions into Eq. (9) gives f g - - 2 0 [-Mw~ + qwo + 21TRNxxu~ + Nw ~ w~l dx = 2F,uolend + 2Frwo lend + 2Tw~lend + 2 f pWo dx (18) - 2 f [21TRmiiouo + mWowol dx. (18) p(x, t). Wo = 2R r Pow(x, 0, t) dO, (10) Integrating Eq. (18) by parts, one gets Integrating Eq. (18) by parts, one gets (10) (Rw'-M'-Fr)wolends + Lg[inWo + MI/ -(Nw~)' + q-plwo dx+ (21TRNxx - FJuolends f g - - + 2R1T( - N'xx + mUo)uo dx + (M - T)w ~Iends = 0 o (19) (Rw'-M'-Fr)wolends + Lg[inWo + MI/ an equivalent mass per unit length, m(x, t). String-on-Foundation When the plastic deformation of an axially re- stained shell are finite, axial membrane forces can no longer be ignored. The axial bending moment dominates the shell response during infinitesimal deflection, but it becomes less significant when compared to the axial membrane force as the shell deflections increase. Haythornwaite (1961) dem- onstrated that a fully clamped rigid-plastic beam will enter a membrane state when the beam de- flections are of the order of the thickness of the beam. A similar phenomenon is assumed to take place for the rigid-plastic shell. Neglecting the axial bending moment in Eq. (24) gives p An equivalent impulse load I is then given by I(x) = f' 2R r p(x, 8, t)w(x, 8, t) d8 dt (23) (23) Equivalent Functions Wo Wo = 2Rm r ww(x, 0, t) dO, o (11) -(Nw~)' + q-plwo dx+ (21TRNxx - FJuolends f g - - + 2R1T( - N'xx + mUo)uo dx + (M - T)w ~Iends = 0 o (19) an equivalent ring crushing resistance, (19) q(x, t) . Wo = 2R r (Meei<ee + Neeeee)(x, 0, t) dO, (12) q(x, t) . Wo = 2R r (Meei<ee (12) q(x, t) . Wo = 2R r (Meei<ee The above equations are valid for all virtual veloc- ities. Therefore, (12) (12) + Neeeee)(x, 0, t) dO, mwo + MI/ - (Nw~)' + q = p (20) (20) an equivalent axial bending moment, with boundary conditions Nw' - M' = Fr or Wolend = 0 and M = Tor W~lend = 0, and M(x, t) . w~ = 2R r Mu wl/(x, 0, t) dO, (13) (21) (21) an equivalent axial membrane force, with boundary condition 21TRNxx Fx or ito lend = O. N(x, t)· w~w~ = 2R r Nxxw'w'(x, 0, t) dO, (14) Equations (20) and (21) describe the equations of motion in the radial direction of the leading generator (0 = 0) and axial directions, respec- tively. Each equation is subjected to either force or displacement boundary conditions; i.e., mo- ments and equivalent shear forces or slopes and radial deflections are specified in Eq. (20), while an equivalent applied axial force, Fx· Uolend = 2R r Fxulend dO, (15) o (15) (15) an equivalent applied shear force, an equivalent applied shear force, 173 Plastic Deformation of Cylindrical Shells axial forces or axial displacements are specified in Eq. (21). The above equation of motion is recognized as the rigid-plastic beam equation with finite deflec- tions. No rigorous theoretical methods have been proposed for solving finite deflections of a rigid- plastic beam. Several approximate solutions, which are in good agreement with experimental results, were proposed by Jones (1971) and Vaziri et al. (1987). Rigorous analytical solutions for lim- iting cases of Eq. (24) do exist and these will be explained below. Impulsive loading In elasticity, the pressure loading can be ex- pressed as an impulsive loading if the pulse dura- tion of the pressure load impinging on the shell is much less than the fundamental period of vi bra- tion of the shell. In plasticity, the pressure loading is considered as an impulsive loading if the pulse duration is much shorter than the response time of the shell. (Vibrations do not occur because plastic work is dissipated instead of being stored as in an elastic body.) The magnitude and distribu- tion of the initial velocity of the shell is governed by the impulse intensity, the shell impedance (ph), and the shell geometry, outer radius, and length. Beam-on-Foundation When deflections are infinitesimal, the equivalent axial membrane force is negligible compared to the equivalent axial bending moment. Omitting the (Nw~)' term in Eq. (24) gives (25) To find the corresponding impulsive loading I(x, 8), the pressure pulse is integrated in time and the loading to the shell is introduced as an initial shell velocity Vo(x, 8). The magnitude of the initial velocity is calculated from the transfer of linear momentum to the shell, (25) subject to the boundary conditions Nw~ - M' = Fr and M = T. Yu and Stronge (1990) used the above equation to derive solutions for the tran- sient deformation of cylinders subject to projec- tile impact. When shell deformations become large, they included a membrane factor to account for axial membrane forces in an approximate way. f' p(x, 8, t) dt = I(x, 8) = mVo(x, 8), (22) (22) where I is the specific impulse with units pressure time (not to be confused with an impulse with units force time) and m = ph is the mass per unit area or material impedance. Neglect of Axial Deformation If we confine our analysis to a plastic foundation with an infinite shear resistance but finite com- pressive resistance q, u = 0 and consequently, if = u = O. Setting if = 0 in Eq. (19) gives only an equation of motion for the radial deflection of the leading generator mw" - (Nw~)' + q = p, (26) (26) subject to the boundary condition Nw ~ = Fr. In the following sections, we present a solution for the cylinder in a purely membrane state and com- pare the analytical predictions to DYNA 3D re- sults. mwo + Mil - (Nw~)' + q = p, (24) (24) with boundary conditions Nw~ - M' = Fr and M=T. with boundary conditions Nw~ - M' = Fr and M=T. 174 Hoo Fatt et al. (32) STRING-ON-FOUNDATION SUBJECT TO IMPULSIVE LOADING (32) with boundary conditions with boundary conditions As an example, consider a fully-clamped cylinder of finite length 2L. The cylinder is subject to asymmetric impulsive loading. The circumferen- tial distribution of the impulsive load is a cosine function on the upper half of the shell circumfer- ence. Two impulsive load distributions in the axial direction are considered, a parabolic distri- bution and a uniform distribution. The shell un- dergoes radial deformation w(x, (), t), where x and () denote the axial and circumferential coordinates and t denotes time. Wi = 0 at x = 0 and (33) Wi = 0 at x = 0 (33) Wi = 0 at x = 0 (33) and W = 0 at x = 1, (34) (34) and initial condiions W = 0 at I = 0 (35) W = 0 at I = 0 (35) and Based on the string-on-foundation analogy for the shell, the equation for the plastic deformation of the leading generator of the shell «() = 0) under impulsive loading is Wi = Vex) at I = 0, (36) (36) where the normalized velocity is V (eDi)· where the normalized velocity is V (eDi)· Eigenfunction Expansion cos An = 0 or An = (211 - 1)rr/2, for n = 1, 2, 3,. . . (44) (44) The solution is thus The present solution methodology, based on the eigenvalue expansion method, covers a fixed length of the beam and precludes the application of the rigorous unloading analysis. Instead, an approximate and more restrictive unloading crite- rion is proposed. The eigenfunction coefficients are found from the initial conditions, Egs. (41) and (42), The eigenfunction coefficients are found from the initial conditions, Egs. (41) and (42), ~ Bn COS(AnX) = !(1 - X2) (46) n~l (46) p p The solution in the loading region, Eg. (50), is expressed in terms of an infinite series of modal function COS(AnX) and variable coefficients wn(i) and and ~ AnA" COS(AnX) = iJ(x). (47) n~l x w(x, l) = ~ wn(i) COS(AnX). (51) n~l (47) x w(x, l) = ~ wn(i) COS(AnX). (51) n~l (51) The value of Bn is For impulsive loading with expansion coefficients specified by Egs. (48) and (49), the amplitudes wn(i) are diminishing functions oftime. Each am- plitude reaches a maximum value when the corre- sponding velocity vanishes (48) (48) The value of An depends on the axial distribu- tion of the impulse, !(X) , (52) (52) ifj(x) = 1 - X2 (49) if!(x) = 1. It can be shown that higher modes decay more rapidly than the lower ones. It is assumed that each mode contributes to the final deflection of the structure only in the time 0 < i < in' (49) { [An sin(Ani) + Bn[COS(Ani) - 1]], wn(i) = forO < i < in o for i > in' (53) Using!(1 - X2) = L~~l Bn COS(AnX), we can also express Eg. (45) as (53) W = ~ [An Sin(AJ) + Bn[COS(Ani) - 1]] COS(AnX). (50) n=l where in is defined by Eg. (52). According to the above criterion, all modes contributes to the shell response early on. Later only the lower modes survive and the motion ends with the fundamental mode. This criterion of progressive switching off of higher modes has many advantages. It confirms the property of mode convergence proven by Martin and Sy- monds (1966) for general rigid, perfectly plastic structures. It also leads to a much desired closed- form solution for the final deflection, giving realis- tic permanent shapes of the deformed shell. Fi- Eigenfunction Expansion The shell is clamped at both ends so that Eq. (27) is subject to the boundary conditions The shell is clamped at both ends so that Eq. (27) is subject to the boundary conditions Reduce the problem into a homogeneous system of equations by assuming a solution of the form w~ = 0 atx = 0 (28) w~ = 0 atx = 0 (28) 28) (37) (28) (37) and and The homogeneous system is The homogeneous system is W 0 = 0 at x = ± L. (29) (38) (29) (38) The initial-boundary partial differential equation is also subject to the initial conditions with boundary conditions with boundary conditions <l>i = 0 at x = 0 (39) (39) Wo = 0 at! = 0 (30) (31) Wo = 0 at! = 0 (30) (30) and and and <I> = 0 at x = 1, (40) (31) <I> = 0 at x = 1, (40) (31) (40) and initial conditions and initial conditions where Vo(x) is the initial velocity distribution along the length of the shell. (41) (41) It is assumed for simplicity that equivalent functions are constant. The initial-boundary value problem can then be expressed in terms of the following normalized variables: and and <l>i = H(x) at I = 0, (42) (42) 1. x = xlL, axial coordinate 1. x = xlL, axial 2. I = telL, time where j is the amplitude of the dimensionless impUlse andf(x) is the axial distribution of the im- pulse. 2. I = telL, time 2. I = telL, time 3. W = woN/L2q, transverse deflection p Equation (39) is automatically satisfied if where e = YNlm is the plastic wave speed in the string. Denoting derivatives with respect to the nor- malized variables gives oc <I> = L [An sin(Ani) + Bn COS(Ant)] cos (An x). (43) n=\ (43) Plastic Deformation of Cylindrical Shells 175 The eigenvalues are determined from condition (40) so that 0, ei 7"- 0). Full unloading takes place when both components are zero. An exact unloading analysis was performed by Suliciu et al. (1995) in a related problem of an infinite string on a plastic foundation loaded im- pulsively. Using the method of characteristics, several regions were identified in the phase plane (x, 1). The analysis was complicated and involved propagation of rigid zones into an already de- formed plastic string. Unloading and Final Deformation The generalized strain rates, i.e., strain rate and velocity rate, in the string-on-foundation model are ei, Wi, where Si = Wi W ii. The unloading condi- tion is formulated in the space of the strain rate vector. One can distinguish two cases of partial unloading and full unloading. Partial unloading occurs when only one component of the general- ized strain rate vanishes (e t = 0, Wi 7"- 0 or Wi = Hoo Fatt et al. 176 nally, it eliminates formation and propagation of rigid zones because all points of the structure are brought to rest at the same time in the terminal phase of the shell motion. The unloading criterion was first formulated by Wierzbicki (1972) for vis- coplastic structures, and later modified by Wierz- bicki (1974) for rigid, perfectly plastic structures. The solution with an initial velocity distribution that is parabolic converges more rapidly than that with a uniform distribution. The rate of conver- gence for the parabolic distribution is of the order l/A~ while that for the uniform is only of the order of IIA~. Simple closed-form expressions for very large impulses may be derived for the series solu- tions in Eqs. (60) and (61). g p y p Differentiating Eq. (50) with respect to i gives x Wi = 2: [AnAn cos(Anl) - BnAn sin(Ani)] cos(Anx). n~1 (54) x Wi = 2: [AnAn cos(Anl) - BnAn sin(Ani)] cos(Anx). n~1 (54) Setting Eq. (54) equal to zero signifies that each mode unloads when Approximation Solutions Previous analysis shows that a one-term approxi- mation for the central deflection of the shell is within 5% of an eigenfunction solution if 1 > 1. (The eigenfunction solution was set to be within a convergence tolerance of 10-6 , see Liao, 1993.) Assuming a one-term approximation, we get Al = 7T12 and (55) (55) or at a characteristic time in or at a characteristic time in (56) (56) The above unloading criterion satisfies both con- ditions (ei = 0 and Wi = 0). A closed-form expres- sion for the final deformation profile is obtained by using trigonometric relations to give The above unloading criterion satisfies both con- ditions (ei = 0 and Wi = 0). A closed-form expres- sion for the final deformation profile is obtained by using trigonometric relations to give The corresponding closed-form expression for the final central deflection is (57) (57) The series solution in Eq. (61) does not con- verge as rapidly as that for a parabolically distrib- uted impulse. However, for very large impUlses, the term V(A~j2 + 1) - 1-'> A), and Eq. (61) may be rewritten as and and (58) (58) (64) (64) Substituting these into Eq. (50) gives an expres- sion for the final deformation profile When 1 > 10, an approximation to the 50-term series solution in Eq. (64) is When 1 > 10, an approximation to the 50-term series solution in Eq. (64) is x wf = 2: [V(A~ + B~) - Bnlcos(Anx). (59) n~1 (59) (65) (65) For the parabolic distribution, the above equation reduces to and the central deflection is It is interesting to note that when the above ex- pression is rewritten in unnormalized quantities, the solution for Of is independent q. It is interesting to note that when the above ex- pression is rewritten in unnormalized quantities, the solution for Of is independent q. for the uniform distribution, we get 177 Plastic Deformation of Cylindrical Shells SUMMARY OF DYNA 3D RESULTS Moussouros and Koenig (1994) have produced DYNA 3D solutions for the impulsively loaded 6061-T6 aluminum shell using the Belytschko- Tsai 5 degree-of-freedom elements with no in- plane torsional components. Equivalent functions m, N, and q depend on the mode of plastic collapse of the ring. Cline and lahsman (1967) found that a ring subjected to a cosine impulse distribution over its upper half collapses in two stages: a "short-time response," when the plastic work dissipated is predominantly due to membrane compression, and a "long-time response," when the plastic work dissipated is predominantly due to bending at plastic hinges. As shown in Fig. 2, the ring is a membrane mode of plastic collapse during the short-time response and a bending mode during the long-time re- sponse. The transition from one mode to the other and lor the interaction of both modes were not addressed by Cline and lahsman, and will be a topic for future research. The shell geometry was assumed to be a per- fectly circular cylinder loaded with an asymmet- ric radial impUlsive load. The (T - e relationship of 6061-T6 Al material was assumed to be bilinear, i.e., linear elastic, linear strain hardening with the following material properties: 1. E = 10.8(106) psi, Young's modulus 2. (Ty = 41,600 psi, yield strength 3. Ep = 161,000 psi, linear strain-hardening modulus p For the particular load cases considered in the DYNA 3D analysis, the ring collapses in the mem- brane mode. A flow stress of (To = 45,000 psi is used in evaluating equivalent functions for the 6061-T6 aluminum shell in Appendix A. The cor- responding values of the equivalent functions are 4. p = 2.6(10-4) lb m/in. 3 , density. The bilinear approximation for the (T - e curve is an idealization of the actual (T - e. For ductile materials, the plastic modulus decreases with in- creasing strain and the material fractures at a fi- nite value of strain. The plastic strain was ex- tended to 80% (corresponding to a stress of 170,400 psi), even though the shell would fracture before attaining a strain at this value. No fracture criterion was introduced in the numerical simula- tion because the sole purpose of the numerical exercise was to test the rigid-plastic prediction of the large plastic deformation of the shell. 1. m = 6.1(10)-4Ib m/in., equivalent mass 2. N = 106,030 lb, equivalent tensile force 3. SUMMARY OF DYNA 3D RESULTS q = 22,500 lb/in., equivalent ring crush- ing resistance. 1. m = 6.1(10)-4Ib m/in., equivalent mass 2. N = 106,030 lb, equivalent tensile force 3. q = 22,500 lb/in., equivalent ring crush- ing resistance. 1. m = 6.1(10)-4Ib m/in., equivalent mass 2. N = 106,030 lb, equivalent tensile force 3. q = 22,500 lb/in., equivalent ring crush- ing resistance. The permanent centerline deflection of the shell for all five numerical test cases are compared to the rigid-plastic approximations of them in Table 2. In all cases the analytical predictions are within 25% of numerical results. Test 1 in- volving a parabolic load distribution of Vo = 14,505 in./s has a much higher deformation than Test 4, involving a uniform load distribution with the same total initial impulse. This was expected because the parabolic load is greatest at the cen- terline. Several test cases were examined and these are summarized in Table 1. In all cases the peak velocity was 9,670 in./s, except for the first test case where it was set to be 50% higher than the rest, i.e., Vo = 14,505 in./s. The first test case has a parabolically distributed impUlse and setting the peak velocity at 14,505 in./s results in the same "total" impulse (area under the mass times velocity curve) for Tests 1 and 4. Table 1. Description of Numerical Test Cases Radius Thickness Half-Length Test R (in.) h (in.) L (in.) 1 6 0.25 2 2 6 0.25 2 3 6 0.25 4 4 6 0.25 2 5 6 0.25 4 Axial Load Impulse Velocity Dlh Distribution Vo (in. Is) 48 Parabolic 14,505 48 Parabolic 9,670 48 Parabolic 9,670 48 Uniform 9,670 48 Uniform 9,670 Table 1. Description of Numerical Test Cases 178 Hoo Fatt et al. Membrane Mode (short-time response) Bending Mode (long-time response) FIGURE 2 Plastic collapse of ring in membrane and bending modes. Membrane Mode (short-time response) Bending Mode (long-time response) Bending Mode (long-time response) Membrane Mode (short-time response) FIGURE 2 Plastic collapse of ring in membrane and bending modes. Calculations show that, for the given diameter to thickness ratio of the shells considered by DYNA 3D, the equivalent ring crushing resis- tance, i.e., the equivalent plastic foundation force, in a bending mode is 2 orders of magnitude less than that in a membrane mode. SUMMARY OF DYNA 3D RESULTS An equivalent ring crushing resistance that assumes both mem- brane and bending modes would therefore be lower than one calculated with only a membrane mode. This explains why analytical predictions were less than the numerical solutions. In calcu- lating equivalent functions, the bending mode during the long-time response of the shell was ignored, and the resulting foundation force is stif- fer. A plastic collapse ring model that assumes both membrane and bending modes should bring analytical predictions closer to the numerical re- sults. in the circumferential and longitudinal directions at t = 46.9 fLS and 92.8 fLS, respectively, are within 5% of the DYNA 3D predictions. To compare analytical and numerical predictions of the final deformed profiles, it was assumed that the elastic vibrations in the DYNA 3D analysis attenuated completely at t = 138.6 fLS. The rigid-plastic ap- proximation for the final deformation is within 25% of the DYNA 3D prediction of the shell de- formation at t = 138.6 fLS. A comparison between the rigid-plastic ap- proximation and the numerical solution of the transient response of the maximum centerline de- flection for Tests 2 and 4 are also shown in Fig. 5. Plastic unloading in the rigid-plastic approxi- mation occurs near the first overshoot of the elas- tic shell vibrations. The predicted plastic re- sponse times of the shell are within 10% of the numerical predictions (measured by the time at the first overshoot of the elastic shell vibrations). The analytical predictions of the transient and final deflection profiles of the shell in the circum- ferential and longitudinal direction for Test 4 are compared to the numerical predictions in Figs. 3 and 4. The shell undergoes about 10% circumfer- ential compression on the upper half of its circum- ference so that a considerable amount of plastic work is dissipated during hoop compression. The analytical predictions of the transient deflections CONCLUDING REMARKS A theoretical approach for predicting the plastic deformation of a cylindrical shell subject to asym- metric dynamic loads was developed and com- Table 2. Comparison of DYNA 3D and Analytical Results Test j = VoNlcLq Analytical or = BfL2qlN (in.) 1 2.61 1.08 2 1.74 0.62 3 0.87 0.86 4 1.74 0.75 5 0.87 1.06 Numerical of (in.) % U nderprediction 1.20 -10 0.77 -19 1.14 -25 0.97 -23 1.30 -18 Table 2. Comparison of DYNA 3D and Analytical Results Plastic Deformation of Cylindrical Shells 179 t = 92.8 microsec [in] -2 - - - - Rigid-Plastic Approximation -- DYNA 3D Prediction -4 -60~~--------~----------~10~--------~15 [in] FIGURE 3 Circumferential deflection profile at x = o of the shell. Test 4: uniform velocity = 9670 in. Is, 2R = 12 in., 2L = 4 in., h = 0.25 in. t = 92.8 microsec [in] -2 - - - - Rigid-Plastic Approximation -- DYNA 3D Prediction -4 -60~~--------~----------~10~--------~15 [in] 1.6 1.4 1.2 ~1 ~ ~ 0.8 ~ is ~0.6 a: - - - - Rigid-Plastic Approximation -- DYNA 3D Prediction t = 138.6 microsec ------~ 7 = - = = ~ , " ~nal Deformation t = 92.8 microsec ' : : - -\ , _ 0.4 ~-~-~-~-~-~-~-c;=\~--.--=--~-~--~-~-~- - - - ~ ~ _ t = 46.9 microsec 0.2 , , , oL-~ __ ~ __ ~ __ ~ __ ~ __ ~ __ L-~ __ ~ __ ~ o 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Axial Length [in] FIGURE 4 Transient and final deflection profile of the leading generator of the shell. Test 4: uniform veloc- ity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. 1.6 1.4 1.2 ~1 ~ ~ 0.8 ~ is ~0.6 a: - - - - Rigid-Plastic Approximation -- DYNA 3D Prediction t = 138.6 microsec ------~ 7 = - = = ~ , " ~nal Deformation t = 92.8 microsec ' : : - -\ , _ 0.4 ~-~-~-~-~-~-~-c;=\~--.--=--~-~--~-~-~- - - - ~ ~ _ t = 46.9 microsec 0.2 , , , oL-~ __ ~ __ ~ __ ~ __ ~ __ ~ __ L-~ __ ~ __ ~ o 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Axial Length [in] FIGURE 4 Transient and final deflection profile of the leading generator of the shell. is described by a cosine distribution pared to numerical results from DYNA 3D. The plastic deformation of the shell was found by solv- ing for the transverse deflections of a rigid-plastic beam/string-on-foundation. As an example, the deformation of a cylindrical shell subject to impul- sive loading was predicted by finding the solution of the transient and final deformations of a string- on-foundation. Equivalent functions for the string-on-foundation were evaluated using a membrane mode plastic collapse mechanism for the shell. The analytical predictions of the center- line shell deflection underpredicted the DYNA 3D results by 25%. The discrepancy between the analytical and numerical solutions was attributed to neglecting the bending mode or long-time re- sponse of the shell. A plastic collapse ring model that assumes both membrane and bending phases would result in a reduction of the equivalent crushing resistance of the shell and should in- crease analytical predictions, bringing them closer to the numerical results. Analyzing plastic collapse of a ring in combined membrane and bending modes will be the subject of future re- search. pared to numerical results from DYNA 3D. The plastic deformation of the shell was found by solv- ing for the transverse deflections of a rigid-plastic beam/string-on-foundation. As an example, the deformation of a cylindrical shell subject to impul- sive loading was predicted by finding the solution of the transient and final deformations of a string- on-foundation. Equivalent functions for the string-on-foundation were evaluated using a membrane mode plastic collapse mechanism for the shell. The analytical predictions of the center- line shell deflection underpredicted the DYNA 3D results by 25%. The discrepancy between the analytical and numerical solutions was attributed to neglecting the bending mode or long-time re- sponse of the shell. A plastic collapse ring model that assumes both membrane and bending phases would result in a reduction of the equivalent crushing resistance of the shell and should in- crease analytical predictions, bringing them closer to the numerical results. Analyzing plastic collapse of a ring in combined membrane and bending modes will be the subject of future re- search. (A. I) (A. I) (A. I) Velocity and accelerations fields are similarly de- scribed by cosine distributions, w(6) = wocos 6 and w(6) = wocos 6. CONCLUDING REMARKS Test 4: uniform veloc- ity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. FIGURE 3 Circumferential deflection profile at x = o of the shell. Test 4: uniform velocity = 9670 in. Is, 2R = 12 in., 2L = 4 in., h = 0.25 in. is described by a cosine distribution To evaluate the fully plastic bending moments and tensile forces, assume a limited interaction yield curve 1.2 l' ~ 0.4 o Test 4: Uniform Velocity = 9670 in/s, 2R = 12 in, 2L = 4 in, h = 0.25 in , , -------------------------\--r-------------------- Test 2: Parabolic Velocity = 9670 in/s, 2R = 12 in, 2L = 4 in, h = 0.25 in - - - - Rigid-Plastic Approximation -- OYNA 3D Prediction Time [sec] FIGURE 5 Transient centerline deflection of the leading generator of the shell for shells. Test 2: para- bolic velocity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. Test 4: uniform velocity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. 1.2 l' ~ 0.4 o Test 4: Uniform Velocity = 9670 in/s, 2R = 12 in, 2L = 4 in, h = 0.25 in , , -------------------------\--r-------------------- Test 2: Parabolic Velocity = 9670 in/s, 2R = 12 in, 2L = 4 in, h = 0.25 in - - - - Rigid-Plastic Approximation -- OYNA 3D Prediction Time [sec] APPENDIX A: EQUIVALENT FUNCTIONS Equivalent functions for the impulsively loaded shell are calculated for the membrane mode plas- tic collapse mechanism shown in Fig. 2. FIGURE 5 Transient centerline deflection of the leading generator of the shell for shells. Test 2: para- bolic velocity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. Test 4: uniform velocity = 9670 in./s, 2R = 12 in., 2L = 4 in., h = 0.25 in. Following Cline and lahsman (1967), the defor- mation in the upper half of the shell circumference Hoo Fatt et al. 180 Laboratory, J. G. Eng. Res Assoc., Technical Report No. 10, February. Laboratory, J. G. Eng. Res Assoc., Technical Report No. 10, February. where Mp[ = U'oh2/4 is the fully plastic bending moment per unit length and Np[ = U'oh is the fully plastic axial force per unit length. Haythornwaite, R. M., 1961, "Mode Change During the Plastic Collapse of Beams and Plates," in Devel- opments in Mechanics, Proceedings of the 7th Mid- western Mechanics Conference,Vol. 1, J. E. Lay and L. E. Malvern, pp. 203-215. p p g With the above distribution, Eq. (A. I), and value for N p[' the equivalent mass and tensile force, defined by Eqs. (11) and (14), are Jiang, J., and Olson, M. D., 1991, "Nonlinear Dynamic Analysis of Blast Loaded Cylindrical Shell Struc- tures," Computers & Structures, Vol. 41, No.1, pp. 41-52. m = 2Rm r cos2e de = nRph/2 equivalent mass per unit length and Jones, N., 1971, "A Theoretical Study of the Dynamic Plastic Behavior of Beams and Plates with Finite Deflections," International Journal of Solids and Structures, Vol. 7, pp. 1007-1029. N = 2RNp[ r cos2e de = 7TRU'oh12 o Liao, S.-W., 1993, "Dynamic Failure of Cylindrical Shells," Final report for the Master of Science de- gree program in the Naval Architecture and Offshore Engineering Department, University of California, Berkeley. equivalent tensile force. The equivalent ring resistance depends on the plastic work dissipated in membrane compression y Martin, J. B., and Symonds, P. S., 1966, "Mode Ap- proximation for Impulsively Loaded Rigid-Plastic Structures," Journal of the Engineering Mechanics Division, Proceedings ASCE, Vol. 92, No. EM5, pp.43-66. (A.3) (A.3) Neglecting tangential component and higher or- der terms, one can approximate the hoop strain rate as {i pp Moussouros, M., and Hoo Fatt, M. APPENDIX A: EQUIVALENT FUNCTIONS S., 1995, "Effect of Shear on Plastic Denting of Cylinders," Interna- tional Journal of Mechanical Sciences, Vol. 37, No. 4, pp. 355-371. . _ {i cos e, for lei < 7T/2 e88 - 0, otherwise. (A.4) Moussouros, M., and Koenig, J., 1994, "Validation of Rigid-Plastic Model for the Dynamic Response of a 6061-T6 Aluminum Shell Subject to ImpUlsive Load- ing Using DYNA 3D," Personal Communications, NSWC, White Oak. (A.4) Substituting Eq. (A.4) into Eq. (A.3) and integrat- ing give Reid, S. R., 1978, "Influence of Geometrical Parame- ters on the Mode of Collapse of a "Pinched" Rigid- Plastic Cylindrical Shell," International Journal of Solids and Structures, Vol. 14, pp. 1027-1043. (A.S) Stricklin, J. A., Haisler, W. E., and von Riesemann, W. A., 1974, "Large Deflection Elastic-Plastic Dy- namic Response of Stiffened Shells of Revolution," Journal of Pressure Vessel Technology," May, pp. 87-95. The authors would like to thank Dr. Geoffrey Main for supporting this project under ONR Grant NOOO 14-94- 1-1026 to the Massachusetts Institute of Technology. pp g p j -1026 to the Massachusetts Institute of Technology. Suliciu, M. M., Suliciu, 1., Wierzbicki, T., and Hoo Fatt, M. S., 1996, "Transient Response of an Impul- sively Loaded Plastic String on a Plastic Founda- tion," Quarterly Applied Mathematics, to appear. pp Witmer, E. A., Herrmann, W., Leech, J. W., and Pian, Wierzbicki, T., 1974, "Application of an Eigenfunction Expansion Method in Plasticity," Journal of Applied Mechanics, Vol. 41, No.2, pp. 448-452. Wierzbicki, T. and Suh, M. S., 1988, "Indentation of Tubes Under Combined Loading," International Journal of Mechanical Science, Vol. 30, No. 3/4, pp. 229-248. Witmer, E. A., Herrmann, W., Leech, J. W., and Pian, pp Wierzbicki, T. and Suh, M. S., 1988, "Indentation of Tubes Under Combined Loading," International Journal of Mechanical Science, Vol. 30, No. 3/4, pp. 229-248. g p , International Journal of Impact Engineering, Vol. 9, pp. 115-126. T. H. H., 1960, "Responses of Plates and Shells to Intense External Loads of Short Duration," Massa- chusetts Institute of Technology Technical Report WADD TR 60-433, April. Yu, T. X., and Stronge, W., 1990, "Large Deflection of a Rigid-Plastic Beam-on-Foundation from Impact," International Journal of Impact Engineering, Vol. 9, pp. 115-126. REFERENCES Underwood, P., 1972, "Transient Response of Inelas- tic Shells of Revolution," Computers & Structures, Vol. 2, pp. 975-989. Brush, D.O., and Almroth, B. 0.,1975, "Strain Local- ization and Fracture in Metal Sheets and Thin- Walled Structures," in Buckling of Bars , Plates, and Shells, McGraw-Hill, New York. Vaziri, R., Olson, M. D., and Anderson, D. L., 1987, "Dynamic Response of Axially Constrained Plastic Beams to Blast Loads," International Journal of Solids and Structures, Vol. 23, No.1, pp. 153-174. Cline, G. B., and Jahsman, W. E., 1967, "Response of a Rigid-Plastic Ring to Impulsive Loading," Jour- nal of Applied Mechanics, Vol. 89, pp. 329-336. Wierzbicki, T., 1972, "An Approximate Linear Theory of Thin Viscoplastic Shells," Archives of Mechanics, Vol. 24, No. 5-6, pp. 941-953. Greenspon, J. E., 1970, "Theoretical Calculation of Iso-Damage Characteristics, " Ballistic Research Plastic Deformation of Cylindrical Shells 181 T. H. H., 1960, "Responses of Plates and Shells to Intense External Loads of Short Duration," Massa- chusetts Institute of Technology Technical Report WADD TR 60-433, April. T. H. H., 1960, "Responses of Plates and Shells to Intense External Loads of Short Duration," Massa- chusetts Institute of Technology Technical Report WADD TR 60-433, April. Wierzbicki, T., 1974, "Application of an Eigenfunction Expansion Method in Plasticity," Journal of Applied Mechanics, Vol. 41, No.2, pp. 448-452. pp Wierzbicki, T. and Suh, M. S., 1988, "Indentation of Tubes Under Combined Loading," International Journal of Mechanical Science, Vol. 30, No. 3/4, pp. 229-248. p Yu, T. X., and Stronge, W., 1990, "Large Deflection of a Rigid-Plastic Beam-on-Foundation from Impact," International Journal of Impact Engineering, Vol. 9, pp. 115-126. pp Witmer, E. A., Herrmann, W., Leech, J. W., and Pian,
https://openalex.org/W1604474099	https://hal.science/hal-00144234/document	English	null	Dynamic Clustering in Object-Oriented Databases: An Advocacy for Simplicity	Lecture notes in computer science	2,001	cc-by	6,924	Dynamic Clustering in Object-Oriented Databases: An Advocacy for Simplicity Jérôme Darmont, Christophe Fromantin, Stéphane Régnier, Le Gruenwald, Michel Schneider Dynamic Clustering in Object-Oriented Databases: An Advocacy for Simplicity Jérôme Darmont, Christophe Fromantin, Stéphane Régnier, Le Gruenwald, Michel Schneider e Darmont, Christophe Fromantin, Stéphane Régnier, Le Gruenwald, Michel Schneider To cite this version: Jérôme Darmont, Christophe Fromantin, Stéphane Régnier, Le Gruenwald, Michel Schneider. Dy- namic Clustering in Object-Oriented Databases: An Advocacy for Simplicity. ECOOP 2000 Sympo- sium on Objects and Databases, Jun 2000, Sofia Antipolis, France. pp.71-85. hal-00144234 Distributed under a Creative Commons Attribution 4.0 International License J. Darmont1, C. Fromantin2, S. Régnier2+3, L. Gruenwald3, M. Schneider2 3School of CS University of Oklahoma Norman, OK 73019, US ggruenwald@ou.edu 2L.I.M.O.S. Université Blaise Pascal 63177 Aubière Cedex, France michel.schneider@isima.fr 1E.R.I.C. Université Lyon 2 69676 Bron Cedex, France jdarmont@univ-lyon2.fr Abstract. We present in this paper three dynamic clustering techniques for Object-Oriented Databases (OODBs). The first two, Dynamic, Statistical & Tunable Clustering (DSTC) and StatClust, exploit both comprehensive usage statistics and the inter-object reference graph. They are quite elaborate. However, they are also complex to implement and induce a high overhead. The third clustering technique, called Detection & Reclustering of Objects (DRO), is based on the same principles, but is much simpler to implement. These three clustering algorithm have been implemented in the Texas persistent object store and compared in terms of clustering efficiency (i.e., overall performance increase) and overhead using the Object Clustering Benchmark (OCB). The results obtained showed that DRO induced a lighter overhead while still achieving better overall performance. Keywords: Object-Oriented Databases, Dynamic Object Clustering, Performance Comparison. Keywords: Object-Oriented Databases, Dynamic Object Clustering, Performance Comparison. HAL Id: hal-00144234 https://hal.science/hal-00144234v1 Submitted on 2 May 2007 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 1. Introduction Object-Oriented Database Management Systems (OODBMSs) always showed performance problems when compared to Relational Database Management Systems (RDBMSs). They really won an edge over RDBMSs only in niche markets, mainly engineering and multimedia applications. This performance problem is essentially caused by secondary storage Input/Output (I/O). Despite numerous advances in hard drive technology, I/Os still require much more time than main memory operations. Several techniques have been devised to minimize I/O transfers and improve the performances of OODBMSs, like query optimization, indexing, buffering, or clustering. Object clustering is a collaborative research topic at Blaise Pascal University (BPU) and the University of Oklahoma (OU) since the early 90’s. The principle of clustering is to store related objects close to each other in order to maximize the amount of relevant information returned when a disk page is loaded into the main memory. Early clustering methods were static [1, 2, 12, 13, 14, 15], i.e., objects were clustered only once at creation time. With these methods, modifying the placement to suit changes in data usage necessitates reorganizing the whole database. This is a heavy task that can only be performed manually when the system is idle. To support databases that are intended to be accessible on a 7 days a week / 24 hours a day basis (e.g., web-accessed databases), dynamic clustering techniques that cluster and recluster objects automatically and incrementally have been designed both by researchers and OODBMS vendors. However, since publications by the latter are very few and research proposals are not always implemented or evaluated, it is hard to select the best technique in a given context. The objectives of this paper are to propose an overview of the research dealing with dynamic object clustering techniques; to present two methods designed at BPU and OU called DSTC and StatClust, as well as a new one called Detection & Reclustering of Objects (DRO); and to compare these techniques in terms of efficiency and clustering overhead. These comparisons have been performed on the Texas system using the OCB benchmark [9], which has been specially designed to evaluate clustering algorithms. The remainder of this paper is organized as follows. Section 2 establishes a state of the art regarding dynamic clustering techniques. Section 3 presents DSTC [3], StatClust [10], and eventually details DRO. Section 4 presents the performance evaluations we performed on Texas. We finally conclude the paper and discuss future research issues. 2. Related Work: Dynamic Object Clustering Methods Most dynamic object clustering methods have been motivated by needs in engineering applications like CAD, CAM, or software engineering applications. A first class of clustering strategies is based on the analysis of database usage statistics. Chang and Katz [5] proposed a physical clustering method based on a particular inheritance link called instance to instance and the declaration of estimated access frequencies associated with three types of relationships (aggregation, equivalence, version). The idea is allowing inheritance of data along any type of attribute and particularly along inter-object links. For instance, it is interesting, when a new version of an object is created, to automatically make it inherit from its ancestor’s aggregation links toward other objects. Inherited data are stored only once, which allows an important gain in terms of disk space, but forces a physical object to be placed as close to inherited data as possible. The access frequencies and the computation of inherited attributes costs help identifying the destination page of a newly created object. If the target page is full, the system can either split the page or elect the next best page as a target page. Dynamic clustering is also coupled with an appropriate buffering strategy that is a variation of Least Recently Used (LRU) allowing a better usage of existing clustering. It is based on prioritizing all pages in memory. Frequently used pages have their priority increased along with their structurally related pages, while unused pages have their priority decreased with time. This method has never been implemented, except within simulation models [5, 8, 10] that hint a potential increase in performance of 200% under certain conditions. Another method based on statistics has been proposed by McIver and King [17], who advocate that object placement determination phases must be independent of the actual placement. The strategy leans on the exploitation of three modules running concurrently. The statistics collection module collects general database usage statistics and also selective database usage statistics concerning depth-first or breadth- first traversals, which are assimilated to navigational and associative accesses, respectively. The cluster analysis module uses a variation of the Cactis algorithm [12]. It first finds out the most referenced object in the database. Then, objects linked to it are grouped on the same disk page in depth-first, by decreasing order of co-usage frequency. 3.1. DSTC DSTC is actually both a dynamic object clustering policy and its associated buffering policy, which aims at clustering together objects that are used together at near instants in time [3]. It measures object usage statistics, while respecting the following constraints: minimize the amount of data managed, maximize the pertinence of collected statistics, reduce the cost of persistent storage for these data, and minimize perturbations on running transactions. This goal is achieved by scaling collected data at different levels and using gradual filters on main memory-stored statistics. Hence, it is possible to store on disk only presumably significant statistics. Database usage statistics concern object access frequencies and inter-object reference usage frequencies. All types of links are considered as physical references, whether they are structural links built at the schema level or logical links depending on applications or induced by physical object fragmentation. All physical accesses from one object toward another are detected and counted. Physical object reorganization is started by a trigger mechanism. Object disk storage is organized through an ordering algorithm that builds linear sequences of objects that capture “attraction forces” between objects. This sequence is sequentially transcribed in a cluster, i.e., a contiguous disk segment of variable size. The underlying algorithm was inspired by [7]. Flexibility in this approach is achieved through various parameters allowing the adaptation of system reactivity to database behavior. These parameters are set up by the database administrator. The DSTC strategy is organized into five phases. 1. Observation phase: During a predefined observation period, object usage statistics are collected and stored in an observation matrix in main memory. 2. Selection phase: Data stored in the observation matrix are sorted and filtered. Only significant statistics are retained. 3. Consolidation phase: Results from the selection phase are used to update data collected in previous observation phases, which are stored in a persistent consolidated matrix. 4. Dynamic cluster reorganization: Statistics from the consolidated matrix are exploited to suggest a reorganization of the physical space. Existing clustering units can be modified and new clustering units can be created. 5. Physical database reorganization: Clustering units are eventually used to consider a new object placement on disk. This phase is triggered when the system is idle. 5. Physical database reorganization: Clustering units are eventually used to consider a new object placement on disk. This phase is triggered when the system is idle. 2. Related Work: Dynamic Object Clustering Methods An advised variation is to use depth-first traversals when navigational accesses are preponderant and breadth-first traversals when associative accesses are preponderant. The type of access to select is provided by usage statistics. Clustering analysis is triggered after collection of a significant amount of statistics. The reorganization module rearranges objects on disk so that the database physical organization corresponds to the page assignments suggested by clustering analysis. A reorganization phase is not always necessary after each clustering analysis phase. When a reorganization phase is triggered, it deals only with objects that have not been clustered. The performance of this method has been evaluated by simulation using the Trouble Ticket Benchmark [16]. This study shows that the collected statistics and the proposed clustering are pertinent, and that a high overhead is caused by the database reorganization phases, where the entire database is locked and the transactions are postponed. Cheng and Hurson state that existing strategies are generally based on one single clustering criterion [7]. Their multi-level clustering allows clustering objects using several criteria at once. The method associates a criterion to each of three types of relationships identified by [5]: equivalence, aggregation, and version. A proximity degree between two objects can be elaborated using the values of these criteria. Clustering is recommended when this proximity degree is sufficiently small. The clustering algorithm actually orders objects on the basis of their proximity degree. Clustering is performed by the system, without any external intervention. Furthermore, this strategy is backed up by a cost model that evaluates the benefit of a possible dynamic reorganization. This proposal has never been implemented. Finally, an innovative strategy has been proposed to handle object clustering in the EOS distributed system [11]. This method exploits the system’s garbage collector and induces a very low overhead. Clustering specifications are provided by the database administrator, who weights arcs in the class aggregation graph according to estimated access probabilities. Objects are clustered with their stronger weighted parent when created. Placements are re-evaluated afterward by the disk garbage collection process and may be modified asynchronously. This proposal has not been implemented. The authors do provide elements regarding feasibility and low cost, but this technique is intimately related to the presence of a disk garbage collector continuously working, which is costly and thus not much used in existing OODBMSs. 3.1. DSTC The principle of the buffering management associated with DSTC is the following. When an object belonging to a cluster is accessed, the whole cluster is loaded. This avoids useless I/Os since objects in the cluster have a good probability to be used by the current transaction. A page replacement algorithm named LRU-C is also proposed. Its principle is to date clusters in the buffer rather than pages. The DSTC strategy has been implemented in Texas [18] on Sun workstations and PCs under Linux. Performance studies have been performed with a benchmark based on OO1 [4] and baptized DSTC-CluB. They showed the efficiency of DSTC compared to a no-clustering policy on simple cases. 3.2. StatClust (Statistical Clustering) This method extends Chang and Katz’ method (see Section 2) [5]. Its authors advocate replacing user-estimated access frequencies by more reliable usage statistics [10], for each of the considered types of links (aggregation, equivalence, version). Statistics regarding read or write accesses have also been added. Clustering is automatic at object creation or update time and when a bad clustering is detected. The user can influence the clustering process through a set of parameters. A bad clustering is detected when the ratio between the number of blocks (set of contiguous pages) read in the buffer and the number of blocks read on disk is smaller than a threshold computed by the system, and the amount of collected statistics is sufficient. The detection of a bad clustering ends the collection of statistics and starts up a reclustering phase that specifies which objects might be reclustered (i.e., which objects show satisfying usage statistics). The physical placement of objects uses an algorithm close to [5], but also supports object duplication. Objects may be duplicated to increase reference locality. An object that is more read than updated is a candidate for duplication. p StatClust has been compared by simulation to static clustering techniques (ORION and Cactis) [10], but not to dynamic clustering techniques, including Chang and Katz’ method, on which it is based. The results are actually very similar to those reported in [8]. Usage Statistics. DRO stores and exploits two principal types of indicators. They are updated dynamically when the database is in use. Th bj t f th b f ti h bj t i Usage Statistics. DRO stores and exploits two principal types of indicators. They are updated dynamically when the database is in use. !"The object access frequency measures the number of times each object is accessed. During the clustering phase, only the objects with the highest access frequencies are taken into account. 3.3. DRO Overview. The design of DRO makes use of the experience accumulated with both the DSTC and StatClust clustering methods, especially at the implementation level. Since these methods were quite sophisticated, they were also very difficult to implement properly and lots of problems occurred in the development process. Furthermore, though they attempt to minimize the amount of usage statistics stored, they use various statistical data that are not easy to manage and whose size often increases drastically. DRO is much easier to implement. It exploits both basic usage statistics and the graph of inter-object references (derived from the schema) to dynamically cluster the database. Its principle is to store together the objects that are the most frequently accessed overall. DRO has been implemented in Texas. Usage Statistics. DRO stores and exploits two principal types of indicators. They are updated dynamically when the database is in use. !"The object access frequency measures the number of times each object is accessed. During the clustering phase, only the objects with the highest access frequencies are taken into account. !"The page usage rate is the ratio between the size of the data effectively stored in the page and the page size, a page being the unit of transfer between disk and memory. This ratio helps determining which pages degrade the system performance. The mean usage rate for all pages and the number of pages loaded are also computed. !"The page usage rate is the ratio between the size of the data effectively stored in the page and the page size, a page being the unit of transfer between disk and memory. This ratio helps determining which pages degrade the system performance. The mean usage rate for all pages and the number of pages loaded are also computed. The data structure presented in Fig. 1 as a UML static structure diagram is used to store DRO’s usage statistics. The PageStat class concerns page statistics. It has three attributes: a page identifier, the number of times this page has been loaded into memory, and its page usage rate. The ObjectStat class concerns object statistics. It also has three attributes: an object identifier, the object access frequency, and a boolean usage indicator. The PageObjectStat class allows large objects to be stored on several pages. It has only one attribute: the size occupied by a given object in a given page. 3.3. DRO To proceed to step 2, two conditions must be met: a) the number of pages to cluster must be greater than one, and b) the ratio between the number of pages to cluster and the number of pages actually used is greater than the page clustering rate parameter (PCRate). Step 2: Clustering Setup. This step helps defining a sequential placement order of objects on disk. The algorithm input is the list of objects to cluster sorted by decreasing access frequency. This step is subdivided into three phases. !"Object clustering using inter-object references. This first phase links objects regarding reference links. The algorithm shown in Figure 2 runs up to a user- defined maximum distance MaxD, i.e., the first iteration considers all the objects referenced by the starting object (distance 1), then the process reiterates for each object found, up to distance MaxD. When linking together objects Oi and Oj of access frequencies AFi and AFj, the dissimilarity rate \|AFi – AFj\| / max(AFi, AFj) must be lower than the maximum dissimilarity rate MaxDR not to link objects that are too weakly bound. Objects are sorted by descending order of access frequency to generate a list defining a placement order of objects so that they can be sequentially written on disk. q y !"Linking of placement order lists. This phase links together the list parts made up in the first phase to obtain a single list. The list parts are considered in their generation order and simply concatenated. !"Resemblance rate computation. The third phase establishes a resemblance rate between the current object placement and the new placement proposed by the clustering algorithm. This resemblance rate helps evaluating how different the new clustering proposal is from the current physical placement of the objects. If the new cluster is found similar (for instance, if the considered objects have already been clustered), no action is undertaken. The resemblance rate is the number of objects in the proposed cluster that are not moved regarding current object placement divided by the number of objects in the cluster. Step 3: Physical Object Clustering. Physical clustering is performed if the resemblance rate computed at step 2 is lower than a user-defined maximum resemblance rate (MaxRR). This operation clusters objects identified in the previous steps, but must also reorganize the database in order to retrieve space made available by movement or deletion of objects. Step 4: Statistics Update. 3.3. DRO 1..* 1 ObjectStat Object_ID : Integer Access_Frequency : Integer Usage_Indicator : Boolean 0..1 PageStat Page_ID : Integer Nb_Load : Integer Usage_Rate : Real 1..* PageObjectStat Object_Size : Integer 1..* 1 Is_Split 0..1 1..* Contains_Object Fig. 1. DRO usage statistics 1..* 1 ObjectStat Object_ID : Integer Access_Frequency : Integer Usage_Indicator : Boolean 0..1 PageStat Page_ID : Integer Nb_Load : Integer Usage_Rate : Real 1..* PageObjectStat Object_Size : Integer 1..* 1 Is_Split 0..1 1..* Contains_Object Fig. 1. DRO usage statistics Fig. 1. DRO usage statistics Whenever an object is accessed, its access frequency is incremented by 1 and its usage indicator is set to true. Page statistics are updated whenever a page moves from the main memory to disk. The statistics attached to all the objects on this page are used to compute the size occupied on the page by objects that have actually been used. The page usage rate is then computed and Nb_Load is increased by 1. If an object is deleted from the database, the corresponding usage statistics are also deleted. If the page that contains this object does not have any more objects in its associated PageObjectStat object, its statistics are also deleted. If an object is merely moved from one page to another, its usage indicator is reset to false and its link to the starting page is deleted. Its statistics will then be linked to the destination page’s statistics when the object is used again. Clustering. The clustering phase can be triggered manually or automatically. It is subdivided into four steps. Until physical object placement, a control procedure checks out after each step whether clustering must abort or resume. Step 1: Determination of Objects to Cluster. This step helps defining the objects belonging to pages with usage rate lower than the minimum usage rate (MinUR) and that have been loaded in memory more times than the minimum loading threshold (MinLT). MinUR and MinLT are user-defined parameters. MinUR helps selecting pages containing a majority of unused objects or objects that are not used together. Objects stored into these pages and whose usage statistics (i.e., an ObjectStat object) are instantiated are selected for clustering. They are attached to instances of the Clustering class. Objects of class Clustering are linked together by two bi-directional relations called Object_Sort and Object_Placement, which store objects sorted by access frequency and a placement order of objects on disk, respectively. 3.3. DRO Sample access frequencies OID Access Frequency 6 60 5 60 4 60 7 40 1 20 2 20 3 20 10 18 1 6 3 10 5 4 2 9 8 7 8 17 Fig. 3. Sample inter-object reference graph Table 2. Sample access frequencies Let MaxD be 1. Objects are considered by the order of access frequency. The dissimilarity rates between object couples (6, 5) and (5, 4) are both 0. The dissimilarity rates of the (6, 3), (5, 3), (5, 8), and (4, 7) couples are all greater than MaxDR, so the first sub-list we obtain is (6, 5, 4). The dissimilarity rate for the (7, 8) couple is 0.575 and hence greater than MaxDR, so (7) remains a singleton. The dissimilarity rates for the (1, 3), (3, 2), and (3, 10) couples are 0, 0, and 0.1, respectively (links to already treated objects are not considered), so the third sub-list is (1, 3, 2, 10). (8) forms the last sub-list since object #9 has never been accessed and thus must not be clustered. Now if MaxD is 2, we have to consider dissimilarity rates up to a “distance” (in number of objects) of 2 from the starting object. For instance, we must consider the (6, 10) couple. Its dissimilarity rate is 0.7, greater than MaxDR. The only change regarding the sub-lists obtained with MaxD set to 1 is the integration of object #8 in the (1, 3, 2, 10) sequence, because the dissimilarity rate of the (10, 8) couple is 0.05, lower than MaxDR. Eventually, the sub-lists are merged in one list by the order of creation OID Access Frequency 6 60 5 60 4 60 7 40 1 20 2 20 3 20 10 18 8 17 Table 2. Sample access frequencies Fig. 3. Sample inter-object reference graph Fig. 3. Sample inter-object reference graph Table 2. Sample access frequencies Table 2. Sample access frequencies Let MaxD be 1. Objects are considered by the order of access frequency. The dissimilarity rates between object couples (6, 5) and (5, 4) are both 0. The dissimilarity rates of the (6, 3), (5, 3), (5, 8), and (4, 7) couples are all greater than MaxDR, so the first sub-list we obtain is (6, 5, 4). The dissimilarity rate for the (7, 8) couple is 0.575 and hence greater than MaxDR, so (7) remains a singleton. 3.3. DRO This update depends on a user-defined statistics update indicator (SUInd). If SUInd is set to true, all statistics are deleted. Otherwise, only statistics regarding pages containing objects that have been moved are deleted. DRO Parameters. The parameters defining the behavior of the DRO strategy are set- up by the database administrator. They are recapitulated in Table 1. We obtained the default values through many experiments on Texas. D = 0 End = false While D < MaxD and not End do D = D + 1 // Browse objects to cluster Starting_object = Clustering.Sort_first While Starting_object ≠ NIL and Starting_object.Placement_previous ≠ NIL do Starting_object = Starting_object.Sort_next End While While Starting_object ≠ NIL do Object_to_link = Starting_object While Object_to_link ≠ NIL and Object_to_link.Placement_previous ≠ NIL do Object_to_link = Object_to_link.Placement_next End while Found = TRUE While Found do // Find an object to cluster different from Starting_object, // referenced on a distance lower than MaxD, with a // dissimilarity rate lower than MaxDR, and attribute // Clustering.Placement_previous set to NIL Found_object = Research_procedure_result() If Found_object ≠ NIL then Object_to_link.Placement_next = Found_object Object_found.Placement_previous = Object_to_link Object_to_link = Object_found Else Found = FALSE End if End while While Starting_object ≠ NIL and Starting_object.Placement_previous ≠ NIL do Starting_object = Starting_object.Sort_next End while End while Fig. 2. Object clustering Parameter Name Type Default value Minimum usage rate MinUR Real 0.8 Minimum loading threshold MinLT Real 1 Page clustering rate PCRate Real 0.05 Maximum distance MaxD Integer 1 Maximum dissimilarity rate MaxDR Real 0.05 Maximum resemblance rate MaxRR Real 0.9 Statistics update indicator SUInd Boolean True ble 1 DRO parameters Table 1. DRO parameters Example of Clustering with DRO. Let us consider the graph of inter-object references from Fig. 3 and the associated access frequencies from Table 2. With the MaxDR parameter set up to 0.1, Fig. 4 shows how the clustering algorithm builds an ordered sequence of objects that will be sequentially written on disk. 1 6 3 10 5 4 2 9 8 7 Fig. 3. Sample inter-object reference graph OID Access Frequency 6 60 5 60 4 60 7 40 1 20 2 20 3 20 10 18 1 6 3 10 5 4 2 9 8 7 8 17 Fig. 3. Sample inter-object reference graph Table 2. 4.1. Experiment Scope Our initial goal was to compare the performances of StatClust, DSTC, and DRO. However, StatClust proved exceedingly difficult to implement in Texas. Since Texas exploits the operating system’s virtual memory, it considers the memory buffer to be of infinite size. Thus, it is impossible to implement StatClust’s module for detecting a bad clustering, because it needs to count the number of pages accessed from the disk and the buffer. Furthermore, substantial additions to Texas would be necessary to support the object replication process advocated by StatClust. Eventually, the object clustering algorithm initially builds a list of candidate pages containing objects related to the current object. To build this list, the database schema must be known. Techniques can be devised to automatically infer the schema, but none of them is easy to implement. In addition, when implementing StatClust, we found that Texas could not handle numerous transactions and the associated statistics on reasonably large databases and invariably crashed. Thus, we were not able to properly compare StatClust to the other algorithms. Hence, we only compare DSTC and DRO here. g y p To compare the performances of DSTC and DRO, we used a mid-sized OCB database composed of 50 classes and 100,000 objects, for a size of about 62 MB. The other OCB parameters defining the database were set to default. Two series of standard OCB transactions (1000 transactions and 10,000 transactions) were executed on this database, before and after object clustering. System performance was measured in terms of I/Os, response time, and relative performance improvement due to clustering. Only the results concerning I/Os are presented in this paper because response time plots present exactly the same tendencies and do not bring additional insight. Eventually, these experiments have been performed in several memory configurations. Since Texas makes an intensive use of virtual memory, it was interesting to see how the system behaved when the ratio main memory size / database size varied. The whole process was reiterated 100 times so that mean tendencies could be achieved. In each iteration, the same random seed was selected for the DSTC and DRO experiments so that they were rigorously identical. 3.3. DRO The dissimilarity rates for the (1, 3), (3, 2), and (3, 10) couples are 0, 0, and 0.1, respectively (links to already treated objects are not considered), so the third sub-list is (1, 3, 2, 10). (8) forms the last sub-list since object #9 has never been accessed and thus must not be clustered. Now if MaxD is 2, we have to consider dissimilarity rates up to a “distance” (in number of objects) of 2 from the starting object. For instance, we must consider the (6, 10) couple. Its dissimilarity rate is 0.7, greater than MaxDR. The only change regarding the sub-lists obtained with MaxD set to 1 is the integration of object #8 in the (1, 3, 2, 10) sequence, because the dissimilarity rate of the (10, 8) couple is 0.05, lower than MaxDR. Eventually, the sub-lists are merged in one list by the order of creation. 6 5 4 7 1 3 2 10 8 6 5 4 7 1 3 2 10 8 Distance = 1 Distance = 2 6 5 4 7 1 3 2 10 8 Final placement order Fig. 4. Sample execution of the DRO clustering algorithm Final placement order Fig. 4. Sample execution of the DRO clustering algorithm 4.2. Experiment Hardware and Software The version of Texas we used is a prototype (version 0.5) running on a PC Pentium 166 with 64 MB of RAM, and version 2.0.30 of Linux. The swap partition size was 64 MB. StatClust, DSTC and DRO are integrated in Texas as a collection of new modules, and a modification of several Texas modules. Texas and the additional StatClust, DSTC and DRO modules were written in GNU C++ version 2.7.2.1. 4.3. Experiment Results DSTC. Fig. 5 and 6 show that clustering with DSTC indeed allows a significant gain in performance, especially when the amount of main memory available is small. Clustering is definitely more useful when the database does not fit wholly within the main memory, since its effects are felt as soon as the system swaps and not only at page load time. This assumption is neatly confirmed by the clustering gain factor graph in Fig. 6. Clustering gain factor is equal to the number of I/Os necessary to execute the transactions after clustering divided by the number of I/Os necessary to execute the transactions before clustering. A discrepancy appears between Fig. 5 and 6 due to the fact that 1000 transactions are not enough: objects are used, clustered, but rarely reused following the same patterns, thus provoking an useless clustering) on small memory configurations. On the other hand, the 10,000 transaction workload appears more representative of actual database usage, allowing an average gain factor of about 2.5. 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After Clustering 0 1 2 3 4 5 6 7 8 9 10 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 5. DSTC results – 1000 transactions 0 50000 100000 150000 200000 250000 300000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering 0 1 2 3 4 5 6 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 6. DSTC results – 10,000 transactions 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After Clustering Fig. 5. DSTC results – 1000 transactions 0 1 2 3 4 5 6 7 8 9 10 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 5. DSTC results – 1000 transactions Fig. 5. DSTC results – 1000 transactions 0 50000 100000 150000 200000 250000 300000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering Fig. 6. DSTC results – 10,000 transactions 0 1 2 3 4 5 6 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 6. 4.3. Experiment Results DSTC results – 10,000 transactions DRO. Fig. 7 and 8 show that DRO bears the same overall behavior as DSTC. However, the gain factor achieved with DRO on the 10,000 transaction workload looks much better. It is indeed about 15. The comparison is unfair, though, because we selected the optimal set of parameters for DRO clustering, while we could not do it for DSTC. Due to technical problems with big databases, we had to parameterize DSTC so that clustering was not the best possible. There was a threshold effect on a set of DSTC parameters. Below this “threshold”, everything worked out fine but clustering was average. Beyond the “threshold”, clustering units were too big for Texas to manage and the system invariably crashed. set of DSTC parameters. Below this “threshold”, everything worked out fine but clustering was average. Beyond the “threshold”, clustering units were too big for Texas to manage and the system invariably crashed. 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering 0 2 4 6 8 10 12 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 7. DRO results – 1000 transactions 0 50000 100000 150000 200000 250000 300000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering 0 5 10 15 20 25 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 8. DRO results – 10,000 transactions 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering 0 2 4 6 8 10 12 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 7. DRO results – 1000 transactions 0 50000 100000 150000 200000 250000 300000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering 0 5 10 15 20 25 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 8. DRO results – 10,000 transactions 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering Fig. 7. 4.3. Experiment Results DRO results – 1000 transactions 0 2 4 6 8 10 12 8 12 16 24 32 64 Memory size (MB) Clustering gain factor Fig. 7. DRO results – 1000 transactions Fig. 7. DRO results – 1000 transactions 0 5 10 15 20 25 8 12 16 24 32 64 Memory size (MB) Clustering gain factor 0 50000 100000 150000 200000 250000 300000 8 12 16 24 32 64 Memory size (MB) Mean number of I/Os Before clustering After clustering Fig. 8. DRO results – 10,000 transactions Fig. 8. DRO results – 10,000 transactions Comparison of DSTC and DRO. To eventually compare DSTC and DRO on a fair ground, we used a smaller database so that DSTC could properly work. We used OCB’s default database (50 classes, 20,000 instances, about 20 MB) and ran two series of typical transactions that were likely to benefit from clustering: depth-3 hierarchy traversals (that always follow the same type of reference) and depth-2 simple traversals (depth-first traversals). The depth of traversals was reduced regarding OCB’s default parameters so that the generated clusters were not too big and the effects of clustering were clear. The traversals have been performed from 100 predefined root objects and each of them was executed 10 times. Table 3 displays the mean number of I/Os concerning database usage before and after clustering. It shows that DSTC and DRO both achieve a substantial increase in performance (factor 6-7). DRO looks even better, though more tuning with DSTC should bring this method on the same level. Unfortunately, such tuning still provoked execution errors in Texas. The big difference between DSTC and DRO lies in clustering overhead (the number of I/Os necessary for an algorithm to cluster the database). DSTC induces a high overhead, which renders it difficult to implement truly dynamically. Its authors actually advocate its triggering when the database is idle. On the contrary, DRO, which is much simpler, present a lower overhead (about 4 times lower) and is certainly better suited to a dynamic execution. Hierarchy traversals Simple traversals DSTC DRO DSTC DRO Pre-clustering usage 1682.6 1686 1682 1683 Post-clustering usage 270.8 226 281.75 236.75 Clustering gain factor 6.21 7.46 5.97 7.11 Clustering overhead 12219.4 3286.8 12174 2804.5 Table 3. Clustering efficiency comparison between DSTC and DRO (I/Os) Table 3. Clustering efficiency comparison between DSTC and DRO (I/Os) 5. Conclusion We have presented in this paper a representative panel of dynamic object clustering techniques, including our first effort in this field: the DSTC and StatClust techniques, which both make an intensive use of statistical counters and include clustering mechanisms with elaborated features. We have also presented a new clustering method, DRO, whose principles are based on those of DSTC and StatClust, but that is much simpler and deals with fewer statistical counters. The idea behind DRO is to provide a clustering method equivalent to or better than DSTC and StatClust while achieving simplicity of implementation. We validated the idea that a simple dynamic clustering technique could provide better results than an elaborated one by comparing DSTC and DRO. Our results showed that DRO indeed performed better than DSTC, which could not be set up in an optimal fashion due to its inherent complexity. Furthermore, the clustering overhead induced by DRO was much lower than that induced by DSTC, definitely proving that a simple approach is more viable in dynamic context than a complex one. To summarize, we showed that DRO was a better choice than DSTC in all circumstances. We also underlined the fact that a dynamic clustering technique is perfectly viable in an OODBMS and could achieve significant gains in performances. Since DRO is based on usage statistics, it fits well with the concept of autoadmin databases that is currently researched in major companies to automate the database tuning process [6]. The perspectives opened by this study are divided into two axes. First, the evaluation of DRO should be carried on on other systems besides Texas, which is a persistent object store rather than a full OODBMS. Such evaluations could be conducted on real OODBMSs like O2, or achieved by simulation. Second, DRO itself could be improved so that clustering overhead is minimized. Some optimizations can be achieved in its code itself (at the list manipulation level, for instance), while others relate more to tuning DRO’s parameters, which could also be achieved by simulation. References 1. T. Andrews, C. Harris, K. Sinkel: ONTOS: A Persistent Database for C++. In: Object-Oriented Databases with Applications to CASE, Networks and VLSI CAD (1991) 387–406 2. V. Benzaken, C. Delobel: Enhancing Performance in a Persistent Object Store: Clustering Strategies in O2. 4th International Workshop on Persistent Object Systems (1990) 403–412 3. F. Bullat, M. Schneider: Dynamic Clustering in Object Database Exploiting Effective Use of Relationships Between Objects. ECOOP ’96, Linz, Austria. LNCS Vol. 1098 (1996) 344–365 4. R.G.G. Cattell: An Engineering Database Benchmark. In: The Benchmark Handbook for Database Transaction Processing Systems. Morgan Kaufmann (1991) 247–281 5. E.E. Chang, R.H. Katz: Exploiting Inheritance and Structure Semantics for Effective Clustering and Buffering in an Object-Oriented DBMS. ACM SIGMOD International Conference on Management of Data (1989) 348–357 6. S. Chaudhuri, V. Narasayya: AutoAdmin “What-if” Index Analysis Utility. ACM SIGMOD International Conference on Management of Data, Seattle, Washington (1998) 367–378 7. J.R. Cheng, A.R. Hurson: Effective clustering of complex objects in object- oriented databases. ACM SIGMOD International Conference on Management of Data (1991) 22–31 8. J. Darmont, L. Gruenwald: A Comparison Study of Clustering Techniques for Object-Oriented Databases. Information Sciences, Vol. 94, No. 1-4 (1996) 55–86 9. J. Darmont, B. Petit, M. Schneider: OCB: A Generic Benchmark to Evaluate the Performances of Object-Oriented Database Systems. 6th International Conference on Extending Database Technology (EDBT ’98), Valencia, Spain. LNCS Vol. 1377 (1998) 326–340 10.J.-Y. Gay, L. Gruenwald: A Clustering Technique for Object Oriented Databases. 8th International Conference on Database and Expert Systems Applications (DEXA ’97), Toulouse, France. LNCS Vol. 1308 (1997) 81–90 11. O. Gruber, L. Amsaleg: Object Grouping in EOS. IWDOM '92 Workshop on Distributed Object Management, University of Alberta, Canada (1992) 117–131 12.S.E. Hudson, R. King: Cactis: A Self-Adaptive Concurrent Implementation of an Object-Oriented Database Management System. ACM Transactions on Database Systems, Vol. 14, No. 3 (1989) 291–321 13.W. Kim, J. Banerjee, H.-T. Chou, J.F. Garza, D. Woelk: Composite Object Support in an Object-Oriented Database System. International Conference on OOPSLA (1987) 118–125 14.C. Lamb, G. Landis, J. Orenstein, D. Weinreb: The ObjectStore Database System. Communications of the ACM, Vol. 34, No. 10 (1991) 50–63 15.D. Maier, J. Stein, A. Otis, A. Purdy: Development of an Object-Oriented DBMS. ACM OOPSLA ’86 (1986) 472–482 16.C. McGlenaghan: OODBMS Benchmark Specification. Technical Report No. At- 12/99-001523-00.01, US West Advanced Technologies (1991) g 17.W.J. Mc Iver Jr., R. King: Self-Adaptive, On-Line Reclustering of Complex Object Data. References ACM SIGMOD Conference (1994) 407–418 18.V. Singhal, S.V. Kakkad, P.R. Wilson: Texas: An Efficient, Portable Persistent Store. 5th International Workshop on Persistent Object Systems (1992)
https://openalex.org/W2888309746	https://europepmc.org/articles/pmc6102890?pdf=render	English	null	Pre-pregnancy cardiovascular risk factors and racial disparities in birth outcomes: the Bogalusa Heart Study	BMC pregnancy and childbirth	2,018	cc-by	7,029	Abstract Background: Racial disparities in birth outcomes are mirrored in cardiovascular health. Recently there have been calls for more attention to preconception and interconceptional health in order to improve birth outcomes, including as a strategy to reduce black-white disparities. Methods: As part of a larger study of cardiovascular and reproductive health (“Bogalusa Babies”), female participants were linked to their children’s birth certificates for Louisiana, Mississippi, and Texas births from 1982 to 2009. Three thousand and ninety-five women were linked to birth certificate data. Birth outcomes were defined as low birthweight (LBW) birthweight < 2500 g; preterm birth (PTB), > 3 weeks early; small for gestational age (SGA), <10th percentile for gestational age (percentiles based on study population); large for gestational age (LGA) >90th percentile for gestational age]. Cardiovascular measures (blood pressure, lipids, glucose, insulin) at the visit closest in time but prior to the pregnancy was examined as predictors of birth outcomes using logistic models adjusted for covariates. Results: Only a few cardiovascular risk factors were associated with birth outcomes. Triglycerides were associated with higher risk of LBW among whites (aOR 1.05, 95% 1.01–1.10). Higher glucose was associated with a reduction in risk of SGA for black women (aOR 0.85, 95% CI 0.76–0.95), but not whites (p for interaction = 0.02). Clear racial disparities were found, but they were reduced modestly (LBW/SGA) or not at all (PTB/LGA) after CVD risk factors were adjusted for. Results: Only a few cardiovascular risk factors were associated with birth outcomes. Triglycerides were associated with higher risk of LBW among whites (aOR 1.05, 95% 1.01–1.10). Higher glucose was associated with a reduction in risk of SGA for black women (aOR 0.85, 95% CI 0.76–0.95), but not whites (p for interaction = 0.02). Clear racial disparities were found, but they were reduced modestly (LBW/SGA) or not at all (PTB/LGA) after CVD risk factors were adjusted for. Conclusions: This analysis does not provide evidence for preconception cardiovascular risk being a strong contributor to racial disparities. Conclusions: This analysis does not provide evidence for preconception cardiovascular risk being a strong contributor to racial disparities. Keywords: Birth weight, low, Premature birth, Infant, small for gestational age, Cholesterol, Glucose, Continental population groups Keywords: Birth weight, low, Premature birth, Infant, small for gestational age, Cholesterol, Glucose, Continental population groups groups [5], including higher blood pressure and diabetes [6]. Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 https://doi.org/10.1186/s12884-018-1959-y Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 https://doi.org/10.1186/s12884-018-1959-y Open Access © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abstract In NHANES, black women of childbearing age had higher systolic blood pressure, diastolic blood pressure, and glycated hemoglobin than other groups [7]. Pre-pregnancy cardiovascular risk factors and racial disparities in birth outcomes: the Bogalusa Heart Study Emily W. Harville1* , Leann Myers2, Tian Shu1, Maeve E. Wallace3 and Lydia A. Bazzano1 Background In the United States, blacks have roughly double the risk of infant mortality of other ethnic groups [1], due largely to preterm birth and fetal growth restriction. This dis- parity persists across populations with comparable ac- cess to health care and in otherwise low-risk groups, such as military populations, those with health insur- ance, and women with higher education and low initial medical risk [2–4]. Disparities in birth outcomes are mirrored in cardiovascular health. Black women have higher cardiovascular risk than other racial/ethnic Recently there have been calls for more attention to preconception and interconceptional health in order to improve birth outcomes [8–11], including as a strategy to reduce black-white disparities [12, 13], in part because black women have an increased risk of preconception hypertension and diabetes [14]. Women with chronic hypertension who become pregnant are at higher risk of preterm birth and small-for-gestational-age [15, 16], while women with diabetes who become pregnant are at higher risk of preterm birth [16–18] and large-for-gestational-age/ macrosomia [19]. Some studies indicate that preconception * Correspondence: eharvill@tulane.edu 1Department of Epidemiology, Tulane School of Public Health and Tropical Medicine, 1440 Canal St. Ste. 2001, #8318, New Orleans, LA 70112-2715, USA Full list of author information is available at the end of the article * Correspondence: eharvill@tulane.edu 1Department of Epidemiology, Tulane School of Public Health and Tropical Medicine, 1440 Canal St. Ste. 2001, #8318, New Orleans, LA 70112-2715, USA Full list of author information is available at the end of the article Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Page 2 of 10 Five thousand, nine hundred and- ten women/girls participated in BHS at least once during the study. Of those, 3263 were linked to birth certificate data, repre- senting a linkage of approximately 65% of the likely births (as approximately 10–15% of women do not give birth). Participants linked to birth records were more likely to be black (37% vs. 35%), had more study visits, had a younger average age at first visit (8.8 vs. 10.5), were less likely to smoke (35% vs. 38%), and had statisti- cally lower mean BMIs (19.9 vs. 20.3) and blood pressure (101.8 vs. 102.8) (Harville et al., under review). Among white women, those whose parents had a high school edu- cation, rather than more or less education, were most likely to be linked, while among black women, likelihood of link- age fell with increased parental education. Among the smaller group with information about adult education, higher education was associated with increased likelihood of linkage among both black and white women. Three thou- sand and ninety-five women had at least one screening as a child, which occurred prior to the first pregnancy in the dataset. Two thousand, seven hundred and sixty-three screenings included a fasting blood samples; 2691 of these had data for analysis of lipids, insulin, and glucose. (More detail on missing data for covariates or individual risk factors is found in the tables). If a woman had multiple pregnancies, the first linked pregnancy was used, and analysis was limited to singleton births. cardiovascular health outside clinical disease also impacts pregnancy health: higher preconception blood pressure has been associated with lower birth weight [20], while both low and high cholesterol have been associated with poor pregnancy outcome [20, 21]. In addition, preconcep- tion glucose levels are associated with increased birth- weight [20]. In our analysis of the Cardiovascular Risk in Young Finns Study [22], higher pre-pregnancy triglycer- ides were associated with a higher risk of hypertensive dis- orders, pre-eclampsia, and gestational diabetes, while higher blood pressure was associated with preterm birth and small-for-gestational-age. Still, few studies have examined preconception cardio- vascular health and birth outcomes [21], and some of the largest are limited to Scandinavian populations [20, 22]. While some studies have examined clinical comorbidities as predictors of birth outcomes in black women [23], few have assessed subclinical measures. Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 In this study, we exam- ine how pre-pregnancy cardiovascular risk factors are asso- ciated with birth outcomes in the Bogalusa Heart Study, a biracial study of cardiovascular health. The research ques- tions are a) are pre-pregnancy cardiovascular risk factors associated with birth outcomes in this cohort; b) do any as- sociations between these risk factors and the outcomes dif- fer between African-American and white women? Based on the results of the analyses of those questions, we then examined whether pre-pregnancy cardiovascular health contribute to racial disparities in birth outcomes. Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Methods Blood pressure levels were measured on the right arm of subjects in a relaxed, sitting position by two trained ob- servers (3 replicates each). Systolic blood pressure and diastolic blood pressure were recorded using a mercury sphygmomanometer. The fifth Korotkoff phase was used for diastolic blood pressure. The mean values of the readings were used for analysis. The Bogalusa Heart Study is a long-running study of child- hood, adolescent, and now adult cardiovascular health, founded by Dr. Gerald Berenson in 1973 [24]. Participants were initially recruited from schools in the Bogalusa, Lou- isiana, area at ages 3–18. Over time, additional waves of data collection were performed, adding additional partici- pants up to adulthood. Currently, participants are largely in their 40s through 60s, and follow-up for cardiovascular and early aging parameters continues. The average age at first visit was 8.5 for black women and 8.6 for white women, p = 0.32. Fasting blood samples were drawn for lipids and glucose analysis. Prior to 1987, serum total cholesterol and trigly- ceride levels were determined by a Technicon AutoAnaly- zer II (Technicon Instrument, Tarrytown, NY). From 1987 to 1996, cholesterol and triglyceride levels were deter- mined by enzymatic procedures on the Abbott VP instru- ment (Abbott Laboratories, Chicago, IL) and on the Hitachi 902 Automatic Analyzer (Roche Diagnostics, In- dianapolis, IN) after 1996. Serum lipoprotein cholesterols were analyzed by using a combination of heparin-calcium precipitation and agar-agarose gel electrophoresis proce- dures. Both chemical and enzymatic procedures met the performance requirements of the Lipid Standardization Program of the Centers for Disease Control and Preven- tion, which has routinely monitored the precision and ac- curacy of cholesterol and triglyceride measurements since 1973. Measurements on CDC-assigned quality control As part of a larger study of cardiovascular and repro- ductive health (“Bogalusa Babies”), female participants were linked to their children’s birth certificates for Lou- isiana, Mississippi, and Texas births from 1982 to 2009. The 1982–1989 records contained fewer variables for linkage than later years; observations that matched on four variables (year of birth, last name, Soundex code for last name, and race) were confirmed by visual com- parison of addresses. From 1990 to 2009, a three stage linkage process was used, including deterministic record linkage based on maternal social security number (SSN), and probabilistic linkage when SSN was unavailable. Harville et al. Methods BMC Pregnancy and Childbirth (2018) 18:339 Page 3 of 10 Page 3 of 10 factors, results are presented adjusted for BMI alone, then with adjustment for all factors. In addition to strati- fied models, combined models were run with an inter- action term for race to examine differences between white and black participants. Results are presented stratified by race. As some studies have found non-linear associations between cardiovascular risk factors and birth outcomes [21], quadratic terms were also exam- ined. Quadratic BMI was examined as a potential con- founder, to assess possible non-linear confounding by BMI, but this term was neither statistically predictive nor did it change the effect estimate. We attempted to determine whether results were different when limited to those whose study visit was closer in time to the preg- nancy, but sample size was relatively small (n = 586, 217 black, 369 white for those with a measure within 5 years of pregnancy) and results became too imprecise to judge whether this group differed from the main sample. samples showed no consistent bias over time within or be- tween surveys. From 1981 to 1991, plasma glucose was measured by a glucose oxidase method using a Beckman Glucose Analyzer (Beckman Instruments, Palo Alto, CA). Since then, it has been measured enzymatically as part of a multichemistry (SMA20) profile. Measurements were made by laboratory technicians blinded to participants’ risk factors. Ten percent blind duplicate samples are selected, prior to blood drawing. The intraclass correlation coefficient for blind duplicate samples ranged from 0.92 for glucose to 0.99 for total cholesterol. Birth outcomes were taken from measures on the birth certificates. Low birthweight (LBW) was defined as birth- weight < 2500 g, preterm birth (PTB) as birth < 37 weeks (obstetric estimate, when available); small for gestational age (SGA) was defined as <10th percentile for gestational age (percentiles based on study population) and large for gestational age (LGA) as >90th percentile for gestational age. Report of both birthweight and gestational age on birth certificates have been found to be highly accurate [25, 26]. A secondary analysis looked at birthweight as a continuous outcome. Racial disparities in birth outcomes were examined, first, unadjusted, then adjusted for the same confounders identified in the previous analysis. Results The study population was approximately one-third black and two-thirds white (Table 1). The majority of the linked pregnancies were first births, although 30% were not. Smoking during pregnancy was listed for 13%. The average age at visit prior to pregnancy was 13 yrs. (range 4–38) and average time between study visit and preg- nancy was 10.8 years (range 0.8–33). Mean and median age at included pregnancy were 24.2 and 23.0. Methods Finally, results were presented adjusted for these confounders as well as car- diovascular risk factors identified as predictors of birth outcomes in this or other analyses, including quadratic associations. As the associations found in this analysis were relatively weak, no formal mediation analysis was performed. Covariates were chosen a priori based on factors known to be associated with exposure and outcome [27, 28]. Race was recorded as white or black at the initial BHS visit. Later follow-ups and birth certificate data with more ex- tensive options indicated that this was consistent with the racial/ethnic self-categorization of almost all participants (< 10 women had Hispanic ethnicity marked on the birth certificate). Participants’ parents provided informed consent for child visits and adult participants provided their own informed consent for BHS measures. The Institutional Review Boards (IRB) of Tulane University (IRB ID#256406), the State Department of Health and Hospitals of Louisiana, and the Texas Department of State Health Services ap- proved the linkage protocol (Mississippi deferred to the Tulane IRB). The linkage was conducted under a waiver of consent, as it was deemed minimal risk and infeasible with- out the waiver. Age at pregnancy was calculated from participant’s date of birth. Data on parity, weight gain, maternal edu- cation, and smoking were taken from the birth certificate data. BMI at time of visit was calculated from measured height and weight: participants were measured in dupli- cate and the average of the measures was used. Adequacy of prenatal care was assessed using the Kotelchuck index [29], and categorized as inadequate, intermediate, ad- equate, and adequate plus. Analysis d Cardiovascular measures at the visit closest in time but prior to the pregnancy (based on an estimated last men- strual period 40 weeks before birthdate) were examined as predictors. Logistic regression was used for dichotom- ous measures. Models were adjusted for variables that were associated with the outcome at p < 0.20: these var- ied by outcome but included age, race, smoking, parity, maternal education, adequacy of prenatal care, BMI at time of visit, age at time of Bogalusa screening, and years between Bogalusa screening and birth. Due to the strong correlation between BMI and cardiovascular risk Only a few cardiovascular risk factors were associ- ated with birth outcomes. Triglycerides were associ- ated with higher risk of LBW among whites (aOR 1.05, 95% 1.01–1.10; Table 2); the effect estimate was very similar although not statistically significant for blacks (1.07, 0.97–1.19). Higher glucose was associ- ated with a reduction in risk of SGA for black women Harville et al. Analysis d BMC Pregnancy and Childbirth (2018) 18:339 Page 4 of 10 Table 1 Participants in the Bogalusa Heart Study linked to birth certificates (N = 3095) Overall study population Black (n = 1139) White (n = 1956) p- value N % N % N % parity < 0.01 1 2116 68.4 760 65.5 1384 70.3 2 630 20.4 229 19.7 410 20.8 3 243 7.9 106 9.1 137 7.0 4+ 104 3.3 66 5.7 38 1.9 education < 0.01 < high school 780 25.2 343 29.6 448 22.7 high school graduate 1190 38.5 474 40.9 732 37.2 some college 575 18.6 219 18.9 361 18.3 college & beyond 547 17.7 124 10.7 429 21.8 married < 0.01 yes 1638 53.0 256 22.1 1398 71.0 no 1452 47.0 903 77.9 571 29.0 smoking < 0.01 yes 376 12.8 380 12.8 328 17.4 no 2566 87.2 2595 87.2 1559 82.6 N mean + SD N mean + SD N mean + SD age at screening 3095 13.0 + 6.7 1139 12.5 + 6.4 1956 13.3 + 6.8 < 0.01 time from screening (yrs) 3095 10.8 + 5.8 1139 10.1 + 5.3 1956 11.1 + 6.0 < 0.01 maternal age 3095 24.2 + 5.4 1139 23.0 + 5.4 1956 24.9 + 5.3 < 0.01 pregnancy weight gain (lbs) 2645 30.8 + 13.6 930 29.0 + 13.8 1715 31.7 + 13.4 < 0.01 BMI 3089 20.1 + 5.5 1136 20.4 + 5.9 1953 20.0 + 5.2 0.08 systolic blood pressure 3090 102.0 + 10.5 1136 102.4 + 11.1 1954 101.8 + 10.1 0.11 diastolic blood pressure 3090 63.1 + 9.7 1136 62.7 + 10.5 1954 63.3 + 9.2 0.13 fasted only: cholesterol 2691 170.4 + 30.9 976 173.2 + 31.7 1715 168.7 + 30.3 < 0.01 triglycerides 2691 77.2 + 45.9 976 66.1 + 29.9 1715 83.5 + 51.8 < 0.01 insulin 2098 11.2 + 8.4 777 12.4 + 10.6 1321 10.5 + 6.6 < 0.01 glucose 2447 78.8 + 16.0 888 77.5 + 17.7 1559 79.4 + 14.9 0.01 LDL-C 2688 103.0 + 28.1 974 103.9 + 28.5 1714 102.4 + 27.9 0.21 HDL-C 2690 56.7 + 16.4 975 60.4 + 16.0 1715 54.5 + 16.3 < 0.01 N % N % N % Low birthweight (full-term only) 91 3.3 52 5.2 39 2.2 < 0.01 Preterm birth 297 9.6 131 11.5 166 8.5 < 0.01 Small for gestational age 353 11.4 182 16.0 171 8.7 < 0.01 Large for gestational age 298 9.6 64 5.6 234 12.0 < 0.01 A quadratic association was found with systolic blood pressure, indicating higher risk for LBW and SGA at lower and higher levels (p < 0.05 for quadratic term). Analysis d (aOR 0.86, 95% CI 0.79–0.95; Table 3), but not whites (p for interaction = 0.02). No associations were found with PTB (Table 4) or LGA (Table 5). Results were similar for birthweight as a continuous outcome (Additional file 1: Table S1). Finally, we directly assessed whether cardiovascular risk might contribute to racial disparities. Racial Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Page 5 of 10 Table 2 Relationship between preconception cardiovascular risk factors and low birthweight White Black p for race* risk factor interaction unadjusted adjusted¶ adjusted unadjusted adjusted¶ adjusted un- adjusted adjusted¶ adjusted** OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI systolic BP 1.17 0.86–1.60 0.91 0.78–1.07 1.41 0.93–2.13 1.02 0.79–1.31 1.02 0.78–1.32 0.93 0.65–1.34 0.49 0.51 0.29 diastolic BP 1.01 0.72–1.43 0.94 0.79–1.12 0.99 0.60–1.63 0.95 0.73–1.24 0.98 0.75–1.29 0.79 0.54–1.15 0.77 0.76 0.54 cholesterol* 0.97 0.87-1.09 1.01 0.96–1.05 1.03 0.90–1.18 1.00 0.91–1.11 1.00 0.91–1.09 1.00 0.88–1.13 0.68 0.69 0.74 triglycerides* 1.03 0.99-1.07 1.01 0.98–1.03 1.05 1.01–1.10 1.05 0.97–1.15 0.98 0.88–1.09 1.07 0.97–1.19 0.62 0.42 0.56 LDL* 0.95 0.83-1.08 0.99 0.94–1.04 1.00 0.85–1.17 0.97 0.87–1.09 1.01 0.91–1.11 0.94 0.81–1.09 0.74 0.74 0.43 HDL* 1.01 0.82-1.24 1.05 0.96–1.15 0.94 0.72–1.23 1.10 0.91–1.32 0.95 0.79–1.14 1.15 0.91–1.46 0.54 0.53 0.18 glucose* 1.03 0.80-1.34 1.01 0.92–1.11 1.16 0.78–1.73 0.90 0.77–1.06 1.05 0.89–1.23 0.91 0.72–1.16 0.39 0.38 0.46 insulin* 1.03 0.59-1.80 0.92 0.68–1.23 1.39 0.79–2.44 1.13 0.90–1.41 0.77 0.44–1.34 1.10 0.87–1.40 0.78 0.88 0.91 fasted ¶adjusted for BMI at last screening adjusted models include risk factor, cigarettes, Kotelchuck index, maternal education, parity, mother’s age at child’s birth, year of birth, BMI at last screening, and time (in years) between screening and birth Harville et al. Discussion In this study, we attempted to determine the relation- ship between pre-pregnancy cardiovascular risk factors and disparities in birth outcomes. Although the ex- pected racial disparities in cardiovascular risk factors and birth outcomes were found, there were only limited relationships among those factors. We found some evi- dence of inverse risk for associations between lipids and birthweight, largely among whites; previous studies have tended to find positive associations between pre-pregnancy lipids and birthweight [20]. Although not statistically significant, the size of the effect esti- mate for systolic blood pressure and birthweight was also consistent with previous studies in Scandinavian populations [20, 22]. Although previous studies of pre-pregnancy blood pressure have not found quadratic associations, such associations have been found in stud- ies of blood pressure during pregnancy [30]. When as- sociations were seen in black women, they tended to be protective, with higher glucose being associated with a reduced risk of LBW and SGA, consistent with previ- ous studies [20]. A previous study of preterm birth found increased risk with both low and high levels of pre-pregnancy cholesterol, but did not find racial differ- ences [21]. Strengths of the study include the prospective data collection; well-characterized cardiovascular risk fac- tors; a fairly large, biracial cohort; and linkage to par- ticipants regardless of later participation. Limitations include the variation in time between the pregnancy and the measure, and the lack of information on car- diovascular risk during the pregnancy, which would assist in determining whether preconception cardio- vascular risk provides any additional information be- yond that determined during pregnancy. In addition, we are limited to those who were able to be linked to vital statistics data. Comparisons of those who are linked and who were not does not point to a strongly high- or low-risk profile in those who were excluded; still, this set of potential participants represents a group that has the potential to change the results if they had been able to be included. On balance, there are indicators that both high- and low-risk women may have been less likely to be included; this may have reduced the variability in the sample and thus limited our ability to find differences. Measurement error per se – of the included mea- sures - should be limited. BHS has rigorous quality control methods, including measurement in duplicate, throughout. Analysis d BMC Pregnancy and Childbirth (2018) 18:339 Page 6 of 10 Table 3 Relationship between preconception cardiovascular risk factors and small-for-gestational-age white black p for raceRF interaction unadjusted adjusted¶ adjusted‡ unadjusted adjusted¶ adjusted‡ unadjusted adjusted¶ adjusted‡ OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI systolic BP 0.97 0.83–1.14 1.06 0.88–1.27 1.08 0.88–1.32 0.96 0.83–1.11 0.99 0.84–1.17 1.03 0.85–1.25 0.87 0.88 0.51 diastolic BP 0.85 0.72–1.01 0.89 0.73–1.08 0.84 0.66–1.06 0.93 0.80–1.09 0.96 0.80–1.13 1.02 0.84–1.26 0.44 0.45 0.69 cholesterol* 1.00 0.95–1.06 1.01 0.95–1.07 1.01 0.95–1.08 0.96 0.91–1.02 0.96 0.91–1.02 0.96 0.90–1.02 0.30 0.28 0.15 triglycerides* 1.00 0.96–1.03 1.00 0.97–1.04 0.99 0.95–1.04 1.02 0.97–1.08 1.04 0.98–1.11 1.04 0.97–1.11 0.32 0.24 0.32 LDL* 1.00 0.94–1.06 1.00 0.94–1.07 0.99 0.92–1.06 0.94 0.89–1.01 0.96 0.90–1.02 0.94 0.87–1.01 0.36 0.35 0.22 HDL* 1.04 0.93–1.15 1.02 0.92–1.14 1.11 0.98–1.26 0.97 0.87–1.08 0.95 0.85–1.06 0.99 0.87–1.12 0.33 0.33 0.28 glucose* 1.05 0.92–1.20 1.04 0.91–1.19 1.11 0.95–1.29 0.86 0.79–0.95 0.85 0.77–0.94 0.85 0.76–0.95 0.02 0.02 0.02 insulin* 0.94 0.69–1.27 1.02 0.74–1.41 0.99 0.68–1.43 1.06 0.90–1.25 1.08 0.91–1.28 1.06 0.89–1.27 0.48 0.59 0.55 * fasted ¶adjusted for BMI ‡adjusted for smoking, Kotelchuck index, maternal education, parity, married, maternal age, time since screening, BMI Table 3 Relationship between preconception cardiovascular risk factors and small-for-gestational-age adjusted for BMI ‡adjusted for smoking, Kotelchuck index, maternal education, parity, married, maternal age, time since screening, BMI [33], or producing epigenetic changes that carry into pregnancy [34, 35]. It could also increase the risk of complications such as pre-eclampsia and gestational diabetes [36–38]; vital records data often have limited validity for those complications and record them in a man- ner problematic for this type of analysis (e.g., grouping pre-pregnancy and gestational diabetes) [25, 26]. Racial dis- parities are also apparent in particularly pregnancy-induced hypertension and pre-eclampsia, and thus these complica- tions could mediate an effect of cardiovascular risk on birth outcomes [39–41]. disparities were clearly present in the sample: black women were at increased risk for LBW (Table 6; ad- justed for non-cardiovascular risk factors, OR 3.84, 95% CI 2.07–7.12), PTB (1.66, 1.24–2.23), and SGA (2.18, 1.61–2.96), and reduced risk for LGA (0.35, 0.25–0.48). Further adjustment for cardiovascular risk factors atten- uated the LBW and SGA estimates by approximately 10%. Analysis d disparities were clearly present in the sample: black women were at increased risk for LBW (Table 6; ad- justed for non-cardiovascular risk factors, OR 3.84, 95% CI 2.07–7.12), PTB (1.66, 1.24–2.23), and SGA (2.18, 1.61–2.96), and reduced risk for LGA (0.35, 0.25–0.48). Further adjustment for cardiovascular risk factors atten- uated the LBW and SGA estimates by approximately 10%. Discussion Birth certificate data is generally good quality for preterm birth and low birthweight [42], Preconception cardiovascular risk could lead to poorer birth outcomes by affecting placentation in the first trimester [21, 31, 32], increasing inflammation Page 7 of 10 Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Table 4 Relationship between preconception cardiovascular risk factors and preterm birth white black p for race* risk factor interaction unadjusted adjusted¶ adjusted† unadjusted adjusted¶ adjusted† OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI Unadjusted adjusted¶ adjusted† systolic BP 0.97 0.83–1.13 1.09 0.91–1.31 1.18 0.96–1.45 0.90 0.76–1.06 1.01 0.83–1.22 1.09 0.87–1.35 0.56 0.54 0.48 diastolic BP 0.90 0.76–1.07 1.00 0.82–1.21 1.21 0.97–1.52 0.97 0.81–1.15 1.10 0.90–1.34 1.21 0.96–1.53 0.57 0.60 0.89 Cholesterol* 1.01 0.95–1.07 1.02 0.97–1.08 1.01 0.95–1.08 0.94 0.88–1.01 0.95 0.89–1.01 0.95 0.89–1.03 0.12 0.10 0.25 Triglyce-rides* 0.95 0.90–1.00 0.96 0.91–1.01 0.98 0.93–1.03 0.97 0.91–1.04 0.99 0.92–1.07 0.99 0.91–1.08 0.59 0.50 0.82 LDL* 1.00 0.94–1.07 1.02 0.96–1.08 1.01 0.94–1.08 0.93 0.87–1.00 0.94 0.87–1.02 0.92 0.85–1.00 0.14 0.12 0.14 HDL* 1.10 1.00–1.22 1.08 0.97–1.20 1.03 0.92–1.17 1.00 0.89–1.13 0.98 0.86–1.11 1.08 0.93–1.24 0.23 0.25 0.89 glucose* 0.98 0.87–1.10 0.96 0.85–1.09 0.94 0.82–1.07 0.89 0.80–0.99 0.86 0.77–0.96 0.86 0.75–0.99 0.21 0.19 0.15 insulin* 0.93 0.68–1.28 1.08 0.80–1.48 0.89 0.58–1.38 0.82 0.60–1.13 0.88 0.64–1.22 0.96 0.70–1.33 0.60 0.51 0.94 * fasted ¶adjusted for BMI †adjusted for smoking Kotelchuck index married maternal education time since screening BMI year of birth Harville et al. Discussion BMC Pregnancy and Childbirth (2018) 18:339 Page 8 of 10 Table 5 Relationship between preconception cardiovascular risk factors and large-for-gestational-age white black unadjusted adjusted adjusted§ unadjusted adjusted¶ adjusted§ unadjusted adjusted¶ adjusted§ OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI systolic BP 1.04 0.91–1.19 0.91 0.78–1.07 0.86 0.72–1.02 0.99 0.79–1.24 1.02 0.78–1.32 0.97 0.71–1.31 0.72 0.69 0.66 diastolic BP 1.07 0.92–1.25 0.94 0.79–1.12 0.91 0.75–1.10 0.96 0.76–1.22 0.98 0.75–1.29 0.96 0.70–1.31 0.45 0.48 0.49 cholesterol* 1.02 0.98–1.07 1.01 0.96–1.05 1.02 0.97–1.08 1.00 0.91–1.09 1.00 0.91–1.09 1.00 0.91–1.10 0.69 0.71 0.57 triglycerides* 1.02 0.99–1.04 1.01 0.98–1.03 1.00 0.98–1.03 0.97 0.88–1.08 0.98 0.88–1.09 1.00 0.89–1.11 0.43 0.33 0.48 LDL* 1.01 0.96–1.06 0.99 0.94–1.04 1.01 0.95–1.06 1.00 0.91–1.11 1.01 0.91–1.11 1.02 0.92–1.14 0.96 0.96 0.93 HDL* 1.01 0.93–1.11 1.05 0.96–1.15 1.05 0.95–1.15 0.96 0.80–1.15 0.95 0.79–1.14 0.88 0.71–1.09 0.61 0.58 0.22 glucose* 0.99 0.89–1.09 1.01 0.92–1.11 0.97 0.88–1.07 1.05 0.89–1.23 1.05 0.89–1.23 1.07 0.90–1.26 0.51 0.59 0.30 insulin* 1.10 0.88–1.38 0.92 0.68–1.23 0.85 0.62–1.16 0.86 0.55–1.34 0.77 0.44–1.34 0.74 0.42–1.31 0.32 0.35 0.40 * fasted ¶adjusted for BMI §adjusted for smoking, parity, time since screening, BMI, maternal age Table 5 Relationship between preconception cardiovascular risk factors and large-for-gestational-age although gestational age measurements may be of limited quality for the oldest women, as ultrasound dating was less consistent during the time period of their pregnancies [43]. However, measurement error still exists in the sense that a single risk factor meas- urement is representing a large time period, and the measurements were not taken at the same time point for every woman, compounding the degree of error. This error would tend to bias the results towards the null, and it is possible that we would see a greater ef- fect with measurements nearer in time to the pregnancy; our sample size for this analysis was too small to provide much information on this point. Measurement of non-cardiovascular covariate data is potentially more problematic. Tobacco use tends to be underreported on birth certificates [44]. Adjustment for covariates is fairly limited, although the strongest risk factors that were likely to be associated with ex- posure and outcomes were included. Overall, residual confounding would likely bias away from the null. Discussion Similarly, if cardiovascular risk factors are acting as confounders of the race-birth outcome relationship, imperfect measurement could be leading to residual con- founding and preventing full adjustment for those factors, suggesting that better measures would more fully attenu- ate the relationships. Although prior preterm birth and low birthweight are associated with the outcome, in most cases the cardiovascular measure would have occurred prior to that birth as well, so it is unlikely that adjustment for this risk factor would be valid [45]. Competing interests d EWH is an associate editor of this journal. Additional file Additional file 1: Table S1. Relationship between preconception cardiovascular risk factors and continuous birthweight, the Bogalusa Heart Study. (DOCX 14 kb) Funding 7. Arbour MW, Corwin EJ, Salsberry PJ, Atkins M. Racial differences in the health of childbearing-aged women. MCN Am J Matern Child Nurs. 2012;37(4):262–8. The Bogalusa Heart Study is supported by grants R01HD069587, R01HL016592, R01AG041200, P50HL015103, R01HD032194. This work was supported in part by U54 GM104940 from the National Institute of General Medical Sciences of the National Institutes of Health, which funds the Louisiana Clinical and Translational Science Center. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. 8. Posner SF, Broussard DL, Sappenfield WM, Streeter N, Zapata LB, Peck MG. Where are the data to drive policy changes for preconception health and health care? Womens Health Issues. 2008;18(6 Suppl):S81–6. 8. Posner SF, Broussard DL, Sappenfield WM, Streeter N, Zapata LB, Peck MG. Where are the data to drive policy changes for preconception health and health care? Womens Health Issues. 2008;18(6 Suppl):S81–6. 9. Berends AL, de Groot CJ, Sijbrands EJ, Sie MP, Benneheij SH, Pal R, Heydanus R, Oostra BA, van Duijn CM, Steegers EA. Shared constitutional risks for maternal vascular-related pregnancy complications and future cardiovascular disease. Hypertension. 2008;51(4):1034–41. 9. Berends AL, de Groot CJ, Sijbrands EJ, Sie MP, Benneheij SH, Pal R, Heydanus R, Oostra BA, van Duijn CM, Steegers EA. Shared constitutional risks for maternal vascular-related pregnancy complications and future cardiovascular disease. Hypertension. 2008;51(4):1034–41. Availability of data and materials 10. Ashton DM, Lawrence HC 3rd, Adams NL 3rd, Fleischman AR. Surgeon General's conference on the prevention of preterm birth. Obstet Gynecol. 2009;113(4):925–30. The datasets analyzed during the current study are not publicly available due to confidentiality/research subject protection but are available from the corresponding author on reasonable request complying with applicable ethical standards. 11. Shapiro-Mendoza CK, Barfield WD, Henderson Z, James A, Howse JL, Iskander J, Thorpe PG. CDC grand rounds: public health strategies to prevent preterm birth. MMWR Morb Mortal Wkly Rep. 2016;65(32):826–30. Conclusions Mechanisms underlying the persistence of racial dispar- ities in birth outcomes continue to elude public health researchers. The findings presented in this study do not include strong patterns of association between lifetime cardiovascular risk profiles and racial differences in inci- dence of adverse birth outcomes despite sound theoret- ical plausibility. However, the ambiguity in our results underscores the need for more research that considers pre-pregnancy health status and biological pathways Table 6 Racial disparities in birth outcomes and cardiovascular risk factors unadjusted adjusted additionally adjusted for cardiovascular risk factors OR 95% CI OR 95% CI OR 95% CI Low birthweight 2.44 1.69–3.73 3.84a 2.07-7.12 3.32e 1.52-7.21 Preterm birth 1.40 1.10–1.79 1.66b 1.24–2.23 1.65f 1.19-2.30 Small-for-gestational-age 1.99 1.59–2.48 2.18c 1.61-2.96 1.95g 1.37-2.76 Large-for-gestational-age 0.44 0.33–0.58 0.35d 0.25-0.48 0.32h 0.22-0.45 Adjusted for: a(same adjustments as previous tables) smoking, Kotelchuck index, married, maternal education, time since screening, BMI, year of birth bsmoking, Kotelchuck index, maternal education, parity, maternal age, year of birth, time since screening, BMI csmoking, Kotelchuck index, maternal education, parity, married, maternal age, time since screening, BMI dsmoking, parity, time since screening, BMI, maternal age eprevious column + triglycerides, insulin, systolic blood pressure, systolic blood pressure-squared fprevious column + glucose, LDL gprevious column + glucose, systolic blood pressure, systolic blood pressure-squared hprevious column + glucose Page 9 of 10 Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Page 9 of 10 underlying racial disparities in birth outcomes. Future studies should aim for more precise measures and bio- logical indicators of mechanism to improve our under- standing of these outcomes. 2Department of Global Biostatistics and Data Science, Tulane School of Public Health and Tropical Medicine, New Orleans, LA, USA. 3Department of Global Community Health and Behavioral Sciences, Tulane School of Public Health and Tropical Medicine, New Orleans, LA, USA. Received: 10 February 2017 Accepted: 31 July 2018 Abbreviations BHS: Bogalusa Heart Study; BMI: Body mass index; HDL: High-density lipoprotein; LBW: Low birthweight; LDL: Low-density lipoprotein; LGA: Large- for-gestational-age; OR: Odds ratio; PTB: Preterm birth; SGA: Small-for- gestational-age; SSN: Social Security Number 3. Whitehead N, Helms K. Racial and ethnic differences in preterm delivery among low-risk women. Ethn Dis. 2010;20(3):261–6. 4. Alexander GR, Kogan MD, Himes JH, Mor JM, Goldenberg R. Racial differences in birthweight for gestational age and infant mortality in extremely-low-risk US populations. Paediatr Perinat Epidemiol. 1999;13(2):205–17. Acknowledgements 5. Finkelstein EA, Khavjou OA, Mobley LR, Haney DM, Will JC. Racial/ethnic disparities in coronary heart disease risk factors among WISEWOMAN enrollees. J Womens Health (Larchmt). 2004;13(5):503–18. Richard Johnson and Judy Moulder at the Mississippi State Department of Health Chris Simmons and Jamie Huang at the Texas Department of State Health Services. 6. Kurian AK, Cardarelli KM. Racial and ethnic differences in cardiovascular disease risk factors: a systematic review. Ethn Dis. 2007;17(1):143–52. Authors’ contributions 12. Lu MC, Kotelchuck M, Hogan V, Jones L, Wright K, Halfon N. Closing the black-white gap in birth outcomes: a life-course approach. Ethn Dis. 2010; 20(1 Suppl 2):S2-62-76. EWH conceived and wrote the paper, and supervised data linkage and analysis. LM advised on biostatistics and performed data analysis. TS constructed relevant datasets and participated in data analysis. MEW performed the data linkage and assisted in paper conceptualization. LAB supervised data collection and study design. All authors revised the paper critically for content and contributed to study and analysis design. All authors read and approved the final manuscript. 13. Ramey SL, Schafer P, DeClerque JL, Lanzi RG, Hobel C, Shalowitz M, Chinchilli V, Raju TN. The preconception stress and resiliency pathways model: a multi-level framework on maternal, paternal, and child health disparities derived by community-based participatory research. Matern Child Health J. 2015;19(4):707–19. p y g p p critically for content and contributed to study and analysis design. All authors read and approved the final manuscript. 14. D’Angelo D, Williams L, Morrow B, Cox S, Harris N, Harrison L, Posner SF, Hood JR, Zapata L. Preconception and interconception health status of women who recently gave birth to a live-born infant--pregnancy risk assessment monitoring system (PRAMS), United States, 26 reporting areas, 2004. MMWR Surveill Summ. 2007;56(10):1–35. 14. D’Angelo D, Williams L, Morrow B, Cox S, Harris N, Harrison L, Posner SF, Hood JR, Zapata L. Preconception and interconception health status of women who recently gave birth to a live-born infant--pregnancy risk assessment monitoring system (PRAMS), United States, 26 reporting areas, 2004. MMWR Surveill Summ. 2007;56(10):1–35. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 18. Melamed N, Chen R, Soiberman U, Ben-Haroush A, Hod M, Yogev Y. Spontaneous and indicated preterm delivery in pregestational diabetes mellitus: etiology and risk factors. Arch Gynecol Obstet. 2008;278(2):129–34. References h 1. Matthews TJ, Mac Dorman MF. Infant mortality statistics from the 2010 period linked birth/infant death data set. Natl Vital Stat Rep. 2013;62(8):1–26. 2. Hatch M, Berkowitz G, Janevic T, Sloan R, Lapinski R, James T, Barth WH Jr. Race, cardiovascular reactivity, and preterm delivery among active-duty military women. Epidemiology. 2006;17(2):178–82. Ethics approval and consent to participate pp p p Participants’ parents provided informed consent for child visits and adult participants provided their own informed consent for BHS measures. The Institutional Review Boards (IRB) of Tulane University (IRB ID#256406), the State Department of Health and Hospitals of Louisiana, and the Texas Department of State Health Services approved the linkage protocol (Mississippi deferred to the Tulane IRB). The linkage was conducted under a waiver of consent, as it was deemed minimal risk and infeasible without the waiver. 15. Odell CD, Kotelchuck M, Chetty VK, Fowler J, Stubblefield PG, Orejuela M, Jack BW. Maternal hypertension as a risk factor for low birth weight infants: comparison of Haitian and African-American women. Matern Child Health J. 2006;10(1):39–46. 15. Odell CD, Kotelchuck M, Chetty VK, Fowler J, Stubblefield PG, Orejuela M, Jack BW. Maternal hypertension as a risk factor for low birth weight infants: comparison of Haitian and African-American women. Matern Child Health J. 2006;10(1):39–46. Consent for publication Not applicable Competing interests EWH is an associate editor of this journal. Competing interests EWH is an associate editor of this journal. Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 20. Romundstad PR, Davey Smith G, Nilsen TI, Vatten LJ. Associations of prepregnancy cardiovascular risk factors with the offspring’s birth weight. Am J Epidemiol. 2007;166(12):1359–64. 21. Catov JM, Ness RB, Wellons MF, Jacobs DR, Roberts JM, Gunderson EP. Prepregnancy lipids related to preterm birth risk: the coronary artery risk development in young adults study. J Clin Endocrinol Metab. 2010; 95(8):3711–8. 22. Harville EW, Viikari JS, Raitakari OT. Preconception cardiovascular risk factors and pregnancy outcome. Epidemiology. 2011;22(5):724–30. 23. Ehrenthal DB, Jurkovitz C, Hoffman M, Kroelinger C, Weintraub W. A population study of the contribution of medical comorbidity to the risk of prematurity in blacks. Am J Obstet Gynecol. 2007;197(4):409. e401–406 p y y 24. Berenson GS. Bogalusa heart study: a long-term community study of a rural biracial (black/white) population. Am J Med Sci. 2001;322(5):293–300. 25. Vinikoor LC, Messer LC, Laraia BA, Kaufman JS. Reliability of variables on the North Carolina birth certificate: a comparison with directly queried values from a cohort study. Paediatr Perinat Epidemiol. 2010;24(1):102–12. 26. Zollinger TW, Przybylski MJ, Gamache RE. Reliability of Indiana birth certificate data compared to medical records. Ann Epidemiol. 2006;16(1):1–10. 27. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet. 2008;371(9606):75–84. 28. Kramer MS. The epidemiology of low birthweight. Nestle Nutr Inst Workshop Ser. 2013;74:1–10. 29. Kotelchuck M. The adequacy of prenatal care utilization index: its US distribution and association with low birthweight. Am J Public Health. 1994; 84(9):1486–9. 30. Zhang J, Klebanoff MA. Low blood pressure during pregnancy and poor perinatal outcomes: an obstetric paradox. Am J Epidemiol. 2001;153(7):642–6. 31. Thorp JM. Placental vascular compromise: unifying the etiologic pathways of perinatal compromise. Curr Prob Obstet Gynecol. 2001;24(6):197–220. 32. Kroener L, Wang ET, Pisarska MD. Predisposing factors to abnormal first trimester placentation and the impact on fetal outcomes. Semin Reprod Med. 2016;34(1):27–35. 33. McDade TW, Borja JB, Largado F, Adair LS, Kuzawa CW. Adiposity and chronic inflammation in young women predict inflammation during normal pregnancy in the Philippines. J Nutr. 2016;146(2):353–7. 34. St Fleur M, Damus K, Jack B. The future of preconception care in the United States: multigenerational impact on reproductive outcomes. Ups J Med Sci. 2016:1–5. 35. Steegers-Theunissen RP, Twigt J, Pestinger V, Sinclair KD. The periconceptional period, reproduction and long-term health of offspring: the importance of one-carbon metabolism. Hum Reprod Update. 2013; 19(6):640–55. 36. 43. Campbell S. A short history of Sonography in obstetrics and Gynaecology. Facts Views Vis Obgyn. 2013;5(3):213–29. 44. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs. 2006;35(1):3–12. 45. Howards PP, Schisterman EF, Poole C, Kaufman JS, Weinberg CR. “Toward a clearer definition of confounding” revisited with directed acyclic graphs. Am J Epidemiol. 2012;176(6):506–11. Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Author details 1 1Department of Epidemiology, Tulane School of Public Health and Tropical Medicine, 1440 Canal St. Ste. 2001, #8318, New Orleans, LA 70112-2715, USA. 19. Simmons D. Diabetes and obesity in pregnancy. Best Pract Res Clin Obstet Gynaecol. 2011;25(1):25–36. Page 10 of 10 Page 10 of 10 Harville et al. BMC Pregnancy and Childbirth (2018) 18:339 Egeland GM, Klungsoyr K, Oyen N, Tell GS, Naess O, Skjaerven R. Preconception cardiovascular risk factor differences between gestational hypertension and preeclampsia: cohort Norway study. Hypertension. 2016; 67(6):1173–80. 37. Baumfeld Y, Novack L, Wiznitzer A, Sheiner E, Henkin Y, Sherf M, Novack V. Pre-conception dyslipidemia is associated with development of preeclampsia and gestational diabetes mellitus. PLoS One. 2015;10(10): e0139164. 38. Harville EW, Juonala M, Viikari JS, Raitakari OT. Preconception metabolic indicators predict gestational diabetes and offspring birthweight. Gynecol Endocrinol. 2014;30(11):840–4. 39. Baraban E, McCoy L, Simon P. Increasing prevalence of gestational diabetes and pregnancy-related hypertension in Los Angeles County, California, 1991-2003. Prev Chronic Dis. 2008;5(3):A77. 39. Baraban E, McCoy L, Simon P. Increasing prevalence of gestational diabetes and pregnancy-related hypertension in Los Angeles County, California, 1991-2003. Prev Chronic Dis. 2008;5(3):A77. 40. Cabacungan ET, Ngui EM, McGinley EL. Racial/ethnic disparities in maternal morbidities: a statewide study of labor and delivery hospitalizations in Wisconsin. Matern Child Health J. 2012;16(7):1455–67. 40. Cabacungan ET, Ngui EM, McGinley EL. Racial/ethnic disparities in maternal morbidities: a statewide study of labor and delivery hospitalizations in Wisconsin. Matern Child Health J. 2012;16(7):1455–67. 41. Mbah AK, Alio AP, Marty PJ, Bruder K, Wilson R, Salihu HM. Recurrent versus isolated pre-eclampsia and risk of feto-infant morbidity outcomes: racial/ ethnic disparity. Eur J Obstet Gynecol Reprod Biol. 2011;156(1):23–8. 41. Mbah AK, Alio AP, Marty PJ, Bruder K, Wilson R, Salihu HM. Recurrent versus isolated pre-eclampsia and risk of feto-infant morbidity outcomes: racial/ ethnic disparity. Eur J Obstet Gynecol Reprod Biol. 2011;156(1):23–8. 42. Martin JA, Wilson EC, Osterman MJ, Saadi EW, Sutton SR, Hamilton BE. Assessing the quality of medical and health data from the 2003 birth certificate revision: results from two states. Natl Vital Stat Rep. 2013;62(2):1–19. 42. Martin JA, Wilson EC, Osterman MJ, Saadi EW, Sutton SR, Hamilton BE. Assessing the quality of medical and health data from the 2003 birth certificate revision: results from two states. Natl Vital Stat Rep. 2013;62(2):1–19.
https://openalex.org/W1947983253	https://retrovirology.biomedcentral.com/counter/pdf/10.1186/1742-4690-12-S1-O27	English	null	Host determinants of HTLV-1 integration site preference	Retrovirology	2,015	cc-by	578	McCallin et al. Retrovirology 2015, 12(Suppl 1):O27 http://www.retrovirology.com/content/12/S1/O27 Open Access Open Access Host determinants of HTLV-1 integration site preference From 17th International Conference on Human Retroviruses: HTLV and Related Viruses Trois Ilets, Martinique. 18-21 June 2015 Published: 28 August 2015 Published: 28 August 2015 Integration of the provirus into the host genome is an essential step in the establishment of a chronic retroviral infection. Retroviruses do not integrate at random. Each genus of retrovirus preferentially integrates in genomic sites associated with certain genomic features. This pre- ference is thought to result from an interaction of the pre-integration complex with a specific host co-factor. Such an interaction has been well documented between HIV-1 integrase and lens epithelium-derived growth fac- tor (LEDGF) and between MLV and members of the Bromoextraterminal Domain Protein (BET) family. HTLV-1 also has a reproducible preference to integrate near various genomic features; however, the putative host co-factor of HTLV-1 integrase remains unknown. We present evidence that demonstrates that the HTLV- 1 integrase co-factor may be the protein phosphatase 2A (PP2A) complex. Two independent protein interac- tion screens (yeast-2-hybrid and co-immunoprecipita- tion) both revealed an interaction between HTLV-1 integrase and the PP2A holoenzyme. We used co- culture with HTLV-1 producer cell line MT2 to infect rat cell lines containing an inducible knockdown of PP2A. We found that PP2A-knockdown altered the integration profile of these de novo integration sites. In contrast to infection in PP2A-competent cells, integra- tion sites in PP2A-knockdown cells lacked the charac- teristic proximity to genomic features associated with transcriptional activity (genes, transcription start sites, CpG islands) and to epigenetic marks associated with active chromatin. We are in the process of investigating how PP2A inhibitors and shRNAs targeting specific reg- ulatory subunits of the PP2A holoenzyme influence inte- gration site selection. doi:10.1186/1742-4690-12-S1-O27 Cite this article as: McCallin et al.: Host determinants of HTLV-1 integration site preference. Retrovirology 2015 12(Suppl 1):O27. doi:10.1186/1742-4690-12-S1-O27 Cite this article as: McCallin et al.: Host determinants of HTLV-1 integration site preference. Retrovirology 2015 12(Suppl 1):O27. * Correspondence: amy.mccallin08@imperial.ac.uk Division of Infectious Diseases, St Marys Hospital, Imperial College London, W2 1PG, UK © 2015 McCallin et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http:// creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/ zero/1.0/) applies to the data made available in this article, unless otherwise stated. Host determinants of HTLV-1 integration site preference Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit cle distributed under the terms of the Creative Commons Attribution License (http:// rmits unrestricted use, distribution, and reproduction in any medium, provided the mons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/ this article, unless otherwise stated. Submit your next manuscript to BioMed Central and take full advantage of: © 2015 McCallin et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http:// creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/ zero/1.0/) applies to the data made available in this article, unless otherwise stated.
https://openalex.org/W3094669679	https://zenodo.org/records/6523954/files/Resende%20et%20al.%202020%20-%20Decision%20Support%20Systems%20to%20Promote%20Health%20and%20Well-being%20of%20People%20during%20their%20Working%20Age%20the%20Case%20of%20the%20WorkingAge%20EU%20Project.pdf	English	null	Decision Support Systems to Promote Health and Well-Being of People During Their Working Age: The Case of the WorkingAge EU Project	Lecture notes in computer science	2,020	cc-by	5,772	⋆This project has received funding from the European Union’s Horizon 2020 research and innovation programme, under grant agreement N. 826232. 1 Introduction The increasing longevity of population in industrialized countries is shifting the age distribution of the workforce. However, the aging process can lead to physio- logical and cognitive changes that can aﬀect the well-being and quality of life of workers. Workplace design and organizational strategy can play a pivotal role in maximizing the comfort and performance of occupants, as well as in promotion healthy habits in working/living environments. The ongoing EU project WorkingAge (Smart Working environments for all Ages - https://www.workingage.eu/) – WA for short – focuses on these issues and promotes healthy habits in working environments targeting people aged over 45. The aim of the project is to achieve a better understanding of well-being at work and of factors that may inhibit or deteriorate prolonged employment [9]. The Working Age Of Wellbeing (WAOW) tool is under development to provide workers with assistance in their everyday routine in the form of recommenda- tions, risks avoidance and reminders. The WAOW tool monitors the worker’s behavior, health data and preferences through continuous data collection using Internet of Things (IoT) devices –sensors, interaction tools, wearable devices– to provide recommendations on working habits, physical activities and social re- lations. The core of the WAOW tool environment is a Decision Support System (DSS), a reasoning system about the worker conditions and for early detection of possible issues. The DSS supports workers in taking eﬀective and personalized decisions and interventions based on a set of measured indicators. The paper is organized as follows. Section 2 reports related work; Section 3 provides an overview of the WAOW tool; Section 4 introduces the interven- tion strategies to improve workers’ well being; Section 5 introduces the WAOW Ontology-based data model; Section 6 shows how the DSS provide recommen- dations to the workers; ﬁnally, Section 7 outlines conclusions and future work. Decision Support Systems to Promote Health and Well-being of People during their Working Age: the Case of the WorkingAge EU Project⋆ Rosa Maria Resende de Almeida5[0000−0003−3721−7214], Adriana Grau Aberturas5, Yolanda Bueno Aguado5, Maurizio Atzori1[0000−0001−6112−7310], Alessandro Barenghi3[0000−0003−0840−6358], Gianluca Borghini2[0000−0001−8560−5671], Carlos Alberto Catalina Ortega5[0000−0003−1788−0596], Sara Comai3[0000−0002−9554−8815], Raquel Losada Dur´an5, Mariagrazia Fugini3[0000−0002−0692−0153], Hatice Gunes4[0000−0003−2407−3012] Basam Musleh Lancis5[0000−0002−7485−7029], Gerardo Pelosi3[0000−0002−3812−5429], Vincenzo Ronca2[0000−0002−7174−6331], Licia Sbattella3, Roberto Tedesco3[0000−0002−2830−4247], and Tian Xu4 1 Universit`a di Cagliari - DMI, Cagliari, Italy atzori@unica.it 2 BrainSigns s.r.l., Roma, Italy {gianluca.borghini, vincenzo.ronca}@brainsigns.com 3 Politecnico di MIlano - DEIB, Milano, Italy {alessandro.barenghi, sara.comai, mariagrazia.fugini, gerardo.pelosi, licia.sbattella, roberto.tedesco}@polimi.it 4 University of Cambridge - DCST, Cambridge, U.K. {hatice.gunes, tian.xu}@cl.cam.ac.uk 5 Instituto Tecnol´ogico de Castilla y Le´on, Burgos, Spain {carlos.catalina, basam.musleh}@itcl.es Fundaci´on INTRAS, Spain {rra, aga, rld, yba}@intras.es 1 Universit`a di Cagliari - DMI, Cagliari, Italy atzori@unica.it 2 BrainSigns s.r.l., Roma, Italy {gianluca.borghini, vincenzo.ronca}@brainsigns.com 3 Politecnico di MIlano - DEIB, Milano, Italy {alessandro.barenghi, sara.comai, mariagrazia.fugini, gerardo.pelosi, licia.sbattella, roberto.tedesco}@polimi.it 4 University of Cambridge - DCST, Cambridge, U.K. {hatice.gunes, tian.xu}@cl.cam.ac.uk 5 Instituto Tecnol´ogico de Castilla y Le´on, Burgos, Spain {carlos.catalina, basam.musleh}@itcl.es Fundaci´on INTRAS, Spain {rra, aga, rld, yba}@intras.es } 4 University of Cambridge - DCST, Cambridge, U.K. {hatice.gunes, tian.xu}@cl.cam.ac.uk 5 Instituto Tecnol´ogico de Castilla y Le´on, Burgos, Spain {carlos.catalina, basam.musleh}@itcl.es Fundaci´on INTRAS Spain {carlos.catalina, basam.musleh}@itcl.es Fundaci´on INTRAS, Spain {rra, aga, rld, yba}@intras.es Abstract. The WorkingAge project aims at improving the psycho-physical condition of workers, with a special focus on ageing subjects. In this context, a Decision Support System, based on a hybrid data-driven / model-driven approach, and fed with data coming from a plethora of en- vironmental and wearable sensors, provides personalised advises to the worker. In this paper we brieﬂy present the WorkingAge project and ar- chitecture, and then focus on the decision-making pipeline that, starting from raw data, generates the advises. Keywords: Occupational Safety · Decision Support Systems for Health at Work · Suggestion of Health Strategies · Occupational Ontology. ⋆This project has received funding from the European Union’s Horizon 2020 research and innovation programme, under grant agreement N. 826232. R. M. Resende de Almeida et al. 2 6 https://www.bsigroup.com/en-GB/Occupational-Health-and-Safety-ISO-45001/ 2 Related Work Safety and well-being at work is recognized to be highly related to social, eco- nomic, and environmental conditions [19]. These should be enhanced to improve health and well aging [12], in the line of what recommended by the Occupational Health and Safety (OHS) management system standard6 described in [6]. An in- teresting overview of problems related to aging and well-being at work and in societal environments is given in [13]. Working, even at a relatively advanced age, is considered good to maintain physical and mental health [10]. However, not all workers beneﬁt of favorable conditions, due to dangerous or stressing work environments. Moreover, ageing and shrinking of work forces [2] worsen the work conditions of many classes of workers, due to staﬀshortage and lack of expertise in young generations [1]. New working and living conditions are encouraged in many studies and experiences, 3 WorkingAge EU Project 3 centered around concepts of autonomy and quality of life [4,8] The WAOW tool can be classiﬁed as a Self-Management Occupational Safety and Health Supervi- sion System (SMOSHS System) [3]. The tool is being designed considering DSS issues, in order to adapt its interface and functions to the user, in the line of what proposed in [15] and considering that the partners involved in experiments apply procedures compliant with GDPR. Coming to the DSS, in the last years, DSSes have become quite popular in health services, to enhance interaction between patients and their physicians [16], in social care [5], and in many other areas, using technologies which commonly draw on an existing knowledge base of evidence and guidelines to provide logical reasoning-based expert advice. In [21], a DSS is described for workers with dis- abilities who are more sensitive for stress at work and for the injuries connected with non-adequate workplaces. Based on ontologies, many DSSes have been de- veloped for business decision making, or for cyber-physical systems [17,18], which are very similar to our application area. In fact, these systems tightly integrate software, human, and physical components with the need to satisfy constraints on performance and safety of the monitored activities, exhibiting a high degree of automation for the management of their functionalities and decision making processes. In general, there is a strong need for semantic models of the involved application domains. 2 Related Work In such a context, the clinical domain [14] witnesses an increasing demand to develop knowledge-based DSSs employing medical knowl- edge and expertise coming from diﬀerent ontologies. 3 The WAOW Tool The main goal of the WA project is designing and developing the WAOW tool. Such tool provides feedback to the employers by means of warnings and advises. Furthermore, the WAOW tool collects information useful to perform changes in the workplace that may lead to an increased level of comfort for the work- ers and new solutions to promote ﬂexible and sustainable job longevity mea- sures. The WAOW tool is focused on the usage of innovative Human Com- puter Interaction (HCI) methods, including augmented reality, virtual reality, gesture/voice recognition and gaze tracking. It will measure the user’s psycho- logical/emotional/health state, collect data about the environmental conditions of her/his workplace, and will provide advises taking into consideration also gender, ethics and security aspects. The WAOW tool architecture can be di- vided into three parts (see Fig. 1). Firstly, there are numerous sensors that are deployed in the workplace or worn by the worker in order to acquire all the in- formation (Raw data) required by the system; notice that, another input source is provided by questionnaires administered to the worker by means of the WA App, running on the worker’s phone. These pieces of information are then sent via Wi-Fi or Bluetooth to the Edge Cloud (the second part of the architecture) to be processed by means of diﬀerent algorithms to the end of computing a rel- atively small set of High-Level information regarding the well-being status of the workers. Finally, High-Level information is sent to the WA App, where it is 4 R. M. Resende de Almeida et al. Classiﬁer Sensors Trained model Facts Data-driven approach Model-driven approach High-Level info Rules Advice Decision Support System Rule updater Worker Feedback Ontology Edge Cloud servers WA App Wi-Fi BT Wi-Fi Question -naires WA App Raw data Fig. 1. The WAOW tool decision making pipeline. Classiﬁer Sensors Trained model Data-driven approach Edge Cloud servers Wi-Fi BT Question -naires WA App Raw data Facts Model-driven approach Rules Advice Decision Support System Rule updater Worker Feedback Ontology WA App i Fig. 1. The WAOW tool decision making pipeline. stored into the Ontology and processed by the DSS, which generates advises to the worker (see Section 6 for details). 4 The Intervention Strategy The WA project intends to develop an e-coach system of interaction to implement an intervention program with workers over 45 years to prolong their work ability and autonomy, reduce strain in diﬀerent working conditions and promote well- being throughout diﬀerent areas of their lives, promoting an inclusive approach to the health at work. It seeks to increase awareness of workers on aspects such as ergonomic conditions, occupational hazards, stress and its eﬀects on health, providing the worker with the knowledge and skills to reduce negative eﬀects or adopt healthier and safer behaviors. The WA Intervention Framework is pillared on well-known models, such as the CDC Workplace Health Model, the Plan-Do-Check-Act (PDCA) cycle adopted by the Occupational Safety and Health Administration for the certiﬁ- cation process of the OHSAS 18001 management standard, and the Integrated Risk Management (IRM) model, a set of practices and processes supported by a risk-aware culture. It considers the multi-factorial nature of subjective well- being and work ability to ensuring a holistic and comprehensive approach to the individual. The WA Intervention Strategy designed so far considers wellness on physical, emotional, intellectual, occupational, social and environmental dimen- sions. The integrated approach fosters awareness on psycho-social and ergonomic risk prevention measures, seeking to promote attitudinal and behavioral changes towards healthy lifestyles. From this perspective, WA addresses four major areas of intervention: Ergonomics & Physical environment (which promotes adequate work postures and encourages healthy work habits), Worker’s health and per- sonal characteristics (which encourages the user to follow and/or acquire health promotion and disease prevention strategies), Psycho-social factors (which aims at increasing the self-awareness about cognitive and emotional states of activa- tion or discomfort) and Lifestyle & Health habits (which promotes the acquisition and maintenance of healthy lifestyle habits). The goal of the WA Interventions 5 WorkingAge EU Project WorkingAge EU Project is not limited to leveraging the power of technology for recommend practices for safety and health but also for general well.being. Taking advantage of the technological concept of WA the proposed intervention approach aims to change the way the workers see and take care of their health, helping to better un- derstand their own health, improving self-awareness of well-being, and changing how workers access to assessment data and advice. For employers, WA provides many opportunities for smarter health and well-being support. 4 The Intervention Strategy In the develop- ment of the WA Intervention Framework and Strategies, a multidisciplinary team of expert professionals designed the intervention around the following pillars. Framework for stress and strain assessment and intervention. Strain is an essential aspect for assessing the eﬀects of work on a given person describing her/his reactions to the set of working conditions. Stress describes the external characteristics of a work situation that inﬂuences the working person. These include, for example, physical and organizational working conditions. Within the WAOW tool, there will be diﬀerent sensors used to measure physical (muscular, skeletal, cardiovascular and somatic) and psychological (mental and emotional) strain, as deﬁned in [20]. The analysis of these measures lets the WAOW tool to generate a personalized advice, which aims at improving the worker’s condition, based on the strain typology. WA focuses the design of the intervention in a goal-oriented approach, where coaching is “essentially about helping individuals regulate and direct their inter- personal and interpersonal resources to better attain their goals” [11]. Speciﬁc, Measurable, Achievable, Relevant Time bound (SMART) goals will be progres- sively tuned around the user proﬁle, promoting behavioral changes, improving the healthy habits of the worker, taking into account behavioral, environmental and personal factors. Moreover, a speciﬁc set of interventions will be developed to meet the diﬀerent SMART goals along the WA tool usage. The Recommendation System. The DSS, incorporating the framework de- picted above, will generate the appropriate recommendation. Relevant informa- tion for the DSS will come from measurements, from SMART goals and from questionnaires (see Section 6). The DSS will propose an advice at any moment, and/or in response to a speciﬁc measure, indicative of a possible risk to the user’s health. The recommendations are focused on ergonomics, psycho-educational and emotional activation control techniques, and behavioural modiﬁcation. 5 The Ontology The WA Ontology represents the data model for the High-level information sam- ples generated by Edge Cloud servers. Starting from such pieces of information, the DSS will generate personalized advices to the workers. The Ontology is represented by means of the OWL2 DL language. In the OWL language, a concept is represented as a class, if it can be seen as a set; the relationships between classes are named object properties; the attributes belonging to a class are named datatype properties; ﬁnally, individuals deﬁne the R. M. Resende de Almeida et al. 6 Worker UserID: string AddOnDate: YYYYMMDD hasMeasurement Sensor hasSocialEnvironment SocialEnvironment hasActivity Activity receivesAdvice Advice providesFeedback Feedback refersTo hasSmartGoals SmartGoal Timestamp: YYYYMMDDHHMMSS Timestamp: YYYYMMDDHHMMSS Typology: string Typology: {Positive, Negative} Category: {ErgonomicHaz, PhysicalHaz, SecurityHaz, Environmental, Cognitive} 0..N 0..N 0..N 0..N 0..N 0..N Probability: 0..1 hasProﬁle Proﬁle Gender: {Male, Female, Other} Birthday: YYYYMMDD GeneralHealthBegin: { Much_better_now, Somewhat_better_now, About_the_same, Somewhat_worse, Much_worse} measuredDuringActivity GeneralHealthEnd: {Much_better_now, Somewhat_better_now, About_the_same, Somewhat_worse, Much_worse} WorkPosition: {Production_worker, Ofﬁce_worker, Teleworker} Height: nonNegativeInteger Weight: nonNegativeInteger Fig. 2. Description of the worker, and related classes. Birthday: YYYYMMDD AddOnDate: YYYYMMDD Gender: {Male, Female, Other} UserID: string Proﬁle WorkPosition: {Production_worker, Ofﬁce_worker, Teleworker} Fig. 2. Description of the worker, and related classes. instances of a class, i.e., elements of a set. For providing the best privacy, the resulting OWL ﬁle is stored on the worker’s phone, into the WA App. In our Ontology each datatype property corresponds to a speciﬁc High-level information, and whenever a new High-level information is added to the Ontol- ogy, a new individual is created. In the following, we describe the ﬁve parts that compose our Ontology, using a simple graphical language, where squares repre- sent classes, arcs show object properties and “pins” depict datatype properties. A dotted square denotes a “link” to a class that is described into other drawings. Worker. This part is the “core” of the Ontology and describes the proﬁle of the worker, as well as the connections to all the relevant information provided by other parts of the Ontology (see Fig. 2). The Worker class, together with the Profile class, describe the basic characteristics of the worker. Notice that UserID is described as a key for the class Worker, meaning that such a datatype property will be used by the WAOW tool to identify a speciﬁc worker. A worker is associated with zero or more advices, and can provide a corresponding feedback. 5 The Ontology advices are associated with a probability; this is due to the fact that any High- level information is, in general, generated by data-driven classiﬁers. We leverage that characteristic by providing not only the classiﬁer output (which corresponds to a High-level information to be stored into the corresponding datatype property of a speciﬁc class) but also the related probability. Thus, the ProbLog engine can provide a probabilistic advice; such probability can be seen as a reliability score associated with the advice. Finally, the worker is associated with several sensor measurements, social environments, activities, and SMART goals. Note that a sensor measurement happens in the context of a given activity; in fact, a worker can be measured at the workplace or during free time. Sensor. This is the most complex part of the Ontology and describes the High-level information samples calculated from the Edge Cloud servers, about a 7 WorkingAge EU Project SHQVRU EQYLURQPHQWaOSHQVRU LVA BRG\SHQVRU LVA OWKHUSHQVRU LVA MLFURSKRQH LVA FaFLaOE[SUHVV LVA E\HMRYH LVA BRG\PRVWXUH LVA SPaUWEaQG LVA LVA TKHUPRH\JURPHWULF LVA LLJKW LVA UVHULRFaWLRQ LVA QXHVWLRQQaLUHV LVA MHQWaOSWUaLQLHYHO:¬^HLJK, MHGLXP, LRZ` TLPHVWaPS: YYYYMMDDHHMMSS VaOHQFH:¬^PRVLWLYH, NHXWUaO,¬NHJaWLYH` AURXVaO:¬^HLJK,¬NHXWUaO, LRZ` IVTKHLaVWMHaVXUH: ERRO EQYMLFURSKRQH EQYNRLVH: ^HLJK, MHGLXP, LRZ` EPRSWaWH:¬^JR\, AQJHU, SaGQHVV, NHXWUaO` PUREaELOLW\: 0..1 LRFaWLRQ: ^HRPH, WRUN, OWKHU` HXPaQNRLVH: ^HLJK, MHGLXP, LRZ` NHXURSK\VLRORJLFaO LVA SWUHVVLHYHO:¬^HLJK, MHGLXP, LRZ` EPRSWaWH: ^PRVLWLYH, NHXWUaO, NHJaWLYH` MHQWaOWRUNORaG:¬^HLJK, MHGLXP, LRZ` LLJKWLQJDLVFRPIRUW: ¬^NRQH, E[FHVVLYH_LLJKWLQJ, IQVXIÀFLHQW_LLJKWLQJ` THPSHUaWXUHDLVFRPIRUW: ^NRQH, HLJK_THPS, LRZ_THPS` CO2DLVFRPIRUW: ^NRQH, UQFRPIRUWaEOH_CO2, DaQJHURXV_CO2_LHYHOV`¬ RULA_SLWWLQJRLVN: ^HLJK, MHGLXP, LRZ, NRQH` REBA_SWaQGLQJRLVN: ^HLJK, MHGLXP, LRZ, NRQH` HHaUWRaWH: QRQNHJaWLYHIQWHJHU SWHSV: QRQNHJaWLYHIQWHJHU CaORULHCRQVXPSWLRQ: QRQNHJaWLYHIQWHJHU MRYLQJDLVWaQFH: QRQNHJaWLYHIQWHJHU DHHSSOHHSDXUaWLRQ: QRQNHJaWLYHIQWHJHU LLJKWSOHHSDXUaWLRQ: QRQNHJaWLYHIQWHJHU BRG\SFaOH LVA BMI:¬QRQNHJaWLYHIQWHJHU WHLJKW:¬QRQNHJaWLYHRHaO HXPLGLW\DLVFRPIRUW: ¬^NRQH, HLJK_HXPLGLW\, LRZ_HXPLGLW\` QXHVWLRQQaLUHV iVA iVA WRUNAELOLW\MHQWaODHPaQGV: ^VHU\_JRRG, RaWKHU_JRRG, MRGHUaWH, RaWKHU_SRRU, VHU\_SRRU` WRUNAELOLW\PK\VLFaODHPaQGV: ^VHU\_JRRG, GHQHUaOHHaOWK: ^E[FHOOHQW, VHU\_JRRG, GRRG, FaLU, PRRU` MSD_QHFN: ERRO MSD_VKRXOGHU: ERRO HHaUWRaWH: QRQNHJaWLYHIQWHJHU Fig. 3. Information derived from several body and environmental sensors (Questionnaires class omitted). QXHVWLRQQaLUHV i A iVA WRUNAELOLW\MHQWaODHPaQGV: ^VHU\_JRRG, RaWKHU_JRRG, MRGHUaWH, RaWKHU_SRRU, VHU\_SRRU` WRUNAELOLW\PK\VLFaODHPaQGV: ^VHU\ JRRG GHQHUaOHHaOWK: ^E[FHOOHQW, VHU\_JRRG, GRRG, FaLU, PRRU` MSD_QHFN: ERRO MSD VKRXOGHU: ERRO Fig. 3. Information derived from several body and environmental sensors (Questionnaires class omitted). 5 The Ontology Organisational isA Social isA isA isA Interactions isA isA WorkingTime Typology: {Fabric_related, Abstract, Human_related} Typology: {Regular_daily, Shift_system, Flexible, Irregular} Typology: {Human-Human, Human-Machine} Typology: {None, Low, Medium, High} Timestamp: YYYYMMDDHHMMSS SocialEnvironment IsTheLastMeasure: bool TypeOfWork Conﬂicts Fig. 5. Information about social characteristics of the working environment. Task Ofﬁce isA Assembly isA Teleworking isA FreeTime isA Typology: {rough, medium-ﬁne, precise, very_delicate} Structure: {No_structure, Little_structure, Medium_structure, High_structure, Very_high_structure} Repetitiveness: {Very_low, Low, Moderate, High, Very_high} Timestamp: YYYYMMDDHHMMSS IsTheLastMeasure: bool Activity isA Typology: {manager, prevention_engineer, informatics, clerk} Typology: {teleoperator, informatics, clerk} Fig. 4. Information about activities of the worker. Task Ofﬁce isA Assembly isA Teleworking isA FreeTime isA Typology: {rough, medium-ﬁne, precise, very_delicate} Structure: {No_structure, Little_structure, Medium_structure, High_structure, Very_high_structure} Repetitiveness: {Very_low, Low, Moderate, High, Very_high} Timestamp: YYYYMMDDHHMMSS IsTheLastMeasure: bool Activity isA Typology: {manager, prevention_engineer, informatics, clerk} Typology: {teleoperator, informatics, clerk} Fig. 4. Information about activities of the worker. Activity isA Teleworking Fig. 4. Information about activities of the worker. Fig. 4. Information about activities of the worker. Fig. 4. Information about activities of the worker. Organisational isA Social isA isA isA Interactions isA isA WorkingTime Typology: {Fabric_related, Abstract, Human_related} Typology: {Regular_daily, Shift_system, Flexible, Irregular} Typology: {Human-Human, Human-Machine} Typology: {None, Low, Medium, High} Timestamp: YYYYMMDDHHMMSS SocialEnvironment IsTheLastMeasure: bool TypeOfWork Conﬂicts Fig. 5. Information about social characteristics of the working environment. Organisational isA Social isA isA isA Interactions isA isA WorkingTime Fabric_related, uman_related} Typology: {Regular_daily, Shift_system, Flexible, Irregular} Typology: {Human-Human, Human-Machine} Typology: {None, Low, Medium, High} Timestamp: YYYYMMDDHHMMSS SocialEnvironment IsTheLastMeasure: bool peOfWork Conﬂicts Timestamp: YYYYMMDDHHMMSS IsTheLastMeasure: bool isA Typology: {Human-Human, Human-Machine} Typology: {None, Low, Medium, High} Fig. 5. Information about social characteristics of the working environment. about social characteristics of the working environme ig. 5. Information about social characteristics of the ers improve their working and living conditions. The Class GoalState shows the worker’s progress towards the set SMART goal, which is updated weekly. These goals are mostly related to nutrition, sleep, and social relations, and aim at challenging the workers in terms of behavioural change, and of motivating them to adopt a healthier lifestyle. 5 The Ontology DaLO\QXHVWLRQQaLUHBHIRUHWRUN MRQWKO\QXHVWLRQQaLUH WRUNAELOLW\: ^E[FHOOHQW, VHU\_JRRG, GRRG, FaLU, PRRU` FRRGHaELWV: ^VHU\_JRRG, GRRG, FaLUO\_EaG, BaG` MRRG: ^VHU\_JRRG, GRRG, NHXWUaO, BaG, VHU\_EaG` SOHHSQXaOLW\: ^VHU\_JRRG, GRRG, FaLUO\_EaG, BaG` \ ^ , _ , _ J , J ` THPSRUaODHPaQGV: ^LRZ, MHGLXP_ORZ, MHGLXP_KLJK, HLJK` FUXVWUaWLRQ: ^LRZ, MHGLXP_ORZ, MHGLXP_KLJK, HLJK` EIIRUW: ^LRZ, MHGLXP_ORZ, MHGLXP_KLJK, HLJK` PHUIRUPaQFH: ^GRRG, MHGLXP_JRRG, MHGLXP_SRRU, PRRU` MRRG: ^VHU\_JRRG, GRRG, NHXWUaO, BaG, VHU\_EaG` WaWHUIQWaNHDa\: ^PRRU, FaLU` SS MSD_ORZHU_EaFN: ERRO MSD_KLSV_WKLJKV: ERRO MSD_NQHHV: ERRO RFLaOFUHTXHQF\:¬QRQNHJaWLYHIQWHJHU MSD_aQNOHV_IHHW: ERRO SRFLaORHOaWLRQV: QRQNHJaWLYHIQWHJHU worker. The taxonomy in Fig. 3 shows the sensor typologies that the WAOW tool leverages for collecting information about the worker and her/his environment. Each measurement is provided with a timestamp and a probability value; over the time, we collect an ordered sequence of High-level information samples, of various typologies, associated with various reliability values (adopting the same interpretation described for advices). The DSS will reason on this sequence. Sensors are subdivided into environmental sensors, body sensors, and other ty- pologies (which include questionnaires and user location). Questionnaires, which will be administered to the worker by means of the WA App, are considered as a special case of sensors, and provide daily, weekly or monthly “measurements”. Activity and social environment. The class named “Activity” describes the relevant characteristics of the activity associated with a worker. Worker’s activities are split into two main parts, as Fig. 4 shows: Task and FreeTime. In fact, the WAOW tool will measure workers at the working place, and during their activities at home or in other places. By “social environment” we mean the relevant social characteristics of the working environment as captured by classes in Fig. 5. SMART Goals. This part describes the so-called SMART goals about healthy habits that the worker is encouraged to adopt (see Fig. 6). According to the worker’s context, a SmartGoal will be set for each individual worker. The Ontol- ogy stores the current state of the individual worker as well as the goals that are identiﬁed to be met by the worker. These personalized goals can help the work- 8 R. M. Resende de Almeida et al. Task Ofﬁce isA Assembly isA Teleworking isA FreeTime isA Typology: {rough, medium-ﬁne, precise, very_delicate} Structure: {No_structure, Little_structure, Medium_structure, High_structure, Very_high_structure} Repetitiveness: {Very_low, Low, Moderate, High, Very_high} Timestamp: YYYYMMDDHHMMSS IsTheLastMeasure: bool Activity isA Typology: {manager, prevention_engineer, informatics, clerk} Typology: {teleoperator, informatics, clerk} Fig. 4. Information about activities of the worker. 6 The Decision Support System One of the central components of the WAOW tool is the DSS; such a system supports the worker while dealing with working activities. The DSS we developed is based on a model-driven approach – in particular, it’s a rule-based engine – but it is able to work with the results of the Edge Cloud algorithms, which generate a probabilistic output. Thus, the whole “decision-making pipeline”, as Fig. 1 shows, is actually based on a hybrid data-driven/model-driven approach. We opted for this design as for the models working on Raw data we found feasible corpora that permitted to leverage data-driven models. For generating advices, however, no feasible corpora existed, and creating our own dataset was not an WorkingAge EU Project 9 PhysActGoal isA NutriHabGoal isA Timestamp: YYYYMMDDHHMMSS SmartGoal RecommStepDay: nonNegativeInteger BMImin: nonNegativeInteger BMImax: nonNegativeInteger Timestamp: YYYYMMDDHHMMSS GoalState Reached: bool PhysActState NutriHabState isA isA ActualStep: nonNegativeInteger ActualBMI: nonNegativeInteger hasPhysState hasNutriState SleepTimeGoal isA SocialRelationGoal isA SleepTimeState isA MobileBeforeSleepState isA MinDayNumToMeet: nonNegativeInteger MobileBeforeSleepGoal isA RecommSleepTime: nonNegativeInteger ActualSleepTime: nonNegativeInteger hasSleepState hasMobileState SocialRelationState isA ActualSocialFrequency: nonNegativeInteger hasSocialState ActualPeriodWithoutPhone: nonNegativeInteger 7 7 7 7 0..N 0..N 0..N 0..N 0..N RecommPeriodWithoutPhone: nonNegativeInteger RecommSocialFrequency: nonNegativeInteger RecommWaterIntakeDay: nonNegativeInteger ActualWaterIntakeDay: nonNegativeInteger 7 Fig. 6. Information about the SMART goals. Fig. 6. Information about the SMART goals. option. Thus, we decided for a model-based approach. Model-driven approaches refer to methods that are able to reason over existing and infer new knowledge based on a given model of the domain at hand. These solutions are often referred to as expert systems. A typical model is often expressed as a set of rules, which describe how to derive new facts, as it happens in well-known declarative pro- gramming languages like Prolog. Declarative model-driven approaches are also used for case-based reasoning, to provide insights based on previously occurred instances. Given a model, the computation is performed by the reasoning engine, that applies the model to user-provided data to infer new knowledge regarding the domain. Research in model-based reasoning focus in various areas, including the core engine optimizations, debugging, expressivity of rules and knowledge representa- tion in general. Recent research included uncertainty in their proposed models, leading to engines able to handle probabilistic rules [7]. In the design of the WAOW tool, we determined that uncertainty was needed in order to represent aspects of the model which cannot be predicted in a deterministic way. 6 The Decision Support System For in- stance the conﬁdence that an expert may give to a speciﬁc suggestion, or the measure of a sensor, may not be always equal to 100%. This is why we chose ProbLog as the inference engine for the DSS. ProbLog. ProbLog [7] is a probabilistic logic programming proposed by the DLAI group at the KU Leuven University (Belgium), heavily inspired by Prolog. It reuses the Prolog syntax but allows probabilistic labeling of rules and facts, allowing not only inference of certainly-true facts (according to the usual declar- ative models), but also derive the level of conﬁdence for that speciﬁc inferred fact. An interesting property of ProbLog is that it does not require to specify an exact Bayesian Network, relying instead on a sort of best-eﬀort approximation whenever some nodes and edges of the full causal graph are missing. It can also R. M. Resende de Almeida et al. 10 compute marginal probabilities of any number of ground atoms in the presence of evidence, therefore recomputing probabilities whenever new data is provided. This is important since in many real applications, including those in the context of the WA project, many aspects of the causal dependencies cannot be modeled exactly, and so a level of approximation must be handled by the tool. As the original Prolog, ProbLog uses procedural interpretation of Horn clauses, in this case labeled with probabilities. As we show next, we use these probabilistic rules to model scenarios in the context of working environments. The DSS. As Fig. 1 shows, the DSS is fed by High-Level information generated by the data-driven models running into the Edge Cloud. The Ontology acts as a data model with the aim of describing all the possible worker’s physical, cognitive and emotional states. The DSS leverages the OWL deﬁnition of the Ontology to check for correct- ness and consistency the coming High-level information. Individuals, however, are not described in terms of OWL syntax elements: for best eﬃciency, they are directly created as probabilities facts, following the ProbLog syntax. The outputs of the DSS will be highly supported by the Interventions platform, described in Section 4. The suggestions and interventions implemented in the WAOW tool are provided by experts. Such suggestions and interventions proposed to the user will be implemented as probabilistic rules by the DSS: the selection is carried out taking into account the probability associated to each worker’s state. 6 The Decision Support System Thus, the associated probability helps in ranking them and select the best one. Moreover, if the selected suggestion is associated with a probability lower than a given threshold, the DSS could simply give up, avoiding to ﬂood the worker with useless advices. Finally, the generated suggestion –for example, take pause– is translated into human language: “The working area is too noisy and you look tired, it’s time to take a pause.” generates the take pause suggestion whenever a worker in an excessively loud environment is stressed despite using safety ear muﬀs, giving to the rule a weight of 0.3. The ﬁnal probability associated to the suggestion takes into account the probability associated to each proposition in the rule body. More than one suggestion could be generated by the DSS. Thus, the associated probability helps in ranking them and select the best one. Moreover, if the selected suggestion is associated with a probability lower than a given threshold, the DSS could simply give up, avoiding to ﬂood the worker with useless advices. Finally, the generated suggestion –for example, take pause– is translated into human language: “The working area is too noisy and you look tired, it’s time to take a pause.” 7 Conclusion The paper has described the WA project and tools, aimed at providing smart assistance to elder workers. The paper has focused on the Ontology and the DSS portions of WA, which will generate recommendations about risk avoidance and well-being at work. Experiments about DSS are being conducted via in- company use cases, testing the proposed solutions by collecting sensor data and by improving the learning and smart capabilities of the DSS. 6 The Decision Support System Such a probability is modulated by the High-Level information provided by each sensor integrated into the WAOW tool and by the feedback provided by each worker toward the interaction with the DSS. The rules leveraged by the DSS will consider diﬀerent factors that may af- fect the well-being of the worker, such as the environment (e.g., loudly, incorrect lighting), wrong behaviour at work (e.g., use of wrong tools or used in a wrong way, posture), worker’s habits and state (including out-of-work elements such as food, sleep). The DSS will interact with diﬀerent entities integrated in the WAOW tool. In particular, the DSS will receive in input the descriptions of the worker’s physical and mental states, with the support of the Ontology, and will produce recommendations based on the worker’s current state and on her/his history (both stored by the DSS as probabilistic facts). Notice that the worker can provide feedback to the DSS which, in turn, can update the rules to cus- tomise them to the worker’s preferences. For instance, if a given suggestion in the past (e.g. the use of earmuﬀs) did not produce an expected beneﬁt (e.g. “decrease in user stress recorded through sensors”), the DSS may then explore other suggestions with the given user (e.g., “take a pause of X minutes”). Action Rules: a simple example for user suggestions. Let’s assert, in Prolog, that Jack is a worker and every worker is also a person: worker(jack). person(X) :- worker(X). The left hand side of the rule is called head, the right hand side is the body. Up- percase X is a variable. Whenever the body holds, the head holds too. With this 11 WorkingAge EU Project simple representation, the DSS can automatically infer that Jack is a person, that is, person(jack) holds. In our experiments, we modelled many possible aspects of diﬀerent working environments, from worker proﬁle –for example, ed- ucation or activity typology– to time constraints, and even complex behavioural knowledge; as an example, the following rule: 0.3::suggestion(take_pause,W) :- worker(W), isShown(stressed,W), env(env_discrete_volume_serious_hazard), isUsing(W,use_noise_reduction_safety_ear_muffs). generates the take pause suggestion whenever a worker in an excessively loud environment is stressed despite using safety ear muﬀs, giving to the rule a weight of 0.3. The ﬁnal probability associated to the suggestion takes into account the probability associated to each proposition in the rule body. More than one suggestion could be generated by the DSS. References 1. Birkel, H.S., et al.: Development of a risk framework for industry 4.0 in the context of sustainability for established manufacturers. Sustainability 11(2), 384 (2019) 1. Birkel, H.S., et al.: Development of a risk framework for industry 4.0 in the context of sustainability for established manufacturers. Sustainability 11(2), 384 (2019) 2. Calzavara, M., Battini, D., Bogataj, D., Sgarbossa, F., Zennaro, I.: Ageing work- force management in manufacturing systems: state of the art and future research agenda. International Journal of Production Research 58(3), 729–747 (2020) 2. Calzavara, M., Battini, D., Bogataj, D., Sgarbossa, F., Zennaro, I.: Ageing work- force management in manufacturing systems: state of the art and future research agenda. International Journal of Production Research 58(3), 729–747 (2020) 3. Chen, H., Hou, C., Zhang, L., Li, S.: Comparative study on the strands of research on the governance model of international occupational safety and health issues. Safety science 122 (2020) 3. Chen, H., Hou, C., Zhang, L., Li, S.: Comparative study on the strands of research on the governance model of international occupational safety and health issues. Safety science 122 (2020) 4. Cheng, Z., Nielsen, I., Cutler, H.: Perceived job quality, work-life interference and intention to stay. International Journal of Manpower (2019) 4. Cheng, Z., Nielsen, I., Cutler, H.: Perceived job quality, work-life interference and intention to stay. International Journal of Manpower (2019) 5. Cresswell, K., et al.: Investigating the use of data-driven artiﬁcial intelligence in computerised decision support systems for health and social care: A systematic review. Health Informatics Journal (2020) R. M. Resende de Almeida et al. 12 6. Darabont, D.C., Bejinariu, C., Baciu, C., Bernevig-Sava, M.A.: Modern approaches in integrated management systems of quality, environmental and occupational health and safety. Calitatea 20(S1), 105 (2019) 7. Fierens, D., den Broeck, G.V., Renkens, J., Shterionov, D., B. Gutmann, I.T., Janssens, G., Raedt, L.D.: Inference and learning in probabilistic logic programs using weighted boolean formulas. Theor. Pract. Log. Prog. 15(3), 358–401 (2015) 8. Fitzpatrick, M.D., Moore, T.J.: The mortality eﬀects of retirement: Evidence from social security eligibility at age 62. J. Public Econ. 157, 121–137 (2018) 9. Fugini, M., Barenghi, A., Comai, S., Pelosi, G., Tedesco, R., van Gasteren, M., Catalina, C., Arribas Leal, E., Losada Dur´an, R., Resende de Almeida, R., Mertens, A., Rick, V.: Workingage: Providing occupational safety through pervasive sensing and data driven behavior modeling. References In: Proceedings of the 30th European Safety and Reliability Conference and the 15th Probabilistic Safety Assessment and Man- agement Conference. Research Publishing (2020) 10. Golightley, M., Goemans, R.: Social work and mental health. SAGE Pub. (2020) 11. Grant, A., Stober, D.: Evidence Based Coaching Handbook: Putting Best Practices to Work for Your Clients. Wiley (2006) 12. Grosch, J.W., Scholl, J.C.: Designing age-friendly workplaces: An occupational health perspective. In: Current and Emerging Trends in Aging and Work, pp. 429–451. Springer (2020) 13. Kalteh, H.O., Mortazavi, S.B., Mohammadi, E., Salesi, M.: The relationship be- tween safety culture and safety climate and safety performance: a systematic re- view. International journal of occupational safety and ergonomics pp. 1–11 (2019) 14. Lodh, N., Sil, J., Bhattacharya, I.: Graph based clinical decision support system using ontological framework. In: International Conference on Computational Intel- ligence, Communications, and Business Analytics. pp. 137–152. Springer (2017) 15. Mabkhot, M.M., Amri, S.K., Darmoul, S., Al-Samhan, A.M., Elkosantini, S.: An ontology-based multi-criteria decision support system to reconﬁgure manufacturing systems. IISE Transactions 52(1), 18–42 (2020) 16. Moreira, M., Rodrigues, J., Korotaev, V., Al-Muhtadi, J., Kumar, N.: A compre- hensive review on smart decision support systems for health care. IEEE Systems Journal 13(3), 3536–3545 (2019) 17. Pease, S.G., et al.: An interoperable semantic service toolset with domain ontol- ogy for automated decision support in the end-of-life domain. Future Generation Computer Systems 112, 848–858 (2020) p y , ( ) 18. Petnga, L., Austin, M.: An ontological framework for knowledge modeling and decision support in cyber-physical systems. Adv. Eng. Inf. 30(1), 77–94 (2016) 19. Roughton, J., Crutchﬁeld, N., Waite, M.: Safety culture: An innovative leadership approach. Butterworth-Heinemann (2019) ( ) 20. Schlick, C., Bruder, R., Luczak, H.: Arbeitswissenschaft. Springer-Verlag (2018) 21. Vujica Herzog, N., Harih, G.: Decision support system for designing and assign- ing ergonomic workplaces to workers with disabilities. Ergonomics 63(2), 225–236 (2020)
https://openalex.org/W2809660117	https://europepmc.org/articles/pmc6161011?pdf=render	English	null	Chemical Vapor Deposition of Photocatalyst Nanoparticles on PVDF Membranes for Advanced Oxidation Processes	Membranes	2,018	cc-by	10,569	Received: 21 May 2018; Accepted: 19 June 2018; Published: 21 June 2018 Abstract: The chemical binding of photocatalytic materials, such as TiO2 and ZnO nanoparticles, onto porous polymer membranes requires a series of chemical reactions and long puriﬁcation processes, which often result in small amounts of trapped nanoparticles with reduced photocatalytic activity. In this work, a chemical vapor deposition technique was investigated in order to allow the nucleation and growth of ZnO and TiO2 nanoparticles onto polyvinylidene diﬂuoride (PVDF) porous membranes for application in advanced oxidation processes. The thickness of obtained surface coatings by sputtered nanoparticles was found to depend on process conditions. The photocatalytic efﬁciency of sputtered membranes was tested against both a model drug and a model organic pollutant in a small continuous ﬂow reactor. Keywords: chemical vapor deposition; polyvinylidene diﬂuoride; titanium dioxide; zinc oxide; photocatalysis membranes membranes membranes membranes Article Membranes 2018, 8, 35; doi:10.3390/membranes8030035 Chemical Vapor Deposition of Photocatalyst Nanoparticles on PVDF Membranes for Advanced Oxidation Processes Giovanni De Filpo 1, Elvira Pantuso 2, Katia Armentano 2, Patrizia Formoso 2 ID , Gianluca Di Proﬁo 3, Teresa Poerio 3 ID , Enrica Fontananova 3, Carmen Meringolo 3, Alexander I. Mashin 4 ID and Fiore P. Nicoletta 1,* ID 1 Department of Chemistry and Chemical Technologies, University of Calabria, Via P. Bucci Cubo 15/C, 87036 Rende (CS), Italy; giovanni.deﬁlpo@unical.it 1 Department of Chemistry and Chemical Technologies, University of Calabria, Via P. Bucci Cubo 15/C, 87036 Rende (CS), Italy; giovanni.deﬁlpo@unical.it y g p 2 Department of Pharmacy, Health and Nutritional Sciences, University of Calabria, Via P. Bucci Ediﬁcio Polifunzionale, 87036 Rende (CS), Italy; elvirapnt.ep@gmail.com (E.P.); katiarmentano@hotmail.it (K.A.); patrizia.formoso@unical.it (P.F.) 3 National Research Council of Italy (CNR)—Institute on Membrane Technology (ITM), Via P. Bucci Cub 17/C, 87036 Rende (CS), Italy; g.diproﬁo@itm.cnr.it (G.D.P.); t.poerio@itm.cnr.it (T.P.); 3 National Research Council of Italy (CNR)—Institute on Membrane Technology (ITM), Via P. Bucci Cu 17/C, 87036 Rende (CS), Italy; g.diproﬁo@itm.cnr.it (G.D.P.); t.poerio@itm.cnr.it (T.P.); e.fontananova@itm.cnr.it (E.F.); c.meringolo@itm.cnr.it (C.M.) 17/C, 87036 Rende (CS), Italy; g.diproﬁo@itm.cnr.it (G.D.P.); t.poerio@itm.cnr.it (T.P.); e.fontananova@itm.cnr.it (E.F.); c.meringolo@itm.cnr.it (C.M.) 4 g 4 Applied Physics & Microelectronics, Lobachevsky St 603950 Nizhni Novgorod, Russia; mashin@unn.ru 603950 Nizhni Novgorod, Russia; mashin@unn.ru * Correspondence: ﬁore.nicoletta@unical.it; Tel.: +39-0984-493194 1. Introduction Thin ﬁlms are material layers that have thicknesses varying from tens of nanometers to a few micrometers [1]. They are generally obtained by deposition processes on the surface of given substrates. The thin ﬁlm growth is generally a two-step process where an initial random nucleation step is followed by an ordered growth. Nucleation and growth—and consequently the ﬁlm structure—depend on the substrate chemistry (surface composition and structure), the method used [2], and deposition conditions [3]. Thin ﬁlm deposition methods are classiﬁed in solid, liquid, and gas phase deposition methods according to the physical state of the deposited material. A further classiﬁcation of gas deposition methods distinguishes chemical vapor deposition (CVD) and physical vapor deposition (PVD) processes. Both methods involve the deposition of atoms or molecules carried in their vapor phase Membranes 2018, 8, 35; doi:10.3390/membranes8030035 www.mdpi.com/journal/membranes 2 of 15 y with [4]. Membranes 2018, 8, 35 onto the substrate the substrate, whil onto the substrate surface. In a CVD process, the deposited (target) material reacts chemically with the substrate, while in PVD processes, the deposited molecules and substrate are still distinct [4]. industries. Efficient separation and purification are also important for food and pharmaceutical plants in order to guarantee high‐quality water after removal of toxic components from industrial Ph i l i d b id d h d b b Separation, concentration, and puriﬁcation processes present some challenges for chemical industries. Efﬁcient separation and puriﬁcation are also important for food and pharmaceutical plants in order to guarantee high-quality water after removal of toxic components from industrial wastewater. Pharmaceutical active compounds can be considered as hazardous substances because of their potential threat to health and the environment. They are also considered an emerging pollutant due to the failure of classical treatments (such as ﬁltration, adsorption, bio-oxidation, sedimentation, coagulation, chlorination, and UV-irradiation) to effectively remove them [5–7]. In addition to common chemical pollutants, over 80 pharmaceutical active compounds have been detected in wastewater efﬂuents and surface water across the world [8], with concentration ranging from few ng L−1 to several g L−1. The most important sources of pharmaceutical active compounds are incorrect disposal of unused drugs and efﬂuents of wastewater treatment plants (including pharmaceutical industries, hospital wastewater, aqua-farming, and cattle-breeding) [9–11]. 1. Introduction In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron-hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. controls the exchange of matter and energy between the regions” [14]. Today, membranes are efficiently used for water desalinization, wastewater purification, recovery of valuable constituents from production waste, gas separation in petrochemical processes, concentration and purification in food and drug applications, artificial organs and therapeutic systems, energy conversion, and storage systems [15–17]. In addition to their technical simplicity and energy efficiency, membrane processes can be easily upscaled from batchwise treatment of small quantities to large‐scale continuous operations. An advanced oxidation process (AOP) is a simple technique that allows an efficient degradation of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron‐hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. More recently, polymer membranes have been used as innovative separation materials [13]. A membrane can be deﬁned as “a barrier that separates and/or contacts two different regions and controls the exchange of matter and energy between the regions” [14]. Today, membranes are efﬁciently used for water desalinization, wastewater puriﬁcation, recovery of valuable constituents from production waste, gas separation in petrochemical processes, concentration and puriﬁcation in food and drug applications, artiﬁcial organs and therapeutic systems, energy conversion, and storage systems [15–17]. In addition to their technical simplicity and energy efﬁciency, membrane processes can be easily upscaled from batchwise treatment of small quantities to large-scale continuous operations. controls the exchange of matter and energy between the regions” [14]. Today, membranes are efficiently used for water desalinization, wastewater purification, recovery of valuable constituents from production waste, gas separation in petrochemical processes, concentration and purification in food and drug applications, artificial organs and therapeutic systems, energy conversion, and storage systems [15–17]. 1. Introduction In addition to their technical simplicity and energy efficiency, membrane processes can be easily upscaled from batchwise treatment of small quantities to large‐scale continuous operations. An advanced oxidation process (AOP) is a simple technique that allows an efficient degradation f i ll t t ll f d i t t I AOP i ll t t i li d y p q g p An advanced oxidation process (AOP) is a simple technique that allows an efﬁcient degradation of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron-hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron‐hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. An advanced oxidation process (AOP) is a simple technique that allows an efﬁcient degradation of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. 1. Introduction While the measured concentrations can result in water that is low or below drinking water guidelines and health criteria [12], their continuous accumulation in aquatic environment can also represent a real hazard. wastewater. Pharmaceutical active compounds can be considered as hazardous substances because of their potential threat to health and the environment. They are also considered an emerging pollutant due to the failure of classical treatments (such as filtration, adsorption, bio‐oxidation, sedimentation, coagulation, chlorination, and UV‐irradiation) to effectively remove them [5–7]. In addition to common chemical pollutants, over 80 pharmaceutical active compounds have been detected in wastewater effluents and surface water across the world [8], with concentration ranging from few ng L−1 to several g L−1. The most important sources of pharmaceutical active compounds are incorrect disposal of unused drugs and effluents of wastewater treatment plants (including pharmaceutical industries, hospital wastewater, aqua‐farming, and cattle‐breeding) [9–11]. While the measured concentrations can result in water that is low or below drinking water guidelines and health criteria [12], their continuous accumulation in aquatic environment can also represent a real hazard. More recently, polymer membranes have been used as innovative separation materials [13]. A membrane can be defined as “a barrier that separates and/or contacts two different regions and t l th h f tt d b t th i ” [14] T d b More recently, polymer membranes have been used as innovative separation materials [13]. A membrane can be deﬁned as “a barrier that separates and/or contacts two different regions and controls the exchange of matter and energy between the regions” [14]. Today, membranes are efﬁciently used for water desalinization, wastewater puriﬁcation, recovery of valuable constituents from production waste, gas separation in petrochemical processes, concentration and puriﬁcation in food and drug applications, artiﬁcial organs and therapeutic systems, energy conversion, and storage systems [15–17]. In addition to their technical simplicity and energy efﬁciency, membrane processes can be easily upscaled from batchwise treatment of small quantities to large-scale continuous operations. An advanced oxidation process (AOP) is a simple technique that allows an efﬁcient degradation of organic pollutants generally found in wastewater. In an AOP, organic pollutants are mineralized by the generation of highly reactive hydroxyl radicals [18]. Among the advanced treatment technologies, UV photocatalysis by nanoparticles (e.g., ZnO and TiO2), has attracted great interest in recent years [19]. 1. Introduction I h t id ti (Fi 1) l t t d f l b d t d ti In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron-hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. In a photooxidation process (Figure 1), electrons are promoted from valence band to conduction band—resulting in the formation of electron‐hole pairs—when the catalyst nanoparticle (CNp) is irradiated by UV light with energy intensity larger than the characteristic band gap (3.37 eV for ZnO and 3.2 eV for TiO2, respectively) [20,21]. Figure 1. Schematization of the photoactivity of a catalyst nanoparticle (CNp). Figure 1. Schematization of the photoactivity of a catalyst nanoparticle (CNp). Figure 1. Schematization of the photoactivity of a catalyst nanoparticle (CNp). Figure 1. Schematization of the photoactivity of a catalyst nanoparticle (CNp). Both electrons and holes can move to the semiconductor surface and produce radicals, which can oxidize organic compounds (OC), whereas electrons can reduce them, according to the reactions Both electrons and holes can move to the semiconductor surface and produce radicals, which can oxidize organic compounds (OC), whereas electrons can reduce them, according to the reactions 3 of 15 Membranes 2018, 8, 35 reported [19,22]. The degradation of organic compounds (OC) by photocatalyst nanoparticles (CNp) is shown below. h h+ CNp + hν →e− cb + h+ vb O2 + e− cb →O−• 2 H2O + h+ vb →OH• + H+ O−• 2 + H2O →H2O2 →2OH• OH• + OC →OCox OC + e− cb →OCred ZnO and TiO2 are the most commonly used photocatalysts due to their redox ability, chemical stability, reduced toxicity towards the environment and health, and low cost [23]. In addition to the mineralization of organic compounds, the reactive redox species—such as hydroxyl radicals (OH•), superoxide anions (O−• 2 ), and hydrogen peroxide molecules (H2O2) generated by UV irradiation—can damage the cell membrane of microorganisms [24] and kill bacteria, viruses, fungi, and algae [25], thus conferring long-term antibacterial and antifungal properties [26–28] to photocatalysts. More recently, submerged membranes have been integrated by semiconductor photocatalysts in order to photomineralize membrane fouling [29,30], thereby reducing cleaning and maintenance costs. In particular, Ho et al. 2 Materials and Methods 2. Materials and Methods 2. Materials and Methods The substrates used were PVDF membrane disks with a diameter of 47 mm, a porosity of 70%, and a mean pore size of 0.10 μm (Durapore ©, Merck KGaA, Darmstadt, Germany). Prior to use, membranes were washed in methanol (Sigma‐Aldrich, Milan, Italy) by an ultrasonic bath (model M1800H‐E, Bransonic, Danbury, CT, USA). The deposition of nanostructured photocatalysts on PVDF membranes was obtained by sputtering of appropriate targets by process inert gas ions (argon) The substrates used were PVDF membrane disks with a diameter of 47 mm, a porosity of 70%, and a mean pore size of 0.10 µm (Durapore ©, Merck KGaA, Darmstadt, Germany). Prior to use, membranes were washed in methanol (Sigma-Aldrich, Milan, Italy) by an ultrasonic bath (model M1800H-E, Bransonic, Danbury, CT, USA). The deposition of nanostructured photocatalysts on PVDF membranes was obtained by sputtering of appropriate targets by process inert gas ions (argon) in a Edwards AUTO-306 sputtering system (Edwards, Burgess Hill, UK). in a Edwards AUTO‐306 sputtering system (Edwards, Burgess Hill, UK). ZnO was deposited on PVDF membranes by Argon (purity 99.999%) sputtering of a ZnO target (purity 99.99%, Goodfellow Cambridge Ltd., Huntingdon, England). The deposition of nanostructured TiO2 was obtained by reactive sputtering using a Ti target (purity 99.999%, Goodfellow Cambridge Ltd., Huntingdon, England) in presence of a gaseous mixture of argon and oxygen (purity 99.999%, pressure of gas mixture: p(Ar) = 2.8 × 10−3 mbar and p(O2) = 1.2 × 10−3 mbar, p(Ar)/p(O2) = 2.3). The reactive gas mixture reacts with the substrate and sputtered atoms, forming a thin film of desired compound onto the substrate. The particular pressure ratio between Ar and O2 h d f h l h f O h h h h ZnO was deposited on PVDF membranes by Argon (purity 99.999%) sputtering of a ZnO target (purity 99.99%, Goodfellow Cambridge Ltd., Huntingdon, England). The deposition of nanostructured TiO2 was obtained by reactive sputtering using a Ti target (purity 99.999%, Goodfellow Cambridge Ltd., Huntingdon, England) in presence of a gaseous mixture of argon and oxygen (purity 99.999%, pressure of gas mixture: p(Ar) = 2.8 × 10−3 mbar and p(O2) = 1.2 × 10−3 mbar, p(Ar)/p(O2) = 2.3). The reactive gas mixture reacts with the substrate and sputtered atoms, forming a thin ﬁlm of desired compound onto the substrate. 1. Introduction obtained an enhancement in the ﬁltration ﬂux of a submerged membrane reactor by integration of photooxidation process and membrane ﬁltration [31], while Mendez-Arriaga et al. [32] studied the TiO2 photocatalytic degradation of pharmaceutical compounds such as diclofenac, naproxen, and ibuprofen. The combination of membrane separation and advanced oxidation processes is an emerging technology for the complete removal of pollutants because each technique complements the advantages of the other. In particular, the AOP eliminates membrane fouling and allows the remediation of the concentrate while, at the same time, the membrane process ﬁlters the feed and concentrates pollutants to an optimal level for AOP [33–35]. Nevertheless, the chemical binding of photocatalysts onto porous polymer membranes requires a series of chemical reactions and long cleaning processes, which often result in small amounts of trapped nanoparticles with reduced photocatalytic activity. In addition, some polymers, such as polytetraﬂuoroethylene and polyvinylidene ﬂuoride, are hardly functionalizable in order to trap photocatalyst molecules. In this work, a chemical vapor deposition process was investigated in order to allow the nucleation and growth of ZnO and TiO2 photocatalytic nanoparticles onto polyvinylidene diﬂuoride (PVDF) porous membranes for applications in AOP. The purpose of this work was the coupling of ﬁltration properties of polymer membranes with the photocatalytic activity of ZnO and TiO2 nanoparticles nucleated and grown on PVDF porous membranes using the CVD technique. The substrates used were membrane disks in PVDF, which is a thermoplastic material characterized by high strength and nontoxicity and, consequently, widely used in membrane processes and food applications. Moreover, PVDF is characterized by high chemical and UV stability, which renders this material particularly interesting for photocatalytic applications. CVD is a well-known chemical process for the production of high-purity, high-performance solid thin ﬁlms. In a typical CVD process, the substrate is exposed to one or more volatile precursors, which react on the substrate surface to produce the desired layer (Figure 2). 4 of 15 4 of 15 Membranes 2018, 8, 35 Membranes 2018 8 x FO Figure 2. Schematization of the chemical vapor deposition (CVD) technique. The substrate is exposed to one or more volatile precursors, which react on the substrate surface to produce the desired thin film Figure 2. Schematization of the chemical vapor deposition (CVD) technique. The substrate is exposed to one or more volatile precursors, which react on the substrate surface to produce the desired thin ﬁlm. Figure 2. 1. Introduction Schematization of the chemical vapor deposition (CVD) technique. The substrate is exposed to one or more volatile precursors, which react on the substrate surface to produce the desired thin Figure 2. Schematization of the chemical vapor deposition (CVD) technique. The substrate is exposed to one or more volatile precursors, which react on the substrate surface to produce the desired thin ﬁlm. The photocatalytic efficiency of sputtered membranes was tested against a model drug (diclofenac sodium) and a model pollutant (methylene blue) in a small continuous flow reactor The photocatalytic efﬁciency of sputtered membranes was tested against a model drug (diclofenac sodium) and a model pollutant (methylene blue) in a small continuous ﬂow reactor. The photocatalytic efficiency of sputtered membranes was tested against a model drug (diclofenac sodium) and a model pollutant (methylene blue) in a small continuous flow reactor The photocatalytic efﬁciency of sputtered membranes was tested against a model drug (diclofenac sodium) and a model pollutant (methylene blue) in a small continuous ﬂow reactor. 2 Materials and Methods 2. Materials and Methods The particular pressure ratio between Ar and O2 was chosen in order to form the anatase polymorph of TiO2 [36], which is more photoactive than rutile polymorph [37]. was chosen in order to form the anatase polymorph of TiO2 [36], which is more photoactive than rutile polymorph [37]. A microRaman spectrometer (Labram, Horiba Jobin Yvon) equipped with an Olympus microscope and interfaced to a color camera was used to confirm the presence of TiO2 anatase thin l Th R ll d h h 100 bj i i H N l ( i i A microRaman spectrometer (Labram, Horiba Jobin Yvon) equipped with an Olympus microscope and interfaced to a color camera was used to conﬁrm the presence of TiO2 anatase thin layers. The Raman spectra were collected through a 100× objective using a He–Ne laser (emission wavelength at 632.8 nm, power 5 mW). layers. The Raman spectra were collected through a 100× objective using a He–Ne laser (emission wavelength at 632.8 nm, power 5 mW). Membrane and nanoparticle morphology was investigated by scanning electron microscopy (SEM). Observations were performed on membrane top surfaces coated with a thin gold or graphite film by a Leica LEO 420 (Leica Microsystems, Cambridge, England) or a Quanta 200 (FEI/Philips, Eindhoven, Netherlands) scanning electron microscope equipped with a backscatter electron dete to E e y di e i e X ay (EDX) a e e obtai ed ith a Phe o P oX SEM (Phe o Membrane and nanoparticle morphology was investigated by scanning electron microscopy (SEM). Observations were performed on membrane top surfaces coated with a thin gold or graphite ﬁlm by a Leica LEO 420 (Leica Microsystems, Cambridge, England) or a Quanta 200 (FEI/Philips, Eindhoven, Netherlands) scanning electron microscope equipped with a backscatter electron detector. Energy-dispersive X-ray (EDX) maps were obtained with a Phenom ProX SEM (Phenom-World, Eindhoven, The Netherlands). Transmission electron microscope (TEM) images were collected with a 5 of 15 Membranes 2018, 8, 35 JEM 1400 Plus transmission electron microscope operating at 100 kV (Jeol, Akishima, Tokyo, Japan). The shape and size of nanoparticles was obtained by software analysis of TEM pictures. The number of measured nanoparticles—taken from different pictures of the same sample—was at least 100, and their size was evaluated with an image software (Motic Images Plus 2.0, MoticEurope S.L.U., Barcelona, Spain). 2 Materials and Methods 2. Materials and Methods Static contact angles to pure water were measured with a CAM 200 contact angle meter (KSV Instruments LTD, Helsinki, Finland) at 25 ◦C. A drop (2 µL) of water was put onto the sample surface by a microsyringe, and measurements were carried out by setting the tangents on both visible edges of the droplet on ﬁve different positions of each sample and calculating the average value of the measurements. The permeation tests were carried out with distilled water using a tangential ﬂow ﬁltration cell having an active area of 14.5 cm2. The feed solution (at 25 ± 1 ◦C) was pumped parallel to the membrane surface by a gear pump at the transmembrane pressure of 0.4 bar. The feed ﬂow rate was 250 mL min−1. Permeate samples were collected every 5 min in order to determine the transmembrane ﬂux, J, deﬁned as: V J = Vp A ∆t (1) J = Vp A ∆t (1) where Vp is the permeate volume passed through the membrane in the ﬁxed time interval ∆t and A is the effective membrane area. The photoactivity of ZnO and TiO2 layers was tested in a small continuous plant where either a diclofenac sodium (9.3 × 10−5 M, Sigma Aldrich, Milan, Italy) or methylene blue (1.3 × 10−5 M, Sigma Aldrich, Milan, Italy) water solution was recirculated by a peristaltic system through a round cell, which was equipped with a quartz window to allow UV irradiation and divided into two compartments by membrane [38]. The sputtered side of membranes was exposed to the UV light from a medium-high pressure mercury vapor lamp (ZS lamp, Helios Italquartz, Italy) with an irradiance of 6 W m−2. At the cell exit, the solution passed through a quartz ﬂow cuvette placed inside a spectrophotometer able to read the absorbance value at either 275 nm or 665 nm, which are the wavelength of maximum absorption of diclofenac sodium and methylene blue, respectively. The absorbance was measured at regular intervals of 5 or 15 min. 3. Results and Discussion The quality of the obtained thin ﬁlm is strongly dependent on the process parameters. In particular, the sputtering yield (Y) is deﬁned as the number of sputtered atoms per impinging ion. Consequently, a higher yield results in a higher sputtering deposition rate. The sputtering yield depends on several parameters [39], including the energy of the incident ions, the masses of the ions and target atoms, the binding energy of atoms in the solid, and the incident angle of ions. The sputtering yield can be expressed as: Y = α Mm Em (M + m)2UM (2) (2) where m and M are the mass of the bombing ion and target atom, respectively, Em is the kinetic energy of bombing ion, and Um is the bonding energy of target metal. α takes into account the incident angle of ions. where m and M are the mass of the bombing ion and target atom, respectively, Em is the kinetic energy of bombing ion, and Um is the bonding energy of target metal. α takes into account the incident angle of ions. It is important to recall that magnetic ﬁeld strength, CVD chamber volume, power density, gas composition and pressure are other factors that can affect yield values [40]. In addition, the ﬁlm deposition rate decreases for increasing target-substrate distances. Therefore, under the chosen sputtering process parameters (obtained starting from values based on the previous works [41,42] using the same sputtering source), an optimal deposition rate for ZnO and TiO2 of about 2 and 1 nm min−1, respectively, was gained. The sputtering time used for the ZnO target was twice that of the Ti one in order to have similar layer thicknesses. The different sputtering conditions (sputtering power and time, target distance, and gas pressure) were tested in order to ﬁnd the optimal set of parameters able to give a homogeneous membrane 6 of 15 Membranes 2018, 8, 35 coverage with no polymer substrate damage and very small photocatalyst nanoparticles. This last condition ensures a high photoactivity, with catalysis being a surface process. Due to inadequate sputtering conditions, typical sample drawbacks were inhomogeneous coverage, pore occlusion, and presence of cracks (damage of thin ﬁlm), as shown in Figure 3. The best results—in terms of both coverage quality and nanoparticle size—were obtained with the sputtering conditions (sputtering power, target distance, gas pressure, sputtering time) shown in Table 1. 3. Results and Discussion Membranes 2018, 8, x FOR PEER REVIEW 6 of 15 This last condition ensures a high photoactivity, with catalysis being a surface process. Due to inadequate sputtering conditions, typical sample drawbacks were inhomogeneous coverage, pore occlusion, and presence of cracks (damage of thin film), as shown in Figure 3. The best results—in terms of both coverage quality and nanoparticle size—were obtained with the sputtering conditions (sputtering power, target distance, gas pressure, sputtering time) shown in Table 1. Figure 3. Morphology of virgin and TiO2 sputtered polyvinylidene difluoride (PVDF) membranes after different sputtering times (t): (A) virgin PVDF membrane; t = 0, (B) t = 1 h; (C) t = 2 h; (D) t = 3 h; (E) t = 4 h. Figure 3. Morphology of virgin and TiO2 sputtered polyvinylidene diﬂuoride (PVDF) membranes after different sputtering times (t): (A) virgin PVDF membrane; t = 0, (B) t = 1 h; (C) t = 2 h; (D) t = 3 h; (E) t = 4 h. Figure 3. Morphology of virgin and TiO2 sputtered polyvinylidene difluoride (PVDF) membranes after different sputtering times (t): (A) virgin PVDF membrane; t = 0, (B) t = 1 h; (C) t = 2 h; (D) t = 3 h; (E) t = 4 h. Figure 3. Morphology of virgin and TiO2 sputtered polyvinylidene diﬂuoride (PVDF) membranes after different sputtering times (t): (A) virgin PVDF membrane; t = 0, (B) t = 1 h; (C) t = 2 h; (D) t = 3 h; (E) t = 4 h. Table 1. Optimal sputtering parameters able to give a homogeneous coverage with no substrate damage and small nanoparticles. Photocatalyst Sputtering Target Pressure/10−6 Sputtering Table 1. Optimal sputtering parameters able to give a homogeneous coverage with no substrate damage and small nanoparticles. Target y Layer p g Power/W g Distance/10−2 m bar p g Time/min ZnO ZnO 35 8 P(Ar) = 4.5 30 Ti TiO2 65 7 P(Ar) = 2.8 1 60 1 P(O2) = 1.2 × 10−6 bar. 3. Results and Discussion Figure 4 shows the morphology of the top surface in virgin and sputtered PVDF membranes d h i l di i d i T bl 1 B h h l i (Fi 4B C f Target Photocatalyst Layer Sputtering Power/W Target Distance/10−2 m Pressure/10−6 bar Sputtering Time/min ZnO ZnO 35 8 P(Ar) = 4.5 30 Ti TiO2 65 7 P(Ar) = 2.8 1 60 1 P(O2) = 1.2 × 10−6 bar. of the top surface in vi 1 P(O2) = 1.2 × 10−6 bar. under the experimental conditions reported in Table 1. Both photocatalyst coatings (Figure 4B,C for ZnO and TiO2, respectively) were homogeneous with no evident alteration/damage of the virgin PVDF membrane (Figure 4A). In addition, no occlusion of membrane pores was present. Coatings Figure 4 shows the morphology of the top surface in virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1. Both photocatalyst coatings (Figure 4B,C for ZnO and TiO2, respectively) were homogeneous with no evident alteration/damage of the virgin PVDF membrane (Figure 4A). In addition, no occlusion of membrane pores was present. Coatings had 7 of 15 7 f 15 Membranes 2018, 8, 35 a cauliﬂower structure with aggregate diameters of around 100 nm and formed by agglomeration of smaller primary nanoparticles (see later). had a cauliflower structure with aggregate diameters of around 100 nm and formed by agglomeration of smaller primary nanoparticles (see later). p y p Figure 4. Morphology of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1: (A) virgin PVDF membrane; (B) ZnO sputtered PVDF membrane; and (C) TiO2 sputtered PVDF membrane. I d t f th fi th t PVDF b h l d i Figure 4. Morphology of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1: (A) virgin PVDF membrane; (B) ZnO sputtered PVDF membrane; and (C) TiO2 sputtered PVDF membrane. Figure 4. Morphology of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1: (A) virgin PVDF membrane; (B) ZnO sputtered PVDF membrane; and (C) TiO2 sputtered PVDF membrane. Figure 4. Morphology of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1: (A) virgin PVDF membrane; (B) ZnO sputtered PVDF membrane; and (C) TiO2 sputtered PVDF membrane. 3. Results and Discussion In order to further confirm that PVDF membranes were homogeneously covered with photocatalysts, SEM backscattering electron micrographs and spot EDX analysis on sputtered PVDF membranes were performed. Figure 5 shows SEM backscattering electron micrographs and EDX maps for ZnO and TiO2 sputtered membranes. In order to further conﬁrm that PVDF membranes were homogeneously covered with photocatalysts, SEM backscattering electron micrographs and spot EDX analysis on sputtered PVDF membranes were performed. Figure 5 shows SEM backscattering electron micrographs and EDX maps for ZnO and TiO2 sputtered membranes. 8 of 15 8 of 15 Membranes 2018, 8, 35 Membranes 2018, 8, x FOR PEER REVIEW 8 of 15 Figure 5. (A,B) Scanning electron microscopy (SEM) backscattering electron micrographs, (C,D) elemental mapping, and (E–H) energy‐dispersive X‐ray (EDX) analysis performed on TiO2 (on the left) and ZnO (on the right) sputtered PVDF membranes. At larger magnifications, TEM analysis allows characterizing the morphology of primary nanoparticles grown on PVDF membranes. As shown in Figure 6A,B, both ZnO and TiO2 primary nanoparticles were rather spherical in shape with similar average diameters of 11.6 ± 4.2 and 12.1 ± 3.4 nm, respectively (Figure 7). Primary nanoparticles agglomerated into larger aggregates. Figure 5. (A,B) Scanning electron microscopy (SEM) backscattering electron micrographs, (C,D) elemental mapping, and (E–H) energy-dispersive X-ray (EDX) analysis performed on TiO2 (on the left) and ZnO (on the right) sputtered PVDF membranes. At larger magniﬁcations, TEM analysis allows characterizing the morphology of primary nanoparticles grown on PVDF membranes. As shown in Figure 6A,B, both ZnO and TiO2 primary nanoparticles were rather spherical in shape with similar average diameters of 11.6 ± 4.2 and 12.1 ± 3 4 nm respectively (Figure 7) Primary nanoparticles agglomerated into larger aggregates Figure 5. (A,B) Scanning electron microscopy (SEM) backscattering electron micrographs, (C,D) elemental mapping, and (E–H) energy‐dispersive X‐ray (EDX) analysis performed on TiO2 (on the left) and ZnO (on the right) sputtered PVDF membranes. Figure 5. (A,B) Scanning electron microscopy (SEM) backscattering electron micrographs, (C,D) elemental mapping, and (E–H) energy-dispersive X-ray (EDX) analysis performed on TiO2 (on the left) and ZnO (on the right) sputtered PVDF membranes. At larger magnifications, TEM analysis allows characterizing the morphology of primary nanoparticles grown on PVDF membranes. As shown in Figure 6A,B, both ZnO and TiO2 primary nanoparticles were rather spherical in shape with similar average diameters of 11.6 ± 4.2 and 12.1 ± 3.4 nm, respectively (Figure 7). 3. Results and Discussion 10 of 15 10 of 15 Membranes 2018, 8, 35 Membranes 2018 8 x FO Figure 8. Raman spectrum of TiO2 nanoparticles sputtered on PVDF membrane. The peaks at 399, 516, and 639 cm−1 are associated to the Raman active modes B1g, A1g, and Eg, respectively, confirming the anatase structure of TiO2. Figure 8. Raman spectrum of TiO2 nanoparticles sputtered on PVDF membrane. The peaks at 399, 516, and 639 cm−1 are associated to the Raman active modes B1g, A1g, and Eg, respectively, conﬁrming the anatase structure of TiO2. Figure 8. Raman spectrum of TiO2 nanoparticles sputtered on PVDF membrane. The peaks at 399, 516, and 639 cm−1 are associated to the Raman active modes B1g, A1g, and Eg, respectively, confirming the anatase structure of TiO2. Figure 8. Raman spectrum of TiO2 nanoparticles sputtered on PVDF membrane. The peaks at 399, 516, and 639 cm−1 are associated to the Raman active modes B1g, A1g, and Eg, respectively, conﬁrming the anatase structure of TiO2. The Raman spectrum of TiO2 nanoparticles consisted of three peaks with strong intensities at 399, 516, and 639 cm−1, which can be associated to the Raman active modes B1g, A1g, and Eg of anatase structure of TiO2 thin layers. These values are in good agreement with the Raman bands reported in literature [43]. The fourth active Raman mode of anatase structure of TiO2, which is generally placed at 196 cm−1 (Eg mode), was out of the instrument range. It is expected that the deposition of ZnO and TiO2 thin films could change the hydrophilicity of virgin PVDF membranes. Table 2 shows the static contact angles to pure water measured for virgin and sputtered The Raman spectrum of TiO2 nanoparticles consisted of three peaks with strong intensities at 399, 516, and 639 cm−1, which can be associated to the Raman active modes B1g, A1g, and Eg of anatase structure of TiO2 thin layers. These values are in good agreement with the Raman bands reported in literature [43]. The fourth active Raman mode of anatase structure of TiO2, which is generally placed at 196 cm−1 (Eg mode), was out of the instrument range. It is expected that the deposition of ZnO and TiO2 thin ﬁlms could change the hydrophilicity of virgin PVDF membranes. Table 2 shows the static contact angles to pure water measured for virgin and sputtered membranes. 3. Results and Discussion Both ZnO and TiO2 thin films drastically reduced the contact angle, i.e., increased the hydrophilicity of virgin PVDF membrane from 61° to 27° and 26°, respectively, with a consequent decrease in fouling. It is important to recall that the photoactivity of ZnO and TiO2 layers is able to further overcome this drawback. Moreover, the increase in hydrophilicity is expected to have a positive effect in the membrane permeation properties. Table 2 also shows the transmembrane fluxes of virgin and sputtered membranes. The performant effect of catalyst thin film increased the transmembrane flux of virgin PVDF from a value of 200 to 760 and 710 L m−2 h−1 for ZnO and TiO2 thin films, respectively, as a result of the increased hydrophilicity. These results also confirm that the sputtered thin layer did not occlude the membrane pores. Table 2 shows the static contact angles to pure water measured for virgin and sputtered membranes. Both ZnO and TiO2 thin ﬁlms drastically reduced the contact angle, i.e., increased the hydrophilicity of virgin PVDF membrane from 61◦to 27◦and 26◦, respectively, with a consequent decrease in fouling. It is important to recall that the photoactivity of ZnO and TiO2 layers is able to further overcome this drawback. Moreover, the increase in hydrophilicity is expected to have a positive effect in the membrane permeation properties. Table 2 also shows the transmembrane ﬂuxes of virgin and sputtered membranes. The performant effect of catalyst thin ﬁlm increased the transmembrane ﬂux of virgin PVDF from a value of 200 to 760 and 710 L m−2 h−1 for ZnO and TiO2 thin ﬁlms, respectively, as a result of the increased hydrophilicity. These results also conﬁrm that the sputtered thin layer did not occlude the membrane pores. Table 2. Contact angle and transmembrane flux of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1. Photocatalyst Contact Angle/deg Transmembrane Flux/L m−2 h−1 Virgin PVDF 61 ± 1 200 ± 15 ZnO 27 ± 2 760 ± 15 TiO2 26 ± 2 710 ± 15 I d t t t th i h t t l ti ti it i t h ti l ti d Table 2. Contact angle and transmembrane ﬂux of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1. 3. Results and Discussion Primary nanoparticles agglomerated into larger aggregates. At larger magniﬁcations, TEM analysis allows characterizing the morphology of primary nanoparticles grown on PVDF membranes. As shown in Figure 6A,B, both ZnO and TiO2 primary nanoparticles were rather spherical in shape with similar average diameters of 11.6 ± 4.2 and 12.1 ± 3.4 nm, respectively (Figure 7). Primary nanoparticles agglomerated into larger aggregates. 9 of 15 f Membranes 2018, 8, 35 Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 6. Morphology of catalyst nanoparticles grown on PVDF membranes under the experimental conditions reported in Table 1: (A) ZnO sputtered PVDF membrane; and (B) TiO2 sputtered PVDF membrane. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. Figure 7. Distribution of diameters shown by primary nanoparticles present on ZnO and TiO2 sputtered PVDF membranes. MicroRaman spectrum of PVDF membrane obtained by reactive sputtering using a Ti target is reported in Figure 8. MicroRaman spectrum of PVDF membrane obtained by reactive sputtering using a Ti target is reported in Figure 8. MicroRaman spectrum of PVDF membrane obtained by reactive sputtering using a Ti target is reported in Figure 8. 3. Results and Discussion Indeed, ZnO samples have a larger band gap, which leads to the production of less radicals and, consequently, to a lower dye photodegradation. On the contrary, TiO2 has a higher quantum yield engendered by a relatively slower electron‐hole pair recombination, faster electron‐hole pair migration to the surface, fewer defects, and exciton traps in the crystal lattice [47]. Figure 9. Photodegradation of diclofenac sodium salt by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of diclofenac. Figure 9. Photodegradation of diclofenac sodium salt by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of diclofenac. Figure 9. Photodegradation of diclofenac sodium salt by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of diclofenac. Figure 9. Photodegradation of diclofenac sodium salt by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of diclofenac. Even if kinetic rate constants are strongly dependent on membrane composition, investigated pollutants and the UV lamp power used [48], it is important to note that the obtained values of kinetic rate constants are the same order of magnitude as those found in other literature works where the photoactive nanomaterial was either directly synthesized or immobilized on polymer substrates [49– 51]. These rate values make the CVD of photocatalyst nanoparticles on porous polymer membranes a suitable technique for applications in the field of advanced oxidation processes. In fact, the CVD of photocatalysts is a fast process that avoids expensive and time‐consuming syntheses and cleaning post‐treatments, as it is possible to directly sputter photocatalysts onto commercially available membranes while keeping good photocatalytic activities at the same time. Recycling of a catalyst is a very important property in practical applications In order to assess the recycling properties of Even if kinetic rate constants are strongly dependent on membrane composition, investigated pollutants and the UV lamp power used [48], it is important to note that the obtained values of kinetic rate constants are the same order of magnitude as those found in other literature works where the photoactive nanomaterial was either directly synthesized or immobilized on polymer substrates [49–51]. 3. Results and Discussion Nevertheless, the methylene blue photodegradation stopped after 4 h with a plateau of 33% and 8% for PVDF membranes with sputtered ZnO and TiO2 nanoparticles, respectively. Similarly, Figure 10 shows the activity of ZnO and TiO2 sputtered membranes in the methylene blue photodegradation. Also in this case, the photodegradation kinetics of the organic pollutant by PVDF membranes with sputtered ZnO and TiO2 nanoparticles was a first order kinetics but with larger rate constants of 2.2 × 10−2 min−1 and 2.8 × 10−2 min−1, respectively. Nevertheless, the methylene blue photodegradation stopped after 4 h with a plateau of 33% and 8% for PVDF membranes with sputtered ZnO and TiO2 nanoparticles, respectively. Th b h i f i i PVDF b d i Fi 9 d 10 k i h UV The behavior of virgin PVDF membrane reported in Figures 9 and 10 takes into account the UV photolysis of diclofenac sodium and methylene blue, respectively. The behavior of virgin PVDF membrane reported in Figures 9 and 10 takes into account the UV photolysis of diclofenac sodium and methylene blue, respectively. S l t ff t th d d ti ffi i f h t t l t i l di th Several parameters can affect the degradation efﬁciency of photocatalysts including the particular dye/drug, the pH of the solution, the presence of oxygen, the addition of hydrogen peroxide, the nanoparticle average size, and the amount and type of catalyst [44]. In particular, several experimental investigations have found that TiO2 nanoparticles show a photocatalytic efﬁciency higher than ZnO nanoparticles due to their band gap values [45,46]. Indeed, ZnO samples have a larger band gap, which leads to the production of less radicals and, consequently, to a lower dye photodegradation. On the contrary, TiO2 has a higher quantum yield engendered by a relatively slower electron-hole pair recombination, faster electron-hole pair migration to the surface, fewer defects, and exciton traps in the crystal lattice [47]. Several parameters can affect the degradation efficiency of photocatalysts including the particular dye/drug, the pH of the solution, the presence of oxygen, the addition of hydrogen peroxide, the nanoparticle average size, and the amount and type of catalyst [44]. In particular, several experimental investigations have found that TiO2 nanoparticles show a photocatalytic efficiency higher than ZnO nanoparticles due to their band gap values [45,46]. 3. Results and Discussion Photocatalyst Contact Angle/deg Transmembrane Flux/L m−2 h−1 Virgin PVDF 61 ± 1 200 ± 15 ZnO 27 ± 2 760 ± 15 TiO2 26 ± 2 710 ± 15 Table 2. Contact angle and transmembrane flux of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1. Table 2. Contact angle and transmembrane ﬂux of virgin and sputtered PVDF membranes under the experimental conditions reported in Table 1. I o e o e ei p o o a a y i a i i y agai p a a eu i a a i e o pou a organic pollutants, sputtered membranes were placed in a small continuous flow reactor where a water solution of either diclofenac sodium—a well‐known anti‐inflammatory drug—or methylene blue—a well‐known organic dye—was circulated. Figure 9 shows the activity of ZnO and TiO2 sputtered membranes in the diclofenac sodium salt In order to test their photocatalytic activity against pharmaceutical active compounds and organic pollutants, sputtered membranes were placed in a small continuous ﬂow reactor where a water solution of either diclofenac sodium—a well-known anti-inﬂammatory drug—or methylene blue—a well-known organic dye—was circulated. g y p photodegradation. The drug photodegradation by PVDF membranes with sputtered ZnO and TiO2 nanoparticles followed a first order kinetics with similar rate constants of 6.8 × 10−3 min−1 and 8.3 × Figure 9 shows the activity of ZnO and TiO2 sputtered membranes in the diclofenac sodium salt photodegradation. The drug photodegradation by PVDF membranes with sputtered ZnO and 11 of 15 Membranes 2018, 8, 35 TiO2 nanoparticles followed a ﬁrst order kinetics with similar rate constants of 6.8 × 10−3 min−1 and 8.3 × 10−3 min−1, respectively. An almost complete photodegradation of diclofenac sodium salt was obtained within 6 h in both cases. Membranes 2018, 8, x FOR PEER REVIEW 11 of 15 Si il l Fi 10 h h i i f Z O d TiO d b i h h l Similarly, Figure 10 shows the activity of ZnO and TiO2 sputtered membranes in the methylene blue photodegradation. Also in this case, the photodegradation kinetics of the organic pollutant by PVDF membranes with sputtered ZnO and TiO2 nanoparticles was a ﬁrst order kinetics but with larger rate constants of 2.2 × 10−2 min−1 and 2.8 × 10−2 min−1, respectively. 3. Results and Discussion These rate values make the CVD of photocatalyst nanoparticles on porous polymer membranes a suitable technique for applications in the ﬁeld of advanced oxidation processes. In fact, the CVD of photocatalysts is a fast process that avoids expensive and time-consuming syntheses and cleaning post-treatments, as it is possible to directly sputter photocatalysts onto commercially available membranes while keeping good photocatalytic activities at the same time. Recycling of a catalyst 12 of 15 Membranes 2018, 8, 35 is a very important property in practical applications. In order to assess the recycling properties of photocatalysts, ten photodegradation cycles were performed using the same ﬁlm and fresh methylene blue. Both ZnO and TiO2 sputtered membranes were reused in successive runs without performing any cleaning procedure and gave percentages of degraded methylene blue similar to those obtained after the ﬁrst run. These results demonstrated the ability of the ZnO and TiO2 sputtered membranes to fully preserve/restore their initial photocatalytic efﬁciency. In addition, the long-term stability of photocatalyst deposit onto polymer substrates was checked after ten cycles of successive photocatalysis processes. No evident damage was revealed in the nanoparticles layer morphology, conﬁrming the main advantage of an easy reuse of membranes with sputtered catalysts over the homogeneous and heterogeneous catalysis processes, which suffer the drawbacks of catalyst recovery and damage of polymers used to functionalize membranes or bind catalysts [52]. Further work is in progress in order to control the primary particle size in a ﬁner way. Membranes 2018, 8, x FOR PEER REVIEW 12 of 15 sputtered membranes to fully preserve/restore their initial photocatalytic efficiency. In addition, the long‐term stability of photocatalyst deposit onto polymer substrates was checked after ten cycles of successive photocatalysis processes. No evident damage was revealed in the nanoparticles layer morphology, confirming the main advantage of an easy reuse of membranes with sputtered catalysts over the homogeneous and heterogeneous catalysis processes, which suffer the drawbacks of catalyst recovery and damage of polymers used to functionalize membranes or bind catalysts [52]. Further work is in progress in order to control the primary particle size in a finer way. Figure 10. Photodegradation of methylene blue by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of methylene blue. Figure 10. Photodegradation of methylene blue by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. 3. Results and Discussion The behavior of virgin PVDF membrane takes into account the UV photolysis of methylene blue. Figure 10. Photodegradation of methylene blue by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of methylene blue. Figure 10. Photodegradation of methylene blue by PVDF membranes with sputtered ZnO and TiO2 nanoparticles. The behavior of virgin PVDF membrane takes into account the UV photolysis of methylene blue. 4. Conclusio In this 4. Conclusions , p y p photocatalytic nanoparticles, and an application of their use in AOP were shown. The overall sputtering process took less than 1 h, which is lower than conventional times employed in organic synthesis processes of similar materials. The thickness of obtained surface coating by sputtered nanoparticles was found to depend on process conditions. The membranes functionalized with ZnO and TiO2 nanoparticles were characterized by contact angles lower than that shown by virgin membrane, making these composite membranes suitable for the filtration of aqueous solutions. The deposition of a thin layer of nanoparticles increased the transmembrane fluxes as hydrophilicity increased, and no pore occlusion occurred. In addition to its long‐term stability and solvent‐free features, the proposed process of membrane functionalization can be easily upscaled to manufacture membrane modules for the efficient degradation of organic pollutants generally found in wastewater. Author Contributions: E.P., K.A., P.F., T.P., E.F. and C.M. performed the experiments and the data analysis. G.D.F., G.D.P., A.I.M. and F.P.N. provided expert input into the experiments and discussions. G.D.F. and F.P.N. In this work, the results of a CVD functionalization of polymer porous membranes with photocatalytic nanoparticles, and an application of their use in AOP were shown. The overall sputtering process took less than 1 h, which is lower than conventional times employed in organic synthesis processes of similar materials. The thickness of obtained surface coating by sputtered nanoparticles was found to depend on process conditions. The membranes functionalized with ZnO and TiO2 nanoparticles were characterized by contact angles lower than that shown by virgin membrane, making these composite membranes suitable for the ﬁltration of aqueous solutions. The deposition of a thin layer of nanoparticles increased the transmembrane ﬂuxes as hydrophilicity increased, and no pore occlusion occurred. In addition to its long-term stability and solvent-free features, the proposed process of membrane functionalization can be easily upscaled to manufacture membrane modules for the efﬁcient degradation of organic pollutants generally found in wastewater. supe vised t e study a d p ovided scie ti ic discussio . A t e aut o s co t ibuted to t e w iti g a d evisio of the article. Funding: This research was funded by MIUR, the Italian Ministry for University and Research, grant number EX 60% 2017 P 12 Author Contributions: E.P., K.A., P.F., T.P., E.F. and C.M. performed the experiments and the data analysis. G.D.F., G.D.P., A.I.M. and F.P.N. Conflicts of Interest: The authors declare no conflict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Chidichimo, G.; Cupelli, D.; De Filpo, G.; Formoso, P.; Nicoletta, F.P. Nanoparticles as a smart technology for remediation. In Sustainable Development in Chemical Engineering, Innovative Technologies; Piemonte, V., De Falco, M., Basile, A., Eds.; John Wiley & Sons, Ltd.: Chichester, UK, 2013; pp. 297–329. ISBN 9781119953524. 2. Kudryashov, M.A.; Mashin, A.I.; Logunov, A.A.; Chidichimo, G.; De Filpo, G. Frequency dependence of the electrical conductivity in Ag/PAN nanocomposites. Tech. Phys. 2012, 57, 965–970. [CrossRef] 2. Kudryashov, M.A.; Mashin, A.I.; Logunov, A.A.; Chidichimo, G.; De Filpo, G. Frequency dependence of the electrical conductivity in Ag/PAN nanocomposites. Tech. Phys. 2012, 57, 965–970. [CrossRef] . Löbl, P.; Huppertz, M.; Mergel, D. Nucleation and growth in TiO2 ﬁlms prepared by sputtering evaporation. Thin Solid Films 1994, 251, 72–79. [CrossRef] 3. Löbl, P.; Huppertz, M.; Mergel, D. Nucleation and growth in TiO2 ﬁlms prepared by sputtering and evaporation. Thin Solid Films 1994, 251, 72–79. [CrossRef] 4. Carter, C.B.; Norton, M.G. Thin ﬁlms and vapor deposition. In Ceramic Materials; Carter, C.B., . Carter, C.B.; Norton, M.G. Thin ﬁlms and vapor deposition. In Ceramic Materials; Carter, C Norton, M.G., Eds.; Springer: New York, NY, USA, 2007; ISBN 978-0-387-46270-7. 5. Drioli, E.; Ali, A.; Macedonio, F. Membrane distillation: Recent developments and perspectives. Desalination 2015, 356, 56–84. [CrossRef] 5. Drioli, E.; Ali, A.; Macedonio, F. Membrane distillation: Recent developments and perspectives. Desalination 2015, 356, 56–84. [CrossRef] 6. Bhattacharyya, D. Functionalized membranes and environmental applications. Clean Technol. Environ. Policy 2007, 9, 81–83. [CrossRef] 6. Bhattacharyya, D. Functionalized membranes and environmental applications. Clean Technol. Environ. Policy 2007, 9, 81–83. [CrossRef] 7. Di Proﬁo, G.; Polino, M.; Nicoletta, F.P.; Belviso, B.D.; Caliandro, R.; Fontananova, E.; De Filpo, G.; Curcio, E.; Drioli, E. Tailored hydrogel membranes for efﬁcient protein crystallization. Adv. Funct. Mater. 2014, 24, 1582–1590. [CrossRef] 8. Heberer, T. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: A review of recent research data. Toxicol. Lett. 2002, 131, 5–17. [CrossRef] . Jones, O.A.H.; Voulvoulis, N.; Lester, J.N. Human pharmaceuticals in the aquatic environment a rev Environ. Technol. 2001, 22, 1383–1394. [CrossRef] [PubMed] 10. Santos, J.L.; Aparicio, I.; Alonso, E. Occurrence and risk assessment of pharmaceutically active compounds in wastewater treatment plants. A case study: Seville city (Spain). Environ. Int. 2007, 33, 596–601. [CrossRef] [PubMed] 11. Ziylan, A.; Ince, N.H. The occurrence and fate of anti-inﬂammatory and analgesic pharmaceuticals in sewage and fresh water: Treatability by conventional and non-conventional processes. J. 4. Conclusio In this 4. Conclusions provided expert input into the experiments and discussions. G.D.F. and F.P.N. supervised the study and provided scientiﬁc discussion. All the authors contributed to the writing and revision of the article. EX 60%‐2017.Page: 12 C fli f I Th h d l fl f Funding: This research was funded by MIUR, the Italian Ministry for University and Research, grant number EX 60%-2017. Conflicts of Interest: The authors declare no conflict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 13 of 15 13 of 15 Membranes 2018, 8, 35 References Hazard. Mater. 2011, 187, 24–36. [CrossRef] [PubMed] 12. Kolpin, D.W.; Furlong, E.T.; Meyer, M.T.; Thurman, E.M.; Zaugg, S.D.; Barber, L.B.; Buxton, H.T. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. Streams, 1999–2000: A national reconnaissance. Environ. Sci. Technol. 2002, 36, 1202–1211. [CrossRef] [PubMed] 13. Salehi, S.M.; Di Proﬁo, G.; Fontananova, E.; Nicoletta, F.P.; Curcio, E.; De Filpo, G. Membrane distillation by novel hydrogel composite membranes. J. Membr. Sci. 2016, 504, 220–229. [CrossRef] 4. Strathmann, H. Introduction to Membrane Science and Technology; Wiley-VCH Verlag GmbH & Co. KG Weinheim, Germany, 2011; ISBN 978-3-527-32451-4. 15. Drioli, E.; Giorno, L. Encyclopedia of Membranes; Springer: Berlin/Heidelberg, Germany, 2016; ISBN 978-3-662-44325-5. 16. Nicoletta, F.P.; Cupelli, D.; Formoso, P.; De Filpo, G.; Colella, V.; Gugliuzza, A. Light responsive polymer membranes: A review. Membranes 2012, 134, 134–197. [CrossRef] [PubMed] 17. Formoso, P.; Pantuso, E.; De Filpo, G.; Nicoletta, F.P. Electro-conductive membranes for permeation enhancement and fouling mitigation: A short review. Membranes 2017, 7, 39. [CrossRef] [PubMed] 18. Pelizzetti, E.; Serpone, N. Photo-Catalysis Fundamentals and Applications; John Wiley & Sons, Inc.: New York, NY, USA, 1989. 19. Arana, J.; Melian, J.A.H.; Rodriguez, J.M.D.; Pena, P.P. TiO2-photocatalysis as a tertiary treatment of naturally treated wastewater. Catal. Today 2002, 76, 279–289. [CrossRef] 20. Hoffmann, M.R.; Martin, S.T.; Choi, W.Y.; Bahnemann, D.W. Environmental applications of semiconductor photo-catalysis. Chem. Rev. 1995, 95, 69–96. [CrossRef] 21. Minero, C. A rigorous kinetic approach to model primary oxidative steps of photo-catalytic degradations. Sol. Energy Mater. Sol. Cells 1995, 38, 421–430. [CrossRef] 22. Mahmoodi, N.M.; Arami, M.; Limaee, N.Y.; Tabrizi, N.S. Kinetics of heterogeneous photo-catalytic degradation of reactive dyes in an immobilized TiO2 photo-catalytic reactor. J. Colloid Interface Sci. 2006, 295, 159–164. [CrossRef] [PubMed] 14 of 15 14 of 15 Membranes 2018, 8, 35 23. Bhatkhande, D.S.; Pangarkar, V.G.; Beenackers, A. Photocatalytic degradation for environmental applications: A review. J. Chem. Technol. Biotechnol. 2002, 77, 102–116. [CrossRef] 24. Cho, M.; Chung, H.; Choi, W.; Yoon, J. Linear correlation between inactivation of E. coli and OH radical concentration in TiO2 photocatalytic disinfection. Water Res. 2004, 38, 1069–1077. [CrossRef] [PubMed] 25. Huang, Z.; Maness, P.C.; Blake, D.; William, A.J. Bactericidal mode of titanium dioxide photocatalysis. J. Photochem. Photobiol. A 2000, 130, 163–170. [CrossRef] 26. De Filpo, G.; Palermo, A.M.; Tolmino, R.; Formoso, P.; Nicoletta, F.P. Gellan gum hybrid hydrogels for the cleaning of paper artworks contaminated with Aspergillus versicolor. Cellulose 2016, 23, 3265–3279. [CrossRef] 27. References De Filpo, G.; Palermo, A.M.; Rachiele, F.; Nicoletta, F.P. Preventing fungal growth in wood by titanium dioxide nanoparticles. Int. Biodeterior. Biodegrad. 2013, 85, 217–222. [CrossRef] 28. De Filpo, G.; Palermo, A.M.; Munno, R.; Molinaro, L.; Formoso, P.; Nicoletta, F.P. Gellan gum/titanium dioxide nanoparticle hybrid hydrogels for the cleaning and disinfection of parchment. Int. Biodeterior. Biodegrad. 2015, 103, 51–58. [CrossRef] 29. Shon, H.K.; Vigneswaran, S.; Ngo, H.H.; Kim, J.H. Chemical coupling of photocatalysis with ﬂocculation and adsorption in the removal of organic matter. Water Res. 2005, 39, 2549–2558. [CrossRef] [PubMed] 30 Mozia, S ; Morawski, A W Hybridization of photocatalysis and membrane distillation for puriﬁcation of 9. Shon, H.K.; Vigneswaran, S.; Ngo, H.H.; Kim, J.H. Chemical coupling of photocatalysis with ﬂoccula and adsorption in the removal of organic matter. Water Res. 2005, 39, 2549–2558. [CrossRef] [PubMed] 29. Shon, H.K.; Vigneswaran, S.; Ngo, H.H.; Kim, J.H. Chemical coupling of photocatalysis with ﬂocculation and adsorption in the removal of organic matter. Water Res. 2005, 39, 2549–2558. [CrossRef] [PubMed] 30. Mozia, S.; Morawski, A.W. Hybridization of photocatalysis and membrane distillation for puriﬁcation of wastewater. Catal. Today 2006, 118, 181–188. [CrossRef] 0. Mozia, S.; Morawski, A.W. Hybridization of photocatalysis and membrane distillation for puriﬁcatio wastewater. Catal. Today 2006, 118, 181–188. [CrossRef] 31. Ho, D.P.; Vigneswaran, S.; Ngo, H.H. Integration of photocatalysis and microﬁltration in removing efﬂuent organic matter from treated sewage efﬂuent. Sep. Sci. Technol. 2010, 45, 155–162. [CrossRef] 32. Mendez-Arriaga, F.; Esplugas, S.; Gimenez, J. Photocatalytic degradation of nonsteroidal anti-inﬂammatory drugs with TiO2 and simulated solar irradiation. Water Res. 2008, 42, 585–594. [CrossRef] [PubMed] 33. Ganiyu, S.O.; Van Hullebusch, E.D.; Cretin, M.; Esposito, G. Coupling of membrane ﬁltration and advanced oxidation processes for removal of pharmaceutical residues: A critical review. Sep. Purif. Technol. 2015, 156, 891–914. [CrossRef] 34. Molinari, R.; Caruso, A.; Argurio, P.; Poerio, T. Degradation of the drugs Gemﬁbrozil and Tamoxifen in R. pressurized and de-pressurized membrane photoreactors using suspended polycrystalline TiO2 as catalyst. J. Membr. Sci. 2008, 319, 54–63. [CrossRef] 35. Sarasidis, V.C.; Plakas, K.V.; Patsios, S.I.; Karabelas, A.J. Investigation of Diclofenac degradation in a continuous photo-catalytic membrane reactor: Inﬂuence of operating parameters. Chem. Eng. J. 2014, 239, 299–311. [CrossRef] 36. Zeman, P.; Takabayashi, S. Effect of total and oxygen partial pressures on structure of photocatalytic TiO2 ﬁlms sputtered on unheated substrate. Surf. Coat. Technol. 2002, 153, 93–99. [CrossRef] 37. References Luttrell, T.; Halpegamage, S.; Tao, J.; Kramer, A.; Sutter, E.; Batzill, M. Why is anatase a better photocatalyst than rutile?—Model studies on epitaxial TiO2 ﬁlms. Sci. Rep. 2014, 4, 4043. [CrossRef] [PubMed] 38. Mastropietro, T.F.; Meringolo, C.; Poerio, T.; Scarpelli, F.; Godbert, N.; Di Proﬁo, G.; Fontananova, E. Multistimuli activation of TiO2/α-Alumina membranes for degradation of methylene blue. Ind. Eng. Chem. Res. 2017, 56, 11049–11057. [CrossRef] 39. Plummer, J.D.; Deal, M.; Grifﬁn, P.D. Silicon VLSI Technology; Prentice Hall Press: Upper Saddle River, NJ, USA, 2008. 40. Tavakolmoghadam, M.; Mohammadi, T.; Hemmati, M.; Naeimpour, F. Surface modiﬁcation of PVDF membranes by sputtered TiO2: Fouling reduction potential in membrane bioreactors. Desalin. Water Treat. 2014, 57, 3328–3338. [CrossRef] 41. Marques, S.M.; Tavares, C.J.; Lanceros-Mendez, S.; Denchev, Z. X-ray scattering experiments on sputtered titanium dioxide coating onto PVDF polymers for self-cleaning application. J. Appl. Polym. Sci. 2011, 119, 726–731. [CrossRef] 42. Tavares, C.J.; Marques, S.M.; Lanceros-Mendez, S.; Robouta, L.; Alves, E.; Barradas, N.P.; Munnik, F.; Girardeau, T.; Riviere, J.P. N-doped photocatalytic titania thin ﬁlms on active polymer substrates. J. Nanosci. Nanotechnol. 2010, 10, 1072–1077. [CrossRef] [PubMed] 43. Ohsaka, T.; Izumi, F.; Fujiki, Y. Raman spectrum of anatase, TiO2. J. Raman Spectrosc. 1978, 7, 321–324. [CrossRef] 44. Qamar, M.; Muneer, M. A comparative photocatalytic activity of titanium dioxide and zinc oxide by investigating the degradation of vanillin. Desalination 2009, 249, 535–540. [CrossRef] 15 of 15 Membranes 2018, 8, 35 45. Colombo, E.; Li, W.; Bhangu, S.K.; Ashokkumar, M. Chitosan microspheres as a template for TiO2 and ZnO microparticles: Studies on mechanism, functionalization and applications in photocatalysis and H2S removal. RSC Adv. 2017, 7, 19373–19383. [CrossRef] 46. Barnes, R.J.; Molina, R.; Xu, J.; Dobson, P.J.; Thompson, I.P. Comparison of TiO2 and ZnO nanoparticles for photocatalytic degradation of methylene blue and the correlated inactivation of gram-positive and gram-negative bacteria. J. Nanopart. Res. 2013, 15, 1432. [CrossRef] 47. Bennett, S.W.; Keller, A.A. Comparative photoactivity of CeO2, γ-Fe2O3, TiO2 and ZnO in various aqueous systems. Appl. Catal. B Environ. 2011, 102, 600–607. [CrossRef] 48. Martins, P.M.; Miranda, R.; Marques, J.; Tavares, C.J.; Botelho, G.; Lanceros-Mendez, S.; Leong, S.; Razmijou, A.; Wang, K.; Hapgood, K.; et al. TiO2 based photocatalytic membranes: A review. J. Membr. Sci. 2014, 472, 167–184. [CrossRef] 49. Martins, P.M.; Miranda, R.; Marques, J.; Tavares, C.J.; Botelho, G.; Lanceros-Mendez, S. Comparative efﬁciency of TiO2 nanoparticles in suspension vs. immobilization into (PVDF–TrFE) porous membranes. RSC Adv. © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References 2016, 6, 12708–12716. [CrossRef] 50. Fischer, K.; Grimm, M.; Meyers, J.; Dietrich, C.; Gläser, R.; Schulze, A. Photoactive microﬁltration membranes via directed synthesis of TiO2 nanoparticles on the polymer surface for removal of drugs from water. J. Membr. Sci. 2015, 478, 49–57. [CrossRef] 51. Starr, B.J.; Tarabara, V.V.; Herrera-Robledo, M.; Zhou, M.; Roualdès, S.; Ayral, A. Coating porous membranes with a photocatalyst: Comparison of LbL self-assembly and plasma-enhanced CVD techniques. J. Membr. Sci. 2016, 514, 340–349. [CrossRef] 52. Song, H.; Shao, J.; He, Y.; Liu, B.; Zhong, X. Natural organic matter removal and ﬂux decline with PEG–TiO2-doped PVDF membranes by integration of ultra-ﬁltration with photocatalysis. J. Membr. Sci. 2012, 405–406, 48–56. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2887899875	https://europepmc.org/articles/pmc6141291?pdf=render	English	null	Post-epidemic awareness and knowledge of Lassa fever among residents in affected community in Ibadan, Oyo State, Nigeria	Veterinary world/Veterinary World	2,018	cc-by	4,325	Copyright: Awosanya, Open Access. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. RESEARCH ARTICLE Open Access RESEARCH ARTICLE Open Access Veterinary World, EISSN: 2231-0916 Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf Abstract Aim An outbreak of Lassa fever occurred in Ibadan with a case fatality rate of 50% in 2012. Awareness creation and sensitization is a known disease prevention and control strategy. An assessment of the awareness level and knowledge of Lassa fever in the affected community and a nearby university community was done to aid the development of effective information, education, and communication (IEC) material adaptable to the affected community. Materials and Methods: A semi-structured questionnaire was used to obtain the data about awareness and knowledge of Lassa fever from 130 respondents. Descriptive statistics and statistical differences between categorical variables were done using Fisher’s exact test at 5% significant level. Results: Respondents’ age was 29.9 ± 10.9 years. Awareness level in the affected and university communities was 42 (65%) and 55 (85%), respectively (p=0.02). The most reported source of awareness was the television and radio (59.8%). Only 33.1% of all respondents had good knowledge of the clinical symptoms. Most (68.5%) of the respondents knew rat as the reservoir: However, 56.9% and 80.0% of respondents from the affected and university communities, respectively, had this knowledge (p=0.01). About one-third (30.0%) of the respondents had good knowledge of preventive measures: 18.5% and 41.5% from affected and university communities, respectively (p=0.01). Conclusion: Knowledge of respondents on Lassa fever symptoms, reservoir, and preventive measures was low in the affected community; the IEC material was developed to address the knowledge gaps. Awareness was also intensified in the affected community. Keywords: IEC materials, knowledge, Lassa fever, outbreaks. E. J. Awosanya Veterinary Public Health and Preventive Medicine, Faculty of Veterinary Medicine, University of Ibadan, Ibadan, Oyo State, Nigeria. Department of Veterinary Public Health and Preventive Medicine, Faculty of Veterinary Medicine, Univ Ibadan, Oyo State, Nigeria. Corresponding author: E. J. Awosanya, e-mail: emmafisayo@yahoo.com Received: 08-03-2018, Accepted: 05-07-2018, Published online: 04-08-2018 doi: 10.14202/vetworld.2018.1059-1063 How to cite this article: Awosanya EJ (2018) Post-epidemic awareness and knowledge of Lassa fever among residents in affected community in Ibadan, Oyo State, Nigeria, Veterinary World, 11(8): 1059-1063. Post-epidemic awareness and knowledge of Lassa fever among residents in affected community in Ibadan, Oyo State, Nigeria E. J. Awosanya Veterinary World, EISSN: 2231-0916 Introduction in nosocomial cases through contact with secretions and excretions of infected persons [1]. About 80% of human infections are asymptomatic [1,3]. Lassa fever is one of the viral hemorrhagic fevers caused by a single-stranded RNA virus beloning to the virus family Arenaviridae. Lassa fever presents with gradual onset of fever, malaise, headache, sore throat, muscle pain, chest pain, nausea, vomiting, diarrhea, cough, and abdominal pain, which may progress to facial edema, mucosal bleeding, disorientation, coma, and death in the late stages. The incubation period of Lassa fever ranges from 6 to 21 days [1]. The case fatality rate (CFR) may range from 1% to 15% [1] but could exceed 50% depending on the patient sta tus [2-4]. There could be a direct transmission of the disease to humans through ingestion of food and food materials contaminated by the feces and/or urine of the mammalian reservoir of the Lassa virus, a peridomestic multimammate rat, Mastomys natalensis. The disease could also spread from person to person, especially, The first outbreak of Lassa fever in Nigeria was in 1969 in a village called Lassa in Borno State [5], and the disease has assumed an endemic status [1]. About two-third of the 36 states in Nigeria are endemic. There appears to be a seasonal pattern in the outbreak of Lassa fever in Nigeria, with most cases occurring in the dry season. Within the past 7 years, Nigeria reported between 18 and 201 confirmed cases of Lassa fever annually, with annual CFR of between 24% and 79% among confirmed cases [6]. A single case of Lassa fever is regarded as an outbreak, and a suspected case of Lassa fever is defined as illness with gradual onset with one or more of the following: Malaise, fever, headache, sore throat, cough, nausea, vomiting, diarrhea, myalgia, chest pain hearing loss, and a history of contact with excreta of rodents or with a case of Lassa fever, while a confirmed case of Lassa fever is a suspected case that is laboratory confirmed (positive IgM antibody, PCR, or virus isolation) or epidemiologically linked to a laboratory-confirmed case [7]. In August 2012, Oyo State had four suspected cases of Lassa fever; two were laboratory-confirmed cases and one fatality [6]. Results The mean age of the respondents was 29.9±10.9 years, 77 (59.2%) were female. Of the 130 respondents, most (66.9%) had tertiary education (Table-1). The overall awareness level of Lassa fever was 74.6%. About 65% (42 of 65) of respondents from the affected local community and about 85% (55 of 65) of respondents from the unaf fected university community were aware of Lassa fever. The difference in awareness level was statis tically significant at p=0.02 (Table-2). Only 33.1% of the total respondents had good knowledge of the clinical symptoms. The knowledge of clinical symp toms from the affected and university communities was 29.2% and 36.9%, respectively. The difference Study population Table-1: Demographic characteristics of respondents in the affected local and unaffected university communities, Ibadan, Nigeria 2012. Table-1: Demographic characteristics of respondents in the affected local and unaffected university communities, Ibadan, Nigeria 2012. The respondents were residents from the affected local community and a university community with no history of Lassa fever outbreak. Half of the total respondents were each from affected local community and a university community without history of Lassa fever outbreak. Variables Affected local community n=65 (%) Unaffected University community n=65 (%) Age (years) <18 0 (0) 1 (1.5) 18‑22 4 (6.2) 35 (53.9) 23‑27 9 (13.8) 21 (32.3) 28‑32 16 (24.6) 8 (12.3) 33‑37 8 (12.3) 0 (0) >38 28 (43.1) 0 (0) Gender Female 37 (56.9) 40 (61.5) Male 28 (43.1) 25 (38.5) Educational level None 1 (1.5) 0 (0) Primary 7 (10.8) 0 (0) Secondary 35 (53.9) 0 (0) Tertiary 22 (33.8) 65 (100) Variables Affected local community n=65 (%) Unaffected University community n=65 (%) Study design This study was a cross-sectional survey. Data collection and reliability A semi-structured questionnaire was used to obtain the data on demography, awareness, and knowl edge of Lassa fever. The questionnaire was pretested and interviewer administered to the respondents. The knowledge scale on Lassa fever has good internal con sistency, with a Cronbach’s alpha coefficient of 0.75. This study, therefore, aimed to determine the awareness level immediately after the outbreak of Lassa fever and to identify knowledge gaps in the affected local community. Study site and setting The study site was Ibadan in Oyo State. Ibadan is located on latitude 7.3877800 and longitude 3.8963900 in decimal coordinates. The human population of Ibadan is about 2,559,853 of the 5,580,894 in Oyo State [8]. Ibadan has two tertiary hospitals, 18 sec ondary and 244 primary health facilities [9]. Ibadan is an agrarian community with some commercial and industrial activities. The inhabitants are diverse with the Yoruba Ethnic Group in the majority. The Yorubas have a predilection for living in high-density urban centers [10]. mean age. mean age. Ethical approval and informed consent The study was part of outbreak response, so the ethical approval was not necessary. However, respondents who do not wish to participate were respected and not denied of the common right on information dissemination. Confidentiality of the col lected data was maintained. Verbal informed consent was obtained from each participant. Data analysis The data were entered into Excel spreadsheet 2010 and analyzed using Epi-Info version 3.5.4. Knowledge of clinical symptoms, general preventive measures, and personal protective measures was on a scale of four; a score of two and above is graded as good. In addition, ability to know that spread of the Lassa fever is through contact with the secretion or excretion from an infected person and that rat can transmit the Lassa virus is graded as good. Descriptive statistics were done, and significant differences between categorical variables were assessed using Fisher’s exact test. The level of significance was 5% at 95% confidence interval. Study period The study was conducted from September to December 2012. Introduction As part of the interventions, Veterinary World, EISSN: 2231-0916 1059 Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf an assessment of the awareness level and knowledge of Lassa fever of the affected local community and a nearby unaffected university community was done to develop an effective information education and com munication material that can be adapted to the affected local community. Sample size and sampling Residents from both the affected (61.9% [26 of 42]) and university (58.2% [32 of 55]) communities reported the television and radio as the most source of information; however, the difference was not statistically significant between the two com munities. Friends (47.3% [26 of 55]) were the second most source of information among residents from the university community and the third most source of information (14.3% [6 of 42]) in the affected local community; the difference was, however, sta than respondents from the university community (Table-3). Discussion The overall awareness level of Lassa fever by respondents was moderately high, with respondents from the university community 3 times more likely to Table-2: Test of significance in awareness level and knowledge of Lassa fever between affected local and unaffected university communities, Ibadan, Nigeria 2012. Variables Affected local community n=65 (%) Unaffected university community n=65 (%) Odds ratio (95% CI) p-value Awareness level Yes 42 (64.6) 55 (84.6) 0.3 (0.1‑0.8) 0.02* No 23 (35.4) 10 (15.4) Knowledge of symptoms Good 19 (29.2) 24 (36.9) 0.7 (0.3‑1.5) 0.46 Fair 46 (70.8) 41 (63.1) Knowledge of spread Good 9 (13.8) 3 (4.6) 3.3 (0.9‑12.9) 0.13 Fair 56 (86.2) 62 (95.4) Knowledge of prevention Good 12 (18.5) 27 (41.5) 0.3 (0.1‑0.7) 0.01* Fair 53 (81.5) 38 (58.5) Knowledge of role of rat in transmission Good 37 (56.9) 52 (80.0) 0.3 (0.2‑0.7) 0.01* Fair 28 (43.1) 13 (20.0) Knowledge of what to do personally to avoid contracting Lassa fever Good 15 (23.1) 10 (15.4) 1.7 (0.7‑4.0) 0.37 Fair 50 (76.9) 55 (84.6) Significant at P<0.05. CI=Confidence interval Figure-1: Source of awareness on Lassa fever among respondents in both affected local and unaffected university communities, Ibadan, Nigeria 2012. Table-2: Test of significance in awareness level and knowledge of Lassa fever between affected local and unaffected university communities, Ibadan, Nigeria 2012. Figure-1: Source of awareness on Lassa fever among respondents in both affected local and unaffected university communities, Ibadan, Nigeria 2012. Figure-1: Source of awareness on Lassa fever among respondents in both affected local and unaffected university communities, Ibadan, Nigeria 2012. was, however, not statistically significant (p>0.05). Most (68.5%) of all respondents had good knowl edge of the reservoir as rat. The knowledge of the reservoir from the affected and university com munities was 56.9% and 80.0%, respectively. The difference was statistically significant (p=0.01). About one-third (30.0%) of the total respondents had a good knowledge of preventive measures. Sample size and sampling The knowledge of preventive measures from the affected and university communities was 18.5% and 41.5%, respectively. The difference was statistically signif icant (p=0.01) (Table-2). Of the 97 respondents who were aware of Lassa fever, the most reported source of awareness was the television and radio (59.8%), while the least reported source was religious houses (1.0%) (Figure-1). Residents from both the affected (61.9% [26 of 42]) and university (58.2% [32 of 55]) communities reported the television and radio as the most source of information; however, the difference was not statistically significant between the two com munities. Friends (47.3% [26 of 55]) were the second most source of information among residents from the university community and the third most source of information (14.3% [6 of 42]) in the affected local community; the difference was, however, sta tistically significant (p=0.02) between the commu nities. On what to do should they have persistent fever, sore throat, persistent headache for more than three days, most respondents both from the affected (72.3% [47 of 65]) and university (80.0% [52 of 65]) communities would visit a health center. However, though not statistically significant, it is noteworthy that respondents from the affected local community are more likely to visit other places, such as pat ent or drug store, herbalist, or use of herbs, while less likely to self-medicate or visit religious houses Figure-1: Source of awareness on Lassa fever among respondents in both affected local and unaffected university communities, Ibadan, Nigeria 2012. than respondents from the university community (Table-3). Sample size and sampling A total of 130 respondents were engaged. The formula for a cross-sectional survey was used [11] with the assumption that 50% of the community mem bers were aware of Lassa fever and had good knowl edge of it and a margin of error of 10%. To minimize sampling bias, the respondents were selected using transect-trained enumerators (2) moved in all direc tions, and in any street, every other household was selected until the sample size was completed. More so, it was ensured that the calculated true population mean age of 24.5 years for Ibadan [8] falls within one standard deviation from the sampled respondents’ 1060 Veterinary World, EISSN: 2231-0916 Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf Table-2: Test of significance in awareness level and knowledge of Lassa fever between affected local and unaffected university communities, Ibadan, Nigeria 2012. Variables Affected local community n=65 (%) Unaffected university community n=65 (%) Odds ratio (95% CI) p-value Awareness level Yes 42 (64.6) 55 (84.6) 0.3 (0.1‑0.8) 0.02 No 23 (35.4) 10 (15.4) Knowledge of symptoms Good 19 (29.2) 24 (36.9) 0.7 (0.3‑1.5) 0.46 Fair 46 (70.8) 41 (63.1) Knowledge of spread Good 9 (13.8) 3 (4.6) 3.3 (0.9‑12.9) 0.13 Fair 56 (86.2) 62 (95.4) Knowledge of prevention Good 12 (18.5) 27 (41.5) 0.3 (0.1‑0.7) 0.01* Fair 53 (81.5) 38 (58.5) Knowledge of role of rat in transmission Good 37 (56.9) 52 (80.0) 0.3 (0.2‑0.7) 0.01* Fair 28 (43.1) 13 (20.0) Knowledge of what to do personally to avoid contracting Lassa fever Good 15 (23.1) 10 (15.4) 1.7 (0.7‑4.0) 0.37 Fair 50 (76.9) 55 (84.6) Significant at P<0.05. CI=Confidence interval was, however, not statistically significant (p>0.05). Most (68.5%) of all respondents had good knowl edge of the reservoir as rat. The knowledge of the reservoir from the affected and university com munities was 56.9% and 80.0%, respectively. The difference was statistically significant (p=0.01). About one-third (30.0%) of the total respondents had a good knowledge of preventive measures. The knowledge of preventive measures from the affected and university communities was 18.5% and 41.5%, respectively. The difference was statistically signif icant (p=0.01) (Table-2). Of the 97 respondents who were aware of Lassa fever, the most reported source of awareness was the television and radio (59.8%), while the least reported source was religious houses (1.0%) (Figure-1). Discussion Friends have been reported to play a vital role in information dissemination among students of secondary and tertiary learning [17,18]. However, it has been reported that information dissemination on a thematic issue is more effective if several media are adopted than just a particular channel [19]. It is worth mentioning that respondents from affected local community were 3 times more likely to have good knowledge of the manner of spread of Lassa fever and 2 times more likely to have good knowledge of how to personally protect oneself from contracting Lassa infection than respondents from the unaffected university community, though not significantly so, maybe due to recent experience of Lassa fever out break. Experience sometimes may positively influence knowledge acquisition [22]. In addition, the more like lihood of respondents in affected local community than the university community to visit other places such as drug or patent stores and herbalist centers aside health- care centers for medical attention should inform health workers or researchers of other places for active case search, surveillance, and route of information dissem ination when working among such study population. Uzochukwu and Onwujekwe [23] observed a similar trend in the health-seeking behavior of respondents in the treatment and diagnosis of malaria, a febrile ill ness which presents with similar symptoms with the early stage of Lassa fever. The observation also lays credence to approach information dissemination on a thematic issue through the use of the combination of channels for IEC to be effective [19]. p j p [ ] More than half of the respondents had knowl edge of the role of rat (Mastomys spp.) in the trans mission of the Lassa virus, similar to the findings of Akinbodewa et al. [20] and Asogun et al. [12]. However, respondents from the university community were 3 times more likely to have good knowledge of the role of rat in Lassa virus transmission than respon dents from the affected local community. The knowl edge of the respondents on general prevention of Lassa fever in both local and university communities was low, similar to the report of Olalekan [14]. In a similar pattern, respondents from the university com munity were 3 times more likely to have good knowl edge of general prevention of Lassa fever than those from local communities. Discussion The overall awareness level of Lassa fever by respondents was moderately high, with respondents from the university community 3 times more likely to be aware than local community. Level of education has been associated with awareness level and knowledge of Lassa fever [12]. Low level of awareness on Lassa fever is often reported from members of local com munities [12-15]. However, Reuben and Gyar [16] reported a higher level of awareness in a local com munity probably due to intensified sensitization and the extent of endemicity of Lassa fever in that local ity. The most reported source of awareness was the radio and television similar to the reports of Ilesanmi et al. [13], Olalekan [14], and Asogun et al. [12] in Veterinary World, EISSN: 2231-0916 1061 Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf e analysis of source of awareness and what action will be taken in case of index of suspicion for Lassa ondents in affected local and unaffected university communities, Ibadan, Nigeria 2012. Table-3: Bivariate analysis of source of awareness and what action will be taken in case of index of suspicion for Lassa fever among respondents in affected local and unaffected university communities, Ibadan, Nigeria 2012. Variables with multiple responses Affected local community n (%) Unaffected university community n (%) Odds ratio (95% CI) p‑value Source of awareness Radio/TV 26 (55.3) 32 (39.5) Ref. Health workers 11 (23.4) 19 (23.5) 0.7 (0.3‑1.9) 0.61 Friends 6 (12.8) 26 (32.1) 0.3 (0.1‑0.9) 0.02 Religious houses, spouse, and others 4 (8.5) 4 (4.9) 1.2 (0.2‑7.3) 1.00 Action to be taken on index of suspicion for Lassa fever Visit to nearby health center 47 (53.4) 52 (61.2) Ref. Visit to a patent/drug store 16 (18.2) 7 (8.2) 2.5 (0.9‑7.9) 0.09 Visit to a herbalist/use herbs 7 (7.9) 2 (2.4) 3.9 (0.7‑39.6) 0.16 Visit to a religious house 3 (3.4) 7 (8.2) 0.5 (0.1‑2.2) 0.47 Self‑medication 13 (14.8) 16 (18.8) 0.9 (0.4‑2.2) 0.97 Others 2 (2.3) 1 (1.2) 2.2 (0.1‑133) 0.94 Significant at P<0.05. CI=Confidence interval Significant at P<0.05. CI=Confidence interval seeking health-care intervention early. Early treatment with ribavirin is effective as a post-exposure prophylac tic though with possible side effects [21]. local communities. However, among respondents from the university community, friends were equally found to play a significant role as a source of awareness on Lassa fever. Discussion Although the questionnaire was interviewer administered to respondents in their local language, the higher level of education among respondents from the university community could have accounted for the differences in knowledge levels [12]. Another possible contributory factor to the differences in knowledge levels could be the reduced frequency with which information on Lassa fever is heard on either the radio or television due to an erratic power supply which is more of a challenge in local commu nities than the university community. In general, low knowledge of the early symptoms of Lassa fever was observed among the respondents similar to the report of Akinbodewa et al. [20]. Although the early symptoms of Lassa fever are similar to that of malaria and other viral hemorrhagic fevers, the knowledge of Lassa clin ical presentations could assist an affected individual in References 15. Oladeinde, B.H., Omoregie, R. and Odia, I. (2015) Public awareness of Lassa fever in three rural communities of Nigeria. Int. J. Health Promot. Educ., 53: 128-135. 1. World Health Organization. (2017a) Lassa Fever. Available: http://www.who.int/mediacentre/factsheets/fs179/en/. Accessed on 02-08-2017. 1. World Health Organization. (2017a) Lassa Fever. Available: http://www.who.int/mediacentre/factsheets/fs179/en/. Accessed on 02-08-2017. 16. Reuben, C.R. and Gyar, S.D. (2016) Knowledge, attitudes and practices of Lassa fever in and around Lafia, Central Nigeria. Int. J. Public Health Epidemiol. Res., 2: 14-19. 2. Fisher-Hoch, S.P., Tomori, O., Nasidi, A., Perez- Oronoz, G.I., Fakile, Y., Hut wagner, L. and McCormick, J.B. (1995) Review of cases of nosocomial Lassa fever in Nigeria: The high price of poor medical prac tice. BMJ,  311: 857-859. 17. Awosanya, E.J. and Akande, H.O. (2015) Animal health care seeking behavior of pets or livestock owners and knowledge and awareness on zoonoses in a university com munity. Vet. World, 8: 841-847. 3. Richmond, J.K. and Baglole, D.J. (2003) Lassa fever: Epidemiology, clinical features, and social conse quences. BMJ, 327: 1271-1275. 18. Awosanya, E.J. and Adebimpe, A.P. (2013) Factors asso ciated with rabies awareness and attitude to dog bite in a University community. Bull. Anim. Health Prod. Afr., 61: 559-70. 4. World Health Organization. (2017b) Epidemic Focus. Available: http://www.who.int/csr/disease/epidemic-focus/ lassa-fever/en/. [Last accessed on 02-08-2017. 19. Laverack, G. and Huy, D.D. (2003) Transforming infor mation, education and communication in Vietnam. Health Educ., 103: 363-369. [ 5. Frame, J.D., Baldwin, J.M. Jr., Gocke, D.J. and Troup, J.M. (1970) Lassa fever, a new virus disease of man from West Africa. I. Clinical description and pathological findings. Am. J. Trop. Med. Hyg., 19: 670-676. 20. Akinbodewa, A.A., Adejumo, O.A., Alli, E.O., Olarewaju, C.A., Akinbodewa, G.O., Adejumo, O.A., Osho, P.O., Akinfiresoye, A.O. and Balogun, F.O. (2016) Knowledge of Lassa fever among students of a college of education: Call for inclusion in curriculum. Br. J. Med. Med. Res., 16: 1-8. 6. Federal Ministry of Health. (2016) Morbidity and mor tality due to Lassa fever virus in Nigeria. FMOH/NCDC, Surveillance. Available from: http://www.health.gov.ng/ index.php/resources/reports/weekly-epidemiology-reports. Accessed on 27-01-2016. 21. Hadi, C.M., Goba, A., Khan, S.H., Bangura, J., Sankoh, M., Koroma, S., Juana, B., Bah, A., Coulibaly, M. and Bausch, D.G. (2010) Ribavirin for Lassa fever post-expo sure prophylaxis. Emerg. Infect. Dis., 16: 2009. 7. Integrated Disease Surveillance and Response/World Health Organization and Centers for Disease Control and Prevention. Conclusion The knowledge gaps identified were the low knowledge level on the role of the rat in the transmis sion of Lassa virus, general preventive measures, and early symptoms of Lassa fever among affected local respondents which were taken into consideration in the development of the IEC materials in the affected local community. The importance of seeking immedi ate health care in health-care facilities should an indi vidual have persistent fever, sore throat, and persistent headache for more than 3 days was also stressed in the jingle developed The message content of IEC 1062 Veterinary World, EISSN: 2231-0916 Available at www.veterinaryworld.org/Vol.11/August-2018/7.pdf materials should be developed not only to acquaint the people but also to provide the needed knowledge for action. Health Management Information System Data, 2016. Available from: https://www.dhis2nigeria.org.ng. Accessed on 03-07-2017. Health Management Information System Data, 2016. Available from: https://www.dhis2nigeria.org.ng. Accessed on 03-07-2017. 10. Oyo State Government. (2016) Our History. Available from: http/www.oyostate.gov.ng/about-oyo-state/our-his tory/. Accessed on 15-07-2016. Acknowledgments g p 12. Asogun, D.A., Adomeh, D.I., Ehimuan, J., Odia, I., Hass, M., Gabriel, M., Ölschläger, S., Becker-Ziaja, B., Folarin, O., Phelan, E. and Ehiane, P.E. (2012) Molecular diagnostics for Lassa fever at Irrua specialist teaching hos pital, Nigeria: Lessons learnt from two years of laboratory operation. PLoS Negl. Trop. Dis., 6: e1839. The author acknowledges the support of the Nigeria Field Epidemiology and Laboratory Training Program and Oyo State Ministry of Health for their role during the outbreak response, data collection on the field, and financial support. p g p 13. Ilesanmi, O.S., Omotoso, B., Alele, F.O. and Adewuyi, P. (2015) Awareness of Lassa fever in a Rural Community in South West Nigeria. J. Comm. Hum. Rights, 4: 1-10. Authors’ Contributions EJA conceived the study, analyzed the data, drafted, and approved the final manuscript. 11. Thrusfield, M. (2007) What sample size should be selected? In: Veterinary Epidemiology. 3rd ed. UK Blackwell Publishing Limited., Ames, IA. p232-238. Competing Interests 14. Olalekan, A.W. (2015) Community awareness and percep tion towards rodent control: Implications for prevention and control of Lassa fever in urban slums of southwestern Nigeria. MJHS, 2: 26-32. The author declare that there is no competing interests. References (2010) Technical Guidelines for Integrated Disease Surveillance and Response in the African Region, Brazzaville, Republic of Congo and Atlanta, USA. p1-398. 22. Huang, F.L. and Moon, T.R. (2009) Is experience the best teacher? A multilevel analysis of teacher characteristics and student achievement in low performing schools. Educ. Assess. Eval. Acc., 21: 209-234. p g p 8. National Population Census (NPC) Nigeria. (2006) Population Distribution by Sex, State, LGAs and Senatorial Districts: 2006 Census Priority Tables. Vol. 3. Available from: http://www.population.gov.ng/index.php/oyo-state. Accessed on 12-05-2016. 23. Uzochukwu, B.S. and Onwujekwe, O.E. (2004) Socio- economic differences and health-seeking behaviour for the diagnosis and treatment of malaria: A case study of four local government areas operating the Bamako initiative programme in south-east Nigeria. Int. J. Equity Health, 3: 6. 9. District Health Information System. (2016) Oyo State ** 9. District Health Information System. (2016) Oyo State ****** Veterinary World, EISSN: 2231-0916 1063
https://openalex.org/W2886175104	https://environmentalevidencejournal.biomedcentral.com/track/pdf/10.1186/s13750-018-0131-5	English	null	What is the effect of prescribed burning in temperate and boreal forest on biodiversity, beyond pyrophilous and saproxylic species? A systematic review	Environmental evidence	2,018	cc-by	21,110	Eales et al. Environ Evid (2018) 7:19 https://doi.org/10.1186/s13750-018-0131-5 Eales et al. Environ Evid (2018) 7:19 https://doi.org/10.1186/s13750-018-0131-5 Environmental Evidence Open Access Abstract Background: While the effects of prescribed burning on tree regeneration and on pyrophilous and/or saproxylic species are relatively well known, effects on other organisms are less clear. The primary aim of this systematic review was to clarify how biodiversity is affected by prescribed burning in temperate and boreal forests, and whether burn- ing may be useful as a means of conserving or restoring biodiversity, beyond that of pyrophilous and saproxylic species. Methods: The review examined primary field studies of the effects of prescribed burning on biodiversity in boreal and temperate forests in protected areas or under commercial management. Non-intervention or alternate levels of intervention were comparators. Relevant outcomes were species richness and diversity, excluding that of pyrophilous and saproxylic species. Relevant studies were extracted from a recent systematic map of the evidence on biodiversity impacts of active management in forests set aside for conservation or restoration. Additional searches and a search update were undertaken using a strategy targeted to identify studies focused on prescribed burning interven- tions. Grey literature and bibliographies of relevant published reviews were also searched for evidence. Studies were assessed for internal and external validity and data were extracted, using validity assessment and data extraction tools specifically designed for this review. Studies were presented in a narrative synthesis and interactive map, and those which were suitable were quantitatively synthesised using meta-analyses, subgroup analysis and meta-regression. Results: Searches generated a total of 12,971 unique records. After screening for relevance, 244 studies (from 235 articles) were included in this review. Most studied forests were located in the USA (172/244), with the rest located in Canada, Europe and Australia. Eighty-two studies reporting 219 comparisons were included in the quantitative synthesis. Within the meta-analyses for each group of taxa, we identified a small to moderate volume of evidence, and heterogeneity was ubiquitous. Prescribed burning had significant positive effects on vascular plant richness, non- native vascular plant richness, and in broadleaf forests, herbaceous plant richness. Time since the burn, forest type and climate zone were significant moderators predicting the effect of burning on herbaceous plant richness. No other significant relationships were identified. © The Author(s) 2018. What is the effect of prescribed burning in temperate and boreal forest on biodiversity, beyond pyrophilous and saproxylic species? A systematic review Jacqualyn Eales1,2†, Neal R. Haddaway1,3† , Claes Bernes1, Steven J. Cooke4, Bengt Gunnar Jari Kouki6, Gillian Petrokofsky7 and Jessica J. Taylor4 Jacqualyn Eales1,2†, Neal R. Haddaway1,3† , Claes Bernes1, Steven J. Cooke4, Bengt Gunnar Jonsson5, Jari Kouki6, Gillian Petrokofsky7 and Jessica J. Taylor4 © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abstract This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Page 2 of 33 Eales et al. Environ Evid (2018) 7:19 Conclusions: Knowledge gaps exist for studies outside North America, in mixed forests and for non-plant organism outcomes. We identify a need to apply study designs consistently and appropriately, minimising the impact of con- founding factors wherever possible, and to provide extensive detail in study reports. We recommend that researchers build long-term datasets charting the impacts of prescribed burning on succession. The lack of consistent findings was likely due to high inter-study heterogeneity, and low numbers of comparable studies in each quantitative synthe- sis. We found no consistent effects of moderators, and were unable to test the effect of many potential moderators, due to a lack of reporting. Rather than making any general recommendations on the use of prescribed burning for biodiversity restoration, we provide an evidence atlas of previous studies for researchers and practitioners to use. We observe that outcomes are still difficult to predict, and any restoration project should include a component of moni- toring to build a stronger evidence base for recommendations and guidelines on how to best achieve conservation targets. Prescribed burning may have harmful effects on taxa that are conservation-dependent and careful planning is needed. Keywords: Fire regime, Disturbance, Forest conservation, Controlled burn, Forest set-aside, Forest reserve, Habitat management Prescribed burn Keywords: Fire regime, Disturbance, Forest conservation, Controlled burn, Forest set-aside, Forest reserve, Habitat management, Prescribed burn and livelihoods. Such practices, which began at least 100 years ago in the United States [12], have been increas- ingly common due to the desire to minimise catastrophic fire events [13]. Fire suppression can halt fires alto- gether, leading to a lack of specific habitats or resources for those species that are associated with fires and other natural disturbances [14]. This anthropogenic fire sup- pression has been shown to affect native forest biodiver- sity negatively [15], notably for pyrophilous (fire-loving) species and several saproxylic species (those dependent on dead wood) [16]. Abstract Furthermore, fire suppression has the potential to change many aspects of forest structure, disturbance dynamics, and succession, with equally clear consequences for forest-dwelling biota. In particular, northern Europe has seen drastic reductions in the extent and severity of forest fires [17, 18]. There has been debate in the literature regarding whether fire suppression has contributed to the accumulation of dense woody vegeta- tion which could have implications for biodiversity and lead to increased fire risk, areas burned and fire inten- sity (debate summarised in [19]). This debate extends to peatlands [20]. Active, policy-driven fire suppression since the late nineteenth century, particularly in managed areas, and changed landscape structure are likely key fac- tors behind changes in fire regimes [21]. Background Environ Evid (2018) 7:19 Page 3 of 33 on native biota [22]. In North America, recognition of the ecological and hazard reduction benefits has been slow, particularly when fire has been publicly viewed as incom- patible with timber production [16]. Thus, the extent and purpose of prescribed burning varies in this region. As acceptance of prescribed burning grows, there is inter- est in investigating how the amount and distribution of fuel will impact forest structural complexity and the biota associated with this complexity, following fires [22]. Pre- scribed burning for wildlife in southern Europe is far less developed than in other areas of the world, and the envi- ronmental implications remain poorly understood [26]. Across all boreo-temperate regions, it is clear that where prescribed burning is undertaken, it requires engagement with local and regional communities, since the practice typically involves potentially contentious trade-offs [21]. on native biota [22]. In North America, recognition of the ecological and hazard reduction benefits has been slow, particularly when fire has been publicly viewed as incom- patible with timber production [16]. Thus, the extent and purpose of prescribed burning varies in this region. As acceptance of prescribed burning grows, there is inter- est in investigating how the amount and distribution of fuel will impact forest structural complexity and the biota associated with this complexity, following fires [22]. Pre- scribed burning for wildlife in southern Europe is far less developed than in other areas of the world, and the envi- ronmental implications remain poorly understood [26]. Across all boreo-temperate regions, it is clear that where prescribed burning is undertaken, it requires engagement with local and regional communities, since the practice typically involves potentially contentious trade-offs [21]. A third suggested review topic was the effects of pre- scribed burning on the diversity of species other than those directly dependent on fire and dead wood. The direct impacts of fire on tree regeneration, pyrophilous and saproxylic species have been well studied, and one of the systematic reviews in progress is investigating the effect of dead-wood manipulation (e.g. through burning) on biodiversity in forests [32]. Furthermore, one recent systematic review investigated the impact of restora- tion burning on tree regeneration in boreal forests [34]. However, the systematic review described herein focuses on the effects of prescribed burning on other aspects of biodiversity. Background It would be valuable to broaden knowledge of how pre- scribed burning affects forest biodiversity, particularly because such effects could be viewed as either negative or positive. Additionally, the practice of prescribed burn- ing is now fairly common in temperate and boreal for- ests worldwide, further indicating the need for thorough investigation of its impacts on species other than those that can be considered as pyrophilous or saproxylic. For example, the Life + Taiga project is a 5-year European Union funded programme (2015–2019) ongoing in Swe- den [35]. The project involves 14 regional County Admin- istrative Boards and aims to perform 120 controlled fires in boreal forests, with the aim of conserving and restor- ing biodiversity. f Forest burning can impact organisms and habitats directly and/or indirectly via beneficial effects on pyroph- ilous or saproxylic species. In general, the direct effects appear to be clear and quick, with overall positive effects on forest biodiversity [27–29]. The immediate effects of fire on pyrophilous and saproxylic species, and also tree regeneration, are well documented [22]. However, the impact of prescribed burning on other components of biodiversity are less clear and/or consistent. The rela- tive importance of the frequency, extent, and intensity of burns on restoration success also remains undetermined. A total of 227 studies in the systematic map of manage- ment interventions in temperate or boreal forests [30] described effects of prescribed burning. Additional stud- ies in the topic area have become available more recently, since the last search for evidence was undertaken by the map authors in 2015. The current literature lacks an up-to-date systematic review assessing the full evidence base on the impact of prescribed burning on biodiversity of temperate and boreal forests worldwide. This review addresses this need by exploring the often-ignored wider impacts of prescribed burning. Identification of review topic A systematic map published in 2015 identified studies on a variety of active management interventions that could be useful for conserving or restoring forest biodiversity in boreal and temperate regions [30]. A total of 812 stud- ies describing a variety of interventions were identified as relevant to the map. Since the map was based on evidence relevant to the Swedish environment, it focused on forest types that are represented in Sweden (i.e. boreal and tem- perate), but such forests exist in many parts of the world (e.g., Russia, northern North America, southern parts of Australia). In accordance with accepted systematic map- ping guidance [31], the map gives an overview of the evidence base by providing a database with descriptions of relevant studies, but it does not synthesise reported results. Background In boreal and temperate regions, the biodiversity of for- ests set aside from forestry practice is often considered best preserved by non-intervention [1]. However, in many protected forests, remaining biodiversity values are legacies of past disturbances, e.g. recurring fires, grazing, or small-scale felling [2]. These forests may require active management to enhance or maintain the biodiversity characteristics that were the reason for protecting them [1, 3]. Such management can be particularly relevant where the aim is to restore lost ecological values, such as to restore particular seral stages or vegetation mosaics, upon which certain taxa depend [4].ii Naturally occurring fires (wildfires) are considered to be an essential part of boreo-temperate forest distur- bance dynamics [5]. It is well documented that in some regions wildfires have always occurred and have long- term patterns (fire regimes), probably related to large- scale and long-term climate and vegetation changes [6–8]. It is also recognised that humans have, for thou- sands of years, managed or altered ecosystems with fire, for example, the Maori colonization of the southern island of New Zealand around 700–800 years ago was characterized by widespread destruction of tropical for- ests by burning [9]. In general, fires modify the structure of a forest in a way that many forest-dwelling species find beneficial and are specifically adapted to [10]. Historical fire regimes are challenging to characterise but are clearly variable in their frequency, extent, and intensity [11]. This inherent variability is likely to have important conse- quences for forest biodiversity, but it also makes it highly challenging to explore the ecological consequences in a systematic and detailed way. i Prescribed burning, also known as controlled burning or planned burning is currently used in some protected areas as an active management tool to enhance and main- tain habitats for biodiversity outcomes in boreo-temper- ate forests [22, 23]. Prescribed burning is also commonly used for the purpose of mitigating wildfire risk by man- aging the accumulation of fuel in forests when and where necessary. Historically, this has been the primary purpose in Australia, where the practice is widely applied [24, 25]. In this region, there is also recognition by management authorities that planned burns can have positive effects Fire suppression is a management practice to mini- mise the negative impacts of wildfires, particularly on commercially managed forests, and on human lives Eales et al. Stakeholder engagement We established the scope and focus of the review in close cooperation with stakeholders, following the outputs provided by the systematic map [30]. The stakeholders were based primarily in Sweden and included research- ers (e.g. academic researchers from the University of Umeå), practitioners and managers, forestry companies (e.g. Bergvik Skog), local and governmental administra- tion boards (e.g. the Swedish Environmental Protection Agency), and global conservation charities (e.g. World Wildlife Fund). Before submission, peer review, and final publication of the protocol, a draft version was open for public review at the website of the Mistra Council for Evidence-Based Environmental Management (Mistra The map identified four potential subtopic areas that were sufficiently covered by existing studies to be included in a full systematic review. The selection of top- ics was also based on their significance for managers of forest reserves and other stakeholders, and on their rel- evance to Swedish forests. Two of the suggested system- atic reviews are currently in progress (the impact of dead wood on biodiversity [32]; the impacts of grazing on bio- diversity [33]). Page 4 of 33 Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 Page 4 of 33 EviEM) in July 2016. The draft was also sent directly to stakeholders. The draft protocol was revised in response to appropriate comments. Objective of the systematic reviewh The primary aim of this systematic review was to clarify if, and how, the diversity and richness of non-pyrophilous and non-saproxylic species in boreal and temperate for- ests is affected by prescribed burning. We searched not only for studies of interventions in actual forest reserves and other kinds of set-asides, but also for appropriate evi- dence from non-protected and commercially managed forests, since some of the practices applied in commercial forestry may be relevant to conservation or restoration. Quantitative synthesis of selected studies and a narrative synthesis were used to fulfil this aim.h Search stringh The search string for the additional literature searches was based on a subset of the search terms used for the systematic map [30], focusing on terms related to pre- scribed burning. We conducted a scoping exercise in May 2016 to assess alternative search terms, testing them against a set of articles suggested by review team members and known to be relevant. Searches were undertaken in July 2016. Details of the scoping exercise and search string development are provided in the pro- tocol for this review [36]. i The secondary aim of this systematic review was to provide an overview of available evidence on how biodi- versity of boreal and temperate forests (apart from that of pyrophilous and saproxylic species) is affected by pre- scribed burning. A systematic map of the evidence base was used to provide this overview. The ultimate purpose of the review was to investigate whether prescribed burning may be used as a means of conserving or restoring biodiversity in forest set-asides, and if so, what conditions increase its effectiveness. During article screening a small number of additional synonyms were added to the search string and used in a set of supplementary searches in December 2016. The additional population terms were “stand”, “plantation”, “wood”, “tree”, “clone”, “tract” and “savanna”. The additional intervention terms were “prescri”, “intro- duce” and “broadcast”. The additional outcome term was “richness”. The search string was adapted to spe- cific databases using appropriate syntax. Details of the July 2016 and December 2016 strings are given in Addi- tional file 1 together with search dates and the number of articles found. The search string is summarised in Table 1. This string differs from that presented in the protocol due to the supplementary searches conducted in December 2016. Searches for literature A subset of the evidence base examined in this system- atic review was identified by a systematic map of man- agement interventions in temperate or boreal forests [30]. Searches for the map were performed in May– August 2014, with an update in March 2015. Of the 812 studies included in the map, 227 reported on impacts of prescribed burning and were therefore potentially relevant to this review. However, we also conducted additional searches for evidence, both to find recently published literature and because the searches for the systematic map were focused on forest types occurring in Sweden, whilst we aimed to be more inclusive in this review. Primary questionf What is the effect of prescribed burning in temperate and boreal forest on biodiversity, not including pyrophilous and saproxylic species? Components of the question Population: boreal and temperate forests Intervention: prescribed burning. Intervention: prescribed burning. Comparator: no burning or alternative levels of burn- ing, before burning. Outcomes: diversity and richness of species (excluding pyrophilous and saproxylic species) as one of a number of measures of biodiversity reported in the literature. Search engines An internet search was performed using Google Scholar (scholar.google.com) and a subset of the search terms described above (see Additional file 1 for details). Search results were extracted using the software Publish or Per- ish [38] (up to 1000 results viewable and extractable). Duplicates within sets of search results were removed within EndNote. Citations were then uploaded to the review management software EPPI Reviewer (eppi.ioe. ac.uk/eppireviewer4) and screened together with biblio- graphic database search results. i 9. International Union for Conservation of Nature (www.iucn.org).i i 9. International Union for Conservation of Nature (www.iucn.org).i 10. Metsähallitus (www.metsa.fi). 10. Metsähallitus (www.metsa.fi). i 11. Natural Resources Canada (www.nrcan.gc.ca).h 12. The Nebraska Prescribed Fire conference (out- doornebraska.gov/prescribedfire). i 13. New Zealand Ministry for the Environment (www. mfe.govt.nz). i 13. New Zealand Ministry for the Environment (www. mfe.govt.nz). 14. Nordic Council of Ministers (www.norden.org). 14. Nordic Council of Ministers (www.norden.org). 15. Norwegian Environment Agency (www.miljødirek- toratet.no). 15. Norwegian Environment Agency (www.miljødirek- toratet.no). Bibliographic databases ymparisto.fi). 8. Finland’s environmental administration (www. ymparisto.fi). 1. Ancient Tree Forum (www.ancient-tree-forum.org. uk). Bibliographic databases This review follows the methods outlined in an a priori protocol [36]. It has been conducted according to CEE’s Guidelines for Systematic Reviews [37]. Due to the large volume of evidence identified that was not suitable for quantitative synthesis we deviate from the protocol in that we added an extra first step before full synthesis: we initially produced a detailed systematic map data- base describing all studies, followed by a quantitative synthesis of all studies that provided sufficient data for meta-analysis. Searches were conducted in the following online biblio- graphic databases: 1. Web of Science Core Collections (Stockholm Univer- sity Library subscription). y y 2. Scopus (Stockholm University Library subscription). 3. CAB abstracts (Oxford University library subscrip- tion). Page 5 of 33 Eales et al. Environ Evid (2018) 7:19 Table 1 The search string to which the combined database searches are equivalent Search string Population terms (forest OR woodland* OR “wood* pasture” OR “wood meadow” OR stand OR plantation* OR wood* OR tree* OR clone* OR tract* OR savanna) AND Intervention terms ((prescribed OR control OR experiment* OR prescri* OR introduce* OR broadcast) AND (burn* OR fire)) AND Outcome terms (diversity OR (species AND (richness OR focal OR target OR keystone OR umbrella OR red-list OR threatened OR endangered OR rare)) OR “species density” OR “number of species” OR indicator* OR abundance OR “forest structure” OR habitat* OR richness) Table 1 The search string to which the combined database searches are equivalent (forest* OR woodland* OR “wood* pasture” OR “wood meadow” OR stand OR plantation* OR wood* OR tree* OR clone* OR tract* OR savanna) ((prescribed OR control OR experiment* OR prescri* OR introduce* OR broadcast) AND (burn* OR fire)) AND (diversity OR (species AND (richness OR focal OR target OR keystone OR umbrella OR red-list OR threatened OR endangered OR rare)) OR “species density” OR “number of species” OR indicator* OR abundance OR “forest structure” OR habitat* OR richness) Searches were made using topic words or title, abstract and keywords. No subject category limitations were used. No language or document type restrictions were applied, but searches were performed using Eng- lish search terms only. 5. European Commission Joint Research Centre (ec. europa.eu/dgs/jrc). 6. European Environment Agency (www.eea.europ a.eu). 6. European Environment Agency (www.eea.europ a.eu). 7. Food and Agriculture Organization of the United Nations (www.fao.org). 7. Food and Agriculture Organization of the United Nations (www.fao.org). 8. Finland’s environmental administration (www. Supplementary searches (377/3764) of the articles retrieved by the July 2016 search were screened for relevance at title and abstract by both reviewers, prior to screening of the full set of results. Reviewers agreed on 80% of decisions. All disa- greements were discussed in detail and inclusion criteria were annotated and further clarified verbally before the title and abstract screening continued. A third reviewer (NH) was brought into discuss borderline studies. During screening of evidence, we identified a number of relevant literature reviews that did not contain primary data for inclusion in the review. We searched for evi- dence in the bibliographies of these reviews to identify potentially relevant studies that had been missed by other targeted searches. We recognise that data and studies from commercially valuable forests held by private companies is a source of potentially relevant evidence. However, we did not make efforts to include this evidence in our review since access is likely to be difficult and unevenly distributed [39]. Moreover, such an approach is unlikely to be repeatable or comprehensive, due to differences between compa- nies in allowing third-party access to data. To establish a rough estimate of the amount of data missed, BGJ contacted two major forest companies in Sweden and was informed that although they do undertake regular prescribed burning, no structured data on the effects is collected. Following title and abstract screening, attempts to retrieve full texts were made. Additional file 2 contains a list of 56 articles (10% of all articles potentially relevant at title and abstract level), that were not found in full text. Each obtained full text was screened by one reviewer following consistency checking, where a random sample of 10% (51/534) of the full texts retrieved were assessed by both reviewers at full text. This consistency checking showed a relatively high consistency rate of 74%. Follow- ing detailed discussion of all agreements it was ascer- tained that one reviewer was overly conservative in their inclusions. Discussions of these discrepancies between reviewers resulted in additional specifications of how the inclusion criteria were to be interpreted. Some doubt- ful cases, where the two reviewers could not include or exclude an article with certainty even after having read the full text, were discussed and decided on by the entire review team (all authors). Following removal of these non-relevant articles the consistency rate increased to > 90% (50/51 agreements). Specialist websitesh 16. Norwegian Forest and Landscape Institute (www. skogoglandskap.no). 16. Norwegian Forest and Landscape Institute (www. skogoglandskap.no). The websites of 28 specialist organisations (listed below) were searched for relevant evidence. These websites were searched using both the built-in search facilities where available and by hand searching for research studies. The search terms used were based on the search string described in Table 1, adjusted for the searching capa- bilities of each website. The search terms used across all websites are listed in Additional file 1. All potentially rel- evant evidence was recorded. Searches were performed in Danish, English, Finnish, French, Norwegian, and Swedish according to the language of the website (see Additional file 1). 17. Norwegian Institute for Nature Research (www. nina.no). 17. Norwegian Institute for Nature Research (www. nina.no). 18. Parks Canada (www.pc.gc.ca). 18. Parks Canada (www.pc.gc.ca). 19. Society for Ecological Restoration (www.ser.org). 20. Swedish County Administrative Boards (www.lanss tyrelsen.se). 21. Swedish Environmental Protection Agency (www. naturvardsverket.se). 21. Swedish Environmental Protection Agency (www. naturvardsverket.se). 22. Swedish Forest Agency (www.skogsstyrelsen.se). 22. Swedish Forest Agency (www.skogsstyrelsen.se). 23. Swedish University of Agricultural Sciences (www. slu.se). 23. Swedish University of Agricultural Sciences (www. slu.se). 24. UK Environment Agency (www.environmen t-agency.gov.uk). 1. Ancient Tree Forum (www.ancient-tree-forum.org. uk). 25. United Nations Environment Programme (www. unep.org). 2. Australian Department of Environment and Energy (www.australia.gov.au/directories/Australia/envir onment). 26. United States Environmental Protection Agency (www.epa.gov). 3. Bureau of Land Management, US Dept. of the Inte- rior (www.blm.gov). 27. United States National Parks Service (www.nps. gov). 4. Environment Canada (www.ec.gc.ca). 28. US Forest Service (www.fs.fed.us). Eales et al. Environ Evid (2018) 7:19 Page 6 of 33 Page 6 of 33 Screening of literatureh The evidence was screened for relevance within EPPI Reviewer. Search results from the bibliographic data- bases and search engines were added to the software. Prior to screening, duplicates were removed using the “fuzzy matching” function followed by additional manual removal (by JE and JT). Articles found using specialist websites (searches undertaken by JT and JK) or bibliographies of reviews (searches undertaken by JE), and those supplied by mem- bers of the review team (JK) were also entered at this stage in the screening process. A list of all articles excluded from the systematic review on the basis of full-text assessment is provided in Addi- tional file 3 together with the reasons for exclusion. Screening process Search results were evaluated for inclusion at two succes- sive levels; title and abstract, and full text. This represents a change from the protocol, where we planned to assess titles and abstracts separately in two successive stages. This change reflected a decision that it was more efficient to screen titles and abstracts in EPPI Reviewer together. Sets of search results were allocated to reviewers (JE and JT) randomly. At no stage was a reviewer responsi- ble for screening an article of which they were an author. In cases of uncertainty about inclusion decisions (for example where information was missing or unclear), the reviewer erred on the side of caution, choosing inclusion rather than exclusion. Estimating comprehensiveness of the search Since our review followed the same basic search strat- egy and used a very similar search string to the original systematic map published by Bernes et al. [30], we have not repeated tests of the comprehensiveness of the search that were originally performed therein. Supplementary searches Of the remaining full texts, 50% were dual screened and discussed prior to the final set of 50% being screened by one reviewer (JE). Study inclusion criteria Every study had to pass each of the following criteria in order to be included, either by providing all the required data itself or by referring to other articles where neces- sary information was presented. Relevant populations Forests in the boreal or temperate vegetation zones. Any habitat with a tree layer (canopy cover at least 10% and canopy height capable of reaching at least 5 m) was regarded as forest [40]. As an approxi- mation of the boreal and temperate vegetation zones we used the cold Köppen–Geiger climate zones (the D zones) and a subset of the temperate zones (Cfb, Cfc and Csb), as defined by Peel et al. [41], shown in Fig. 1. Forest stands Articles were assessed by a single reviewer (JE or JT). As a check of consistency, a random sample of 10% Eales et al. Environ Evid (2018) 7:19 Page 7 of 33 Fig. 1 Köppen–Geiger climate zones relevant to this systematic review. Relevant zones include all of the cold climate types (D) and some of the temperate ones (C) (The map is modified from Peel et al. [41]) Fig. 1 Köppen–Geiger climate zones relevant to this systematic review. Relevant zones include all of the cold climate types (D) and some of the temperate ones (C) (The map is modified from Peel et al. [41]) dominated by ponderosa pine (Pinus ponderosa) were considered relevant even if located outside the climate zones mentioned above. These forests constitute a well- studied North American habitat type that shares several characteristics with the pine forests in boreal and temper- ate regions. Studies of the South African Fynbos region were excluded due to the ecosystem being a shrubland system that generally does not fulfil the tree-layer crite- ria. Studies of stands where authors reported that 75% or more of the basal area or timber volume had been har- vested or naturally lost were also excluded. Relevant types of outcome Diversity (e.g. Shannon and Simpson’s index of diversity) and richness of plants, animals, lichen, and fungi, except pyrophilous and sap- roxylic species. Studies of cavity-nesting birds and tree- roosting bats were included, as these species are not fully dependent on dead wood or fire. Data extraction strategy Extraction of meta‑data Meta-data (descriptive information regarding the study context and methods) were extracted for all studies in the review and used to populate a systematic map data- base of relevant research relating to the impacts of pre- scribed burning on biodiversity. Additional file 5 displays a schema of the meta-data extracted from all studies. Meta-data relating to study location were extracted from the included articles where possible, but if no geographi- cal coordinates were given, we recorded approximate coordinates based on reported site names, maps or tex- tual descriptions of study locations (or coordinates pro- vided in another article describing the same site). Where coordinates given by study authors were clearly incorrect, we recorded coordinates based on other information provided by the study (e.g. distance from a named place or point of interest). y Critical appraisal of study validity was conducted on all quantitatively synthesised studies to ensure that: (1) all data used in meta-analyses was of sufficient quality to be reliable and generalisable across the evidence base; and (2) studies that were of the highest reliability could be identified to examine possible influences of bias on the results of meta-analyses (via sensitivity analysis, see below). The criteria used for study validity assessment are presented in Table 2. These criteria reflect what the review team deemed to be critical variables influencing the reliability of study findings. They relate to both inter- nal validity (methodological quality) and external valid- ity (generalisability), and include: efforts by study authors to measure and control for baseline differences before intervention; the level of replication and representative- ness of samples; allocation of samples and matching of control and intervention sites; the presence of severe confounders; appropriateness and suitability of the appli- cation of the intervention; and, the suitability of the out- come measurement methods. For each of these domains, studies were categorised as to how well they fulfilled the criteria: yes, partly, no, or unclear. Based on these cate- gories for individual domains in Table 2, each study was then given an overall rating of high, medium, medium (unclear), or low validity, using the procedure presented in Table 3. The category of medium (unclear) was given to studies that were assigned “unclear” and not “partly” for one or more domains and “yes” for all other domains, as detailed in Table 3. Critical appraisal of study validity Studies categorised as being of low validity were excluded from meta-analyses. A list of these studies is provided in Additional file 4 together with the reasons for exclusion. Since the focus of this review is a combination of sys- tematic mapping and quantitative synthesis, and since available resources were limited, only studies eligible for meta-analyses were subject to study validity assessment (see “Eligibility for meta-analysis” below). This deviates from the protocol, which stated that all studies would be critically appraised. Study inclusion criteria Environ Evid (2018) 7:19 ognise that this is a topic of interest that may warrant a separate evidence synthesis. our review. Thus, we treat studies without the highest reporting quality in the same way as we do those with- out the highest methodological quality or generalisability. These studies are clearly separated in all reporting within this review. Relevant type of study Primary field studies (obser- vational or manipulative). Based on this criterion, we excluded simulation studies, reviews, commentaries and policy discussions. Where necessary, detailed reasoning concerning valid- ity assessment was recorded alongside the categorisa- tions. Each study undergoing validity assessment was appraised by two reviewers. Cases where reviewers (JE and JT) disagreed were discussed, with a third reviewer (NH) involved in the discussions for cases which were borderline. In no case was a reviewer responsible for crit- ically appraising a study of which they were an author. Language Full text written in English, French, Swedish or Finnish. This selection reflects the language capabilities of the review team and their respective institutions, from which assistance could be provided. Study inclusion criteria Studies which reported a representative list of species in the study area based on standard survey methods suitable for the taxa of study were included in the review, and the outcome was used as a measure of species richness, even if authors did not provide a total of the number of species listed or refer to species richness explicitly. Diversity or richness that was transformed or corrected, for example using jackknife estimates, was also regarded as relevant. In addition to diversity and richness, our review protocol listed abun- dance of communities or species as a relevant outcome [36], but we decided to focus the review on the former outcomes, since these are more direct measures of bio- diversity [42, 43]. The protocol also listed community composition as a relevant outcome, but this was rarely reported in the studies we encountered, and the review team decided to focus on the most commonly reported biodiversity measures. The following specific outcomes were not considered eligible since they are measures of beta diversity: Jaccard’s diversity index (a measure of species turnover rather than diversity); similarity indi- ces, such as Sorensen’s similarity index (not a meas- ure of diversity). Seed bank diversity and richness were excluded because the seed bank represents a source of colonisation, rather than an established plant commu- nity, the latter being the focus of our review. Although we have chosen not to review seed bank diversity, we rec- Relevant types of intervention Prescribed burning. Studies of intentional burning in the field were included, except where the primary purpose of burning was to con- trol invasive species, because the characteristics of such burnings (extent, duration, intensity) are likely to be fun- damentally different from other burns (typically for res- toration or fuel reduction). Studies on wildfires were not included even if relevant control sites were available. Relevant type of comparator Non-intervention or alter- native levels of intervention. Both temporal and spatial comparisons of how prescribed burning affects biodiver- sity were considered to be relevant. This means that we included both ‘BA’ (before/after) studies, i.e. comparisons of the same site prior to and following an intervention, and ‘CI’ (control/impact) studies, i.e. comparisons of treated and untreated sites (or sites that had been subject to dif- ferent kinds of treatment). Studies combining these types of comparison, i.e. those with a ‘BACI’ (before/after/con- trol/impact) design, were also included. Page 8 of 33 Eales et al. Data extraction strategy Extraction of meta‑data This does not relate to study valid- ity directly (unclear studies are not necessarily less valid), but we believe it is dangerous to assume that information that is missing would otherwise relate to high validity in We recorded the number of independent burn/control areas and the number of replicate samples within burn/ control areas. Spatial replication was recorded as the number of samples measured within each independent burn unit (intervention or comparator site). If treated sites and controls were not replicated to the same extent, we recorded each number separately. If the number of replicates within independent burn or control areas var- ied, we recorded the range in the number of replicate samples. In cases where some of the data reported by a study fell outside the scope of our review (e.g. where some of the study sites were located outside relevant vegetation zones), we recorded information only for those parts of the study that fulfilled our inclusion criteria. i The meta-data coding was undertaken by JT and JE. A consistency check was undertaken on 8% (20/244) of Page 9 of 33 Eales et al. Environ Evid (2018) 7:19 Table 2 Study validity assessment criteria Reviewers answered the questions in the left column with ‘Yes’, ‘Partly’, ‘No’ and ‘Unclear’ based on the specifications in the table. The answer ‘n/a’ was used if the criterion was not applicable in a particular instance. Reviewers could also provide comments on each study regarding its external validity Question/criterion Yes Partly No Unclear Did the study have a temporal and/or spatial control? BACI study BA study or CI study N/a (not eligible according to inclu- sion criteria) Lacking sufficient information to judge Degree of replication appropriate and representative? (to outcome measure) Replicated burns, OR a single large burn area, with samples considered to be representative and spatially independent Single burn with limited sampling within burn A single/small burn area with very limited sampling within burn Lacking sufficient information to judge Does treatment allocation account for spatial heterogeneity? (e.g. blocked or randomised) and/or intervention and comparator sites well-matched i.e. similar at baseline (e.g. aspect, soil type, forest type) Treatment allocation (multiple burn) or replication (within a single burn) accounts for spatial heterogeneity i.e. appropriately randomised, ran- domised and blocked or stratified OR stated/clear attempt at match- ing OR highly likely to be similar Spatial heterogeneity not fully accounted for i.e. Table 3 Overall assessment of study validity/risk of bias no “Partly” in any field) it was classed as “Medium (unclear)” If none of the above factors applied, the study was considered to have High validity y y g ( y yi ) ( ) If none of the above factors applied, the study was considered to have High validity of the above factors applied, the study was considered to have High ing). These studies were of limited value because they could not be compared with other studies in a quantitative analysis. the studies, with subsequent discussion to maximise the consistency of coding between reviewers. Meta-data on these studies were extracted by both reviewers. Discrep- ancies were discussed, and the meta-data recording sheet refined to improve clarity before the rest of the meta-data coding was undertaken. • • Multiple interventions were applied concurrently in comparison with no intervention, e.g. thinning and burning compared with no intervention. • • Additional interventions (such as thinning or manipulation of grazing) had been carried out across the study areas (in both burned and unburned plots). Eligibility for meta‑analysis Studies were considered unsuitable for meta-analysis (and no outcome data were extracted from them) if any of the following applied: Two studies reported natural levels of grazing in both burned and unburned plots and were included in the meta-analysis. Some other studies in our review may have included study plots subject to grazing, despite not explicitly reporting it. In such cases, it was assumed that any such grazing was likely to represent natural levels. Studies in which all sites were subject to non- natural/domestic/high grazing were not included in the meta-analysis. • • The study provided quantitative data that were already provided in another relevant article (in cases of such redundant data, studies providing more information were selected for further synthesis, but missing information was filled in from linked stud- ies). • • Measures of outcome variability and/or data on sam- ple sizes were not available (and not possible to cal- culate from raw data)—effect sizes could not be cal- culated. Data extraction strategy Extraction of meta‑data partial randomi- sation/blocking/stratification OR Moderately matched Purposive treatment allocation that clearly does not account for spatial heterogeneity AND/OR Do not match Lacking sufficient information to judge No severely confounding factors present (factors likely to confound the effect of the intervention on outcome)? apart from those present at baseline (e.g. other interventions) Confounding factors likely to be mini- mal, were minimised or lacking Some confounders present, likely to have moderate impact on outcome Subject to confounders with major impact on outcome (such that outcomes are not clearly the effect of the intervention) Lacking sufficient information to judge Intervention was likely appropriately and realistically applied? Typical of a prescribed burn in terms of temperature, area, low intensity/ severity of burn, or described as a prescribed burn and likely to be a typical prescribed burn Partly typical of a prescribed burn in one (or more) aspect/s but not in others Not typical of a prescribed burn (e.g. inappropriate temperature, very high intensity/severity) Lacking any information/description of burn Outcome measure method was appropriate? A reliable or standard measurement method for the outcome, and at a seasonal, temporal and spatial scale that is likely to capture any impact of the intervention on the outcome Outcome measure method was partly appropriate and reliable Outcome measure method was not appropriate or reliable (due to method choice and/or poor analysis of the data) Lacking sufficient information to judge No Page 10 of 33 Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 If a study was classed as Medium solely due to being “Unclear” (i.e. no “Partly” in any field) it was classed as “Medium (unclear)” If none of the above factors applied, the study was considered to have High validity Table 3 Overall assessment of study validity/risk of bias Studies were assigned Low validity if any of the following factors applied Studies were assigned Low validity if any of the following factors applied Any of these questions answered with “No” or “Unclear” • Did the study have a temporal and/or spatial control? • Degree of replication appropriate and representative? Any of these questions answered with “No” • Does treatment allocation account for spatial heterogeneity? and/ or Intervention and comparator sites well-matched • No severely confounding factors present? apart from those present at baseline • Intervention was likely appropriately and realistically applied? • Outcome measure method was appropriate? • Study methodology and results are generalisable to other prescribed burns in temperate or boreal forest l d d ( l ) l d f f h f ll f Any of these questions answered with “No” or “Unclear” • Did the study have a temporal and/or spatial control? • Degree of replication appropriate and representative? OR Any of these questions answered with “No” • Outcome measure method was appropriate? p Studies that were not assigned Low validity were considered to have Medium validity or Medium (unclear) validity if any of the following factors applied Studies that were not assigned Low validity were considered to have Medium validity or Medium (unclear) validity if any of the following factors applied ot assigned Low validity were considered to have Medium validity or Medium (unclear) validity if any of the following factors dies that were not assigned Low validity were considered to have Medium validity or Medium (unclear) validity if any of the fo pplied Any of these questions answered with “Partly” • Did the study have a temporal and/or spatial control? • Degree of replication appropriate and representative? (to outcome measure) OR Any of these questions answered with “Partly” or “Unclear”: • Does treatment allocation account for spatial heterogeneity? and/ or Intervention and comparator sites well-matched • Does treatment allocation account for spatial heterogeneity? and/ or Intervention and comparator sites well-matched • No severely confounding factors present? apart from those present at baseline p • No severely confounding factors present? apart from those present at baseline • Intervention was likely appropriately and realistically applied? • Outcome measure method was appropriate? • Study methodology and results are generalisable to other prescribed burns in temperate or boreal forest If a study was classed as Medium solely due to being “Unclear” (i.e. Potential effect modifiers and reasons for heterogeneity fi g y To the extent that data were available, the following potential effect modifiers were recorded for all studies included in the review: Where data were reported by authors as a range, for example a range of burn frequencies, we used the mid- point value of the range to represent the data. Where a study reported outcomes for multiple time points, we only extracted data from the final sampling, but we recorded cases where time series data were available. • • Geographical coordinates (latitude and longitude). • • Altitude. • • Altitude. • • Forest type (coniferous, broadleaf, or mixed). • • Dominant tree species. • • Forest stand age and origin. • • Forest stand age and origin. The burn season was reported in different ways across studies, and we therefore coded this variable as “dor- mant” (autumn/winter) or “growing” (spring/summer). For studies in the northern hemisphere, autumn/winter started from September and lasted 6 calendar months. For studies undertaken in the southern hemisphere, autumn/winter started from March. • • Fire history. • • Fire history. • • Burning frequency (either single or serial burning). • • Burn season. • • Other details regarding the burn (as described by authors). • • Other interventions at study sites (harvesting, thin- ning, understorey removal, grazing etc.) We recognise that the terms “saproxylic” and “pyroph- ilous” may be used differently by different authors, and whether an organism can be classed as one of the above is also likely to depend upon landscape or regional ele- ments. Where reported in studies included in our meta-analyses, the maximum percentage of saproxylic/ pyrophilous species within a studied community was approximately 25%. Since it was not reported whether these species groups were present in the surveyed com- munities for most comparisons in the quantitative syn- thesis (207/219), the review team decided to include the 12 comparisons that stated that they included saproxylic/ pyrophilous species as part of the surveyed community. As stated in the inclusion criteria, studies where only sap- roxylic/pyrophilous species were recorded were not eligi- ble for this systematic review. The following additional potential effect modi- fiers were recorded for all studies included in the meta-analyses: • • Climate zone. • • Forest disturbance history. • • Study duration. • • Number of burn events during the study. • • Burn frequency (number of burns per year across the study period). Extraction of quantitative data suitable for meta‑analysis For studies with medium or high validity and with out- comes considered suitable for meta-analysis (see “Data • • Effects of burning were compared with effects of alternative levels of burning (rather than no burn- Eales et al. Environ Evid (2018) 7:19 Page 11 of 33 Page 11 of 33 both JE and JT. All discrepancies were discussed, and the data extraction sheet was refined to improve clarity before the rest of the data extraction was undertaken (see Additional file 5). In a deviation from the protocol, extracted data were double-checked, but not always by a different reviewer, due to time constraints. synthesis and presentation”—“Eligibility for meta-analy- sis”) we undertook full data extraction (i.e. we extracted quantitative results and effect modifier data in addition to meta-data). We extracted data relating to compari- sons between burned and unburned sites only in order to focus on the impact of burning as a sole intervention. synthesis and presentation”—“Eligibility for meta-analy- sis”) we undertook full data extraction (i.e. we extracted quantitative results and effect modifier data in addition to meta-data). We extracted data relating to compari- sons between burned and unburned sites only in order to focus on the impact of burning as a sole intervention. Outcome means, measures of variability (standard deviation, standard error, confidence intervals, etc.), and sample sizes were extracted from text, tables and graphs, using image analysis software [44] where necessary. Data on interventions and other potential effect modi- fiers were extracted from the included articles. We also recorded, where reported, the reason for burning, i.e. burn intention. f If raw data (rather than means) were provided, we calculated and recorded summary statistics ourselves. Where data or information were missing or unclear we attempted to contact authors via email to retrieve the missing or unclear data. At no stage was a reviewer responsible for extracting information from a study of which they were an author. Some studies were unclear about the level of replica- tion used. Where possible for these studies, we extracted two measures of sample size: the total number of sub- samples and the number of true replicates (the number of replicates we deemed to represent independent samples). The systematic map database and narrative synthesis Standardised effect sizes were calculated for all outcomes using Hedges’ g statistic [45], i.e. the difference between the mean response to burning and the mean response to no burning, divided by the pooled standard deviation, and with an adjustment for small sample sizes: All relevant studies were included in a systematic map database of evidence relating to the impacts of prescribed burning on biodiversity in boreo-temperate forests. We also produced an evidence atlas, an interactive geograph- ical information system (GIS). The evidence atlas plots study locations on a world map, and data on the studies can be displayed by clicking on the symbols in the map. Both the evidence atlas and the database allow data to be filtered and sorted. The meta-data were used to col- late descriptive statistics and a narrative synthesis of the evidence. Hedges′ g = M1 −M2 SD∗ Pooled × N −3 N −2.25 × N −2 N , where M1 and M2 are the intervention and comparator mean values, respectively, SD∗ Pooled is the pooled standard deviation, and N is the sample size. Positive effect sizes thus indicate that the response parameter (species rich- ness or diversity) was higher in burned areas than in non- burned areas. In addition to the evidence atlas, the evidence base was summarised in a series of tables describing the nature of the study setting and methods, and the type of burning intervention employed.i Members of the review team independently identified key knowledge gaps (underrepresented subtopics that warrant further primary research) and knowledge clus- ters (well-represented subtopics that are amenable to synthesis via systematic review) by independently assess- ing the evidence in the review and discussing gaps and clusters as a team.fi Quantitative synthesis—data preparation y In preparation for meta-analyses, we made a number of initial conversions and transformations of data extracted from included studies. BACI outcomes were converted to CI by subtraction of data sampled before intervention from those sampled after intervention. Measures of vari- ability reported as standard errors or confidence intervals were converted to standard deviations. In cases where study authors had reported data according to taxonomic categories more specific than those used in our analyses, we combined different outcomes from the same plots (e.g. merging separate data on grasses and herbaceous plants to obtain data on understorey plants). In these cases, to maintain biological appropriateness, we com- bined richness data by summing, and combined diversity data by using the arithmetic mean (see Additional file 6: 2b, “Variability measure plan”). Potential effect modifiers and reasons for heterogeneity • • Burn intention (e.g. fuel reduction, habitat mainte- nance). • • Time between last burn and last outcome measure. • • Method of sampling (e.g. point count, litter sam- ples). A further check was undertaken by JT and JE on 9% (8/98) of the studies, with all decisions discussed in order to maximise the consistency of coding between reviewers. Data from these studies were extracted by • • Area of study units (sampling plots). • • Share of saproxylic and/or pyrophilous species in outcome measure (e.g. percentage). Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 Page 12 of 33 Page 12 of 33 Effect size calculationf y p The systematic map database and narrative synthesis Simpson’s index Wh th Where authors reported diversity as “Simpson’s D”, we converted it to “Simpson’s diversity index 1-D”. This was necessary because when using “Simpson’s D”, which ranges from 0 to 1, a positive effect size indicates lower diversity, which is the opposite direction to the other indices used in our meta-analysis, such as Shannon diver- sity. The definition of Simpson’s index used was generally poorly reported. Because Simpson’s can also be reported as a reciprocal, i.e. 1/D, wherever authors reported Simp- son’s index with a value greater than 1, we made the assumption that the authors used the reciprocal. Some studies possessed sufficient data for meta-anal- ysis but could not be meta-analysed because there were too few similar effect size estimates to allow meaningful quantitative synthesis (i.e. < 4 studies). Thus, the effect estimates and their variability for these studies and all other studies in the meta-analyses below were plotted visually using forest plots that combined all related out- come measures (e.g. all vegetation outcomes). Summary effect estimates were not plotted for these forest plots, since no actual meta-analysis was performed. We combined Shannon and Simpson indices from dif- ferent studies in the same meta-analyses, since these indices are standardised and we are comparing differ- ences between scale-free values. Although it would have been informative to determine the influence of the choice of diversity index on the effect size, the low number of studies prevented us from undertaking such a sensitivity analysis. Separation of studies h f For the purposes of this review, we defined a study as an experiment or observation that was undertaken over a specific time period at a particular site or set of sites. If multiple articles reported data for the same study site(s), they were given the same “Site ID” and were essentially considered as reports of the same study. If a single article reported data separately for different sites that we con- sidered to be ecologically independent, we assigned a separate Site ID to each site. For the rest of this report we refer to independent effect estimates used in meta-analy- ses as ‘comparisons’. Hence, one article and one location could be represented in multiple outcomes in the same Page 13 of 33 Eales et al. Environ Evid (2018) 7:19 meta-analysis. Similarly, one study could be represented y multiple comparisons across multiple meta-analyses f different outcomes. Adjustment accounting for pseudoreplication Where we were aware (based on information in publi- cations or from contact with authors) or had reason to assume that published outcomes were based on partly Fig. 2 Flow diagram showing the number of studies at each stage of the review Fig. 2 Flow diagram showing the number of studies at each stage of the review Fig. 2 Flow diagram showing the number of studies at each stage of the review Fig. 2 meta-analysis. Similarly, one study could be represented by multiple comparisons across multiple meta-analyses of different outcomes. meta-analysis. Similarly, one study could be represented by multiple comparisons across multiple meta-analyses of different outcomes. Adjustment accounting for pseudoreplication Adjustment accounting for pseudoreplication Where we were aware (based on information in publi- cations or from contact with authors) or had reason to assume that published outcomes were based on partly Eales et al. Environ Evid (2018) 7:19 Page 14 of 33 Page 14 of 33 Thirdly, we calculated and plotted Cook’s distance for each unmoderated model to identify highly influential studies or groups of studies. Finally, we calculated fail-safe numbers for meta-analyses showing significant summary effect estimates (fsn function within the metafor package in R [47]). The fail-safe number represents the number of studies with null effect necessary to change a model’s significance level to α (0.05) and shows how robust the results would be to additional studies. Separation of studies h f The script used to run models in R is provided in Additional file 7 and the data used in these models is provided in Additional file 8. subsampled data (i.e. averaged samples were not from independent replicates), we calculated effect sizes using a modified equation to avoid overestimation of effect sizes. First, standard errors were converted to standard devia- tions using total numbers of subsamples as sample sizes (so as to be conservative). Hedges’ g effect sizes (based on Equations 4.19 and 4.22 in Borenstein et al. [45]) were also calculated using the total number of subsamples, but each pooled standard error was calculated using both the number of true replicates and the total number of sub- samples as sample sizes. This method gives the most con- servative estimate of variability. Quantitative synthesis—meta‑analysis d ff l We ran random effects meta-analysis models in R [46] using the rma.mv function in the metafor package [47]. For each model, we declared Site ID (a unique code for each independent study site or set of sites) as a random factor to account for multiple outcomes being reported from the same location. We only performed meta-anal- ysis where more than three comparisons could be com- bined. We produced forest plots to visualise effect sizes from individual comparisons and summary effect esti- mates across groups of comparable studies. A total of 108 studies (from 106 articles) came from the systematic map that preceded this review [30]. The remaining 121 studies from the systematic map identified as relating to prescribed burning were not eligible for inclusion, primarily due to ineligible out- comes (n = 116), such as measures of abundance but not diversity or richness. The searches undertaken in July and December 2016 identified a further 117 stud- ies (from 113 articles); 81 studies (79 articles) from the July searches and 36 studies (34 articles) from Decem- ber searches. In review bibliographies we also found 19 relevant studies (from 18 articles) that had not been retrieved by our online searches. No relevant stud- ies were identified through searches of organisational websites. The number of articles excluded after full text screening is presented by exclusion reason in Table 4. All articles excluded from the review at full-text assess- ment are listed in Additional file 3 together with the reason for exclusion. After producing unmoderated models and forest plots, we analysed the influence of the following moderators within studies with sufficient data, also assessing the influence of the moderator on residual heterogeneity: • • Time since burning (time between last burn and out- come measure). • • Burn frequency: the number of burns per year across the study period, defined as the time between first burn and last sampling. A frequency of 1 was used when a study lasted < 1 year. • • Burn season (“dormant” or “growing”). We have produced an evidence atlas (https:// maps.esp.tl/maps/_SR15-Evidence-Atlas/pages/map. jsp?geoMapId=450603&TENANT_ID=198852) that shows the geographical location and meta-data from the systematic map database for each study. Figure 3 is a static image of part of the interactive evidence atlas.h • • Climate zone (Köppen–Geiger zones Cf, Cs, Df, Ds). • • Climate zone (Köppen–Geiger zones Cf, Cs, Df, Ds). • • Forest type (broadleaf, coniferous, mixed). The evidence base Our systematic review included a total of 244 studies from 235 articles. A flow diagram presenting the number of articles (and studies) included and excluded at each stage of this review is presented in Fig. 2. Study design A total of 39 of 244 studies presented before–after (BA) data, 152 presented control–impact (CI) data, and 85 studies included before–after–control–impact (BACI) data. One study did not clearly report its design. Since some studies included data based on more than one study design, the sum of the numbers above exceeds the total number of studies. Study language Almost all of the 244 studies were published in English. The only exceptions were one study in Finnish and one in French. Following validity assessment, 82 studies were deemed to be of sufficient validity for meta-analysis and 16 studies were excluded from the quantitative synthe- sis due to low validity (see Additional file 4 and “Narra- tive synthesis” below). Quantitative synthesis—meta‑analysis d ff l • • Forest type (broadleaf, coniferous, mixed). We investigated the influence of moderators individu- ally rather than combining all moderators in one model because many studies did not report all information. The 244 studies considered relevant for the review are detailed in the systematic map database (Additional file 9). Of these studies in the map, 98 had sufficient data to be eligible for meta-analysis. From the remain- ing studies, 146 did not have sufficient information or data to allow inclusion in the quantitative synthesis. Details of these studies excluded from further synthesis can be found along with all the other included studies in Additional file 9. We examined the robustness of our results in several ways. First, we produced funnel plots to identify cases where publication bias might be present [48]. We did this using 1/(square root of sample size) as a measure of pre- cision, since standard errors are inappropriate for fun- nel plots of standardised effect sizes [18]. Secondly, we examined the influence of the validity of studies as judged during validity assessment. We repeated our unmoder- ated model calculations using only ‘high validity’ studies (where n > 3) and examined whether our findings altered. Page 15 of 33 Page 15 of 33 Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 Table 4 Total numbers of articles excluded listed by primary exclusion reason Articles retrieved by searches Articles from systematic map Exclusion reason Exclude on population 76 3 Exclude on intervention 165 2 Exclude on comparator 11 0 Exclude on outcome 129 58 Exclude on abundance outcome 90 58 Exclude on study type 34 0 Exclude on climate zone 160 0 Exclude on language 27 0 Duplicate 74 0 Total excluded at full text 766 121 Table 4 Total numbers of articles excluded listed by primary exclusion reason America (182/244 studies): 172 in the USA and 10 in Canada (Fig. 3). The other studies were from Europe (28/244 studies), with 12 in Finland, 5 in Sweden, 2 each in Spain, France and Portugal and 1 each in Estonia, Lithuania, Norway, Poland and the UK. The remaining 34 studies were from Australia. Thus, while parts of the temperate and boreal zones were well covered by studies, gaps exist in other areas, particularly Russia, Kazakhstan, Northern China, Eastern Europe and New Zealand. Publication yearh There was a peak in publication of studies on biodiver- sity effects of prescribed burning between 2005 and 2009 (Fig. 4). The data suggest a plateau in the publication of studies since 2012. Narrative synthesis Study location An overview of the 244 studies included in the review is provided in the systematic map database (Additional file 9). Most of the studies were conducted in North Fig. 3 Screenshot of the evidence atlas showing details for one study in a pop-up box Fig. 3 Screenshot of the evidence atlas showing details for one study in a pop-up box Eales et al. Environ Evid (2018) 7:19 Page 16 of 33 Fig. 4 Number of articles published per year in CAB abstracts returned from a search using the forest population terms described in Table 1 (undertaken in November 2017), and the number of articles included in this review of prescribed burning on biodiversity Fig. 4 Number of articles published per year in CAB abstracts returned from a search using the forest population terms described in Table 1 (undertaken in November 2017), and the number of articles included in this review of prescribed burning on biodiversity Investigated forests with limited additional information. Where provided, further details included measures of fire intensity or severity, flame height, or type of ignition used. We found studies focusing on coniferous, broadleaf and mixed forests (Table 5). Coniferous forests were the most commonly represented type (126/244 studies), followed by broadleaf forests (54/244 studies). Further details on forest types and dominant tree species are provided in Additional file 9 and the evidence atlas. Generally, infor- mation regarding stand age and management history was poorly reported (either missing or not clearly described) across the evidence base. l A total of 59 of 244 studies undertook serial burning (i.e. burning an area/site more than once) and recorded data after the final burn. Ninety-four of 244 studies pro- vided time series data (richness or diversity data recorded at multiple time points in a treatment area) with the aim of tracking the response to the treatment over time. Additional interventions alongside burning (either investigated on separate sites or combined with burning on the same site) included: thinning; partial harvesting; understorey harvesting; creation of dead wood; grazing/ grazing exclusion; planting understorey vegetation; and The prescribed burning interventions Details about the burn intervention were typically not reported or reported inconsistently across studies. Often, burns were described only as being “prescribed burning” Table 5 Number of studies of different forest types in our review Forest category Examples of dominant tree species No. of studies Broadleaf (all except Australian) Acer spp., Quercus spp. Fraxinus spp., Acer spp. 54 Broadleaf (Australian), sometimes described as Jarrah, Karri, Eucalypt or Sclerophyll forest Eucalyptus spp. 34 Mixed (broadleaf and conifers) Quercus, Pinus, Populus tremuloides 29 Conifers Abies spp., Pinus spp., Libocedrus decurrens, Pseudotsuga menziesii 126 Not reported 1 Examples of dominant tree species Examples of dominant tree species Eales et al. Environ Evid (2018) 7:19 Page 17 of 33 Eales et al. Environ Evid (2018) 7:19 Page 17 of 33 complete removal of tree layer. These are listed for each study in the database provided in Additional file 9. Of the remaining 82 studies eligible for full quantitative synthesis, only 19 were categorised as having high valid- ity (Additional file 10). The other 63 studies were consid- ered to have medium validity, most commonly because they were either BA or CI studies, not BACI, or because they only partially accounted for spatial heterogene- ity in treatment allocation. Three studies of potentially “high validity” were downgraded to “medium validity (unclear)” because of a lack of information on their meth- ods, warranting a conservative approach. Measured outcomesh The numbers of studies with data for different out- comes are presented in Table 6. The majority of studies (144/244) contained data for plant richness and/or diver- sity. A large number of studies also reported data on rich- ness or diversity of invertebrate groups (60/244), such as arthropods, insects or beetles. Fewer studies reported fungal (16/244), mammal (6/244), amphibian (3/244) or reptile (4/244) richness or diversity. Data on lichens and bryophytes were poorly represented (5 and 2 studies, respectively). Justification for burning We found that for most stud- ies from the USA the burns were conducted for multiple purposes; both for fuel reduction and for promotion of biodiversity. Finnish studies (from two projects) investi- gated burning to promote biodiversity, as did one Cana- dian study. All Australian studies (n = 4) and the Spanish study had the aim of fuel reduction. The remaining studies did not report the intention of the burn. Since some reports included multiple outcome categories, the sum of the numbers exceeds the total number of reports (n = 244) Quantitative synthesis Study validity critical appraisal results Sixteen stud- ies were excluded from full synthesis due to low validity (see Additional file 4). The main reasons for exclusion were: intervention was not externally valid (7 studies, e.g. extremely high intensity burning); likely high heterogene- ity between treatment and control sites (3 studies); inap- propriate outcome measurement method (3 studies) and confounders present (3 studies, confounded by previous burning or pest outbreaks). Quantified outcomes From the 82 studies, we identi- fied 219 comparisons (i.e. effect size estimates) for use in our quantitative synthesis (Additional file 8). Thirty- one comparisons referred to diversity using Shannon Table 6 Outcome categories as defined in this review and number of reports per outcome Since some reports included multiple outcome categories, the sum of the numbers exceeds the total number of reports (n = 244) Code Description Number of studies containing richness data Number of studies containing diversity data Tree Trees (including seedlings and saplings) 20 9 Vasc Vascular plants (including a mix of herbaceous and woody species) 107 45 VascH Herbaceous plants (vascular) 19 4 VascW Woody plants (vascular) 21 3 Bryo Bryophytes 2 1 Lich Lichens 4 2 Fung Fungi 14 4 Mamm Mammals 7 1 Bird Birds 23 3 Amph Amphibians 3 0 Rept Reptiles 4 0 Beets Saproxylic beetles 1 0 Beetg Ground beetles 13 7 Beeto Other beetles (or all beetles) 9 2 Ins Insects (except beetles only) 18 4 Arth Arthropods (except insects only) 19 6 Inver Invertebrates (except arthropods only) 4 3 Nnativ Invasive/exotic/non-native species 16 0 PlanFunBry All plants, fungi and bryophytes 3 0 Nativ Native species 20 0 Table 6 Outcome categories as defined in this review and number of reports per outcome Since some reports included multiple outcome categories, the sum of the numbers exceeds the total number of reports (n = 244) Eales et al. Environ Evid (2018) 7:19 Page 18 of 33 Page 18 of 33 diversity index, 8 used Simpson’s index and 1 employed the Brillouin diversity index. Most of these comparisons referred to species diversity, but 1 comparison was of the diversity of species or genera, 4 comparisons were made at the order level and 2 comparisons referred to family- or order-level diversity. Meta‑analyses All f All outputs of the meta-analyses, including forest plots, funnel plots and Cook’s distance plots, are presented in Additional file 12. We present the key outputs and plots in this section and summarise the main outputs in Table 7. The upper and lower limits provided with Hedges’ g and regression estimates are 95% confidence intervals. Across the evidence base described here, most burns were undertaken in the growing season (120/219 comparisons). Most comparisons referred to short-term impacts of prescribed burning, with 61/219 comparisons measur- ing biodiversity impacts less than 1 year after the most recent fire, and 67/219 comparisons measuring impacts between 1 and 2 years after the last burning event. 19/219 comparisons (from 3 studies) referred to data sampled at least 10 years after the last burn. Quantitative synthesis Most richness comparisons were made at the species level (173 comparisons), but 6 com- parisons were of the richness of species or genera and 2 were of the richness of orders or families. Of the 219 comparisons, 64 also included outcome data sampled at intermediate time points, i.e. prior to the last time point in a time series. The intermediate time point data themselves were not extracted or analysed in this review (although they were described in meta-data in the systematic map database): we only extracted and used the last time point. Summary forest plots The summary forest plots show- ing effect sizes from all studies reporting the richness and diversity of plants and non-plant organisms (including those that could not be meta-analysed) are presented in Additional file 11. There are no clear visual patterns in response to prescribed burning across taxonomic groups, and so it is clear that further quantitative synthesis is nec- essary, where appropriate. Study duration and timing The duration of study (time between the first burn and the last outcome measure- ment) and the time since burning (time between the last burn and the last outcome measurement) for the 219 comparisons in the quantitative syntheses are presented in Figs. 5 and 6. We found a large number of comparisons in studies that covered long time periods, with 71/219 comparisons referring to effects at least 10 years after the initial burning. Shorter-term prescribed burning studies were also common, with 34/219 comparisons from stud- ies lasting less than 1 year, and 33/219 comparisons from studies lasting between 1 and 2 years. 1. All vascular plant richnessh The unmoderated model shows a significant, positive overall effect of burning on total vascular plant rich- ness (Hedges’ g = 0.397 [0.049–0.744], n = 63, p = 0.025, Fig. 5 Duration of comparisons in quantitative synthesis. Data for a total of 219 comparisons across 82 studies Fig. 5 Duration of comparisons in quantitative synthesis. Data for a total of 219 comparisons across 82 studies Fig. 5 Duration of comparisons in quantitative synthesis. Data for a total of 219 comparisons across 82 studies Eales et al. Environ Evid (2018) 7:19 Page 19 of 33 Fig. 6 Time since last burn for comparisons in quantitative synthesis. A total of 219 comparisons across 82 studies Fig. 6 Time since last burn for comparisons in quantitative synthesis. A total of 219 comparisons across 82 studies Fig. 7). Significant heterogeneity was detected across studies (Q62 = 176.368, p < 0.001). showed a significant impact (see Additional file 12 and Table 7).h None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). There is no indication of publication bias in the funnel plot. The fail-safe number is 23, showing that a relatively large number of studies is required to remove significance of the summary effect. The Cook’s distance plot does not indicate any clear outliers (see Additional file 12). There is no clear indication of asymmetry in the funnel plot (see Additional file 12). The fail-safe number is 331, indicating that the significance of the result is robust. The Cook’s distance plot indicates a number of influential effect sizes but no outliers of concern. i The high-validity sensitivity analysis showed a similar effect estimate (0.317 [0.042–0.591], n = 4, p = 0.024), suggesting that the full meta-analysis result was robust (see Additional file 12). f The sensitivity analysis using only high-validity data resulted in a non-significant summary effect estimate (0.097 [− 0.180 to 0.380], n = 11, p = 0.500). This could suggest that the significance of the full unmoderated model was affected by study validity. However, the non- significant result may be, in part, a consequence of the fact that nine of the 11 comparisons were from conifer- ous forest, a group with a non-significant effect size (see Additional file 12). 3. Herbaceous plant richness Prescribed burning was found not to have a signifi- cant effect on herbaceous plant species richness (0.357 [− 0.176 to 0.891], n = 22, p = 0.189, Fig. 9). There was significant heterogeneity (Q21 = 73.377, p < 0.001). i Forest type was found to have a significant effect on the impact of burning (QM2 = 10.167 p = 0.006), with a significant positive impact in broadleaf forests (0.956 [0.4954 to 1.417], n = 9, p < 0.001), but not for other forest types (Fig. 10). 2. Non‑native vascular plant richness Time since burn was found to have a significant effect on herbaceous plant richness (regression slope of − 0.130 [− 0.248 to − 0.011], n = 22, p = 0.032, Fig. 11). This figure suggests a complex relationship clouded by remaining heterogeneity (QE20 = 60.387, p < 0.001). This heterogeneity disguises both positive and negative effects, and some aspect of context thus remains that we cannot account for. Prescribed burning was found to have a significant, posi- tive impact on non-native vascular plant richness (0.386 [0.154–0.619], n = 10, p = 0.001, Fig. 8), and there was no significant heterogeneity (Q9 = 12.936, p = 0.166). i None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) Eales et al. Environ Evid (2018) 7:19 Page 20 of 33 Table 7 Summary of main results of meta-analyses Outcome and included studies/ moderators Unmoderated model outputs Significance of moderator (QM test of moderators and p-value) Summary effect size CI (lower) CI (upper) p-value Burn frequency Time since burn Forest type Burn season Climate zone 1. Vascular plant richness Medium and high valid- ity studies (n = 63) 0.397 0.049 0.744 0.025* QM1 = 1.231 p = 0.267 QM1 = 1.105 p = 0.293 QM2 = 5.598 p = 0.061 QM3 = 1.827 p = 0.609 QM6 = 6.003 p = 0.423 High validity studies only (n = 11) 0.143 − 0.184 0.377 0.500 – – – – – 2. Non-native vascular plant richness Medium and high valid- ity studies (n = 10) 0.386 0.154 0.619 0.001* QM§ = 0.003 p = 0.958 QM1 = 0.001 p = 0.975 NR, all studies coniferous QM2 = 2.232 p = 0.328 QM4 = 2.287 p = 0.683 High validity studies only (n = 4) 0.317 0.042 0.591 0.024* – – – – – 3. 2. Non‑native vascular plant richness Herbaceous plant richness Medium and high valid- ity studies (n = 22) 0.357 − 0.176 0.891 0.189 QM1 = 0.350 p = 0.554 QM1 = 4.619 p = 0.032* QM2 = 10.167 p = 0.006* QM3 = 2.614 p = 0.455 QM3 = 15.434 p = 0.002* Forest type: Broadleaf (n = 9) 0.956 0.495 1.417 < 0.001* – – – – – Forest type: Coniferous (n = 10) 0.372 − 0.270 1.014 0.256 – – – – – Forest type: mixed (n = 3) NR, only 3 comparisons – – – – – High validity studies only (n = 2) NR, only 2 comparisons 4. Woody plant richness Medium and high valid- ity studies (n = 23) − 0.253 − 0.743 0.237 0.312 QM1 = 0.023 p = 0.879 QM1 = 0.138 p = 0.711 QM2 = 0.107 p = 0.948 QM3 = 3.465 p = 0.325 QM3 = 0.474 p = 0.925 High-validity studies only (n = 3) NR, only 3 comparisons – – – – – 5. Tree richness Medium and high valid- ity studies (n = 13) − 1.035 − 2.095 0.026 0.056 QM1 = 0.404 p = 0.525 QM1 = 0.583 p = 0.463 QM2 = 1.023 p = 0.600 QM2 = 0.958 p = 0.619 QM1 = 0.866 p = 0.352 High validity studies only (n = 1) NR, only 1 comparison – – – – – 6. All vascular plant diversity Medium and high valid- ity studies (n = 13) − 0.065 − 0.343 0.214 0.649 QM1 = 0.692 p = 0.406 QM1 = 0.303 p = 0.582 QM2 = 1.132 p = 0.568 QM2 = 3.620 p = 0.164 QM6 = 3.857 p = 0.696 High validity studies only (n = 4) 0.055 − 0.569 0.678 0.863 – – – – – Table 7 Summary of main results of meta-analyses Significance of moderator (QM test of moderators and p-value) Page 21 of 33 Eales et al. Environ Evid (2018) 7:19 Table 7 (continued) Outcome and included studies/ moderators Unmoderated model outputs Significance of moderator (QM test of moderators and p-value) Summary effect size CI (lower) CI (upper) p-value Burn frequency Time since burn Forest type Burn season Climate zone 7. 2. Non‑native vascular plant richness Fungal richness Medium and high valid- ity studies (n = 5) − 1.163 − 2.420 0.095 0.070 QM1 = 0.305 p = 0.581 QM1 = 0.065 p = 0.800 NR, all studies coniferous QM1 = 0.008 p = 0.927 QM2 = 0.497 p = 0.780 High validity studies only (n = 3) NR, only 3 comparisons – – – – – 8. Bird richness Medium and high valid- ity studies (n = 6) − 0.169 − 0.695 0.356 0.528 QM1 = 0.964 p = 0.326 QM1 = 1.764 p = 0.184 QM2 = 1.029 p = 0.598 QM3 = 2.463 p = 0.482 QM2 = 1.789 p = 0.409 High validity studies only (n = 1) NR, only 1 comparison – – – – – 9. Beetle richness Medium and high valid- ity studies (n = 10) 0.398 − 0.097 0.892 0.115 QM1 = 3.519 p = 0.061 QM1 = 2.080 p = 0.149 QM1 = 0.314 p = 0.575 QM1 = 2.123 p = 0.145 QM2 = 0.615 p = 0.735 High validity studies only (n = 1) NR, only 1 comparison – – – – – NR not run * Significant at p < 0.05. CI = 95% confidence interval Table 7 (continued) Significance of moderator (QM test of moderators and p-value) * Significant at p < 0.05. CI = 95% confidence interval There is no indication of publication bias based on the funnel plot. The Cook’s distance plot shows that the only significant positive effect size in the meta-analysis is an outlier (see Additional file 12). There was also a significant difference between studies in different climate zones (QM3 = 15.434, p = 0.002), but this is likely the result of the only study in the Cf zone being a negative outlier. i Sensitivity analysis using only high-validity stud- ies was not conducted due to low number of studies (n = 3). Burn frequency and burn season were not found to have significant effects (see Additional file 12 and Table 7).h There is a slight indication of asymmetry in the fun- nel plot, suggesting possible publication bias and a more positive result for smaller studies. The Cook’s distance plot indicates the presence of one outlier (see Additional file 12); this is also clear in the forest plot (Fig. 9). 5. Tree richness b d b Prescribed burning was not found to have a significant effect on tree species richness (− 1.035 [− 2.095 to 0.026], n = 13, p = 0.056, Fig. 13). There was significant heteroge- neity (Q12 = 54.355, p < 0.001). i Only two studies were high-validity, precluding validity sensitivity analysis. None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). 6. All vascular plant diversityh The impact of prescribed burning on vascular plant diversity was non-significant (− 0.065 [− 0.343 to 0.224], n = 13, p = 0.649, Fig. 14). There was no significant het- erogeneity in effect sizes (Q12 = 8.938, p = 0.708). i The sensitivity analysis with high-validity studies revealed a non-significant summary effect size, indicating a robust result (see Additional file 12). f None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). 4. Woody plant richnesshi There was no significant effect of prescribed burning on woody plant richness (− 0.253 [− 0.743 to 0.237], n = 23, p = 0.312, Fig. 12). There is significant heteroge- neity present (Q22 = 63.882, p < 0.001). There is no indication of publication bias in the fun- nel plot, and the Cook’s distance plot does not indicate any significant outliers (see Additional file 12). None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). Only one study had high validity, precluding sensi- tivity analysis. Page 22 of 33 Eales et al. Environ Evid (2018) 7:19 Fig. 7 Forest plot of the effect of prescribed burning on species richness of vascular plants There is no clear evidence of publication bias in the funnel plot. The Cook’s distance plot indicates that one study may be an outlier (see Additional file 12). 8. Bird richness Prescribed burning was not found to affect bird rich- ness (− 0.169 [− 0.695 to 0.356], n = 6, p = 0.528, Fig. 16). There was no significant heterogeneity (Q5 = 5.857, p = 0.320). None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). The Cook’s distance plot gave no indication of a clear outlier and the funnel plot is uninformative due to the small sample size (see Additional File 12).h There were too few studies (n = 6) to permit a sensitiv- ity analysis of the impact of validity. Fig. 10 Boxplot showing the impact of prescribed burning on herbaceous plant richness by forest type 7. Fungal richness Prescribed burning was found not to have a significant effect on fungal richness (− 1.163 [− 2.420 to 0.095], Eales et al. Environ Evid (2018) 7:19 Page 23 of 33 Fig. 8 Forest plot of the effect of prescribed burning on richness of non-native vascular plants Fig. 8 Forest plot of the effect of prescribed burning on richness of non-native vascular plants Fig. 8 Forest plot of the effect of prescribed burning on richness of non-native vascular plants ig. 8 Forest plot of the effect of prescribed burning on richness of non-native vascular plants Fig. 9 Forest plot of the effect of prescribed burning on richness of herbaceous plants Fig. 9 Forest plot of the effect of prescribed burning on richness of herbaceous plants Page 24 of 33 Eales et al. Environ Evid (2018) 7:19 n = 5, p = 0.070, Fig. 15). There was significant overall heterogeneity (Q4 = 12.956, p = 0.012). None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). The funnel plot is uninformative due to the small sam- Fig. 10 Boxplot showing the impact of prescribed burning on herbaceous plant richness by forest type Fig. 11 Meta-regression of time since burn against herbaceous plant richness Fig. 10 Boxplot showing the impact of prescribed burning on herbaceous plant richness by forest type Fig. 10 Boxplot showing the impact of prescribed burning on herbaceous plant richness by forest type There were too few studies (n = 5) to permit a sensitiv- ity analysis of the impact of validity. 9. Beetle richness Beetle richness was not found to be affected by pre- scribed burning (0.398 [− 0.097 to 0.892], n = 10, p = 0.115, Fig. 17). There was significant heterogeneity among the comparisons (Q10 = 18.850, p = 0.027). Fig. 11 Meta-regression of time since burn against herbaceous plant richness 0 None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). One study can be seen to be a clear outlier on both the forest plot and the Cook’s distance plot. The funnel plot is uninformative due to the small sample size (see Addi- tional file 12). i Only one study was considered to have high validity, precluding sensitivity analysis. Discussion Pyrophilous and saproxylic species are known to ben- efit from prescribed burning in forests, particularly in the context of biodiversity conservation [23]. However, prescribed burning may also have a wide spectrum of effects on other species, implying the presence of effects on ecosystem characteristics [49] that need to be under- stood while planning and evaluating burning. Occasion- ally, such effects on non-pyrophilous species result from deliberate practices, for example to control invasive spe- cies. More often, however, they are side-effects that may also be in conflict with other goals in maintaining biodi- versity or ecosystem services (e.g. [50]).hf n = 5, p = 0.070, Fig. 15). There was significant overall heterogeneity (Q4 = 12.956, p = 0.012). g y 4 None of the moderators (forest type, burn frequency, time since burning, burn season and climate zone) showed a significant impact (see Additional file 12 and Table 7). This review focused on the effects of prescribed burn- ing on species that are not directly fire-associated (pyrophilous or saproxylic). We identified 244 stud- ies on these effects, including 82 studies eligible for meta-analysis. We found significant positive impacts The funnel plot is uninformative due to the small sam- ple size, and the Cook’s distance plot did not indicate clear outliers (see Additional file 12). Page 25 of 33 Eales et al. Environ Evid (2018) 7:19 Fig. 12 Forest plot of the effect of prescribed burning on richness of woody plants ig. 12 Forest plot of the effect of prescribed burning on richness of woody plants Fig. 12 Forest plot of the effect of prescribed burning on richness of woody plants Fig. 13 Forest plot of the effect of prescribed burning on richness of trees Fig. 13 Forest plot of the effect of prescribed burning on richness of trees non-pyrophilous and non-saproxylic species from pre- scribed burning. This was likely due to large inter-study variation in outcomes, due to high heterogeneity between studies, and low numbers of comparable studies in each non-pyrophilous and non-saproxylic species from pre- scribed burning. This was likely due to large inter-study variation in outcomes, due to high heterogeneity between studies, and low numbers of comparable studies in each of burning on vascular plant richness, non-native plant richness and herbaceous plant richness (broadleaved forest). In all other quantitative analyses, we found no consistent effects on species richness and diversity on Page 26 of 33 Eales et al. Discussion Environ Evid (2018) 7:19 Fig. 14 Forest plot of the effect of prescribed burning on diversity of vascular plants Fig. 14 Forest plot of the effect of prescribed burning on diversity of vascular plants Fig. 15 Forest plot of the effect of prescribed burning on richness of fungi associated with burning. However, the evidence base that we have uncovered suggests that there is also significant heterogeneity with respect to how prescribed burning affects different groups of organisms. While burning has often been shown to favour pyrophilous (e.g. [51, 52]) quantitative synthesis. We found no consistent effects of moderators and were unable to test the effect of many potential moderators, due to a lack of reporting.fi Generally, the effect of fire is believed to be marked, and directly related to the intense and abrupt disturbance Page 27 of 33 Eales et al. Environ Evid (2018) 7:19 and saproxylic species (e.g. [32, 53, 54]) either immedi-f We contend that our observation that the effects of Fig. 16 Forest plot of the effect of prescribed burning on species richness of birds Fig. 17 Forest plot of the effect of prescribed burning on beetle richness Fig. 16 Forest plot of the effect of prescribed burning on species richness of birds Fig. 16 Forest plot of the effect of prescribed burning on species richness of birds Fig. 16 Forest plot of the effect of prescribed burning on species richness of birds Fig. 16 Forest plot of the effect of prescribed burning on species richness of birds ig. 16 Forest plot of the effect of prescribed burning on species richness of birds Fig. 17 Forest plot of the effect of prescribed burning on beetle richness We contend that our observation that the effects of burning on species richness and diversity are highly vari- able across studies is ecologically valid. This is supported by studies that have simultaneously analysed multiple species and verified variable responses also at a more detailed scale (e.g. [29, 50]). In ecological systems, there is typically quite a high level of idiosyncrasy, depend- ing on variable biotic and abiotic circumstances as well as historical events. When studies from highly variable contexts are combined in a meta-analysis, the influence on the outcome of a given treatment (burning in our and saproxylic species (e.g. Discussion [32, 53, 54]) either immedi- ately or after a time lag, the effects of burnings on rich- ness and diversity of other species have been previously shown to vary from strongly positive (e.g. [55–57]) to negative (e.g. [58, 59]), depending on which species is studied. Interestingly, based on our results, this variation was not primarily a between-species phenomenon but rather a between-study phenomenon, meaning that sepa- rate studies on the same taxonomic or ecological groups revealed contrasting outcomes. and saproxylic species (e.g. [32, 53, 54]) either immedi- ately or after a time lag, the effects of burnings on rich- ness and diversity of other species have been previously shown to vary from strongly positive (e.g. [55–57]) to negative (e.g. [58, 59]), depending on which species is studied. Interestingly, based on our results, this variation was not primarily a between-species phenomenon but rather a between-study phenomenon, meaning that sepa- rate studies on the same taxonomic or ecological groups revealed contrasting outcomes. Eales et al. Environ Evid (2018) 7:19 Page 28 of 33 fire that is more damaging to less fire-attuned species [64–66]. case) is also expected to be variable [60]. We attempted to account for some contextual moderators, such as forest type and climate zone. Three moderators were found to have a significant impact on the effects of burn- ing on herbaceous plant richness; for instance, richness increased significantly after burning in broadleaf forests but not in coniferous and mixed forests. However, in many of our analyses the small number of comparable studies combined with the substantial heterogeneity lim- ited the power of moderators to explain the variability in outcomes.f Second, a prescribed burn is rarely an event that can be applied in a standardised way even if initiated delib- erately and controlled. Weather conditions, topog- raphy, and the amount of combustible biomass may considerably affect the severity and, thus, the ecologi- cal consequences of prescribed burning. For example, Gundale et al. [67] reported on burnings that were con- ducted using the same procedures but where variation in weather conditions and in the volume and distribu- tion of fuels led to variation in the behaviour and effects of the fire. This variation is an inherent feature of most controlled burns. The prescribed burn area studied by Elliot et al. [68] had recorded temperatures of < 80 °C on lower slopes but > 800 °C on upper slopes and ridges. Discussion The burns with the highest intensity were described as stand- replacing fire which consumed understorey vegetation and ignited crowns. Similarly, there is likely great vari- ability in the training and experience of the fire team and the proximity to human structures which may influence aversion to “risk”. Assessing the comparability of con- trolled burns is hindered by the often brief reporting of burn characteristics. As well as ecological context, differences in the applica- tion of the intervention are likely to have a strong influ- ence on the study results. We attempted to account for the following moderators: season of burning, the time since the area was burned, and the frequency of burn events. There are other factors that we were unable to incorporate into the analysis that may also influence out- comes (e.g. soil type and moisture, humidity, wind, etc.; see below). In addition, the reason for conducting the prescribed burn may also influence outcomes. In our review, only two of 82 studies in the quantitative analy- sis [61, 62] specified that the objective of burning was control of species rather than biodiversity/restoration (33 studies) or fuel reduction (19 studies). The remaining studies did not report the objective, a recognised poten- tial source of bias in reporting ecological research [63]. Third, the review covered species that are both taxo- nomically and ecologically highly heterogeneous and, thus, may be expected to show variable responses to fire. Our meta-analyses of taxonomic groups necessarily com- bined studies focusing on different subgroups which may have responded in different ways to burning, thus reduc- ing the accuracy of the summary effect size and resulting in a less meaningful summary. Pyrophilous and saprox- ylic species may be expected to benefit from fires where there is mortality of trees during and after a fire event, though in situations where trees are not killed, saproxyl- ics can suffer because of a net loss of dead wood. Other species respond in a range of different ways according to, for example, their motility, habitat preference, germina- tion requirements or seed release mechanisms. A meta- analysis of such diverse taxa could benefit from arranging them into groups based on other characteristics than their taxonomic status. For example, separate analyses of species sharing specific life-history characteristics, such as r- or K-selected species, could provide more detailed understanding of causal factors. Discussion Unfortunately, studies usually do not provide such data, and it is often impos- sible to classify species in this way afterwards, particu- larly where studies report richness or diversity of mixed groups. We focused our review on biodiversity outcomes and acknowledge that prescribed burning will have a wider impact on ecosystem services, such as carbon cycling, soil nutrient cycles and water quality. The importance of the impact on these services and their contextual dependence are worthy of exploration, possibly in a sys- tematic review. Reasons for heterogeneity At least four main factors contribute to the high level of heterogeneity in the observed effects of prescribed burning. First, although the review was restricted to boreal and temperate forests, there was noticeable regional varia- tion among the study systems. In terms of regional cover- age, the availability of studies was clearly biased towards North America: in total, out of the 219 comparisons that were eligible for the meta-analysis, 197 were from North America, which encompasses a large area with heteroge- neity in biotic and abiotic composition within our speci- fied climate zones. Some studies (7 comparisons) were conducted in Australian eucalypt forests that appear to have distinctive fire regimes not found in other regions, possibly because of the particular characteristics of euca- lypt trees, such as oil in leaves encouraging an intense Finally, the sampling methods used to quantify the impacts of prescribed burning varied widely across stud- ies. Methods are chosen by authors to be relevant for the focal species groups, but such methods do not necessarily Eales et al. Environ Evid (2018) 7:19 Page 29 of 33 provide comparable data when combined (in a meta- analysis) across different species groups. For example, taxonomic groups that are very diverse (such as beetles) and hence a significant part of biodiversity are often chal- lenging to sample efficiently and representatively [69]. Depending on the exact method used to capture beetles, such as widely used window-traps or pitfall traps, quite different patterns of the community may be revealed even if samples are assumed to represent communi- ties from the same forest stand [69]. A related problem is that our reviewed studies rarely report beta-diversity across samples and sites. Hence, if an effect of the fire is to increase variability in species composition it might not be captured by plot-based species richness or diver- sity estimates. This clearly mirrors the concern raised by Socolar et al. [70] that conservation research needs to better take beta-diversity issues into consideration. plant colonisers originated from long-term seed banks and established only for a short period after fire before entering the next seed bank period [73]. Burn season is another temporal moderator that could be important. However, as the dormant and growing sea- sons may span across the calendar cut-offs (which we used in most cases) and since authors generally did not report burn seasons clearly, the burn season moderator is subject to inconsistencies. Reasons for heterogeneity This is likely to contribute to noise and limits the ability of the moderator to explain heterogeneity in our datasets. Knowledge gaps and clustersi We identified knowledge gaps and clusters across all of the 244 studies in the review (including both meta-ana- lysed studies as well as in the systematic map) to deter- mine the representation of topics in the evidence base. Another methodological issue of concern relates to study duration. Although a fire event is always abrupt, its consequences have long-lasting impacts, and many ecological effects may only be revealed by time series data that cover at least a few decades following a fire [24, 27]. Only rarely have studies been able to assess such long-term effects and hence capture the successional dynamics after the fire [28]. The length of the monitoring period is likely to have a major influence on the hetero- geneity of the patterns observed in the reviewed studies. For example, short-term studies may be able to capture data on how different species colonise burned areas, but to reveal if species are also able to reproduce and estab- lish populations on these sites requires studies that cover multiple generations for species of interest. We have recorded where studies measured time-series, but due to low comparability across studies, we did not extract or analyse such data for this review. Instead, our analysis has focused on the data reported for the maximum time since fire. Forest types and locationsh The low number of mixed forest studies represents a clear knowledge gap. We also note the lack of studies from the relevant climate zones in Russia, Kazakhstan, Northern China, Eastern Europe and New Zealand, and the seemingly low number of studies from Canada and Fennoscandia. Within the included European studies (15 comparisons, from Western Europe), coverage was incomplete. The dominance of North American studies is a clear knowledge cluster (75% of the evidence base in the entire review and 90% of comparisons in the quan- titative synthesis). It is plausible that there is relevant literature from some of the underrepresented regions in other languages that we could not include, or alterna- tively evidence that has not been referenced in broadly accessible literature sources. It appears that studies on prescribed burnings from Fennoscandia mainly focus on pyrophilous and saproxylic species, and such studies were excluded from this review. i Whilst we recognise that the limited number of stud- ies available for meta-analyses limits the ability of mod- erators to explain heterogeneity, the moderators tested generally had little effect. The only exception is for her- baceous plant species richness. In this case, the effect size was found to decrease with time since burning and also differ between climate zones. There was also a sig- nificant effect of forest type, with herbaceous plant rich- ness in broadleaf forests showing a positive response to prescribed burning. Reasons why increased time since burn may have a negative effect on herbaceous plant richness could include gap dynamics leading to initial colonisation followed by competitive exclusion of some species [71, 72], or the influence of early-successional non-native herbs [28]. It is also possible that some early Heterogeneity in methods used to calculate outcomesh Heterogeneity in methods used to calculate outcomes There was notable heterogeneity in diversity and rich- ness estimation methods across studies included in the review. Effect sizes that are calculated based on percent- age cover, species abundance, basal area, etc. can differ in magnitude or direction. Studies also employed sampling at different spatial scales (i.e. plot sizes), which can inher- ently lead to different results. Especially small plot sizes can increase the risk of missing low-abundance species with direct implication for observed differences among treatments. A large number of studies (96/244) could not be included in the quantitative synthesis due to an unfor- tunate lack of replication or reporting of measures of variability within their data. We thus call on researchers to better report variability in summary data or provide access to raw data so that these statistics can be calcu- lated by meta-analysts. Implications for policy, practitioners and researchers Implications for policy, practitioners and researchersi Implications for policy, practitioners and researchers We found that prescribed burning had a significant positive effect on vascular plant richness, non-native plant richness and herbaceous plant richness (in broad- leaved forest). In all other quantitative analyses, we found no consistent positive or negative effects on species richness and diversity of non-pyrophilous and non-saproxylic species. This was likely due to high inter-study heterogeneity, and low numbers of compa- rable studies in each quantitative synthesis. We found no consistent effects of moderators and were unable to test the effect of many potential moderators, due to a lack of reporting. We note that the actual outcomes in any particular case are still difficult to predict, and any forest restoration or management project using burn- ing should include a component of monitoring in order to build a stronger evidence base for recommendations and guidelines on how to best achieve identified con- servation targets. There are situations where prescribed burning can have harmful effects on taxa that are con- servation-dependent, such as epiphytic lichens [77], and these require that prescribed burning is planned carefully to avoid harmful effects. Biodiversity outcome h l f y Whilst 74% of comparisons eligible for our quantita- tive synthesis referred to plant taxa, only 26% referred to other taxa. Some 83% of total comparisons reported on richness, whilst 17% comparisons reported diversity, demonstrating a strong skew towards richness reporting in the evidence base for the quantitative synthesis. Diver- sity studies typically reported Shannon diversity (76% of comparisons) rather than Simpson’s diversity (21% of comparisons). This may reflect the situation in the eco- logical literature, where the Shannon index is the one most commonly used [74]. Eales et al. Environ Evid (2018) 7:19 Page 30 of 33 Prescribed burning intervention cannot rule out the possibility that some studies in this review included these species groups. Only 15% of studies reported data for more than 5 years since burning, representing a knowledge gap on the long-term effects of fire. Long-term studies have often revealed that the effects of prescribed burning in forest ecosystems may become visible only after decades rather than years. The lack of long-term studies limits the ability to explore whether prescribed fires can meet their (often- cited) target to initiate typical post-fire successions and to restore forest structures typical of areas with natural fire regimes. Limitations of the review and evidence base Publication bias We cannot rule out the risk of publication bias, because of the small number of effect sizes in many of the meta- analyses. We did attempt to source grey literature from organisations and web searching, but further grey lit- erature may exist that was not locatable with our search strategy. Lack of reporting of population, interventions or confounders Many study authors provided limited or no documen- tation of stand age, management history, previous fire events or the prescribed burning program. Although, documenting fire severity quantitatively can be challeng- ing, we call on authors to better report this kind of data, since it is expected that severity has a major influence on ecosystems, including biodiversity [75, 76]. f In general, we expect that many non-saproxylic and non-pyrophilous taxa, such as those covered by this review, may be systematically slow to respond to fire, particularly when exposed to low-severity burning. Thus, for these groups especially, a longer monitoring period would be highly justified and we call on funders and researchers to undertake such long-term investiga- tion. We also call for increased research focusing on the impacts of prescribed burning on non-plant organisms, in particular fungi, birds, herpetofauna, and mammals. Competing interests The authors declare that they have no competing interests. Efforts were made to ensure that authors of research studies included in this review were not involved in any decisions regarding their own work. Burning of a forest system is a management practice that requires long-term commitment to implementation and monitoring of outcomes. Given that a large number of relevant prescribed burning experiments have been initiated, the ground has been set for long-term studies. Coming generations of researchers may be best advised to make use of existing study systems, some of which are identified in this review, thereby enabling outcomes to be monitored over very long periods. Acknowledgements We would like to thank Steve Norman, Elsa Field and Julia Koricheva for pro- viding advice and/or comments on our methodological approach. We thank all study authors who provided additional information and/or data on direct request. We also thank all stakeholders who provided input into this review or the systematic map that preceded this review. Influence of the presence of saproxylic/pyrophilous species within data Saproxylic and pyrophilous species were generally not the target of studies included in this review, but study authors may have included such species in estimates of overall richness and diversity without documenting their presence explicitly. This may to some extent have affected our findings, although we focused on non-saproxylic and non-pyrophilous species groups. It is also likely that in some cases the classification of species to saproxylics and pyrophilous may not be well-established. Thus, we We therefore identify three needs, which, if addressed, would improve the usability of results both in a review like this and for management: (1) document burn- ing severity and fire behaviour and, if possible, conduct experimental burnings where the severity of fire can be manipulated; (2) if possible, replicate treatments in units that are independent of each other; and, (3) monitor the response over long time periods, i.e. decades rather than a few years. Though similar recommendations have Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 Page 31 of 33 Page 31 of 33 Page 31 of 33 been proposed by authors of previous reviews of pre- scribed burning impacts [25, 78, 79], it appears that they are yet to be widely heeded. Possible barriers to uptake may include difficulty/danger in documenting fire sever- ity, and management or ecological constraints in the assignment of fully replicated experimental plots. The resources required for long-term monitoring and staffing turnover combined with differences in monitoring capac- ity of staff are also a likely constraint on experimental design and reporting. Co-designed or co-implemented research, whereby the skills and resources of practition- ers and researchers are used to implement robust pre- scribed burning experiments, may be a way to overcome these barriers. of Biology and Institute of Environmental Science, Carleton University, 1125 Colonel By Drive, Ottawa, ON K1S 5B6, Canada. 5 Department of Natural Sci- ences, Mid Sweden University, 851 70 Sundsvall, Sweden. 6 School of Forest Sciences, University of Eastern Finland-Joensuu, Yliopistokatu 7, PO Box 111, 80101 Joensuu, Finland. 7 Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, UK. Authors’ contributions 6. Pyne SJ. Burning bush: a fire history of Australia. Basingstoke: Macmillan; 1991. This review was conducted by JE, JJ and NRH. The report was drafted by JE and NRH. All authors assisted in editing and revising the report. All authors read and approved the final manuscript. 7. Rius D, Vannière B, Galop D. Holocene history of fire, vegetation and land use from the central Pyrenees (France). Quatern Res. 2012;77(1):54–64.ii 8. Stähli M, Finsinger W, Tinner W, Allgöwer B. Wildfire history and fire ecology of the Swiss National Park (Central Alps): new evidence from charcoal, pollen and plant macrofossils. Holocene. 2006;16(6):805–17. References 1. Lindenmayer D, Franklin J, Fischer J. General management principles and a checklist of strategies to guide forest biodiversity conservation. Biol Conserv. 2006;131(3):433–45. 1. Lindenmayer D, Franklin J, Fischer J. General management principles and a checklist of strategies to guide forest biodiversity conservation. Biol Conserv. 2006;131(3):433–45. 2. Bengtsson J, Nilsson SG, Franc A, Menozzi P. Biodiversity, disturbances, ecosystem function and management of European forests. For Ecol Manag. 2000;132(1):39–50. 2. Bengtsson J, Nilsson SG, Franc A, Menozzi P. Biodiversity, disturbances, ecosystem function and management of European forests. For Ecol Manag. 2000;132(1):39–50. Additional file 11. Narrative synthesis plots. g 3. Lindenmayer DB, Franklin JF. Conserving forest biodiversity: a compre- hensive multiscaled approach. Washington, D.C: Island Press; 2002. Additional file 12. Quantitative synthesis outputs (including forest, fun- nel and Cook’s distance plots). 4. Lindenmayer DB, Margules CR, Botkin DB. Indicators of biodiver- sity for ecologically sustainable forest management. Conserv Biol. 2000;14(4):941–50. Additional file 13. ROSES form for the systematic review report. 5. Agee JK. Fire ecology of Pacific Northwest forests. Washington DC: Island press; 1996. Additional files This review report is financed by the Mistra Council for Evidence-Based Envi- ronmental Management (EviEM). Additional file 1. Literature searches. Additional file 2. Unobtainable articles following title and abstract screening. Additional file 3. Articles excluded from the review at full text. Additional file 4. Studies excluded from the review due to low validity. Additional file 5. Meta-data extraction and coding schema. Additional file 6. Preparation of data for quantitative synthesis. Additional file 7. R script for meta-analyses. Additional file 8. Data used in meta-analyses. Additional file 9. Systematic map database. Additional file 10. Data extraction and validity assessment of quantitative synthesis studies. Additional file 11. Narrative synthesis plots. Additional file 12. Quantitative synthesis outputs (including forest, fun- nel and Cook’s distance plots). Additional file 13. ROSES form for the systematic review report. Additional file 1. Literature searches. Additional file 2. Unobtainable articles following title and abstract screening. Additional file 3. Articles excluded from the review at full text. Additional file 4. Studies excluded from the review due to low validity. Additional file 5. Meta-data extraction and coding schema. Additional file 6. Preparation of data for quantitative synthesis. Additional file 7. R script for meta-analyses. Additional file 8. Data used in meta-analyses. Additional file 9. Systematic map database. Additional file 10. Data extraction and validity assessment of quantitative synthesis studies. Additional file 11. Narrative synthesis plots. Additional file 12. Quantitative synthesis outputs (including forest, fun- nel and Cook’s distance plots). Additional file 13. ROSES form for the systematic review report. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Received: 28 February 2018 Accepted: 28 July 2018 Received: 28 February 2018 Accepted: 28 July 2018 Ethics approval and consent to participate Not applicable. Ethics approval and consent to participate Not applicable. Availability of data and materials A list of excluded studies at full text, a list of unobtainable and un-translated articles, the systematic map database, and the raw data and R scripts for ana- lysing them are included as Additional files with this report. A ROSES form [80] for this systematic review report is included as Additional file 13. 10. Abrams MD. Fire and the development of oak forests. Bioscience. 1992;42(5):346–53. 9. Bowman DM, Balch JK, Artaxo P, Bond WJ, Carlson JM, Cochrane MA, D’Antonio CM, DeFries RS, Doyle JC, Harrison SP. Fire in the Earth system. Science. 2009;324(5926):481–4. Eales et al. Environ Evid (2018) 7:19 Davies GM, Kettridge N, Stoof CR, Gray A, Ascoli D, Fernandes PM, Marrs R, Allen KA, Doerr SH, Clay GD. The role of fire in UK peatland and moorland management: the need for informed, unbiased debate. Philos Trans R Soc B. 2016;371(1696):20150342. 45. Borenstein M, Hedges L, Higgins JP, Rothstein H. Introduction to Meta– analysis. Chichester: Wiley; 2009. ISBN 9780470743386. 46. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2016; 2017. 21. Russell-Smith J, Thornton R. Perspectives on prescribed burning. Front Ecol Environ. 2013;11(s1):e3. 47. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36(3):1–48. 22. Burrows N, McCaw L. Prescribed burning in southwestern Australian forests. Front Ecol Environ. 2013;11(s1):e25–34. 48. Light RJ, Pillemer DB. Summing up: the science of reviewing research. Cambridge: Harvard University Press; 1984. 23. Halme P, Allen KA, Auniņš A, Bradshaw RH, Brūmelis G, Čada V, Clear JL, Eriksson A-M, Hannon G, Hyvärinen E. Challenges of ecological restoration: lessons from forests in northern Europe. Biol Conserv. 2013;167:248–56. 49. Shorohova E, Kneeshaw D, Kuuluvainen T, Gauthier S. Variability and dynamics of old-growth forests in the circumbolear zone: implica- tions for conservation, restoration and management. Silva Fennica. 2011;45(5):785–806. 24. Bradstock RA, Williams JE, Gill MA. Flammable Australia: the fire regimes and biodiversity of a continent. Cambridge: Cambridge University Press; 2002. 50. Granath G, Kouki J, Johnson S, Heikkala O, Rodríguez A, Strengbom J. Trade-offs in berry production and biodiversity under prescribed burning and retention regimes in Boreal forests. J Appl Ecol. 2018. https://doi. org/10.1111/1365-2664.13098. 25. Fernandes PM, Botelho HS. A review of prescribed burning effectiveness in fire hazard reduction. Int J Wildland Fire. 2003;12(2):117–28. 26. Fernandes PM, Davies GM, Ascoli D, Fernández C, Moreira F, Rigolot E, Stoof CR, Vega JA, Molina D. Prescribed burning in southern Europe: developing fire management in a dynamic landscape. Front Ecol Environ. 2013;11(s1):e4–14. 51. Heikkala O, Martikainen P, Kouki J. Prescribed burning is an effective and quick method to conserve rare pyrophilous forest-dwelling flat bugs. Insect Conserv Divers. 2017;10(1):32–41. 52. Kouki J, Hyvärinen E, Lappalainen H, Martikainen P, Similä M. Landscape context affects the success of habitat restoration: large-scale colonization patterns of saproxylic and fire-associated species in boreal forests. Divers Distrib. 2012;18(4):348–55. 27. Burrows N. Linking fire ecology and fire management in south-west Australian forest landscapes. For Ecol Manag. 2008;255(7):2394–406. 28. Eales et al. Environ Evid (2018) 7:19 Eales et al. Environ Evid (2018) 7:19 Page 32 of 33 11. Conedera M, Tinner W, Neff C, Meurer M, Dickens AF, Krebs P. Recon- structing past fire regimes: methods, applications, and relevance to fire management and conservation. Quatern Sci Rev. 2009;28(5–6):555–76. 35. Life + Taiga. Om Taiga: Life + Taiga; 2016. http://lifetaiga.se/om-taiga/. Accessed 25 June 2016. 36. Eales J, Haddaway NR, Bernes C, Cooke SJ, Jonsson BG, Kouki J, Petrokof- sky G. What is the effect of prescribed burning in temperate and boreal forest on biodiversity, beyond tree regeneration, pyrophilous and saprox- ylic species? A systematic review protocol. Environ Evid. 2016;5(1):24. 12. Van Wagtendonk JW. The history and evolution of wildland fire use. Fire Ecol. 2007;3(2):3–17. 13. Aplet GH. Evolution of wilderness fire policy. Int J Wilderness. 2006;12(1):9–13. 13. Aplet GH. Evolution of wilderness fire policy. Int J Wilderness. 37. The Collaboration for Environmental Evidence. Guidelines for systematic reviews in environmental management; 2013. 14. Baker WL. Restoration of landscape structure altered by fire suppression. Conserv Biol. 1994;8(3):763–9. 38. Harzing A. Publish or Perish. http://www.harzing.com/pop.htm. 2007. 15. Falcucci A, Maiorano L, Boitani L. Changes in land-use/land-cover patterns in Italy and their implications for biodiversity conservation. Landscape Ecol. 2007;22(4):617–31. 39. Haddaway NR, Collins AM, Coughlin D, Kohl C. Including non-public data and studies in systematic reviews and systematic maps. Environ Int. 2017;99:351–5. 16. Ryan KC, Knapp EE, Varner JM. Prescribed fire in North American forests and woodlands: history, current practice, and challenges. Front Ecol Environ. 2013;11(s1):e15–24. 40. Food and agriculture organisation of the United Nations. Forest Resources Assessment Working Paper 180. 2015. Available from: http:// www.fao.org/forest-resources-assessment/en/. Accessed 14 Aug 2018. 17. Wallenius T. Major decline in fires in coniferous forests-reconstructing the phenomenon and seeking for the cause. Silva Fennica. 2011;45(1):139–55. 41. Peel M, Finlayson B, McMahon T. Updated world map of the Köppen–Gei- ger climate classification. Hydrol Earth Syst Sci. 2007;11:1633–44. g 18. Zackrisson O. Influence of forest fires on the North Swedish boreal forest. Oikos. 1977;29(1):22–32. 42. Magurran AE. Measuring biological diversity. New York: Wiley; 2013. 43. Marrugan A. Ecological diversity and its measurement. London: Croom Helm; 1988. 19. Johnson E, Miyanishi K, Bridge S. Wildfire regime in the boreal forest and the idea of suppression and fuel buildup. Conserv Biol. 2001;15(6):1554–7. 44. Rohatgi A. WebPlotDigitalizer: HTML5 based online tool to extract numerical data from plot images. Version; 2015. 20. Eales et al. Environ Evid (2018) 7:19 Wittkuhn RS, McCaw L, Wills AJ, Robinson R, Andersen AN, Van Heurck P, Farr J, Liddelow G, Cranfield R. Variation in fire interval sequences has minimal effects on species richness and composition in fire- prone landscapes of south-west Western Australia. For Ecol Manag. 2011;261(6):965–78. 53. Hyvärinen E, Kouki J, Martikainen P, Lappalainen H. Short-term effects of controlled burning and green-tree retention on beetle (Coleoptera) assemblages in managed boreal forests. For Ecol Manag. 2005;212(1–3):315–32. 29. Hyvärinen E, Kouki J, Martikainen P. Prescribed fires and retention trees help to conserve beetle diversity in managed boreal forests despite their transient negative effects on some beetle groups. Insect Conserv Divers. 2009;2(2):93–105. 54. Toivanen T, Kotiaho JS. Burning of logged sites to protect beetles in man- aged boreal forests. Conserv Biol. 2007;21(6):1562–72. 55. Bowles ML, Jacobs KA, Mengler JL. Long-term changes in an oak forest’s woody understory and herb layer with repeated burning. J Torrey Bot Soc. 2007;134(2):223–37. 30. Bernes C, Jonsson BG, Junninen K, Lõhmus A, Macdonald E, Müller J, Sandström J. What is the impact of active management on biodiversity in boreal and temperate forests set aside for conservation or restoration? A systematic map. Environ Evid. 2015;4(1):25. 31. James KL, Randall NP, Haddaway NR. A methodology for systematic map- ping in environmental sciences Environ Evid 2016;5(1):7 30. Bernes C, Jonsson BG, Junninen K, Lõhmus A, Macdonald E, Müller J, Sandström J. What is the impact of active management on biodiversity in boreal and temperate forests set aside for conservation or restoration? A systematic map. Environ Evid. 2015;4(1):25. 56. Welch N, Waldrop TA, Buckner E. Response of southern Appalachian table mountain pine (Pinus pungens) and pitch pine (P. rigida) stands to prescribed burning. For Ecol Manag. 2000;136(1–3):185–97. 31. James KL, Randall NP, Haddaway NR. A methodology for systematic map- ping in environmental sciences. Environ Evid. 2016;5(1):7. 57. White AS. The effects of thirteen years of annual prescribed burn- ing on a Quercus ellipsoidalis community in Minnesota. Ecology. 1983;64(5):1081–5. 32. Bernes C, Jonsson BG, Junninen K, Lõhmus A, Macdonald E, Müller J, Sandström J. What are the impacts of dead-wood manipulation on the biodiversity of temperate and boreal forests. A systematic review proto- col. Environ Evid. 2016;4(1):1. 58. Collins BM, Moghaddas JJ, Stephens SL. Initial changes in forest structure and understory plant communities following fuel reduc- tion activities in a Sierra Nevada mixed conifer forest. For Ecol Manag. 2007;239(1–3):102–11. 33. Author details 1 Mistra EviEM, Stockholm Environment Institute, Box 24218, 104 51 Stock- holm, Sweden. 2 Present Address: European Centre for Environment and Human Health, University of Exeter Medical School, Knowledge Spa, Royal Cornwall Hospital, Truro TR1 3HD, UK. 3 Africa Centre for Evidence, University of Johannesburg, Johannesburg, South Africa. 4 Canadian Centre for Evi- dence‑Based Conservation and Environmental Management, Department 9. Bowman DM, Balch JK, Artaxo P, Bond WJ, Carlson JM, Cochrane MA, D’Antonio CM, DeFries RS, Doyle JC, Harrison SP. Fire in the Earth system. Science. 2009;324(5926):481–4. 10. Abrams MD. Fire and the development of oak forests. Bioscience. 1992;42(5):346–53. Eales et al. Environ Evid (2018) 7:19 Bernes C, Jonsson BG, Junninen K, Lõhmus A, Macdonald E, Müller J, Sandström J. What are the impacts of manipulating grazing and brows- ing by ungulates on plants and invertebrates in temperate and boreal forests? A systematic review protocol. Environ Evid. 2016;5(1):17. 59. Ducherer K, Bai Y, Thompson D, Broersma K. Dynamic responses of a British Columbian forest-grassland interface to prescribed burning. West North Am Nat. 2009;69(1):75–87. 34. Matheis M. Does restoration fire enhance the regeneration of deciduous trees in boreal forests? A systematic review. Sundsvall: Mituniversititet; 2015. 60. Gurevitch J, Hedges LV. Statistical issues in ecological meta-analyses. Ecol- ogy. 1999;80(4):1142–9. 61. Nuzzo VA, McClain W, Strole T. Fire impact on groundlayer flora in a sand forest 1990–1994. Am Midl Nat. 1996;136:207–21. Eales et al. Environ Evid (2018) 7:19 Page 33 of 33 Page 33 of 33 62. Larsen AL, Jacquot JJ, Keenlance PW, Keough HL. Effects of an ongoing oak savanna restoration on small mammals in Lower Michigan. For Ecol Manag. 2016;367:120–7. 73. Wienk CL, Sieg CH, McPherson GR. Evaluating the role of cutting treatments, fire and soil seed banks in an experimental framework in ponderosa pine forests of the Black Hills, South Dakota. For Ecol Manag. 2004;192(2–3):375–93. 63. Parker TH, Nakagawa S, Gurevitch J, IIEE. Promoting transparency in evolutionary biology, ecology, and ornithology. Auk. 2016;133(4):779–82. 74. Morris EK, Caruso T, Buscot F, Fischer M, Hancock C, Maier TS, Meiners T, Müller C, Obermaier E, Prati D. Choosing and using diversity indices: insights for ecological applications from the German Biodiversity Explora- tories. Ecol Evol. 2014;4(18):3514–24. 64. Fairman TA, Nitschke CR, Bennett LT. Too much, too soon? A review of the effects of increasing wildfire frequency on tree mortality and regenera- tion in temperate eucalypt forests. Int J Wildland Fire. 2016;25(8):831–48. 65. Cruz M, Sullivan A, Gould J, Sims N, Bannister A, Hollis J, Hurley R. Anatomy of a catastrophic wildfire: the Black Saturday Kilmore East fire in Victoria, Australia. For Ecol Manag. 2012;284:269–85. 75. Salo K, Kouki J. Severity of forest wildfire had a major influence on early successional ectomycorrhizal macrofungi assemblages, including edible mushrooms. For Ecol Manag. 2018;415–416:70–84. https://doi. org/10.1016/j.foreco.2017.12.044. 66. Bowman D. Rainforests and flame forests: the great Australian forest dichotomy. Geogr Res. 2000;38(3):327–31. 76. Martin RE, Sapsis DB, editors. Fires as agents of biodiversity: pyrodiversity promotes biodiversity. In: Proceedings of the conference on biodiversity of northwest California ecosystems Cooperative Extension, University of California, Berkeley; 1992. 67. Eales et al. Environ Evid (2018) 7:19 Gundale MJ, Metlen KL, Fiedler CE, DeLuca TH. Nitrogen spatial heteroge- neity influences diversity following restoration in a ponderosa pine forest, Montana. Ecol Appl. 2006;16(2):479–89. 77. Hämäläinen A, Kouki J, Lõhmus P. The value of retained Scots pines and their dead wood legacies for lichen diversity in clear-cut forests: the effects of retention level and prescribed burning. For Ecol Manag. 2014;324:89–100. 68. Elliott K, Vose J, Hendrick R. Long-term effects of high intensity prescribed fire on vegetation dynamics in the Wine Spring Creek Watershed, western North Carolina, USA. Fire Ecol. 2009;5(2):66–85. 69. Martikainen P, Kouki J. Sampling the rarest: threatened beetles in boreal forest biodiversity inventories. Biodivers Conserv. 2003;12(9):1815–31. p g forest biodiversity inventories. Biodivers Conserv. 2003;12(9):1815–31. 70. Socolar JB, Gilroy JJ, Kunin WE, Edwards DP. How should beta-diversity inform biodiversity conservation? Trends Ecol Evol. 2016;31(1):67–80. 78. Gill AM, McCarthy MA. Intervals between prescribed fires in Australia: what intrinsic variation should apply? Biol Conserv. 1998;85(1–2):161–9.i y 70. Socolar JB, Gilroy JJ, Kunin WE, Edwards DP. How should beta-diversity inform biodiversity conservation? Trends Ecol Evol. 2016;31(1):67–80. 79. Keeley JE. Fire intensity, fire severity and burn severity: a brief review and suggested usage. Int J Wildland Fire. 2009;18(1):116–26. 71. Hansson L. Boreal ecosystems and landscapes: structures, processes and conservation of biodiversity. Ecol Bull. 1997;46:203. 80. Haddaway N, Macura B, Whaley P, Pullin A. ROSES for systematic review reports. Version 1.0. 2017. https://doi.org/10.6084/m9.figshare.5897272. 72. Uotila A, Kouki J. Understorey vegetation in spruce-dominated forests in eastern Finland and Russian Karelia: successional patterns after anthropo- genic and natural disturbances. For Ecol Manag. 2005;215(1–3):113–37. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Eales et al. Environ Evid (2018) 7:19 Learn more biomedcentral.com/submissions Ready to submit your research ? Choose BMC and benefit from:
https://openalex.org/W2945717251	http://bura.brunel.ac.uk/bitstream/2438/18310/5/FullText.pdf	English	null	Cyclic di‐GMP inactivates T6SS and T4SS activity in <i>Agrobacterium tumefaciens</i>	Molecular microbiology	2,019	cc-by	16,332	Cyclic di-GMP inactivates T6SS and T4SS activity in Agrobacterium tumefaciens Ronan R. McCarthy,1,2* Manda Yu, 3 Kira Eilers,1 Yi-Chieh Wang,3 Erh-Min Lai3 and Alain Filloux 1* 1 MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial College London, London SW7 2AZ, UK. 2 Division of Biosciences, Department of Life Sciences, College of Health and Life Sciences, Brunel University London, Uxbridge, UB8 3PH, UK. 3 Institute of Plant and Microbial Biology, Academia Sinica, Taipei 11529, Taiwan. Ronan R. McCarthy,1,2* Manda Yu, 3 Kira Eilers,1 Yi-Chieh Wang,3 Erh-Min Lai3 and Alain Filloux 1* 1 MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial College London, London SW7 2AZ, UK. Ronan R. McCarthy,1,2* Manda Yu, 3 Kira Eilers,1 Yi-Chieh Wang,3 Erh-Min Lai3 and Alain Filloux 1* 1 MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial College London, London SW7 2AZ, UK. 2 Division of Biosciences, Department of Life Sciences, College of Health and Life Sciences, Brunel University London, Uxbridge, UB8 3PH, UK. 3 Institute of Plant and Microbial Biology, Academia Sinica, Taipei 11529, Taiwan. c-di-GMP levels unleash T6SS and T4SS to advance plant colonization. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Introduction 2 Division of Biosciences, Department of Life Sciences, College of Health and Life Sciences, Brunel University London, Uxbridge, UB8 3PH, UK. 3 Institute of Plant and Microbial Biology, Academia Sinica, Taipei 11529, Taiwan. Interspecies bacterial competition plays a key role in shaping microbial populations and determining what bac- terial species are dominant in a given niche (Kapitein and Mogk, 2014; Bernal et al., 2017a; McNally et al., 2017; Chassaing and Cascales, 2018). It is indeed increas- ingly clear that any specific environmental niche, includ- ing within various parts of the human body (Rojo et al., 2017), will profile the establishment of specific microbial populations. How bacteria respond to environmental vari- ations or intrusion of ‘foreign organisms’ by triggering an attack/defence war game with other bacteria is not clearly understood. Yet, despite the numerous protein secretion systems encoded within bacterial genomes, one has emerged as being the ‘weapon’ of choice through which bacteria mediate interspecies competition, for example, in the human gut (Russell et al., 2014b; Chatzidaki-Livanis et al., 2016; Sana et al., 2016; Anderson et al., 2017) or in planta (Ma et al., 2014; Bernal et al., 2017a; 2017b). This system, termed the Type VI Secretion System (T6SS), is found on the genome of a wide variety of Gram-negative bacteria and can deliver a remarkable array of toxins such as nucleases, amidases and phospholipases (Russell et al., 2011; Russell et al., 2012; Ma et al., 2014; Russell et al., 2014a; Alcoforado Diniz et al., 2015). It is noticeable that in Gram-positive bacteria the type VII secretion sys- tem (T7SS) is now emerging as the antibacterial nano- weapon (Cao et al., 2016). Whereas many of these toxins are antibacterial (Russell et al., 2014a), some are also designed to target eukaryotic host cells (Hachani et al., 2016) and fungi (Trunk et al., 2018) or have a dual func- tion like phospholipases (Jiang et al., 2014). These toxins can be encoded as part of a large T6SS cluster and thus genetically linked with genes encoding core T6SS com- ponents, or be found independently in so-called vgrG/hcp islands (Hachani et al., 2014; Whitney et al., 2014). It has recently emerged that T6SS toxins can specifically inter- act with VgrG proteins (Bondage et al., 2016; Cianfanelli et al., 2016a; Flaugnatti et al., 2016), the puncturing Molecular Microbiology (2019) 112(2), 632–648  doi:10.1111/mmi.14279 First published online 4 June 2019 doi:10.1111/mmi.14279 First published online 4 June 2019 Accepted 15 May, 2019. For correspondence. E-mail a.filloux@ imperial.ac.uk; ronan.mccarthy@brunel.ac.uk; Tel. +44(0)20 7594 9651; Fax +44(0)20 7594 3069.Ronan R. McCarthy and Manda Yu were contributed equally. Summary While ample information is now available consistently describing the structural com- ponents of the T6SS nanomachine (Cianfanelli et al., 2016b; Joshi et al., 2017; Nguyen et al., 2018) and its associated arsenal of toxins (Durand et al., 2014; Hachani et al., 2016; Lien and Lai, 2017), the type of regulatory elements that control expression of the system is variable from one species to another and has not always been conclusively determined. Well characterized signalling cascades have been shown to play a role in activating the T6SS such as the Gac/Rsm cascade (Moscoso et al., 2011; Allsopp et al., 2017) and the LasR/MvfR (PqsR) quorum sensing systems in P. aeruginosa (Lesic et al., 2009; Majerczyk et al., 2016; Sana et al., 2012), or, for example, the autoinducer-2 quorum sensing system and QstR in Vibrio cholerae (Ishikawa et al., 2009; Jaskolska et al., 2018). The environmental stimuli that may trigger T6SS expression are also poorly understood but some modulators have been proposed to be bile salts (Reen et al., 2012; Bachmann et al., 2015), contact dependent Agrobacterium tumefaciens is a soil bacterium that causes crown gall disease in a wide range of plants as a result of the delivery of T-DNA to plant cells via the Type IV Secretion System (T4SS) (Pitzschke and Hirt, 2010; Christie et al., 2014; Hwang et al., 2017). Agrobacterium tumefaciens strain C58 has also been shown to carry a single T6SS cluster composed of two divergently transcribed operons and featuring three toxins, Tae, a © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Fig. 1. Impact of c-d-GMP on Tde1, Tae and Hcp levels. A. The T6SS cluster composed of the divergently transcribed imp (Green) and hcp (Red) operons encoding 14 and 9 genes, respectively, and the remote island starting from atu3642 (vgrG-2) encoded by A. tumefaciens strain C58. B. Expression of wild-type SadC and a catalytically inactive SadC (psadC) in wild-type A. tumefaciens grown to early exponential phase in 523 medium. C. Western blots using Tde1, Tae, Hcp and RpoA (Loading Control) antibodies on whole lysates of ΔtssL cells (No secretion control) or wild- type cells transformed with either pBBRMCS4 (vector control), psadC, psadC, pwspR and pPA2133 grown to early exponential phase in 523 medium. D. C-di-GMP levels were quantified via LC-MS/MS. Summary The Type VI secretion system (T6SS) is a bacterial nanomachine that delivers effector proteins into prokaryotic and eukaryotic preys. This secretion system has emerged as a key player in regulating the microbial diversity in a population. In the plant pathogen Agrobacterium tumefaciens, the signal- ling cascades regulating the activity of this secre- tion system are poorly understood. Here, we outline how the universal eubacterial second messenger cyclic di-GMP impacts the production of T6SS toxins and T6SS structural components. We demonstrate that this has a significant impact on the ability of the phytopathogen to compete with other bacterial spe- cies in vitro and in planta. Our results suggest that, as opposed to other bacteria, c-di-GMP turns down the T6SS in A. tumefaciens thus impacting its ability to compete with other bacterial species within the rhizosphere. We also demonstrate that elevated lev- els of c-di-GMP within the cell decrease the activity of the Type IV secretion system (T4SS) and subse- quently the capacity of A. tumefaciens to transform plant cells. We propose that such peculiar control reflects on c-di-GMP being a key second messenger that silences energy-costing systems during early colonization phase and biofilm formation, while low Accepted 15 May, 2019. For correspondence. E-mail a.filloux@ imperial.ac.uk; ronan.mccarthy@brunel.ac.uk; Tel. +44(0)20 7594 9651; Fax +44(0)20 7594 3069.Ronan R. McCarthy and Manda Yu were contributed equally. c-di-GMP and secretion in Agrobacterium tumefaciens 63 device of the T6SS nanomachine, or can be found fused to VgrG thus forming what is referred to as an evolved VgrG (Pukatzki et al., 2007; Durand et al., 2012) or fused to a PAAR domain which loads the toxin onto a cognate VgrG (Whitney et al., 2015; Quentin et al., 2018). Some bacterial species have been shown to encode numerous T6SSs, with experimental evidence highlighting specific target or prey such as the H1-T6SS in Pseudomonas aeruginosa being associated with interbacterial compe- tition while the H2-T6SS was proposed to preferentially target eukaryotic cells (Sana et al., 2012; Hachani et al., 2014; Jones et al., 2014). Now, it seems that a clear- cut system specificity is unlikely to be the case as the H2-T6SS was recently shown to also be antibacterial (Allsopp et al., 2017). putative peptidoglycan amidase, and two DNases, Tde1 and Tde2 (Fig. 1A) (Ma et al., 2014). The Tde effectors have been shown to play a role in interbacterial compe- tition and plant colonization. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Results The P. aeruginosa diguanylate cyclase SadC supresses T6SS activity in A. tumefaciens C58 The P. aeruginosa diguanylate cyclase SadC supresses T6SS activity in A. tumefaciens C58 The universal second messenger c-di-GMP regulates a wide variety of bacterial phenotypes including motility, protein secretion, exopolysaccharide production, viru- lence, or biofilm formation (Romling and Balsalobre, 2012; Romling et al., 2013; Jenal et al., 2017). Across almost all eubacteria, paradigms have emerged that correlate high intracellular levels of c-di-GMP with loss of motility, increased polysaccharide production and biofilm forma- tion (Valentini and Filloux, 2016; Conner et al., 2017). The T6SS has also emerged as being subject to c-di-GMP control, with high levels of c-di-GMP associated with activation and low levels associated with suppression of this secretion system (Moscoso et al., 2011; Romling et al., 2013) as shown clearly with P. aeruginosa and to a lesser extent with Vibrio alginolyticus. In the latter case, it was proposed that the PppA phosphatase has a negative impact on both the activity of the T6SS and the levels of c-di-GMP (Sheng et al., 2013). In the case of Pseudomonas fluorescens it was also proposed that c- di-GMP could directly bind to the ClpB2 ATPase one of the core component in the T6SS (Trampari et al., 2015). In A. tumefaciens C58, it has been shown that high levels of c-di-GMP are associated with increased extracellular polysaccharide (EPS) production and biofilm formation (Heindl et al., 2014; Feirer et al., 2015), but the impact of c-di-GMP on T6SS has not been explored. Agrobacterium tumefaciens has one T6SS cluster which is composed of two divergently transcribed operons, the imp operon which primarily encodes the structural components of the secretion system and the hcp operon that encodes two of the known A. tumefaciens T6SS toxins, i.e. Tae and Tde1 (Wu et al., 2008; Lin et al., 2014). There is also an orphan cluster that encodes Tde2, the only known T6SS toxin not directly encoded in the A. tumefaciens primary cluster (Fig. 1A) (Ma et al., 2014). Here, we investigate the impact of c-di-GMP on the functionality of this system, and introduced plasmids encoding the well characterized P. Results aeruginosa diguanylate cyclases (DGC) SadC and WspR (Guvener and Harwood, 2007; Merritt et al., 2007a; Dahlstrom and O’Toole, 2017; McCarthy et al., 2017a), a mutated version of SadC that has the canonical catalytic site GGEEF domain changed to AAAEF (psadC) and a Response to external stimuli can be further driven by intracellular second messengers and this is the case for the universal eubacterial second messenger molecule cyclic di-GMP (c-di-GMP) (Jenal et al., 2017), which is tightly associated with biofilm development in a number of bacterial species including P. aeruginosa and A. tumefa- ciens (Heindl et al., 2014; Valentini and Filloux, 2016). The c-di-GMP signalling network can impact virulence in both human and plant pathogens as, for example, seen with the DgcP cyclase in controlling infections by P. aeruginosa or Pseudomonas savastanoi pv savastanoi, an olive tree pathogen (Aragon et al., 2015). Remarkably, in P. aeru- ginosa, c-di-GMP signalling has been shown to regulate T6SS, but while high levels correlate with increased lev- els in biofilm formation and T6SS activity (Moscoso et al., 2011), low levels of this second messenger are associ- ated with increased motility and Type III secretion system (T3SS) activity (Moscoso et al., 2011). The biofilm part of this paradigm has been confirmed within A. tumefaciens. Indeed, one of the pioneering studies on c-di-GMP as a signalling molecule has been carried out in this organism, where it was shown that the activity of a cellulose synthase was dependent on the binding of c-di-GMP (Amikam and Benziman, 1989). Subsequent studies have demonstrated the ability of high levels of c-di-GMP to increase the levels of a number of different A. tumefaciens polysaccharides and as a result biofilm formation (Heindl et al., 2014; Feirer et al., 2015). In the present study, we thus investigate the role of c-di-GMP in regulating the T6SS in A. tumefaciens C58 using a non-native diguanylate cyclase (DGC) from P. aeruginosa as well as native A. tumefaciens DGCs. We discover that contrary to the paradigm established in P. aeruginosa, high levels of c-di-GMP represses the T6SS activity in A. tumefaciens. We show that this occurs at the transcriptional level and has a significant impact on the ability of A. tumefaciens to attack and outcompete other bacterial species. We also demonstrate that this is not specific to a single A. tumefaciens DGC but that several of these enzymes can have a significant impact. Summary Transformed cells were grown for 16 h in 523 medium and cells equivalent to an OD600 5 were collected. Samples of interest were compared to a standard curve derived from measurements of known concentrations of pure c-di-GMP to determine the concentration (in nM) of c-di-GMP in the samples. All experiments are the mean of two independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test with p < 0.05, p < 0.01 , p < 0.001. Fig. 1. Impact of c-d-GMP on Tde1, Tae and Hcp levels. A. The T6SS cluster composed of the divergently transcribed imp (Green) and hcp (Red) operons encoding 14 and 9 genes, respectively, and the remote island starting from atu3642 (vgrG-2) encoded by A. tumefaciens strain C58. B. Expression of wild-type SadC and a catalytically inactive SadC (psadC) in wild-type A. tumefaciens grown to early exponential phase in 523 medium. C. Western blots using Tde1, Tae, Hcp and RpoA (Loading Control) antibodies on whole lysates of ΔtssL cells (No secretion control) or wild- type cells transformed with either pBBRMCS4 (vector control), psadC, psadC, pwspR and pPA2133 grown to early exponential phase in 523 medium. D. C-di-GMP levels were quantified via LC-MS/MS. Transformed cells were grown for 16 h in 523 medium and cells equivalent to an OD600 5 were collected. Samples of interest were compared to a standard curve derived from measurements of known concentrations of pure c-di-GMP to determine the concentration (in nM) of c-di-GMP in the samples. All experiments are the mean of two independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test with p < 0.05, p < 0.01 , p < 0.001. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 634 R. R. McCarthy et al.  subsequently impacts the capacity of A. tumefaciens to transform plant cells as determined by a transient trans- formation assay. This duality is quite remarkable and sug- gests that at a specific stage during the A. tumefaciens colonization process interaction with the rhizosphere and with the plant cells should both be kept silent. and counter-attack based induction (Russell et al., 2011; Basler et al., 2013), part of the competence regulon in V. cholerae (Borgeaud et al., 2015), or danger signal released from lysing cells in a P. aeruginosa population (LeRoux et al., 2015). Summary To date in A. tumefaciens C58, only the ExoR/ChvG/ChvI signalling cascade has been shown to be capable of regulating the T6SS, with induction of this system at low environmental pH leading to activation of the T6SS (Wu et al., 2012). This is believed to play a key role in allowing A. tumefaciens dominate around the site of a wound in a plant, as low pH is characteristic of this niche (Wu et al., 2012; Heckel et al., 2014). © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Results Firstly, it was striking to observe that the clumping/ hyperbiofilm phenotype resulting from SadC overexpres- sion and usually associated with high levels of c-di-GMP (Moscoso et al., 2014) was no longer seen when the cat- alytically inactive sadC mutant was expressed indicating as expected that this enzyme was no longer capable of generating c-di-GMP (Fig. 1B). The reduction in activity of this mutated version of SadC was confirmed by LC-MS/ MS and was shown to be comparable to another mutated version of SadC that had the GGEEF domain changed to GGAAF (Supporting Information Fig. S1). Furthermore, it was readily observed that expression of the sadC-en- coded membrane bound DGC dramatically abrogated Tde1, Tae and Hcp production. Levels seen in the strains containing plasmids encoding the SadC catalytic site mutant were comparable to strains carrying the empty vector (pBBRMCS4) (Fig. 1C) suggesting that the intra- cellular levels of c-di-GMP were effectively responsible for the reduction in the level of these T6SS components and toxins. The expression of the wspR-encoded cytoplasmic DGC did not alter the levels of Tde1, Tae or Hcp (Fig. 1C). This suggests that the intracellular activity of WspR in A. tumefaciens might not be strong enough to impact the lev- els of Tde1/Hcp or that membrane association of the DGC is instrumental to the phenotype. Note that the overexpres- sion of the cytoplasmic phosphodiesterase (PDE) PA2133 also did not significantly impact the levels of Tde1, Tae or Hcp (Fig. 1C). The proposed reduced activity of WspR in A. tumefaciens was confirmed by quantifying the levels of c-di-GMP using LC-MS/MS. This analysis confirmed that SadC produced significantly more c-di-GMP than WspR when expressed in A. tumefaciens (Fig. 1D). Overall, these findings are opposite to the established P. aerugi- nosa paradigm of positive regulation of the T6SS by c-di- GMP. Such variation in control is likely depending on how and where c-di-GMP is acting and this may dramatically differ from one bacterial species to the other. For example, in P. aeruginosa, the control exerted by master regulators such as RetS/LadS is observable on all H1-T6SS genes (Allsopp et al., 2017) whereas in Pseudomonas syringae it can only be seen on a subset of T6SS genes (e.g. icmF) but not on others (e.g. hcp) (Records and Gross, 2010). Results Furthermore, we show that elevated levels of c-di-GMP within the cell also impact the expression of key structural components of the T4SS. This transcriptional influence c-di-GMP and secretion in Agrobacterium tumefaciens 63 (Bernal et al., 2017a; Bernal et al., 2017b). To investi- gate if the impact of c-di-GMP on T6SS was biologically significant and capable of influencing the structure of a bacterial population, bacterial competition assays were performed as described in experimental procedures and using as an attacker strain A. tumefaciens which consti- tutively expresses wild-type (psadC) or mutant version of sadC (psadC). The bacterial prey Escherichia coli carries the pRL662 plasmid conferring gentamicin resistance and survivors after A. tumefaciens contact could be recovered as colony forming units (CFU) on gentamicin-containing plates. Expression of wild-type sadC in A. tumefaciens significantly reduced the ability of A. tumefaciens to kill the prey bacterium E. coli (Fig. 2A) when compared to a strain harbouring the vector control or expressing the non-functional version of sadC, sadC. The low levels of SadC-dependent killing were comparable to those seen with an A. tumefaciens mutant strain lacking a core gene for the T6SS machinery, tssL, thus resulting in T6SS inac- tivation and confirming the T6SS downregulation through SadC activity. To fully rule out that the observed reduction in killing was not a consequence of the increased cell to cell aggregation seen when overexpressing psadC, the assay was repeated in P. aeruginosa, where increased expression of sadC is also known to impact cell to cell aggregation (Moscoso et al., 2014). Competition assays between P. aeruginosa PAO1 expressing sadC and E. coli demonstrated increased levels of killing, contrary to what was observed in the A. tumefaciens competition assays (Supporting Information Fig. S2). A similar increase in killing was seen when using a P. aeruginosa ΔrsmA mutant, which is known to have elevated levels of sadC expression (Moscoso et al., 2014). This suggests that the impact seen on cell-to-cell aggregation by the expression of sadC does not impact the activity of the T6SS. known functional phosphodiesterase PA2133 (Ueda and Wood, 2010), into a wild-type strain of A. tumefaciens. The recombinant strains were grown to early exponen- tial phase and the levels of the T6SS DNase effector Tde1, amidase effector Tae and T6SS nanotube building block Hcp were assessed by western blotting (Ma et al., 2014). Results To add further biological significance to this obser- vation and partially recreate a physiologically relevant environment, the competition assays were repeated but bacteria were infiltrated in the leaves of the Nicotiana benthamiana plant as previously described (Ma et al., 2014). This assay more or less phenocopies our in vitro findings with expression of sadC leading to significant reduction in prey killing, and thus higher CFU recovery from the plant leaves of gentamicin resistant E. coli cells (Fig. 2B). These findings demonstrate that the intracel- lular levels of c-di-GMP can have a significant impact on A. tumefaciens ability to engage via the T6SS in inter- species competition in planta. Endogenous A. tumefaciens DGCs suppress the T6SS activity SadC has a significant impact on T6SS-dependent interbacterial competition in vitro and in planta © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity Impact of c-di-GMP on the ability of A. tumefaciens to kill other bacteria. A. A. tumefaciens strains with either an empty vector control, a plasmid expressing wild-type SadC or a plasmid expressing a catalytically inactive SadC (psadC) were incubated for 14 h on I medium with of E. coli DH10B containing pRL662 which harbours a gentamicin cassette. Bacteria were then resuspended in PBS and plated on LB agar containing gentamicin. B. The bacterial strains described above were resuspended in 1/2 Murashige and Skoog (MS) medium (pH 5.5) and immediately injected into the leaves of 6–8-day-old N. benthamiana plants. After 14 h incubation, coupons were cut from these leaves, homogenized in PBS and plated on LB supplemented with gentamicin (50 µg ml−1). All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain to the vector control with p < 0.05 , p < 0.01 , p < 0.001. et al., 2013) but also that Atu2091 and Atu5372 were increasing the levels of exopolysaccharide production, suggesting they are active cyclases, while Atu1207 and Atu2691 had little or no effect (Fig. 3A). tumefaciens does not guarantee that DGCs encoded within the A. tumefaciens genome can have a similar impact on T6SS. To investigate this, we focused on endogenous A. tumefaciens DGCs, 16 of which are encoded within the genome (Romling et al., 2013), of these we selected 6 that encoded transmembrane domains and carry the canonical GGDEF domain, namely Atu2091, Atu2691, Atu1207 and Atu5372 as well as the two previously described DGCs DgcA (Atu1257) and DgcC (Atu2179) both of which have been shown to impact EPS production (Xu et al., 2013) (Supporting Information Fig. S3). The rationale here was that both SadC and WspR have been shown to be active DGCs, but only the membrane bound SadC was capable of impacting c-di-GMP associated phenotypes (Fig. 1C). This could suggest that the localization of the the DGC may have a critical impact on phenotypic outcomes. The genes encoding these DGCs were amplified and cloned under the control of the IPTG-inducible lac pro- moter in the pTrc200 vector. Using agar plates-contain- ing Congo red, it is possible to reveal exopolysaccharide (EPS) production in response of increased intracellular levels of c-di-GMP (Howie and Brewer, 2009; Heckel et al., 2014; Heindl et al., 2014; Feirer et al., 2015; Feirer et al., 2017). © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity The T6SS is known to play a key role in shaping the micro- biota and this is particularly true in the plant rhizosphere The use of a well-characterized P. aeruginosa DGC to increase the intracellular levels of c-di-GMP in A. Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 636 R. R. McCarthy et al.  Fig. 2. Impact of c-di-GMP on the ability of A. tumefaciens to kill other bacteria. A. A. tumefaciens strains with either an empty vector control, a plasmid expressing wild-type SadC or a plasmid expressing a catalytically inactive SadC (psadC) were incubated for 14 h on I medium with of E. coli DH10B containing pRL662 which harbours a gentamicin cassette. Bacteria were then resuspended in PBS and plated on LB agar containing gentamicin. B. The bacterial strains described above were resuspended in 1/2 Murashige and Skoog (MS) medium (pH 5.5) and immediately injected into the leaves of 6–8-day-old N. benthamiana plants. After 14 h incubation, coupons were cut from these leaves, homogenized in PBS and plated on LB supplemented with gentamicin (50 µg ml−1). All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain to the vector control with p < 0.05 , p < 0.01 , p < 0.001*. 636 R. R. McCarthy et al.  Fig. 2. Impact of c-di-GMP on the ability of A. tumefaciens to kill other bacteria. A. A. tumefaciens strains with either an empty vector control, a plasmid expressing wild-type SadC or a plasmid expressing a catalytically inactive SadC (psadC) were incubated for 14 h on I medium with of E. coli DH10B containing pRL662 which harbours a gentamicin cassette Bacteria were then resuspended in PBS and plated on LB agar containing gentamicin. B. The bacterial strains described above were resuspended in 1/2 Murashige and Skoog (MS) medium (pH 5.5) and immediately injected into the leaves of 6–8-day-old N. benthamiana plants. After 14 h incubation, coupons were cut from these leaves, homogenized in PBS and plated on LB supplemented with gentamicin (50 µg ml−1). All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain to the vector control with p < 0.05 , p < 0.01 , p < 0.001*. Fig. 2. Endogenous A. tumefaciens DGCs suppress the T6SS activity It was clear that some of the A. tumefa- ciens transformed with the above recombinant plas- mids yielded colonies displaying a wrinkly phenotype (Fig. 3A). This confirmed that as previously described DgcA and DgcC could impact EPS production (Xu To assess if anyone of these DGCs was capable of shutting down the T6SS, as SadC did, a secretion assay was performed using a Hcp and a Tde1 specific antibody as described above. This assay revealed that Atu2091, Atu5372 and DgcA were all capable of down-regulating the secretion of Hcp and Tde1 to non-detectable levels, and these levels were comparable to those observed with the non-functional T6SS mutant A. tumefaciens ΔtssL (Fig. 3B). DgcC also had a negative impact on Hcp levels but to a lesser extent. The cellular levels of Hcp, TssB (subunit of the T6SS sheath) and ClpV (T6SS AAA+ ATPase), revealed a similar profile to what was seen with Hcp secretion, with expression of Atu2091, Atu5372 and DgcA leading to reduced levels of these proteins and DgcC having a lesser effect (Fig. 3C). As for the EPS assay, Atu1207 and Atu2691, did not shut off T6SS activ- ity, thus suggesting that they may not be active or are gen- erating levels of c-di-GMP too low to trigger observable phenotypic changes. We then performed in vitro killing assays with each of the recombinant strains. Significant reductions in killing, comparable with what is observed with the tssL mutant, were seen upon expression of dgcA and atu5372 while a reduced but not statistically significant killing was observed for strains expressing atu2091 and dgcC (Fig. 4A). No c-di-GMP and secretion in Agrobacterium tumefaciens 637 c-di-GMP and secretion in Agrobacterium tumefaciens 637 ion in killing was observed with strains express- u1207 and atu2691 To confirm that the observed expressing either patu5372 or patu2691. As expected, the strain expressing patu2691 did not display elevated Impact of native DGCs on T6SS. ession of a range of six different DGCs whose genes were cloned into wild-type A. tumefaciens and spotted onto Congo red agar ng 0.5 mM IPTG and Spectinomycin 200 µg ml−1. tern blot using Hcp, Tde1, and RpoA (Loading Control) antibodies on supernatants of cells transformed with either empty vector pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity coli containing pRL662 which harbours a gentamicin cassette. Bacteria were then resuspended in PBS and plated on LB agar-containing gentamicin. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain with the vector control, p < 0.05 , p < 0.01 ** B. C-di-GMP levels were quantified via LC-MS/MS in strains expressing either an empty vector control, atu5372 or atu2691. Native DGSc were induced with 0 5 mM IPTG for 16 h in 523 638 R. R. McCarthy et al.  others might need a higher threshold. This observation might also be in line with the established concept for the need of spatial pools of c-di-GMP (Christen et al., 2010), and specific subcellular localization of some DGCs such as within the cytoplasmic membrane, or their direct inter- action with the effector generating the output (McCarthy et al., 2017a), would overcome the threshold issue. C-di-GMP regulates both the imp and hcp T6SS clusters at the transcriptional level independently of the ExoR signalling cascade & S Ltd M l l Mi bi l 112 632 648 Although T6SS regulation in A. tumefaciens is relatively poorly understood, one pathway involved is the ExoR/ ChvI/ChvG signalling system which is known to impact EPS production, horizontal gene transfer, motility and virulence (Wu et al., 2012; Heckel et al., 2014). The ChvG/ChvI two-component system positively regulates the T6SS while ExoR acts negatively and is a periplas- mic repressor (Wu et al., 2012). In acidic pH conditions such as those found around a plant wound site, ExoR is degraded, this allows the autophosphorylation of ChvG, which can then transfer a phosphoryl group to its cog- nate response regulator ChvI for activation of T6SS (Wu et al., 2012) (see Discussion). To assess if the impact on T6SS through c-di-GMP signalling involves this cas- cade, psadC was introduced into an A. tumefaciens exoR mutant which is known to have elevated activity of T6SS. If the elevated pool of c-di-GMP is acting through the ExoR cascade, then c-di-GMP transmission should be interrupted in the exoR mutant and no longer be able to impact on the T6SS. This was assessed in standard 523 medium, but also in minimal media (AB-MES pH7), a con- dition where ExoR-mediated repression of T6SS has pre- viously been demonstrated (Wu et al., 2012). Endogenous A. tumefaciens DGCs suppress the T6SS activity tern blot using Hcp, ClpV, TssB, Tde1 and RpoA (Loading Control) antibodies on whole cell lysates of cells transformed with either vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. Fig. 3. Impact of native DGCs on T6SS. A. Expression of a range of six different DGCs whose genes were cloned into wild-type A. tumefaciens and spotted onto Congo red agar containing 0.5 mM IPTG and Spectinomycin 200 µg ml−1. B. Western blot using Hcp, Tde1, and RpoA (Loading Control) antibodies on supernatants of cells transformed with either empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. C. Western blot using Hcp, ClpV, TssB, Tde1 and RpoA (Loading Control) antibodies on whole cell lysates of cells transformed with either empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. Fig. 3. Impact of native DGCs on T6SS. A. Expression of a range of six different DGCs whose genes were cloned into wild-type A. tumefaciens and spotted onto Congo red agar containing 0.5 mM IPTG and Spectinomycin 200 µg ml−1. B. Western blot using Hcp, Tde1, and RpoA (Loading Control) antibodies on supernatants of cells transformed with either empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. C. Western blot using Hcp, ClpV, TssB, Tde1 and RpoA (Loading Control) antibodies on whole cell lysates of cells transformed with either empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 or patu5372 from cells grown to early exponential phase in 523 medium. expressing either patu5372 or patu2691. As expected, the strain expressing patu2691 did not display elevated levels of c-di-GMP compared to the vector control, while a strain expressing atu5372 had levels of c-di-GMP higher reduction in killing was observed with strains express- ing atu1207 and atu2691. To confirm that the observed phenotypes were a direct result of altered levels of c-di- GMP, the levels of c-di-GMP were quantified in strains 638 R. R. McCarthy et al.  Fig. 4. Impact of native DGC on interspecies killing and cdi-GMP levels. A. A. tumefaciens strains with either an empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 and patu5372 were incubated for 14 h on a I medium (pH5.5) with E. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity Remarkably, it was observed that in the exoR mutant, expression of sadC and thus high levels of c-di-GMP are still capable of down-regulating the T6SS, as seen by monitoring Hcp, ClpV and TssB production (Fig. 5A) suggesting that c-di-GMP is acting independently of the ExoR network. Since all data presented thus far of c-di-GMP impact- ing the T6SS were at the protein level, we also wanted to establish if this regulation was originally exerted at the transcriptional level. To address this, RNA was isolated from A. tumefaciens transformed with either the cloning vector, a vector expressing a nonspecific protein GFP, psadC, psadC, patu2691, a proposed inactive DGC or patu5372, a native A. tumefaciens DGC that was shown to have a significant impact on clumping, interspecies kill- ing and Hcp secretion. qRT-PCR was performed on cDNA synthesized using RNA extracted from these strains and primers specific to genes belonging to both divergent operons of the A. tumefaciens T6SS cluster, namely tssB, fha, hcp and clpV (Figs 1A and 5B,C). Remarkably, all Fig. 4. Impact of native DGC on interspecies killing and cdi-GMP levels. levels. A. A. tumefaciens strains with either an empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 and patu5372 were incubated for 14 h on a I medium (pH5.5) with E. coli containing pRL662 which harbours a gentamicin cassette. Bacteria were then resuspended in PBS and plated on LB agar-containing gentamicin. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain with the vector control, p < 0.05 , p < 0.01 ** B. C-di-GMP levels were quantified via LC-MS/MS in strains expressing either an empty vector control, atu5372 or atu2691. Native DGSc were induced with 0.5 mM IPTG for 16 h in 523 rich medium and cells equivalent to an OD600 5 were collected. Samples of interest were compared to a standard curve derived from measurements of known concentrations of pure cdi-GMP to determine the concentration (in nM) of c-di-GMP in the samples. All experiments are the mean of two independent biological experiments with standard deviation error bars. levels. A. A. tumefaciens strains with either an empty vector control, pdgcA, patu2091, pdgcC, patu2691, patu1207 and patu5372 were incubated for 14 h on a I medium (pH5.5) with E. coli containing pRL662 which harbours a gentamicin cassette. Endogenous A. tumefaciens DGCs suppress the T6SS activity Bacteria were then resuspended in PBS and plated on LB agar-containing gentamicin. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test comparing each strain with the vector control, p < 0.05 , p < 0.01 * B. C-di-GMP levels were quantified via LC-MS/MS in strains expressing either an empty vector control, atu5372 or atu2691. Native DGSc were induced with 0.5 mM IPTG for 16 h in 523 rich medium and cells equivalent to an OD600 5 were collected. Samples of interest were compared to a standard curve derived from measurements of known concentrations of pure cdi-GMP to determine the concentration (in nM) of c-di-GMP in the samples. All experiments are the mean of two independent biological experiments with standard deviation error bars. than those seen in a strain expressing the known active P. aeruginosa DGC, sadC (Figs 4B and 1D). It was intriguing to observe that some of the DGCs tested were unable to impact the T6SS suggesting that strength in DGC activ- ity might be an issue in signalling and that, such as with a rheostat, once a threshold of intracellular c-di-GMP is reached, some phenotypes could be triggered while uthors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 c-di-GMP and secretion in Agrobacterium tumefaciens 639 c di GMP and secretion in Agrobacterium tumefaciens 639 Fig. 5. c-di-GMP impacts T6SS at the transcriptional level. A. Western blot using Hcp, TssB, ClpV, and RpoA (Loading control) antibodies on whole cell lysate of ΔexoR cells transformed with either Empty Vector control or psadC grown to early exponential phase in 523 medium or AB-MES pH7.0. B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, psadC or psadC* grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV and 16s rRNA which serves as a normalization control. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001** comparing the vector control to psadC. C. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV or 16s rRNA which serves as a normalization control. Endogenous A. tumefaciens DGCs suppress the T6SS activity Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001** comparing each strain to the expression control pGFP. All experiments are the mean of three independent biological experiments with standard deviation error bars. gene encoding the VirG response regulator is activated via the ExoR/ChvI/ChvG signalling system (Heckel et al., 2014). Activation of this cascade subsequently confers the ability of A. tumefaciens to transfer T-DNA into the host plant cell. We thus analysed the expression of a number of different T4SS genes (virB1, virB2) and T4SS effector gene virE2 in cells expressing the native DGCs (Fig. 6A). qRT-PCR was performed on cDNA synthesized using RNA isolated from cells transformed with either a vector control, a vector expressing a non- specific protein GFP, patu2691 or patu5372. Our results suggested that increasing the levels of c-di-GMP within the cell significantly down-regulates the expression of T4SS machinery and effector genes, while expression of patu2691, a proposed inactive DGC or a non-spe- cific protein pGFP did not negatively impact expression (Fig. 6B). Similar results were seen when cells were expressing psadC, although only partial but significant restoration of expression was seen when expressing psadC* (Fig. 6C). of these genes were significantly down-regulated when atu5372 or sadC was expressed, but not upon expression of an inactive sadC, a proposed inactive DGC atu2691 or a nonspecific protein GFP, confirming that the impact of c-di-GMP on T6SS also occurs at the transcriptional level and that all T6SS genes are impacted. C-di-GMP negatively impacts expression of components of the T4SS and the ability of A. tumefaciens to transform plant cells The significant impact of c-di-GMP on the activity of the T6SS and biofilm formation led us to ask whether other determinants involved in infection progression, such as the T4SS, were also responsive to the intracel- lular levels of c-di-GMP. The rationale here was in line with previous reports where protein secretion systems, e.g. the T6SS and T3SS, are antagonistically con- trolled (Moscoso et al., 2011; McCarthy et al., 2017b). As described above for the control of EPS production, the regulation of the T4SS is dependent on activation of the VirA/VirG signalling by acidity or phenols such as acetosyringone (AS), in which the transcription of the To further interrogate this and determine if this impact on transcription is biologically relevant, an AGROBEST assay was performed (Wu et al., 2014). © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001* comparing each strain to the expression control pGFP. All experiments are the mean of three independent biological experiments with standard deviation error bars. Fig. 5. c-di-GMP impacts T6SS at the transcriptional level. A. Western blot using Hcp, TssB, ClpV, and RpoA (Loading control) antibodies on whole cell lysate of ΔexoR cells transformed with either Empty Vector control or psadC grown to early exponential phase in 523 medium or AB-MES pH7.0. B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, psadC or psadC* grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV and 16s rRNA which serves as a normalization control. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001** comparing the vector control to psadC. C. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV or 16s rRNA which serves as a normalization control. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001** comparing each strain to the expression control pGFP. All experiments are the mean of three independent biological experiments with standard deviation error bars. Fig. 5. c-di-GMP impacts T6SS at the transcriptional level. A. Western blot using Hcp, TssB, ClpV, and RpoA (Loading control) antibodies on whole cell lysate of ΔexoR cells transformed with either Empty Vector control or psadC grown to early exponential phase in 523 medium or AB-MES pH7.0. B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, psadC or psadC* grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV and 16s rRNA which serves as a normalization control. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 , p < 0.001** comparing the vector control to psadC. C. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown to early exponential phase. Expression quantified using primers specific to either tssB, fha, hcp, clpV or 16s rRNA which serves as a normalization control. Endogenous A. tumefaciens DGCs suppress the T6SS activity Operon encoding components of the T4SS, virB1-11 represents the structural components of the T4SS machinery while virD4 is involved in coupling DNA transfer. B RT PCR RNA i l t d f ll t f d ith ith t t t l GFP t 2691 t 5372 i AB MES p g B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown in AB-MES medium (pH 5.5) containing 200 μM acetorysingone (AS) for virulence gene induction. Expression quantified using primers specific to either virB1, virB2, virE2 or 16s rRNA which serves as a normalization control. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001** when compared to the protein expression control pGFP. C. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, psadC, psadC* grown to early exponential phase. Expression quantified using primers specific to virB1, virB2, virE2 and 16s RNA which serves as a normalization control. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001*. Agrobacterium-mediated Arabidopsis transformation assay (Agrobacterium-mediated enhanced seedling transformation) that utilizes β-glucuronidase (GUS) as a reporter carried on the T-DNA to determine the impact of native regulatory elements on Agrobacterium transformation activity (Wu et al., 2014). Four-day-old Arabidopsis seedlings were infected with A. tumefa- ciens strain C58C1 (pTiB6S3ΔT)H carrying either an empty vector control, a vector expressing a nonspecific protein GFP, patu2691 or patu5372 and expressing a GUS reporter. The assay was allowed to proceed for 3 days before seedlings were stained with 5-bromo-4- chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS staining. Remarkably, only expression of patu5372 leads to a dramatic reduction in GUS staining (Fig. 7A) indicating that, in concordance with the qRT-PCR data, the activity of the T4SS is significantly impacted by the intracellular levels of c-di-GMP. Similar findings were Agrobacterium-mediated Arabidopsis transformation assay (Agrobacterium-mediated enhanced seedling transformation) that utilizes β-glucuronidase (GUS) as a reporter carried on the T-DNA to determine the impact of native regulatory elements on Agrobacterium transformation activity (Wu et al., 2014). Four-day-old Arabidopsis seedlings were infected with A. Endogenous A. tumefaciens DGCs suppress the T6SS activity This is an 640 R. R. McCarthy et al.  Fig. 6. Impact of c-di-GMP on T4SS. A. Operon encoding components of the T4SS, virB1-11 represents the structural components of the T4SS machinery while virD4 is involved in coupling DNA transfer. B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown in AB-MES medium (pH 5.5) containing 200 μM acetorysingone (AS) for virulence gene induction. Expression quantified using primers specific to either virB1, virB2, virE2 or 16s rRNA which serves as a normalization control. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001* when compared to the protein expression control pGFP. C. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, psadC, psadC* grown to early exponential phase. Expression quantified using primers specific to virB1, virB2, virE2 and 16s RNA which serves as a normalization control. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001*. 640 R. R. McCarthy et al.  Fig. 6. Impact of c-di-GMP on T4SS. A. Operon encoding components of the T4SS, virB1-11 represents the structural components of the T4SS machinery while virD4 is involved in coupling DNA transfer. B. qRT-PCR on RNA isolated from cells transformed with either an empty vector control, pGFP, patu2691 or patu5372 grown in AB-MES medium (pH 5.5) containing 200 μM acetorysingone (AS) for virulence gene induction. Expression quantified using primers specific to either virB1, virB2, virE2 or 16s rRNA which serves as a normalization control. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , * h d h i i l GFP GMP on T4SS. omponents of the T4SS, virB1-11 represents the structural components of the T4SS machinery while virD4 is involve f Fig. 6. Impact of c di GMP on T4SS. A. Operon encoding components of the T4SS, virB1-11 represents the structural components of the T4SS machinery while virD4 is involved in coupling DNA transfer g p A. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Endogenous A. tumefaciens DGCs suppress the T6SS activity Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, pGFP, patu2691 or patu5372. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001** when compared to the protein expression control pGFP. B. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, psadC or psadC. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001* when compared to the vector control. p p B. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, psadC or psadC. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001* when compared to the vector control. regulating T6SS and T4SS in A. tumefaciens. Previously characterized DGCs and PDEs from P. aeruginosa were transformed into A. tumefaciens. The membrane bound DGC called SadC (Merritt et al., 2007a) was shown to sig- nificantly induce a clumping phenotype, characteristic of high levels of c-di-GMP (Heindl et al., 2014) (Fig. 1B), and down-regulate the expression of both the T6SS toxins and T6SS machinery components (Fig. 1C). This is con- trary to the established paradigm in P. aeruginosa where high levels of c-di-GMP are associated with increased levels of T6SS (Supporting Information Fig. S2) (Moscoso et al., 2011). We demonstrated that the ability to impact T6SS was dependent on SadC being active as a version of SadC that features a mutated GGDEF domain had no impact on bacterial clumping or T6SS activity (Fig. Endogenous A. tumefaciens DGCs suppress the T6SS activity 1B and C). The biological relevance of this down-regulation is validated by the inability of a strain of A. tumefaciens transformed with a plasmid expressing sadC to kill a prey bacterium in vitro or in the native environment of a plant (Fig. 2A and B). observed when cells were expressing psadC, although expression of psadC* could not completely abrogate this (Fig. 7B). Endogenous A. tumefaciens DGCs suppress the T6SS activity tumefa- ciens strain C58C1 (pTiB6S3ΔT)H carrying either an empty vector control, a vector expressing a nonspecific c-di-GMP and secretion in Agrobacterium tumefaciens 641 Fig. 7. Impact of c-di-GMP on plant transformation. A. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, pGFP, patu2691 or patu5372. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001** when compared to the protein expression control pGFP. B. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, psadC or psadC. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001* when compared to the vector control. Fig. 7. Impact of c-di-GMP on plant transformation. A. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, pGFP, patu2691 or patu5372. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001** when compared to the protein expression control pGFP. B. Four-day-old Arabidopsis seedlings were infected with Agrobacterium strain C58C1(pTiB6S3ΔT)H carrying pBISN1 and either an empty vector control, psadC or psadC. These cells were pre-incubated in I medium (pH5.5) supplemented with 200 μM AS to induce expression of the vir genes. Seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) to visualize the GUS activity. All experiments are the mean of three independent biological experiments with standard deviation error bars. Statistical significance was determined using students's t-test with p < 0.05 , p < 0.01 , p < 0.001* when compared to the vector control. Fig. 7. Impact of c-di-GMP on plant transformation. A. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Discussion When further assessed, these DGCs were all capable of inhibiting interbacterial killing by A. tumefaciens both in vitro and in planta to differing degrees. These findings confirm the hypothesis that in A. tumefaciens high levels of c-di-GMP down-regu- late T6SS but they also offer an intriguing insight into the specificity exhibited by different DGCs. The differing levels of killing observed in strains expressing different DGCs (Fig. 4A) highlights the importance of c-di-GMP thresh- old levels to trigger specific output, and may also lend support to an emerging phenomenon whereby the spa- tial and temporal localization of a DGC can significantly influence the resulting phenotypic impact (Romling et al., 2013; Dahlstrom and O’Toole, 2017). The regulation of the T6SS is shown to occur at the tran- scriptional level with both divergent transcriptional units in the T6SS cluster being significantly down-regulated (Fig. 5B and C). The only characterized signalling cascade involved in the regulation of T6SS in A. tumefaciens is the ExoR/ChvI/ChvG system (Wu et al., 2012; Heckel et al., 2014); however, we collected data suggesting that the impact of c-di-GMP acts independently of this regulatory network (Fig. 5A). This finding prompted us to explore the influence of intracellular c-di-GMP levels on the T4SS, another ExoR/ChvI/ChvG target. Intriguingly, we observed a significant transcriptional down-regulation in the expres- sion of key components of the T4SS machinery when a DGC was expressed, sadC or atu5372, and this requires c-di-GMP synthesis since an inactive SadC, SadC, displays only traces of inhibition of T-DNA transfer into Arabidopsis plant cells (Fig. 7AB). Intriguingly, it seems here that c-di-GMP signalling is independent of the master regulator for T4SS and T6SS expression, i.e. ChvI/ChvG/ ExoR. This is slightly different from what is observed with P. aeruginosa, in which the master network controlling T6SS activity, Gac/Rsm, is entangled with c-di-GMP sig- nalling. In this case, the gene encoding the DGC SadC, is directly controlled by the translational repressor RsmA (Moscoso et al., 2014), which in turn modulates c-di-GMP levels. Note that the absence of ExoR, relieves the kinase ChvG which can then phosphorylate the response regu- lator ChvI and therefore activate the T6SS and the T4SS genes. In this case if c-di-GMP was binding ChvI or ChvG to inhibit their activity, such inhibition should have been detected in an exoR mutant, which was not the case (Fig. 5A). Alternatively, a plausible explanation is that c-di-GMP In P. Discussion Agrobacterium tumefaciens is a phytopathogen capa- ble of causing tumorigenesis in a wide variety of plant species through T-DNA transfer via the T4SS (Gelvin, 2010). The start of this transfer process is centered on A. tumefaciens sensing characteristic changes in the rhizo- sphere such as a drop in pH and phytochemical release, indicative of a plant wound. Intriguingly, a drop in pH has also been shown to induce not only the T4SS but also the T6SS. The latter is central to A. tumefaciens outcompet- ing other bacteria within this niche whereas the former allows the unimpeded transfer of T-DNA and eventual formation of a crown gall (Wu et al., 2012). Other than low pH, the understanding of what factors regulate the activity of the T6SS in A. tumefaciens is relatively lim- ited. In this study, we investigate the role of c-di-GMP in The use of an active DGC from P. aeruginosa could create an artificial pseudo-circuit inhibiting the activation 642 R. R. McCarthy et al.  could modulate the activity of another protein, e.g. a mem- brane protein interacting with ExoR. This protein could capture the periplasmic ExoR when it is not bound to c-di- GMP. Instead, when such an integral membrane protein is bound to c-di-GMP, interaction with ExoR does not occur and ExoR is free and available to inhibit the ChvI kinase (Fig. 8). This type of control at the interface cytoplasm/ periplasm and across the cytoplasmic membrane was previously shown in the case of the P. fluorescens Lap system (Hinsa and O’Toole, 2006; Newell et al., 2009). In this case, the degenerated LapD cyclase/phosphodiester- ase is an integral cytoplasmic protein, which can bind c-di- GMP and then interact with the periplasmic LapG protease. When LapD is c-di-GMP-bound, it prevents the LapG pro- tease from cleaving and releasing the surface adhesin LapA, which contributes to biofilm formation only when it remains associated with the cell surface (Navarro et al., 2011; Dahlstrom and O’Toole, 2017). This is in agreement with the dogma-high c-di-GMP/high biofilm. Future work could focus on identifying possible missing links such as the one suggested above. Interestingly the active cyclases are membrane embedded, such as SadC in contrast to WspR, and that may reinforce the idea that the signalling could be a transmembrane inside-out signalling, where the membrane cyclase or an interacting partner therein plays a key role in the process (Navarro et al., 2011). Discussion Future work could focus on identifying possible missing links such as the one suggested above. Interestingly the active cyclases are membrane embedded, such as SadC in contrast to WspR, and that may reinforce the idea that the signalling could be a transmembrane inside-out signalling, where the membrane cyclase or an interacting partner therein plays a key role in the process (Navarro et al., 2011). In the inside-out signalling mechanism described by the O’Toole laboratory, the transmission of the signal involves HAMP domains so a directed approach could be used to assess known c-di-GMP-related proteins with a HAMP domain (Navarro et al., 2011). Alternatively a non-biased approach using random transposon mutagenesis in a strain overex- pressing a DGC and carrying a reporter fusion for T6SS or T4SS expression could be used. Screens could be aimed at identifying mutants whose T4SS or T6SS activity goes high up even in the context of DGC expression, thus pin- pointing any players in the circuitry. In any case, c-di-GMP signalling has shown that the combination and sequence of events that lead from synthesis to output is far from a simple and stereotyped trajectory. Examples from within A. tumefaciens have shown that a dual DGC-PDE, called DcpA, can see its activity balanced from DGC to PDE based on its interaction with another protein called PruA, a pterin reductase. In the presence of PruA, DcpA is a PDE while in its absence it is a DGC. This implies that when DcpA is expressed in a heterologous host, such as E. coli in which PruA is absent, the main activity is PDE, while in Agrobacterium it is mostly a DGC (Feirer et al., 2015). This has also lended support to our rationale to analyse the impact of native Agrobacterium DGC versus heterologous ones from P. aeruginosa to fully validate our phenotypic observations. In P. aeruginosa, the switch in lifestyle from biofilm to motile is also tightly linked to a concomittant switch of the T6SS. To rule this out and to explore the capacity of native A. tumefaciens DGCs to down-regulate T6SS, the genes of six different A. tumefaciens DGCs, that have similar features as SadC, e.g. predicted transmembrane domains, were cloned and expressed. Three of these DGCs, DgcA, Atu2091 and Atu5372, were capable of down regulating the secretion of Hcp to non detectable levels, comparable to those observed in an A. tumefaciens T6SS mutant, ΔtssL (Fig. 3B). © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Discussion In the inside-out signalling mechanism described by the O’Toole laboratory, the transmission of the signal involves HAMP domains so a directed approach could be used to assess known c-di-GMP-related proteins with a HAMP domain (Navarro et al., 2011). Alternatively a non-biased approach using random transposon mutagenesis in a strain overex- pressing a DGC and carrying a reporter fusion for T6SS or T4SS expression could be used. Screens could be aimed at identifying mutants whose T4SS or T6SS activity goes high up even in the context of DGC expression, thus pin- pointing any players in the circuitry. In any case, c-di-GMP signalling has shown that the combination and sequence of events that lead from synthesis to output is far from a simple and stereotyped trajectory. Examples from within A. tumefaciens have shown that a dual DGC-PDE, called DcpA, can see its activity balanced from DGC to PDE based on its interaction with another protein called PruA, a pterin reductase. In the presence of PruA, DcpA is a PDE while in its absence it is a DGC. This implies that when DcpA is expressed in a heterologous host, such as E. coli in which PruA is absent, the main activity is PDE, while in Agrobacterium it is mostly a DGC (Feirer et al., 2015). This has also lended support to our rationale to analyse the impact of native Agrobacterium DGC versus heterologous could modulate the activity of another protein, e.g. a mem- brane protein interacting with ExoR. This protein could capture the periplasmic ExoR when it is not bound to c-di- GMP. Instead, when such an integral membrane protein is bound to c-di-GMP, interaction with ExoR does not occur and ExoR is free and available to inhibit the ChvI kinase (Fig. 8). This type of control at the interface cytoplasm/ periplasm and across the cytoplasmic membrane was previously shown in the case of the P. fluorescens Lap system (Hinsa and O’Toole, 2006; Newell et al., 2009). In this case, the degenerated LapD cyclase/phosphodiester- ase is an integral cytoplasmic protein, which can bind c-di- GMP and then interact with the periplasmic LapG protease. When LapD is c-di-GMP-bound, it prevents the LapG pro- tease from cleaving and releasing the surface adhesin LapA, which contributes to biofilm formation only when it remains associated with the cell surface (Navarro et al., 2011; Dahlstrom and O’Toole, 2017). This is in agreement with the dogma-high c-di-GMP/high biofilm. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Discussion aeruginosa, the switch in lifestyle from biofilm to motile is also tightly linked to a concomittant switch c-di-GMP and secretion in Agrobacterium tumefaciens 643 Fig. 8. Regulation of T6SS and T4SS in A. tumefaciens. In acidic pH conditions such as those found around a plant wound site, the periplasmic regulator ExoR is degraded, this allows the activation of ChvG via autophosphorylation, ChvG can then activate its cognate response regulator ChvI, which then activates the T6SS. The activation of native DGCs by environmental stimuli leads to an increase in the local levels of cyclic di-GMP. These increased levels of c-di-GMP are capable of repressing the transcription of the T6SS and the T4SS via an as yet uncharacterized regulator(s). Fig. 8. Regulation of T6SS and T4SS in A. tumefaciens. In acidic pH conditions such as those found around a plant wound site, the periplasmic regulator ExoR is degraded, this allows the activation of ChvG via autophosphorylation, ChvG can then activate its cognate response regulator ChvI, which then activates the T6SS. The activation of native DGCs by environmental stimuli leads to an increase in the local levels of cyclic di-GMP. These increased levels of c-di-GMP are capable of repressing the transcription of the T6SS and the T4SS via an as yet uncharacterized regulator(s). dynamics of c-di-GMP is key (Christen et al., 2010; Abel et al., 2013; Kulasekara et al., 2013; Skotnicka et al., 2016). It will be instrumental to the fate of the commu- nity at a population level but may of course be specific to some individuals, offsprings of which will constitute a new sub-population with dictinct aims, such as, for example, dispersal from the biofilm. Changes in c-di-GMP within a cell have been demonstrated on numerous instances to significantly alter the behavior of that cell, examples of this include the temporally oscilating global pools of c-di- GMP in Caulobacter crescentus which play a key role in regulating the swarmer to stalked cell transition and in Myxococcus xanthus where flucatuations in c-di-GMP levels control the developmental cycle that results in fruit- ing body formation (Abel et al., 2013; Kulasekara et al., 2013; Ozaki et al., 2014; Lori et al., 2015; Rotem et al., 2015; Skotnicka et al., 2016). from T6SS to T3SS respectively (Moscoso et al., 2011; Valentini and Filloux, 2016; McCarthy et al., 2017b). Discussion This makes sense if considering that the T6SS may help killing unwanted members of a biofilm population and instead prefering the T3SS cytotoxicty when adopting an acute infectious style. In the case of Agrobacterium, simultane- ous down-regulation of the T6SS and T4SS when biofilm formation is induced gives an intriguing insight into the phases of A. tumefaciens infection. The upregulation of motility genes, the T6SS and the T4SS by low pH and phytochemical release associated with a plant wound is the first step in A. tumefaciens targeting the wound site and ensuring a mono-infection by killing competing spe- cies (Merritt et al., 2007b; Wu et al., 2012). Once at the site of the wound it is possible that intracellular levels of c-di-GMP are elevated stimulating the production of EPS, attachment and biofilm formation (Heindl et al., 2014). T6SS and T4SS are concurrently downregulated possibly as an economy measure allowing more efficient chan- nelling of resources towards establishing infection at the site of the wound. Once initial sustainable colonization is established, variation in c-di-GMP levels can then allow further colonization and again competition with other bac- teria and modification of the plant cell environment. The orchestration of c-di-GMP levels and its variation in differ- ent cells and at a different localization in each cell, would be in accordance with a model, where the spatio-temporal from T6SS to T3SS respectively (Moscoso et al., 2011; Valentini and Filloux, 2016; McCarthy et al., 2017b). This makes sense if considering that the T6SS may help killing unwanted members of a biofilm population and instead prefering the T3SS cytotoxicty when adopting an acute infectious style. In the case of Agrobacterium, simultane- ous down-regulation of the T6SS and T4SS when biofilm formation is induced gives an intriguing insight into the phases of A. tumefaciens infection. The upregulation of motility genes, the T6SS and the T4SS by low pH and phytochemical release associated with a plant wound is the first step in A. tumefaciens targeting the wound site and ensuring a mono-infection by killing competing spe- cies (Merritt et al., 2007b; Wu et al., 2012). Once at the site of the wound it is possible that intracellular levels of c-di-GMP are elevated stimulating the production of EPS, attachment and biofilm formation (Heindl et al., 2014). Measurement of c-di-GMP levels For c-di-GMP quantification, samples were prepared as described previously (Valentini et al., 2016) and analysed by liquid-chromatography mass spectrometry (LC-MS/MS). In brief, A. tumefaciens C58 strains harbouring pBBRMCS4 vectors constitutively expressing DGCs were grown over- night in 20 mL 523 medium and cells equivalent to an OD600 5 were collected by centrifugation. Strains harbouring pTrC200 vector or derivatives were grown overnight, adjusted to an OD600 0.1 and grown to an OD600 0.7. Native DGCs were induced with 0.5 mM IPTG for 16 h and cells equivalent to an OD600 5 were collected. Collected cells were resuspended in extraction solution (Acetonitrile/methanol/water, 2/2/1, v/v/v), incubated on ice for 15 min and heated for 10 min at 95–99°C. Cells were centrifuged for 10 min at 4°C, 20,800 x g and supernatant fluid was collected. Extraction was repeated twice and supernatant fluids of the 3 extraction steps were combined and incubated at −20°C overnight. Extraction fluids were centrifuged again and supernatant fluid was analysed at the BIOLOG Life Science Institute (Biolog, Bremen) via LC-MS/MS. Samples of interest were compared to a stan- dard curve derived from measurements of known concen- trations of pure c-di-GMP to determine the concentration (in nM) of c-di-GMP in the samples. Interbacterial competition assay in planta Interbacterial competition assays in planta were per- formed as previously described (Ma et al., 2014). The intra-species A. tumefaciens competition assay was per- formed with a 10:1 attacker-to-target ratio by leaf infiltra- tion of N. benthamiana. Briefly, 523 overnight-cultured A. tumefaciens cells were sub-cultured at 28°C in the same medium for further growth to OD600 1.0–1.5. The harvested cells were resuspended in 1/2 Murashige and Skoog (MS) medium (pH 5.5) to an appropriate OD600. The attacker (OD600 5) and target, E. coli DH10B containing pRL662 (OD600 0.5) were mixed equally before infiltration into 2-month-old leaves of N. benthamiana with use of a needleless syringe. After 24-h incubation at room tem- perature, the infiltrated spot was punched out, ground in 0.9% NaCl, serially diluted, and plated in triplicates on LB agar containing appropriate antibiotic to select for the target cells. All experiments are the mean of three inde- pendent biological experiments with standard deviation Bacterial strains and plasmids Strains, plasmids and primer sequences used in this study are shown in Supporting Information Table S1 and S2. Escherichia coli was cultured in Lysogeny Broth, whereas 523 medium (Kado and Heskett, 1970) was routinely used for A. tumefaciens strains unless otherwise indicated. Growth conditions were previously described (Ma et al., 2009). When required, appropriate antibiotics were added to the medium as follows: for E. coli, 50 µg ml−1 ampicillin, 50 µg ml−1 gentamicin (Gm), 50 µg ml−1 kanamycin (Km) and 100 µg ml−1 spectinomycin; and for A. tumefaciens, 200 µg ml−1 spectinomycin. About 1mM Isopropyl-β-D- thiogalactopyranoside (IPTG) was used when necessary. Site directed mutagenesis was performed on pBBRMCS4- sadC using primers (Supporting Information Table S1) targeted to the catalytic site GGDEF site with specific nucleotide changes corresponding to mutation of the active site to AAAEF. These primers were used to amplify the full plasmid, followed by ligation with T4 ligase and subsequent transformation into E. coli DH5α. Congo red assays were performed as previously described (Moscoso et al., 2011). 644 R. R. McCarthy et al.  with P. aeruginosa were carried out as described before (Pissaridou et al., 2018). Briefly, strains harbouring indicated plasmids were grown over night with appropriate antibiotics and competition assays were performed the next day on LB agar plates without antibiotics using a 1:1 ratio of attacker to prey and incubating at 37°C for 24 h. Attacker strains were P. aeruginosa PAO1 and PAO1 ΔrsmA and prey strain was E. coli Top10 pRL662-gfp. Competitions were recovered and serially diluted prior to spot plating on LB agar plates with antibiotics for selection and grown overnight at 37°C. Survival was assessed by quantitative colony counts on selective media (using gentamicin for selection of E. coli and ampicillin for P. aeruginosa). P. aeruginosa biofilms in the cystic fibrosis lung, whereby exposure of the lung to nitric oxide induces biofilm dispersal through the modulation of the activity of phosphodiester- ases, thus manipulating the intracellular levels of c-di-GMP (Barraud et al., 2006; Barraud et al., 2009; Howlin et al., 2017). In all, while the hypothesis that c-di-GMP is fine tun- ing the behaviour of A. tumefaciens as it transitions from the rhizosphere to the plant surface remains to be veri- fied, here we have demonstrated that c-di-GMP is capable of significantly influencing the ability of A. tumefaciens to compete with other bacterial species and thus may shape the microbial diversity within the soil and also significantly impacts the capacity of A. tumefaciens to carry out one of its most defining features, the ability to transform the plant cell with interkingdom DNA transfer. Discussion T6SS and T4SS are concurrently downregulated possibly as an economy measure allowing more efficient chan- nelling of resources towards establishing infection at the site of the wound. Once initial sustainable colonization is established, variation in c-di-GMP levels can then allow further colonization and again competition with other bac- teria and modification of the plant cell environment. The orchestration of c-di-GMP levels and its variation in differ- ent cells and at a different localization in each cell, would be in accordance with a model, where the spatio-temporal The impact seen in A. tumefaciens is to down- regulate both T4SS and T6SS, so this duality may have significant impact on shaping future bio-control efforts in the agri-sector. Exploiting the knowledge that both the T4SS and the T6SS are down regulated in the presence of elevated levels of c-di-GMP (Fig. 8) may also prompt the development of exogenous soil treatments that could influ- ence the accumulation of this signalling molecule within the cell and in turn influence its physiology/pathogenicity. Similar approaches have proven successful in disrupting © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Interbacterial competition on agar plates Interbacterial competition assays were performed as previ- ously described (Ma et al., 2014; Bondage et al., 2016). In brief, overnight cultures of E. coli DH10B containing pRL662 derivative conferring gentamicin resistance were grown in LB at 37°C. The A. tumefaciens cells were adjusted to OD600 0.1, whereas the E. coli DH10B were adjusted to OD600 0.01, mixed at a 10:1 ratio, and 10 μl was spotted on LB (pH7.0) agar and incubated for 16 h at 28°C. Cells were harvested, serially diluted, and plated in triplicates on LB agar with or without gentamicin for colony forming units (CFU) counting. All experiments are the mean of a minimum of three inde- pendent biological experiments with standard deviation error bars. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 *. Bacterial competitions © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 c-di-GMP and secretion in Agrobacterium tumefaciens 64 error bars. Statistical significance was determined using students t-test with p < 0.05 , p < 0.01 **. Ministry of Science and Technology of Taiwan grant no. 104-2311-B-001-025-MY3. Funding is also supported by the BBSRC Taiwan-UK International Partnering Award A.F (BB/ M02735X/1) and E.M.L. (grant no. 105-2911-I-001-503). RNA extraction and quantitative RT-PCR Strains were grown to early exponential phase before cells were pelleted. Total RNA was extracted from pellets by Total RNA Extraction Kit (Arrowtech). First-strand cDNA was syn- thesized from 4 μg of total RNA with SuperScript III Reverse Transcriptase (Invitrogen) and random primers. Quantitative PCR were performed in QuantStudio 12 K Flex Real-Time PCR machine (Applied Biosystems) with Power SYBR Green PCR Master Mix (Invitrogen). Allsopp, L.P., Wood, T.E., Howard, S.A., Maggiorelli, F., Nolan, L.M., Wettstadt, S., et al. (2017) RsmA and AmrZ orchestrate the assembly of all three type VI secretion systems in Pseudomonas aeruginosa. Proceedings of the National Academy of Sciences of the United States of America, 114, 7707–7712. Amikam, D. and Benziman, M. (1989) Cyclic diguanylic acid and cellulose synthesis in Agrobacterium tumefaciens. Journal of Bacteriology, 171, 6649–6655. Production of T6SS components and T6SS-dependent secretion assay Anderson, M.C., Vonaesch, P., Saffarian, A., Marteyn, B.S. and Sansonetti, P.J. (2017) Shigella sonnei encodes a functional T6SS used for interbacterial competition and niche occupancy. Cell Host & Microbe, 21, 769–776.e3. The T6SS components and effectors in whole cell and during secretion assays were performed as described previously (Lin et al., 2014). Strains were grown to early exponential phase in both AB-MES pH7 medium and 523 medium. Whole cell extract and secreted proteins were separated by SDS/ PAGE and transferred onto a PVDF or nitrocellulose mem- brane using a transfer apparatus (Bio-Rad). The membrane was probed with primary antibody against Tde1 epitope (Ma et al., 2014), Hcp (Wu et al., 2008), Tae (Ma et al., 2014), ClpV (Wu et al., 2012) and RpoA (Lin et al., 2014) was used as an internal control then incubated with horseradish perox- idase-conjugated anti-rabbit secondary antibody (1:20,000) and visualized with the ECL system (Perkin-Elmer). Aragon, I.M., Perez-Mendoza, D., Moscoso, J.A., Faure, E., Guery, B., Gallegos, M.T., et al. (2015) Diguanylate cyclase DgcP is involved in plant and human Pseudomonas spp. infections. Environmental Microbiology, 17, 4332–4351. Bachmann, V., Kostiuk, B., Unterweger, D., Diaz-Satizabal, L., Ogg, S. and Pukatzki, S. (2015) Bile salts modulate the mucin-activated type VI secretion system of pandemic Vibrio cholerae. PLoS Neglected Tropical Diseases, 9, e0004031. Barraud, N., Hassett, D.J., Hwang, S.H., Rice, S.A., Kjelleberg, S. and Webb, J.S. (2006) Involvement of nitric oxide in biofilm dispersal of Pseudomonas aeruginosa. Journal of Bacteriology, 188, 7344–7353. Author contributions The procedures for the seedling transient transformation assay were adapted from (Kim et al., 2006; Wu et al., 2014) with modifications. Agrobacterium tumefaciens cells were pre-induced by addition of 200 μM of acetosyringone at 28°C for 16 h and resuspended in appropriate buffer with cell den- sity OD600 0.02. Four-day-old A. thaliana seedlings were then incubated in 1ml of the buffer in a well of a 6-well plate for another 3 days. The seedlings were then removed from the buffer for GUS staining or GUS activity assay. For GUS stain- ing, seedlings were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-Gluc) at 37°C for 6 h as described. GUS activ- ity assay was determined by the conversion of 4-methylum- belliferyl-β-D-glucuronide (4-MUG) to 4-methylumbelliferone (4-MU). 4-MU fluorescence (excitation. 356 nm, emission. 455 nm) was measured with a 96 microtiter-plate reader (Bio- Tek Synergy Mx) and amount of protein in the reaction was determined by a Bradford assay. The relative GUS activity was expressed as OD455 μg−1 protein (Wu et al., 2014). RMC, AF, MY, EML designed research. RMC, MY, KG, YCW performed research. RMC, AF, MY, EML analysed data. RMC and AF wrote manuscript. RMC, AF, MY, EML designed research. RMC, MY, KG, YCW performed research. RMC, AF, MY, EML analysed data. RMC and AF wrote manuscript. Data availability statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References Abel, S., Bucher, T., Nicollier, M., Hug, I., Kaever, V., Abel zur Wiesch, P. and Jenal, U. (2013) Bi-modal distribution of the second messenger c-di-GMP controls cell fate and asymmetry during the caulobacter cell cycle. PLoS Genetics, 9, e1003744. Alcoforado Diniz, J., Liu, Y.C. and Coulthurst, S.J. (2015) Molecular weaponry: diverse effectors delivered by the Type VI secretion system. Cellular Microbiology, 17, 1742–1751. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Acknowledgements and Palmer, T. (2016) The type VII secretion system of Staphylococcus aureus secretes a nuclease toxin that targets competitor bacteria. Nature microbiology, 2, 16183. Hachani, A., Allsopp, L.P., Oduko, Y. and Filloux, A. (2014) The VgrG proteins are “a la carte” delivery systems for bacterial type VI effectors. The Journal of Biological Chemistry, 289, 17872–17884. Chassaing, B. and Cascales, E. (2018) Antibacterial weap- ons: targeted destruction in the microbiota. Trends in Microbiology, 26, 329–338. Hachani, A., Wood, T.E. and Filloux, A. (2016) Type VI secretion and anti-host effectors. Current Opinion in Microbiology, 29, 81–93. Chatzidaki-Livanis, M., Geva-Zatorsky, N. and Comstock, L.E. (2016) Bacteroides fragilis type VI secretion systems use novel effector and immunity proteins to antagonize human gut Bacteroidales species. Proceedings of the National Academy of Sciences of the United States of America, 113, 3627–3632. Heckel, B.C., Tomlinson, A.D., Morton, E.R., Choi, J.H. and Fuqua, C. (2014) Agrobacterium tumefaciens exoR con- trols acid response genes and impacts exopolysaccharide synthesis, horizontal gene transfer, and virulence gene expression. Journal of Bacteriology, 196, 3221–3233. Christen, M., Kulasekara, H.D., Christen, B., Kulasekara, B.R., Hoffman, L.R. and Miller, S.I. (2010) Asymmetrical distribution of the second messenger c-di-GMP upon bac- terial cell division. Science, 328, 1295–1297. Heindl, J.E., Wang, Y., Heckel, B.C., Mohari, B., Feirer, N. and Fuqua, C. (2014) Mechanisms and regulation of sur- face interactions and biofilm formation in Agrobacterium. Frontiers in Plant Science, 5, 176. Christie, P.J., Whitaker, N. and Gonzalez-Rivera, C. (2014) Mechanism and structure of the bacterial type IV secretion systems. Biochimica et biophysica acta, 1843, 1578–1591. Hinsa, S.M. and O'Toole, G.A. (2006) Biofilm formation by Pseudomonas fluorescens WCS365: a role for LapD. Microbiology, 152, 1375–1383. Cianfanelli, F.R., Alcoforado Diniz, J., Guo, M., De Cesare, V., Trost, M. and Coulthurst, S.J. (2016a) VgrG and PAAR proteins define distinct versions of a functional type VI se- cretion system. PLoS Path, 12, e1005735. Howie, A.J. and Brewer, D.B. (2009) Optical properties of amyloid stained by Congo red: history and mechanisms. Micron, 40, 285–301. Cianfanelli, F.R., Monlezun, L. and Coulthurst, S.J. (2016b) Aim, load, fire: the type VI secretion system, a bacterial nanoweapon. Trends in Microbiology, 24, 51–62. Howlin, R.P., Cathie, K., Hall-Stoodley, L., Cornelius, V., Duignan, C., Allan, R.N., et al. (2017) Low-dose nitric oxide as targeted anti-biofilm adjunctive therapy to treat chronic Pseudomonas aeruginosa infection in cystic fibro- sis. Molecular Therapy, 25, 2104–2116. Conner, J.G., Zamorano-Sanchez, D., Park, J.H., Sondermann, H. Acknowledgements Barraud, N., Schleheck, D., Klebensberger, J., Webb, J.S., Hassett, D.J., Rice, S.A., et al. (2009) Nitric oxide signal- ing in Pseudomonas aeruginosa biofilms mediates phos- phodiesterase activity, decreased cyclic di-GMP levels, and enhanced dispersal. Journal of Bacteriology, 191, 7333–7342. We are grateful to Sharon Longford (NTU, Singapore) for the design of the graphical abstract. Work in Alain Filloux’s laboratory was supported by the BBSRC grant BB/L007959/1 and funding for Lai lab is supported by the © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 646 R. R. McCarthy et al.  Basler, M., Ho, B.T. and Mekalanos, J.J. (2013) Tit-for-tat: type VI secretion system counterattack during bacterial cell-cell interactions. Cell, 152, 884–894. Feirer, N., Xu, J., Allen, K.D., Koestler, B.J., Bruger, E.L., Waters, C.M., et al. (2015) A pterin-dependent signal- ing pathway regulates a dual-function diguanylate cy- clase-phosphodiesterase controlling surface attachment in Agrobacterium tumefaciens. mBio, 6, e00156. Bernal, P., Allsopp, L.P., Filloux, A. and Llamas, M.A. (2017a) The Pseudomonas putida T6SS is a plant warden against phytopathogens. The ISME Journal, 11, 972–987. Feirer, N., Kim, D., Xu, J., Fernandez, N., Waters, C.M. and Fuqua, C. (2017) The Agrobacterium tumefaciens CheY- like protein ClaR regulates biofilm formation. Microbiology, 163(11), 1680–1691. Bernal, P., Llamas, M.A. and Filloux, A. (2017b) Type VI secretion systems in plant-associated bacteria. Environmental Microbiology, 20(1), 1–15. Flaugnatti, N., Le, T.T., Canaan, S., Aschtgen, M.S., Nguyen, V.S., Blangy, S., et al. (2016) A phospholipase A1 antibac- terial Type VI secretion effector interacts directly with the C-terminal domain of the VgrG spike protein for delivery. Molecular Microbiology, 99, 1099–1118. Bondage, D.D., Lin, J.S., Ma, L.S., Kuo, C.H. and Lai, E.M. (2016) VgrG C terminus confers the type VI effector trans- port specificity and is required for binding with PAAR and adaptor-effector complex. Proceedings of the National Academy of Sciences of the United States of America, 113, E3931–3940. g Gelvin, S.B. (2010) Plant proteins involved in Agrobacterium- mediated genetic transformation. Annual Review of Phytopathology, 48, 45–68. Borgeaud, S., Metzger, L.C., Scrignari, T. and Blokesch, M. (2015) The type VI secretion system of Vibrio cholerae fos- ters horizontal gene transfer. Science, 347, 63–67. Guvener, Z.T. and Harwood, C.S. (2007) Subcellular location characteristics of the Pseudomonas aeruginosa GGDEF protein, WspR, indicate that it produces cyclic-di-GMP in response to growth on surfaces. Molecular Microbiology, 66, 1459–1473. Cao, Z., Casabona, M.G., Kneuper, H., Chalmers, J.D. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 11 Acknowledgements Methods in Molecular Biology, 1657, 213–224. Jones, C., Hachani, A., Manoli, E. and Filloux, A. (2014) An rhs gene linked to the second type VI secretion cluster is a feature of the Pseudomonas aeruginosa strain PA14. Journal of Bacteriology, 196, 800–810. McNally, L., Bernardy, E., Thomas, J., Kalziqi, A., Pentz, J., Brown, S.P., et al. (2017) Killing by Type VI secretion drives genetic phase separation and correlates with in- creased cooperation. Nature Communications, 8, 14371. Joshi, A., Kostiuk, B., Rogers, A., Teschler, J., Pukatzki, S. and Yildiz, F.H. (2017) Rules of engagement: the type VI secretion system in Vibrio cholerae. Trends in Microbiology, 25, 267–279. Merritt, J.H., Brothers, K.M., Kuchma, S.L. and O'Toole, G.A. (2007a) SadC reciprocally influences biofilm for- mation and swarming motility via modulation of exopoly- saccharide production and flagellar function. Journal of Bacteriology, 189, 8154–8164. Kado, C.I. and Heskett, M.G. (1970) Selective media for isolation of Agrobacterium, Corynebacterium, Erwinia, Pseudomonas, and Xanthomonas. Phytopathology, 60, 969–976. Kapitein, N. and Mogk, A. (2014) Type VI secretion system helps find a niche. Cell Host & Microbe, 16, 5–6. Merritt, P.M., Danhorn, T. and Fuqua, C. (2007b) Motility and chemotaxis in Agrobacterium tumefaciens surface attach- ment and biofilm formation. Journal of Bacteriology, 189, 8005–8014. Kim, K.W., Franceschi, V.R., Davin, L.B. and Lewis, N.G. (2006) Beta-glucuronidase as reporter gene: advan- tages and limitations. Methods in Molecular Biology, 323, 263–273. Moscoso, J.A., Mikkelsen, H., Heeb, S., Williams, P. and Filloux, A. (2011) The Pseudomonas aeruginosa sen- sor RetS switches type III and type VI secretion via c-di-GMP signalling. Environmental Microbiology, 13, 3128–3138. Kulasekara, B.R., Kamischke, C., Kulasekara, H.D., Christen, M., Wiggins, P.A. and Miller, S.I. (2013) c-di- GMP heterogeneity is generated by the chemotaxis ma- chinery to regulate flagellar motility. eLife, 2, e01402. Moscoso, J.A., Jaeger, T., Valentini, M., Hui, K., Jenal, U. and Filloux, A. (2014) The diguanylate cyclase SadC is a central player in Gac/Rsm-mediated biofilm formation in Pseudomonas aeruginosa. Journal of Bacteriology, 196, 4081–4088. LeRoux, M., Kirkpatrick, R.L., Montauti, E.I., Tran, B.Q., Peterson, S.B., Harding, B.N., et al. (2015). Kin cell lysis is a danger signal that activates antibacterial pathways of Pseudomonas aeruginosa. eLife, 4, e05701. Pseudomonas aeruginosa. eLife, 4, e05701. Lesic, B., Starkey, M., He, J., Hazan, R. and Rahme, g , , Lesic, B., Starkey, M., He, J., Hazan, R. and Rahme, Navarro, M.V., Newell, P.D., Krasteva, P.V., Chatterjee, D., Madden, D.R., O'Toole, G.A., et al. Acknowledgements (2011) Structural basis for c-di-GMP-mediated inside-out signaling controlling periplasmic proteolysis. PLoS Biology, 9, e1000588. y L.G. (2009) Quorum sensing differentially regulates Pseudomonas aeruginosa type VI secretion locus I and homologous loci II and III, which are required for patho- genesis. Microbiology, 155, 2845–2855. Newell, P.D., Monds, R.D. and O'Toole, G.A. (2009) LapD is a bis-(3',5')-cyclic dimeric GMP-binding protein that reg- ulates surface attachment by Pseudomonas fluorescens Pf0-1. Proceedings of the National Academy of Sciences of the United States of America, 106, 3461–3466. Lien, Y.W. and Lai, E.M. (2017) Type VI secretion effec- tors: methodologies and biology. Frontiers in Cellular and Infection Microbiology, 7, 254. Lin, J.S., Wu, H.H., Hsu, P.H., Ma, L.S., Pang, Y.Y., Tsai, M.D., et al. (2014) Fha interaction with phosphothreonine of TssL activates type VI secretion in Agrobacterium tume- faciens. PLoS Path, 10, e1003991. Nguyen, V.S., Douzi, B., Durand, E., Roussel, A., Cascales, E. and Cambillau, C. (2018) Towards a complete structural deciphering of Type VI secretion system. Current Opinion in Structural Biology, 49, 77–84. Lori, C., Ozaki, S., Steiner, S., Bohm, R., Abel, S., Dubey, B.N., et al (2015) Cyclic di-GMP acts as a cell cycle os- cillator to drive chromosome replication. Nature, 523, 236–239. Ozaki, S., Schalch-Moser, A., Zumthor, L., Manfredi, P., Ebbensgaard, A., Schirmer, T., et al. (2014) Activation and polar sequestration of PopA, a c-di-GMP effector pro- tein involved in Caulobacter crescentus cell cycle control. Molecular Microbiology, 94, 580–594. Ma, L.S., Lin, J.S. and Lai, E.M. (2009) An IcmF family protein, ImpLM, is an integral inner membrane protein interacting with ImpKL, and its walker a motif is required for type VI se- cretion system-mediated Hcp secretion in Agrobacterium tumefaciens. Journal of Bacteriology, 191, 4316–4329. Pissaridou, P., Allsopp, L.P., Wettstadt, S., Howard, S.A., Mavridou, D.A.I. and Filloux, A. (2018) The Pseudomonas aeruginosa T6SS-VgrG1b spike is topped by a PAAR protein eliciting DNA damage to bacterial competitors. Proceedings of the National Academy of Sciences of the United States of America, 115, 12519–12524. Ma, L.S., Hachani, A., Lin, J.S., Filloux, A. and Lai, E.M. (2014) Agrobacterium tumefaciens deploys a superfamily of type VI secretion DNase effectors as weapons for interbacterial competition in planta. Cell Host & Microbe, 16, 94–104. Pitzschke, A. and Hirt, H. (2010) New insights into an old story: Agrobacterium-induced tumour formation in plants by plant transformation. The EMBO Journal, 29, 1021–1032. Majerczyk, C., Schneider, E. and Greenberg, E.P. (2016). Acknowledgements and Yildiz, F.H. (2017) The ins and outs of cyclic di-GMP signaling in Vibrio cholerae. Current Opinion in Microbiology, 36, 20–29. Hwang, H.-H., Yu, M. and Lai, E.-M. (2017) Agrobacterium- Mediated Plant Transformation: Biology and Applications. Hwang, H.-H., Yu, M. and Lai, E.-M. (2017) Agrobacterium- Mediated Plant Transformation: Biology and Applications. The Arabidopsis Book, e0186. Ishikawa, T., Rompikuntal, P.K., Lindmark, B., Milton, D.L. and Wai, S.N. (2009) Quorum sensing regulation of the two hcp alleles in Vibrio cholerae O1 strains PLoS One 4 e6734 Hwang, H.-H., Yu, M. and Lai, E.-M. (2017) Agrobacterium- Mediated Plant Transformation: Biology and Applications. The Arabidopsis Book, e0186. Ishikawa, T., Rompikuntal, P.K., Lindmark, B., Milton, D.L. and Dahlstrom, K.M. and O'Toole, G.A. (2017) A symphony of cyclases: specificity in diguanylate cyclase signaling. Annual Review of Microbiology, 71, 179–195. The Arabidopsis Book, e0186. Ishikawa, T., Rompikuntal, P.K., Lindmark, B., Milton, D.L. and Durand, E., Derrez, E., Audoly, G., Spinelli, S., Ortiz- Lombardia, M., Raoult, D., et al. (2012) Crystal structure of the VgrG1 actin cross-linking domain of the Vibrio chol- erae type VI secretion system. The Journal of Biological Chemistry, 287, 38190–38199. Wai, S.N. (2009) Quorum sensing regulation of the two hcp alleles in Vibrio cholerae O1 strains. PLoS One, 4, e6734. Jaskolska, M., Stutzmann, S., Stoudmann, C. and Blokesch, M. (2018) QstR-dependent regulation of natural compe- tence and type VI secretion in Vibrio cholerae. Nucleic Acids Research, 46, 10619–10634. y Durand, E., Cambillau, C., Cascales, E. and Journet, L. (2014) VgrG, Tae, Tle, and beyond: the versatile arsenal of Type VI secretion effectors. Trends in Microbiology, 22, 498–507. Jenal, U., Reinders, A. and Lori, C. (2017) Cyclic di-GMP: second messenger extraordinaire. Nature Reviews Microbiology, 15, 271–284. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 c-di-GMP and secretion in Agrobacterium tumefaciens 64 Jiang, F., Waterfield, N.R., Yang, J., Yang, G. and Jin, Q. (2014) A Pseudomonas aeruginosa type VI secretion phospholipase D effector targets both prokaryotic and eu- karyotic cells. Cell Host & Microbe, 15, 600–610. regulates lipopolysaccharide modification and contributes to Pseudomonas aeruginosa immune evasion. Nature Microbiology, 2, 17027. McCarthy, R.R., Valentini, M. and Filloux, A. (2017b) Contribution of cyclic di-GMP in the control of type III and type VI secretion in Pseudomonas aeruginosa. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 Acknowledgements Valentini, M., Laventie, B.J., Moscoso, J., Jenal, U. and Filloux, A. (2016) The diguanylate cyclase HsbD in- tersects with the HptB regulatory cascade to control Pseudomonas aeruginosa biofilm and motility. PLoS Genetics, 12, e1006354. Romling, U., Galperin, M.Y. and Gomelsky, M. (2013) Cyclic di-GMP: the first 25 years of a universal bacterial second messenger. Microbiology and Molecular Biology Reviews, 77, 1–52. Rotem, O., Nesper, J., Borovok, I., Gorovits, R., Kolot, M., Pasternak, Z., et al. (2015) An extended cyclic di-GMP network in the predatory bacterium Bdellovibrio bacteriov- orus. Journal of Bacteriology, 198, 127–137. Whitney, J.C., Beck, C.M., Goo, Y.A., Russell, A.B., Harding, B.N., De Leon, J.A., et al. (2014) Genetically distinct pathways guide effector export through the type VI secretion system. Molecular Microbiology, 92, 529–542. Russell, A.B., Hood, R.D., Bui, N.K., LeRoux, M., Vollmer, W. and Mougous, J.D. (2011) Type VI secretion delivers bac- teriolytic effectors to target cells. Nature, 475, 343–347. Whitney, J.C., Quentin, D., Sawai, S., LeRoux, M., Harding, B.N., Ledvina, H.E., et al. (2015) An interbacterial NAD(P) (+) glycohydrolase toxin requires elongation factor Tu for delivery to target cells. Cell, 163, 607–619. Russell, A.B., Singh, P., Brittnacher, M., Bui, N.K., Hood, R.D., Carl, M.A., et al. (2012) A widespread bacterial type VI secretion effector superfamily identified using a heuris- tic approach. Cell Host & Microbe, 11, 538–549. Wu, H.Y., Chung, P.C., Shih, H.W., Wen, S.R. and Lai, E.M. (2008) Secretome analysis uncovers an Hcp-family protein secreted via a type VI secretion system in Agrobacterium tumefaciens. Journal of Bacteriology, 190, 2841–2850. Russell, A.B., Peterson, S.B. and Mougous, J.D. (2014a) Type VI secretion system effectors: poisons with a pur- pose. Nature Reviews Microbiology, 12, 137–148. Wu, C.F., Lin, J.S., Shaw, G.C. and Lai, E.M. (2012) Acid-induced type VI secretion system is regulated by ExoR-ChvG/ChvI signaling cascade in Agrobacterium tu- mefaciens. PLoS Path, 8, e1002938. Russell, A.B., Wexler, A.G., Harding, B.N., Whitney, J.C., Bohn, A.J., Goo, Y.A., et al. (2014b) A type VI secre- tion-related pathway in Bacteroidetes mediates interbac- terial antagonism. Cell Host & Microbe, 16, 227–236. Wu, H.Y., Liu, K.H., Wang, Y.C., Wu, J.F., Chiu, W.L., Chen, C.Y., et al. (2014) AGROBEST: an efficient Agrobacterium- mediated transient expression method for versatile gene function analyses in Arabidopsis seedlings. Plant Methods, 10, 19. Sana, T.G., Hachani, A., Bucior, I., Soscia, C., Garvis, S., Termine, E., et al. Acknowledgements Quorum sensing control of Type VI secretion factors re- stricts the proliferation of quorum-sensing mutants. eLife, 5, e14712. Pukatzki, S., Ma, A.T., Revel, A.T., Sturtevant, D. and Mekalanos, J.J. (2007) Type VI secretion system trans- locates a phage tail spike-like protein into target cells McCarthy, R.R., Mazon-Moya, M.J., Moscoso, J.A., Hao, Y., Lam, J.S., Bordi, C., et al. (2017a) Cyclic-di-GMP © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648 648 R. R. McCarthy et al.  where it cross-links actin. Proceedings of the National Academy of Sciences of the United States of America, 104, 15508–15513. Skotnicka, D., Smaldone, G.T., Petters, T., Trampari, E., Liang, J., Kaever, V., et al. (2016) A minimal threshold of c-di-gmp is essential for fruiting body formation and sporulation in Myxococcus xanthus. PLoS Genetics, 12, e1006080. Quentin, D., Ahmad, S., Shanthamoorthy, P., Mougous, J.D., Whitney, J.C. and Raunser, S. (2018) Mechanism of load- ing and translocation of type VI secretion system effector Tse6. Nature Microbiology, 3, 1142–1152. Trampari, E., Stevenson, C.E., Little, R.H., Wilhelm, T., Lawson, D.M. and Malone, J.G. (2015) Bacterial rotary export ATPases are allosterically regulated by the nucle- otide second messenger cyclic-di-GMP. The Journal of Biological Chemistry, 290, 24470–24483. Records, A.R. and Gross, D.C. (2010) Sensor kinases RetS and LadS regulate Pseudomonas syringae type VI secre- tion and virulence factors. Journal of Bacteriology, 192, 3584–3596. Biological Chemistry, 290, 24470–24483. Trunk, K., Peltier, J., Liu, Y.C., Dill, B.D., Walker, L., Gow, N.A.R., et al. (2018) The type VI secretion system deploys antifungal effectors against microbial competitors. Nature Microbiology, 3, 920–931. Reen, F.J., Woods, D.F., Mooij, M.J., Adams, C. and O'Gara, F. (2012) Respiratory pathogens adopt a chronic lifestyle in response to bile. PLoS One, 7, e45978. Ueda, A. and Wood, T.K. (2010) Tyrosine Phosphatase TpbA of Pseudomonas aeruginosa controls extracellular DNA via cyclic diguanylic acid concentrations. Environmental Microbiology, 2, 449–455. Rojo, D., Mendez-Garcia, C., Raczkowska, B.A., Bargiela, R., Moya, A., Ferrer, M., et al. (2017) Exploring the human microbiome from multiple perspectives: factors altering its composition and function. FEMS Microbiology Reviews, 41, 453–478. Valentini, M. and Filloux, A. (2016) Biofilms and cyclic di- GMP (c-di-GMP) signaling: lessons from Pseudomonas aeruginosa and other bacteria. The Journal of Biological Chemistry, 291, 12547–12555. Romling, U. and Balsalobre, C. (2012) Biofilm infections, their resilience to therapy and innovative treatment strate- gies. Journal of Internal Medicine, 272, 541–561. Acknowledgements (2012) The second type VI secretion system of Pseudomonas aeruginosa strain PAO1 is reg- ulated by quorum sensing and Fur and modulates in- ternalization in epithelial cells. The Journal of Biological Chemistry, 287, 27095–27105. Xu, J., Kim, J., Koestler, B.J., Choi, J.H., Waters, C.M. and Fuqua, C. (2013) Genetic analysis of Agrobacterium tu- mefaciens unipolar polysaccharide production reveals complex integrated control of the motile-to-sessile switch. Molecular Microbiology, 89, 929–948. Sana, T.G., Flaugnatti, N., Lugo, K.A., Lam, L.H., Jacobson, A., Baylot, V., et al. (2016) Salmonella Typhimurium utilizes a T6SS-mediated antibacterial weapon to establish in the host gut. Proceedings of the National Academy of Sciences of the United States of America, 113, E5044–5051. Sheng, L., Lv, Y., Liu, Q., Wang, Q. and Zhang, Y. (2013) Connecting type VI secretion, quorum sensing, and c-di- GMP production in fish pathogen Vibrio alginolyticus through phosphatase PppA. Veterinary Microbiology, 162, 652–662. Supporting Information Additional supporting information may be found online in the Supporting Information section at the end of the article. © 2019 The Authors. Molecular Microbiology published by John Wiley & Sons Ltd, Molecular Microbiology, 112, 632–648
https://openalex.org/W2769910208	https://research-repository.uwa.edu.au/files/31589007/Chintalapudi_2018_Cacna2d1.pdf	English	null	Systems genetics identifies a role for Cacna2d1 regulation in elevated intraocular pressure and glaucoma susceptibility	Nature communications	2,017	cc-by	13,500	ARTICLE ARTICLE ARTICLE 1 Department of Ophthalmology, The Hamilton Eye Institute, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 2 Department of Pharmaceutical Sciences, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 3 Department of Pharmaceutics, College of Pharmacy, Mansoura University, Mansoura 35516, Egypt. 4 Department of Epidemiology and Biostatistics, Institute of Computational Biology, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 5 Department of Twin Research and Genetic Epidemiology, King’s College London, London WC2R 2LS, UK. 6 Department of Ophthalmology, Harvard Medical School, Massachusetts Eye and Ear Inﬁrmary, Boston, MA 02298, USA. 7 Department of Anatomy and Neurobiology, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 8 Department of Genetics, Genomics and Informatics, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. Correspondence and requests for materials should be addressed to M.M.J. (email: mjablonski@uthsc.edu) Systems genetics identiﬁes a role for Cacna2d1 regulation in elevated intraocular pressure and glaucoma susceptibility Sumana R. Chintalapudi1, Doaa Maria1,2,3, Xiang Di Wang1, Jessica N. Cooke Bailey 4, NEIGHBORHOOD consortium, International Glaucoma Genetics consortium, Pirro G. Hysi5, Janey L. Wiggs6, Robert W. Williams1,7,8 & Monica M. Jablonski 1,2,7 Glaucoma is a multi-factorial blinding disease in which genetic factors play an important role. Elevated intraocular pressure is a highly heritable risk factor for primary open angle glaucoma and currently the only target for glaucoma therapy. Our study helps to better understand underlying genetic and molecular mechanisms that regulate intraocular pressure, and iden- tiﬁes a new candidate gene, Cacna2d1, that modulates intraocular pressure and a promising therapeutic, pregabalin, which binds to CACNA2D1 protein and lowers intraocular pressure signiﬁcantly. Because our study utilizes a genetically diverse population of mice with known sequence variants, we are able to determine that the intraocular pressure-lowering effect of pregabalin is dependent on the Cacna2d1 haplotype. Using human genome-wide association study (GWAS) data, evidence for association of a CACNA2D1 single-nucleotide polymorph- ism and primary open angle glaucoma is found. Importantly, these results demonstrate that our systems genetics approach represents an efﬁcient method to identify genetic variation that can guide the selection of therapeutic targets. 1 Department of Ophthalmology, The Hamilton Eye Institute, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 2 Department of Pharmaceutical Sciences, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 3 Department of Pharmaceutics, College of Pharmacy, Mansoura University, Mansoura 35516, Egypt. 4 Department of Epidemiology and Biostatistics, Institute of Computational Biology, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 5 Department of Twin Research and Genetic Epidemiology, King’s College London, London WC2R 2LS, UK. 6 Department of Ophthalmology, Harvard Medical School, Massachusetts Eye and Ear Inﬁrmary, Boston, MA 02298, USA. 7 Department of Anatomy and Neurobiology, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. 8 Department of Genetics, Genomics and Informatics, The University of Tennessee Health Science Center, Memphis, TN 38163, USA. Correspondence and requests for materials should be addressed to M.M.J. (email: mjablonski@uthsc.edu) A full list of authors and their afﬂiations appears at the end of the paper. 1 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Stringent reﬁnement based on QTL mapping, correlation analyses, and SNPs was performed to identify positional candidate genes. Subcellular localization of candidates in mouse and human eyes were determined by immunohistochemistry. The NEIGHBORHOOD consortium database was used to identify SNPs within the candidate gene associated with elevated IOP and POAG in humans. The IOP-lowering effect of pregabalin, a drug with high afﬁnity for CACNA2D1 was evaluated as eye drops in D2, BXD48 (D allele), B6, and BXD14 (B allele) strains for validation Both POAG and IOP are highly heritable. In humans, IOP heritability is estimated to be ~55%. Moreover, the genetic risk of elevated IOP and POAG are partially shared5, 6, although some loci that are associated with POAG were not associated with IOP7. To date, multiple candidate IOP or POAG loci have been identiﬁed (e.g., TMCO1, CDKN2B-AS1, GAS7, CAV1/CAV2, SIX1/SIX6, TXNRD2, ATXN2, FOXC1 ABCA1, AFAP1, GMDS, PMM2, TGFBR3 ARHGEF12, FAM125B, FNDC3B, and the ABO blood group), however, their physiological roles are not well understood7–16. Identiﬁcation of additional gene variants that modulate IOP both in animals and humans is therefore likely to provide critical insights and new targets for therapeutic intervention. b C57BL/6J DBA/2J X 8 –log (P-value) 6 0 1 17 20 X 15 13 11 9 Chromosomes IOP lowering using candidate gene agonist/antagonist 7 8 6 5 4 3 2 2 4 62 BXDs, F1s, and progenitors BXD strains Phenotype (IOP) Quantitative trait locus Candidate gene Human GWAS Validation Systems genetics is an extension of complex trait analysis that examines large sets of genotypes and phenotypes to investigate the genetic basis of disease traits17–20. The BXD family is cur- rently the largest and best characterized mouse genetic reference population21. This family is an admixture of C57BL/6J and DBA/ 2J genomes, and the family are segregating for roughly 5.5 million sequence variants22. Over the last decade, the BXD family has become a key resource for systems genetics largely because they have been phenotyped so thoroughly and at many levels—from messenger RNA (mRNA) to maternal behavior17, 23–25. There is also extensive phenome data on the visual system of these strains26–30. One of the major advantages of the BXDs relative to the Collaborative Cross and other new resources is that the DBA/ 2J progenitor strain and many of the BXD progeny strains con- sistently develop high IOP between 6 and 10 months of age31. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Female C57BL/6J a b F1 F2 BXD1 C57BL/6J DBA/2J X 8 –log (P-value) 6 0 1 17 20 X 15 13 11 9 Chromosomes IOP lowering using candidate gene agonist/antagonist 7 8 6 5 4 3 2 2 4 62 BXDs, F1s, and progenitors BXD strains Phenotype (IOP) Quantitative trait locus Candidate gene Human GWAS Validation BXD2 BXD102 + + . . . male DBA/2J Female C57BL/6J a F1 F2 BXD1 BXD2 BXD102 + + . . . male DBA/2J P rimary open angle glaucoma (POAG) is a leading cause of blindness worldwide1. The disease is characterized by progressive optic nerve damage arising from apoptotic cell death of retinal ganglion cells. Elevated intraocular pressure (IOP) is one of the most signiﬁcant risk factors contributing to visual ﬁeld loss in this disease. Steady-state IOP is generated by the balance of aqueous humor (AH) production by the ciliary body (CB) and AH drainage through the trabecular meshwork (TM; conventional pathway), and to a lesser degree the uveoscleral or nonconventional pathway. An imbalance between the inﬂow and outﬂow of AH leads to a change in IOP2–4. Gene variants inﬂuence an individual’s likelihood of developing glaucoma, the rate of disease progression, and how a patient responds to treatment. Because IOP can be medically controlled, IOP reduc- tion is the ﬁrst-line therapeutic option in glaucoma. Current medications do not address the underlying pathologies that lead to elevated IOP, nor do they address the many potential sources of variation related to IOP modulation. P a b F1 F2 BXD1 C57BL/6J DBA/2J X 8 –log (P-value) 6 0 1 17 20 X 15 13 11 9 Chromosomes IOP lowering using candidate gene agonist/antagonist 7 8 6 5 4 3 2 2 4 62 BXDs, F1s, and progenitors BXD strains Phenotype (IOP) Quantitative trait locus Candidate gene Human GWAS Validation BXD2 BXD102 + + . . . Fig. 1 Overview of IOP systems genetics analyses. a BXD strains derived by crossing C57BL/6J (B6) and DBA/2J (D2) parents. The resulting heterozygous F1 mice were crossed to generate genetically diverse but non- reproducible F2 animals. These F2 progeny were iteratively inbred until generation F20+, at which point each strain represents a unique mosaic of B and D alleles. b Workﬂow for IOP systems genetics analysis. IOP was measured from 65 BXD strains aged 9.1–13 months. Genomic regions modulating IOP were identiﬁed using QTL analyses. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Therefore, they are an ideal resource to discover gene variants that modulate both IOP and glaucoma. g In this study, we systematically measure IOP across a large subset of the BXD family in multiple age cohorts. Using stringent stepwise reﬁnement based on expression quantitative trait locus (QTL) mapping, correlation analyses (direct and partial Pearson test), and the analysis of single-nucleotide polymorphisms (SNPs), we identify a candidate gene that modulates IOP, and we combine mouse and human genetic data in an effort to validate the candidate gene. Lastly, we evaluate the IOP-lowering effect of a drug speciﬁcally targeted to the candidate protein product. Collectively, our study ﬁnds that Cacna2d1 modulates IOP and that blocking the function of its gene product, CACNA2D1, with pregabalin reduces IOP in a dose-dependent and haplotype- speciﬁc manner. Fig. 1 Overview of IOP systems genetics analyses. a BXD strains derived by crossing C57BL/6J (B6) and DBA/2J (D2) parents. The resulting heterozygous F1 mice were crossed to generate genetically diverse but non- reproducible F2 animals. These F2 progeny were iteratively inbred until generation F20+, at which point each strain represents a unique mosaic of B and D alleles. b Workﬂow for IOP systems genetics analysis. IOP was measured from 65 BXD strains aged 9.1–13 months. Genomic regions modulating IOP were identiﬁed using QTL analyses. Stringent reﬁnement based on QTL mapping, correlation analyses, and SNPs was performed to identify positional candidate genes. Subcellular localization of candidates in mouse and human eyes were determined by immunohistochemistry. The NEIGHBORHOOD consortium database was used to identify SNPs within the candidate gene associated with elevated IOP and POAG in humans. The IOP-lowering effect of pregabalin, a drug with high afﬁnity for CACNA2D1 was evaluated as eye drops in D2, BXD48 (D allele), B6, and BXD14 (B allele) strains for validation ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 this approach in previous studies to identify gene modulators of various traits in BXD mice26, 27, 29, 30, 32, 33. this approach in previous studies to identify gene modulators of various traits in BXD mice26, 27, 29, 30, 32, 33. across its genome (Fig. 1a), which allows for an inﬁnite source of genetically identical test subjects. IOP is one factor that can inﬂuence glaucoma risk. To identify a gene candidate that modulates IOP, we utilized a systems genetics, pharmacological, and translational approach (Fig. 1b). We have successfully used Trait and genetic variation across BXD mice. IOP varied over twofold in 9.1–13-month-old BXD strains (range of 9.6 ± 0.92 a c e g f h b d Average IOP (mmHg) LRS Enrichment score (ES) 22 WT/WT at Tyrp1 and Gpnmb WT/MUT at Tyrp1 and Gpnmb MUT/WT at Tyrp1 and Gpnmb MUT/MUT at Tyrp1 and Gpnmb Progenitors and F1 18 14 20 Megabases Candidate gene selection criteria 1. Location within the confidence interval of the peak eQTL 2. Cis-modulated gene 3. Significant correlation (linear and partial) between expression of the gene and IOP 4. Functions within a network that could explain its role in modulating IOP 5. Presence of sequence variants between parental strains at/near the region of the gene 6. Expression in the eye and localization to a structure associated with modulation of IOP 7. Association with human POAG or IOP through GWAS or standard linkage studies 8. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Biological association with glaucoma or its treatment 1 2 3 4 5 6 7 8 9 10 11 12 14 15 16 17 18 19 X 13 15 Tyrp1 Gpnmb 10 5 0.200 0.225 FDR q-value=0.004 Nominal P-value=0.002 NES=2.43 0.175 0.150 0.125 0.100 0.075 0.050 0.000 0.025 –0.025 10 6 38 20 02 42 48 06 50 71 31 05 39 101 13 65 40 84 32 65b 19 09 12 60 89 62 29 11 36 75 55 102 90 DBA/2J 68 B6D2F1 67 77 34 14 56 98 66 27 99 61 01 18 33 73 51 43 95 63 69 24 08 16 86 100 87 45 70 15 83 85 C57BL/6J Male IOP (mmHg) Average IOP (mmHg) 24 P<0.0001, R2=0.37 P=0.13, R2=0.09 20 20 Female IOP (mmHg) Genotype 18 16 16 8 8 WT/WT WT/MUT MUT/WT 12.5 Ion channel activity 2.24 2.12 1.98 1.91 1.94 1.92 Substrate-specific channel activity Metal ion transmembrane transporter activity Ion transmembrane transporter activity Transmembrane receptor activity Receptor activity 15.0 17.5 Megabases 20.0 Cacna2d1 MUT/MUT 10 1.5 1.0 0.5 0.0 14 8 10 16 14 18 20 22 12 12 12 4 GO cellular compartment: plasma membrane a Average IOP (mmHg) 22 WT/WT at Tyrp1 and Gpnmb WT/MUT at Tyrp1 and Gpnmb MUT/WT at Tyrp1 and Gpnmb MUT/MUT at Tyrp1 and Gpnmb Progenitors and F1 18 14 10 6 38 20 02 42 48 06 50 71 31 05 39 101 13 65 40 84 32 65b 19 09 12 60 89 62 29 11 36 75 55 102 90 DBA/2J 68 B6D2F1 67 77 34 14 56 98 66 27 99 61 01 18 33 73 51 43 95 63 69 24 08 16 86 100 87 45 70 15 83 85 C57BL/6J b Male IOP (mmHg) 24 P<0.0001, R2=0.37 20 Female IOP (mmHg) 16 8 8 10 16 14 18 20 22 12 12 4 b a Female IOP (mmHg) d Average IOP (mmHg) P=0.13, R2=0.09 20 ( g) Genotype 18 16 8 WT/WT WT/MUT MUT/WT MUT/MUT 10 14 12 d Average IOP (mmHg) P=0.13, R2=0.09 20 18 16 8 10 14 12 c LRS 20 Megabases 1 2 3 4 5 6 7 8 9 10 11 12 14 15 16 17 18 19 X 13 15 Tyrp1 Gpnmb 10 5 1.5 1.0 0.5 0.0 d c e Candidate gene selection criteria 1. Results d d Study design. A major advantage of the genetically diverse BXD family over non-inbred strains is that each strain is homozygous NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 2 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 3e) with 1056 SNPs and 30 insertions/deletions between the parent strains (Supplementary Data 2). These variants segregate among the BXD strains based upon the haplotype of the gene. Similarly, in humans, CACNA2D1 is highly polymorphic and has several splice variants36–38. Selection of positional candidates in the chromosome 5 locus. To determine the candidate gene variants that modulate IOP within the Chr 5 locus, we used the following stringent criteria (Fig. 2e): (1) the gene is located within the conﬁdence interval of the peak eQTL; (2) the gene has cis-modulation; (3) the expres- sion level of the gene across BXD strains is signiﬁcantly correlated with elevated IOP using both linear correlation and partial Pearson correlation analyses; (4) the gene functions within a network that could explain its role in modulating IOP; (5) the gene has sequence variants between parental strains at/near the region of the gene; (6) the gene is expressed in the eye and localized to an area associated with modulation of IOP; (7) the gene is associated with human POAG and/or elevated IOP either through GWAS or standard linkage studies; and (8) the gene has a biological association with glaucoma or its treatment. p We further sought to determine if the Cacna2d1 haplotype inﬂuenced the baseline IOP in BXD mice. Similar to the inﬂuence of the parental allele on Cacna2d1 expression (Fig. 3b), we found a distinct segregation of IOP values among the BXD strains that was dependent upon the haplotype of Cacna2d1 that was age- dependent. The Cacna2d1 haplotype had no signiﬁcant effect on IOP in younger BXD strains (P = 0.34, age 1–2.1 months) (Fig. 3f, left). In contrast, in older age (aged 9.1–13 months), strains with Cacna2d1 inherited from the B6 parent (B allele) had signiﬁcantly higher IOP than those with the D allele (P = 0.0008, Fig. 3f, right). Within the QTL peak at Chr 5, there were 25 positional gene candidates that were cis-regulated (Supplementary Table 1). Using our above criteria, calcium channel, voltage-dependent, α2δ1subunit (Cacna2d1) emerged as the single best positional candidate (r = 0.440; P = 0.0003; Fig. 2f). No other positional candidate fulﬁlled our selection criteria (Supplementary Table 2). Gene set enrichment analyses (GSEA; Broad Institute35) determined that nominally signiﬁcant gene correlates of Cac- na2d1 (Supplementary Data 1) were signiﬁcantly enriched for plasma membrane localization (Fig. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 mm Hg in BXD38 to 19.38 ± 0.43 mm Hg in BXD14; Fig. 2a). IOP heritability was 31%, which is comparable to that of other ocular traits to which QTL mapping has been applied successfully27, 29, 32–34. There was no statistically signiﬁcant difference in IOP between the sexes (Fig. 2b). Cacna2d1 likely functions within pathways that critically inﬂuence ion channels and/or their receptors within the eye that affect IOP regulation. Gene haplotypes of Cacna2d1 contribute to variations in IOP. The expression of the Cacna2d1 transcript in the whole eye varied signiﬁcantly among BXD strains (range of 6.9 ± 0.2 in BXD71 to 9.7 ± 0.3 in BXD48a), which is a 5.6-fold difference in mRNA expression (Fig. 3a). BXD strains with the D parental allele have lower expression of Cacna2d1 in the eye than those strains with the B parental allele (Fig. 3a, b, P = 1.2 × 10−32). Cacna2d1 mapped as a signiﬁcant cis-eQTL (LRS = 143; LOD = 31.01) on proximal Chr 5 at 14.3 Mb, which is within 5 Mb of location of the gene itself (Fig. 3c), making it a cis-regulated gene. g The variation in IOP was mapped to a narrow locus on proximal Chr 5 (likelihood ratio statistic (LRS) = 19.6, logarithm of the odds (LOD) = 4.25, Fig. 2c). This QTL was not associated with IOP in studies using younger sex-matched BXD strains, suggesting that the QTL contains gene variants that contribute to variation in IOP at an older age29. In addition, the location of this LRS interval does not overlap with the location of either Tyrp1 or Gpnmb (Fig. 2c). As conﬁrmation that the well-documented mutations in Tyrp1 and Gpnmb do not inﬂuence IOP in the BXD GRP, IOP also maps to the same location on Chr 5 after we exclude strains that harbor both mutations or harbor only one of the mutations (Supplementary Fig. 1). Moreover, the complete overlap of IOP distributions for all possible genotypes at those two loci (P = 0.13; Fig. 2d) further supports this claim. g g g g g Cacna2d1 encodes for a preproprotein that is cleaved into multiple chains that form the alpha-2 and delta subunits of the voltage-dependent calcium channel complex. CACNA2D1 is a glycosylphosphatidylinositol (GPI)-anchored subunit typically associated with the Cavα1 pore within L-type calcium channels in smooth muscle (Fig. 3d). The gene is highly polymorphic (Fig. Fig. 2 Association analyses reveal Cacna2d1 as a candidate for IOP modulation. a IOP levels vary among the BXD strains. IOP of BXD strains carrying wild-type alleles of Tyrp1 and Gpnmb (WT/WT; blue bars), wild-type Tyrp1 and mutant Gpnmb (WT/MUT; green bars), mutant Tyrp1 and WT Gpnmb (MUT/WT; yellow bars), and mutant alleles of both genes (MUT/MUT; red bars) are shown. n = 65 strains, age = 9.1–13 months. Values denote IOP levels on mmHg scale (mean ± SEM). b No differences in IOP between sexes. There is a strong statistical association between the IOP of males and females (P < 0.0001). The relationship between the sexes is positively correlated (R2 = 0.37, Pearson correlation coefﬁcient = 0.62). c Genetic interval mapping revealed a single signiﬁcant eQTL on proximal Chr 5. This is distinct from the locations of Tyrp1 and Gpnmb in the genome (black vertical lines on Chr 4 and Chr 6, respectively). d Tyrp1 and Gpnmb haplotypes do not inﬂuence IOP in BXD mice. The scatter plot shows average IOP of BXD strains carrying wild-type alleles of Tyrp1 and Gpnmb (WT/WT), wild-type allele of Tyrp1 and mutant allele of Gpnmb (MUT/WT), mutant allele of Tyrp1 and wild- type allele of Gpnmb (WT/MUT), and mutant alleles of both genes (MUT/MUT) (n = 65, P = 0.13, age = 9.1–13 months). P-value was calculated using an ANOVA. e Stringent selection criteria for selecting candidate genes. f A signiﬁcant QTL for IOP is present on chromosome 5 between 14–19 Mb. Cacna2d1 is the strongest candidate gene located in the peak QTL. g Gene set enrichment analysis for positive correlates of Cacna2d1. Gene correlates of Cacna2d1 were signiﬁcantly enriched for localization to the plasma membrane. False discovery rate (FDR) cutoff was set as q ≤0.25. h Gene set enrichment analysis of genes positively correlated with Cacna2d1 presented as molecular function groupings. Categories that are statistically over-represented are shown with their normalized enrichment score (NES) listed next to the bars NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Location within the confidence interval of the peak eQTL 2. Cis-modulated gene 3. Significant correlation (linear and partial) between expression of the gene and IOP 4. Functions within a network that could explain its role in modulating IOP 5. Presence of sequence variants between parental strains at/near the region of the gene 6. Expression in the eye and localization to a structure associated with modulation of IOP 7. Association with human POAG or IOP through GWAS or standard linkage studies 8. Biological association with glaucoma or its treatment f 12.5 15.0 17.5 Megabases 20.0 Cacna2d1 f e h Ion channel activity Number of genes 0 10 30 50 2.24 2.12 1.98 1.91 1.94 1.92 Substrate-specific channel activity Metal ion transmembrane transporter activity Ion transmembrane transporter activity Transmembrane receptor activity Receptor activity g Enrichment score (ES) 0.200 0.225 FDR q-value=0.004 Nominal P-value=0.002 NES=2.43 0.175 0.150 0.125 0.100 0.075 0.050 0.000 (Pos. correlated with Cacna2d1) (Neg. correlated with Cacna2d1) 0.025 –0.025 GO cellular compartment: plasma membrane g h NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 a c e b d f 11 Strains with D allele Strains with B allele Parents and F1s 10 9 8 7 6 71 11 38 29 09 48 16 27 89 70 60 75 85 05 42 20 31 43 45 19 62 06 64 02 84 12 40 69 36 98 63 01 90 32 51 68 33 86 66 21 28 15 13 23 99 08 25 14 77 61 22 65 67 44 34 39 24 55 87 18 56 50 83 48a (96) 73 C57BL/6J 73a (80) B6D2F1 D2B6F1 65a (97) 65b (92) DBA/2J Cacna2d1 expression (log2) Cacna2d1 expression (log2) 10 P=1.2 ×10–32 9 8 7 6 Genotype s-s Ca 2+ α2 α1 δ γ β Extracellular Intracellular Young BXDs Old BXDs 10 15 20 P=0.34 P=0.0008 5 Genotype B D B D IOP (mmHg) BXD strains Megabases 140 1 2 3 4 5 6 9 10 11 12 13 14 16 17 18 19 0.7 0.6 0.4 0.2 0.0 X 15 8 7 60 20 InDels (D2) SNPs (D2) RefSeq gene 100 LRS B D g. 3 Association analyses of Cacna2d1 haplotype variants. a Cacna2d1 expression in the whole eye varies across BXD strains. The bars depict a range of xpression values from 6.9 ± 0.03 and 9.7 ± 0.10 (mean±SEM). On the Y-axis, Cacna2d1 expression is on a log2 scale. Parental strains (black), F1s (black), parental allele Cacna2d1 (green), and B parental allele of Cacna2d1 (red) are on X-axis. b Cacna2d1 haplotype inﬂuences Cacna2d1 expression level in eyes BXD strains. BXD strains with the D parental allele have lower expression of Cacna2d1 in the eye, while those with the B parental allele have higher xpression of the gene (n = 70; P < 0.0001). P-value was calculated using an ANOVA. c Genetic mapping revealed a single highly signiﬁcant cis-eQTL for acna2d1 on Chr 5. The purple triangle indicates the location of Cacna2d1 within the mouse genome. d Subunit assembly of voltage-gated calcium channels. raphic representation of the high voltage-activated calcium channel complex consisting of the main pore forming α1 (blue) subunit plus ancillary, β (blue), (blue), and α2 (red) and δ1 (red) subunits. α2 and δ1 subunits have a disulphide bond between them. The δ1 unit is GPI-anchored to the plasma membrane. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 BXD strains with the D parental allele have lower expression of Cacna2d1 in the eye, while those with the B parental allele have higher expression of the gene (n = 70; P < 0.0001). P-value was calculated using an ANOVA. c Genetic mapping revealed a single highly signiﬁcant cis-eQTL for Cacna2d1 on Chr 5. The purple triangle indicates the location of Cacna2d1 within the mouse genome. d Subunit assembly of voltage-gated calcium channels. Graphic representation of the high voltage-activated calcium channel complex consisting of the main pore forming α1 (blue) subunit plus ancillary, β (blue), γ (blue), and α2 (red) and δ1 (red) subunits. α2 and δ1 subunits have a disulphide bond between them. The δ1 unit is GPI-anchored to the plasma membrane. e UCSC Genome Browser illustration of gene structure, and location of SNPs and InDels in Cacna2d1 on Chr 5 of the mouse genome. Reference Sequence mRNA is represented in blue. f Effect of Cacna2d1 haplotype on IOP is age-dependent. In this scatter plot, parental allele of Cacna2d1 does not signiﬁcantly inﬂuence IOP in young BXDs (P = 0.34, n = 63, age = 1–2.1 months; left). In contrast, BXD strains carrying the B parental allele of Cacna2d1 have higher IOP, while those with the D parental allele have lower IOP in older mice (P = 0.0008, n = 64, age = 9.1–13 months; right). P-value was calculated using an ANOVA. F1s were not included in these analyses signiﬁcant association for these eight SNPs was not found in the IOP metadata, possibly due to different genetic effects in the multiple ethnicities included in that study. was highly expressed in the non-pigmented epithelium. Weak labeling was present in the CM. A similar pattern of expression was observed in the TM and CB of the human donor eye (Fig. 4d–f). CACNA2D1 is localized to the CB and TM in mice and humans. To further test Cacna2d1 as a candidate IOP- modulating gene, we performed immunohistochemistry to determine the localization pattern of CACNA2D1 in healthy mouse and human donor eyes. In the mouse eye, CACNA2D1 is prominently localized to the TM, CB, and ciliary muscle (CM) (Fig. 4a–c). CACNA2D1 was observed in a punctate pattern throughout the TM and Schlemm’s canal. In the CB, CACNA2D1 IOP response to pregabalin depends on Cacna2d1 genotype. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 UCSC Genome Browser illustration of gene structure, and location of SNPs and InDels in Cacna2d1 on Chr 5 of the mouse genome. Reference Sequence RNA is represented in blue. f Effect of Cacna2d1 haplotype on IOP is age-dependent. In this scatter plot, parental allele of Cacna2d1 does not signiﬁcantly ﬂuence IOP in young BXDs (P = 0.34, n = 63, age = 1–2.1 months; left). In contrast, BXD strains carrying the B parental allele of Cacna2d1 have higher IOP, hile those with the D parental allele have lower IOP in older mice (P = 0.0008, n = 64, age = 9.1–13 months; right). P-value was calculated using an NOVA. F1s were not included in these analyses a 11 Strains with D allele Strains with B allele Parents and F1s 10 9 8 7 6 71 11 38 29 09 48 16 27 89 70 60 75 85 05 42 20 31 43 45 19 62 06 64 02 84 12 40 69 36 98 63 01 90 32 51 68 33 86 66 21 28 15 13 23 99 08 25 14 77 61 22 65 67 44 34 39 24 55 87 18 56 50 83 48a (96) 73 C57BL/6J 73a (80) B6D2F1 D2B6F1 65a (97) 65b (92) DBA/2J Cacna2d1 expression (log2) BXD strains b Cacna2d1 expression (log2) 10 P=1.2 ×10–32 9 8 7 6 Genotype B D b a c s a s Megabases 140 1 2 3 4 5 6 9 10 11 12 13 14 16 17 18 19 0.7 0.6 0.4 0.2 0.0 X 15 8 7 60 20 100 LRS d s-s Ca 2+ α2 α1 δ γ β Extracellular Intracellular 7 6 4 2 0 d c e InDels (D2) SNPs (D2) RefSeq gene f Young BXDs Old BXDs 10 15 20 P=0.34 P=0.0008 5 Genotype B D B D IOP (mmHg) e InDels (D2) SNPs (D2) RefSeq gene f e Fig. 3 Association analyses of Cacna2d1 haplotype variants. a Cacna2d1 expression in the whole eye varies across BXD strains. The bars depict a range of expression values from 6.9 ± 0.03 and 9.7 ± 0.10 (mean±SEM). On the Y-axis, Cacna2d1 expression is on a log2 scale. Parental strains (black), F1s (black), D parental allele Cacna2d1 (green), and B parental allele of Cacna2d1 (red) are on X-axis. b Cacna2d1 haplotype inﬂuences Cacna2d1 expression level in eyes of BXD strains. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 2g) and for functions related to neuronal function and excitability, including ion channel activity, substrate-speciﬁc channel activity, metal ion transmem- brane transporter activity, ion transmembrane transporter activity, and receptor activity (Fig. 2h). These results suggest that To further assess the strength of this candidate gene, we evaluated human POAG GWAS data from the NEIGHBOR- HOOD consortium (N = 3853 cases and 33 480 controls)16. A total of 1520 SNPs in the human genomic region that includes CACNA2D1 (chromosome 7:81.9–82.4 Mb) were evaluated, identifying nominal association for POAG (P < 0.05) for 44 SNPs (Supplementary Table 3), with the lead SNP (rs2299184 [A], P = 0.001, odds ratio = 1.15) located in intron 1 near DNaseI hypersensitivity sites annotated by the Encyclopedia of DNA Elements (ENCODE) as active in non-pigmented ciliary epithe- lium. Eight of the 44 human CACNA2D1 SNPs showing nominal association in the NEIGHBORHOOD POAG GWAS were also included in a multi-ethnic IOP association study9, however, NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 4 4 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Based upon our data suggesting that Cacna2d1 modulates IOP, we evaluated the ability of pregabalin, a gabapentinoid drug with high speciﬁcity for CACNA2D1, to affect IOP. Our data demonstrate that pregabalin ophthalmic eye drops reduce IOP in mice in a dose-dependent manner (Fig. 5a). The percent reduc- tion in IOP for both treated and control eyes after application of NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Iridocorneal angle Mouse Human a d c b e f Fig. 4 Cellular localization of CACNA2D1 in C57BL/6 J mouse and human donor eyes. Cellular localization of CACNA2D1 in C57BL/6J mouse and human donor eyes. a–f Sections from C57BL/6J mouse (a–c) and human donor (d–f) iridocorneal angle were labeled with anti-CACNA2D1 antibodies. CACNA2D1 (green) is localized in the ciliary body (CB), trabecular meshwork (TM), Schlemm’s canal (SC), and ciliary muscle (CM) in both B6 mice and human donor eyes. CACNA2D1 was also present in the posterior pigmented epithelium of the iris. n = 2 mice and 1 human. Scale: 100 µm. Blue nuclei, CB ciliary body, CM ciliary muscle, CR cornea, IR iris, SC Schlemm’s canal, TM trabecular meshwork Iridocorneal angle Mouse a c b g Mouse a b a c Human d e f Human d e d Human Fig. 4 Cellular localization of CACNA2D1 in C57BL/6 J mouse and human donor eyes. Cellular localization of CACNA2D1 in C57BL/6J mouse and human donor eyes. a–f Sections from C57BL/6J mouse (a–c) and human donor (d–f) iridocorneal angle were labeled with anti-CACNA2D1 antibodies. CACNA2D1 (green) is localized in the ciliary body (CB), trabecular meshwork (TM), Schlemm’s canal (SC), and ciliary muscle (CM) in both B6 mice and human donor eyes. CACNA2D1 was also present in the posterior pigmented epithelium of the iris. n = 2 mice and 1 human. Scale: 100 µm. Blue nuclei, CB ciliary body, CM ciliary muscle, CR cornea, IR iris, SC Schlemm’s canal, TM trabecular meshwork The Cacna2d1 haplotype also inﬂuenced the drug response. Figure 5b illustrates the percent reduction in IOP after application of 0.9% pregabalin eye drops into the eyes of DBA/ 2J, BXD48 (D allele), C57BL/6J, and BXD14 (B allele) mice. Table 2 lists the pharmacodynamic parameters after application of 0.9% pregabalin eye drops and statistical evaluation of the data, respectively. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Mice with the B allele of Cacna2d1 (i.e., B6 and BXD14) were more responsive to pregabalin (0.9%) eye drops than mice with the D allele (i.e., D2 and BXD48). In addition to larger reductions in IOP, mice with the B allele of Cacna2d1 had larger Tmax, Tend, and area under the curve (AUCtotal) than mice with the D allele. Expanding this analysis to an additional species, we observed a similar IOP-lowering response (22.1 ± 2.8%) in Dutch belted rabbits after instillation of 0.9% pregabalin eye drops (Fig. 5c). Table 3 lists the pharmacodynamic parameters and statistical evaluation of the data after application of 0.9% pregabalin eye drops into the eyes of Dutch belted rabbits. our ophthalmic pregabalin eye drops (0.3–1.2%) are shown in Fig. 5a. Table 1 lists the pharmacodynamic parameters after application of pregabalin eye drops and statistical evaluation of the data, respectively. Drops containing 0.3% drug provided no IOP-lowering effect compared to control. All other concentra- tions of drug reduced IOP in a dose-dependent manner. A pla- teau was reached at 0.9% and there was no signiﬁcant difference in drug response between 0.9 and 1.2% drug (P > 0.05). There was no signiﬁcant difference between the time of maximum response (Tmax) values of 0.6–1.2% concentrations of drug (P > 0.05). In contrast, there was a signiﬁcant difference between the time required for IOP to return to baseline (Tend) for all con- centrations of pregabalin eye drops, (P < 0.0001). The 1.2% pregabalin eye drops extended the duration of the IOP-lowering effect of pregabalin above that obtained with 0.6% pregabalin. Because there was no signiﬁcant difference in the percent reduction of IOP between 0.9 and 1.2% pregabalin, we selected 0.9% as the minimal concentration required to produce the maximum reduction in IOP. our ophthalmic pregabalin eye drops (0.3–1.2%) are shown in Fig. 5a. Table 1 lists the pharmacodynamic parameters after application of pregabalin eye drops and statistical evaluation of the data, respectively. Drops containing 0.3% drug provided no IOP-lowering effect compared to control. All other concentra- tions of drug reduced IOP in a dose-dependent manner. A pla- teau was reached at 0.9% and there was no signiﬁcant difference in drug response between 0.9 and 1.2% drug (P > 0.05). There was no signiﬁcant difference between the time of maximum response (Tmax) values of 0.6–1.2% concentrations of drug (P > 0.05). NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Other positional candidates, such as Crygn and Klhl7, have been associated with other ocular diseases42, 43, but not glaucoma. a We have validated Cacna2d1 using multiple stringent criteria and identiﬁed human SNPs in the CACNA2D1 genomic region that are nominally associated with POAG. Furthermore, we have demonstrated that both baseline IOP and steady-state expression of Cacna2d1 are dependent upon the parental allele of the gene in mice. As a corollary, we further show that the response to pregabalin, a gabapentinoid drug with high afﬁnity for CACNA2D1, is also dependent upon the Cacna2d1 haplotype. This is a bidirectional study that has successfully identiﬁed an IOP-modulating candidate gene in glaucoma and demonstrated a differential response to therapy. Our results further underscore the potential value of using the BXD genetic reference panel as a powerful tool to study human disease. Collectively, our data identify CACNA2D1 as a modulator of IOP and provide an avenue for a precision medicine approach to glaucoma therapy. 2 b c 0 0 10 10 –10 –10 –20 –20 –30 –30 –40 –40 –40 0 0 0 5 5 5 Time (h) Time (h) Time (h) 7 6 CTL TX 6 6 CTL CTL CTL CTL TX TX TX TX 4 4 4 3 3 3 2 2 2 1 1 1 IOP reduction (%) IOP reduction (%) D2 D2 BXD48 BXD48 B6 B6 BXD14 BXD14 Fig. 5 Cacna2d1 haplotype effects IOP-lowering potency of pregabalin, a speciﬁc modulator of CACNA2D1, lowers IOP in B6 mice in a dose- dependent manner. a We measured a dose-dependent reduction in IOP compared to control treatment after a single application of ophthalmic formulation containing a range of concentrations of pregabalin from 0.3 to 1.2%. The minimal concentration required to produce the maximum reduction in IOP is 0.9% (mean ± SEM; n = 6). Statistical details are provided in Table 1. b Pregabalin-induced IOP reduction is haplotype- speciﬁc. Mice carrying the B parental allele of Cacna2d1 (i.e., B6 and BXD14) are more responsive to pregabalin (0.9%) than mice carrying the D parental allele (i.e., D2 and BXD48; mean ± SEM; n = 4–6). Statistical details are provided in Table 2. c A single dose of pregabalin eye drops (0.9%) lowers IOP by 22.1% in Dutch belted rabbits. n = 5. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 The minimal concentration required to produce the maximum reduction in IOP is 0.9% (mean ± SEM; n = 6). Statistical details are provided in Table 1. b Pregabalin-induced IOP reduction is haplotype- speciﬁc. Mice carrying the B parental allele of Cacna2d1 (i.e., B6 and BXD14) are more responsive to pregabalin (0.9%) than mice carrying the D parental allele (i.e., D2 and BXD48; mean ± SEM; n = 4–6). Statistical details are provided in Table 2. c A single dose of pregabalin eye drops (0.9%) lowers IOP by 22.1% in Dutch belted rabbits. n = 5. Statistical details are provided in Table 3 We show that CACNA2D1 is expressed in both CB and TM, the sites of AH production, and outﬂow through the traditional pathway, respectively, thus suggesting a possible role in AH dynamics. It is also modestly present in the CM, a structure that directly affects the distention of the TM in an antagonistic manner3. We also demonstrate that the Cacna2d1 likely functions in a network that includes other modulators of metal ion channel and transport activity. Based upon these collective data, we pro- pose a model for the role of CACNA2D1 in regulating IOP (Fig. 6). Pregabalin, or other gabapentinoid drugs, binds to the CACNA2D1 subunit of the calcium channel, the afﬁnity of which varies depending on the haplotype of the gene. The binding of the drug mitigates the ﬂux of Ca2+ through the α1 pore of the calcium channel, reducing the level of intracellular Ca2+. A reduction in Ca2+ could cause a concomitant reduction in ion transport, water NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Statistical details are provided in Table 3 b 0 10 –10 –20 –30 –40 0 5 Time (h) 6 CTL CTL CTL CTL TX TX TX TX 4 3 2 1 IOP reduction (%) D2 D2 BXD48 BXD48 B6 B6 BXD14 BXD14 b Previous studies have demonstrated a link between Ca2+, calcium channels and glaucoma44–53, providing a historical backdrop for our discovery of the role of Cacna2d1 in IOP modulation. Because Ca2+ likely plays a role in the patho- physiology of glaucoma45, systemic calcium channel blockers (CCBs) (e.g., verapamil, diltiazem, or nimodipine) have been evaluated as plausible therapies for POAG. However, the out- comes of these investigations have been inconsistent, with some studies demonstrating that CCBs are effective in lowering IOP, protecting ganglion cells, increasing ocular blood ﬂow and improving visual function while others fail to replicate those results46–53. In all studies performed to date, however, CCBs have targeted the α1 pore of the Ca2+ channel and have used a systemic route of administration. None of the previous investigations evaluated modulators of the Ca2+ channel auxiliary subunits, such as CACNA2D1, nor have they evaluated topical delivery of the drug. In our study, we demonstrate that targeting CACNA2D1 with topical pregabalin lowers IOP up to 30% in mice with the B haplotype of the gene, which is a signiﬁcantly higher effect than that demonstrated using systemically administered traditional CCBs. c 0 10 –10 –20 –30 –40 0 5 Time (h) CTL TX 6 4 3 2 1 IOP reduction (%) c Previous human genomic studies (GWAS and linkage studies) revealed multiple genomic regions that associated with elevated IOP and/or POAG7, 9–11, 16. Two regions of interest lie on Chr 7q near the locations of CACNA2D154–56 and CAV1/CAV27. Other studies have demonstrated that SNPs in Cacna2d1/CACNA2D1 are associated with short QT syndrome38, perifosine cytotoxi- city57, altered sensitivity to acute noxious heat58, and beef quality traits59, indicating that polymorphisms in this gene have direct and distinct physiological consequences, some of which may be associated with modulation of IOP. Fig. 5 Cacna2d1 haplotype effects IOP-lowering potency of pregabalin, a speciﬁc modulator of CACNA2D1, lowers IOP in B6 mice in a dose- dependent manner. a We measured a dose-dependent reduction in IOP compared to control treatment after a single application of ophthalmic formulation containing a range of concentrations of pregabalin from 0.3 to 1.2%. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 In contrast, there was a signiﬁcant difference between the time required for IOP to return to baseline (Tend) for all con- centrations of pregabalin eye drops, (P < 0.0001). The 1.2% pregabalin eye drops extended the duration of the IOP-lowering effect of pregabalin above that obtained with 0.6% pregabalin. Because there was no signiﬁcant difference in the percent reduction of IOP between 0.9 and 1.2% pregabalin, we selected 0.9% as the minimal concentration required to produce the maximum reduction in IOP. NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 6 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 ARTICLE a b c 0 0 0 10 10 10 –10 –10 –10 –20 –20 –20 –30 –30 –30 –40 –40 –40 0 0 0 5 5 5 Time (h) Time (h) Time (h) 7 6 CTL TX 6 6 CTL CTL CTL CTL TX TX TX TX CTL CTL CTL TX CTL TX TX TX 4 4 4 3 3 3 2 2 2 1 1 1 IOP reduction (%) IOP reduction (%) IOP reduction (%) 0.3% 0.3% 0.6% 0.6% 0.9% 0.9% 1.2% 1.2% D2 D2 BXD48 BXD48 B6 B6 BXD14 BXD14 Fig. 5 Cacna2d1 haplotype effects IOP-lowering potency of pregabalin, a speciﬁc modulator of CACNA2D1, lowers IOP in B6 mice in a dose- dependent manner. a We measured a dose-dependent reduction in IOP compared to control treatment after a single application of ophthalmic formulation containing a range of concentrations of pregabalin from 0.3 to 1.2%. The minimal concentration required to produce the maximum reduction in IOP is 0.9% (mean ± SEM; n = 6). Statistical details are provided in Table 1. b Pregabalin-induced IOP reduction is haplotype- speciﬁc. Mice carrying the B parental allele of Cacna2d1 (i.e., B6 and BXD14) (Supplementary Table 1), have been linked to glaucoma39–41, they do not fulﬁll our stringent selection criteria and therefore have been eliminated as candidate genes (Supplementary Table 2). Other positional candidates, such as Crygn and Klhl7, have been associated with other ocular diseases42, 43, but not glaucoma. a 0 10 –10 –20 –30 –40 0 5 Time (h) 7 6 CTL CTL CTL TX CTL TX TX TX 4 3 2 1 IOP reduction (%) 0.3% 0.3% 0.6% 0.6% 0.9% 0.9% 1.2% 1.2% (Supplementary Table 1), have been linked to glaucoma39–41, they do not fulﬁll our stringent selection criteria and therefore have been eliminated as candidate genes (Supplementary Table 2). Discussion b By combining a forward murine genetics approach with cell biology, pharmacology, and analysis of human GWAS data, we were able to identify a genetic locus that signiﬁcantly modulates IOP in the BXD murine genetic reference panel. Furthermore, we have demonstrated that our top candidate gene—Cacna2d1—is critical for modulation of IOP in these strains. Although other positional candidates in our Chr 5 QTL, such as Nos3 and Sema3e 7 7 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Table 1 Pharmacodynamic parameters after application of drops containing four concentrations of pregabalin to the eyes of B6 mice Pharmacodynamic parameters (mean ± SEM) Pregabalin eye drops 0.3% 0.6% 0.9% 1.2% Baseline IOP (mmHg) 15.4 ± 0.3 15.4 ± 0.1 15.3 ± 0.1 15.5 ± 0.1 IOP at Tmax (mmHg) — 14.2 ± 0.7 12.7 ± 0.8 12.6 ± 0.8 ΔIOP (mmHg) — −1.2 ± 0.8 −2.7 ± 0.8 −2.9 ± 0.9 % Reduction in IOP — 7.8 ± 4.8 17.3 ± 5.4 18.4 ± 5.6 Tmax (h) — 3.0 ± 0.4 2.8 ± 0.5 3.4 ± 0.4 Tend (h) — 5.5 ± 0.3 6.0 ± 0.4 6.8 ± 0.2 AUCtotal (% h) — 42.9 ± 11.1 67.1 ± 19.1 71.9 ± 12.1 Statistical comparisons among four concentrations of pregabalin drops applied to the eyes of B6 mice Pharmacodynamic parameters (mean ± SEM) Overall P-value 0.3% vs. 0.6% 0.3% vs. 0.9% 0.3% vs. 1.2% 0.6% vs. 0.9% 0.6% vs. 1.2% 0.9% vs. 1.2% % Reduction in IOP 0.034 >0.05 >0.05 <0.05 >0.05 >0.05 >0.05 Tmax (h) <0.0001 <0.001 <0.001 <0.0001 >0.05 >0.05 >0.05 Tend (h) <0.0001 <0.0001 <0.0001 <0.0001 >0.05 <0.05 >0.05 AUCtotal (% h) 0.002 >0.05 <0.01 <0.01 >0.05 >0.05 >0.05 Overall P-value represents the outcome of the one-way ANOVA analysis. Individual P-values represent the outcome of Tukey–Kramer multiple comparisons tests. AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours Statistical comparisons among four concentrations of pregabalin drops applied to the eyes of B6 mice Statistical comparisons among four concentrations of pregabalin drops applied to the eyes of B6 mice Pharmacodynamic parameters (mean ± SEM) Overall P-value 0.3% vs. 0.6% 0.3% vs. 0.9% 0.3% vs. 1.2% 0.6% vs. 0.9% 0.6% vs. 1.2% 0.9% vs. Discussion b 1.2% % Reduction in IOP 0.034 >0.05 >0.05 <0.05 >0.05 >0.05 >0.05 Tmax (h) <0.0001 <0.001 <0.001 <0.0001 >0.05 >0.05 >0.05 Tend (h) <0.0001 <0.0001 <0.0001 <0.0001 >0.05 <0.05 >0.05 AUCtotal (% h) 0.002 >0.05 <0.01 <0.01 >0.05 >0.05 >0.05 Overall P-value represents the outcome of the one-way ANOVA analysis. Individual P-values represent the outcome of Tukey–Kramer multiple comparisons tests. AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours Table 2 Pharmacodynamic parameters after application of 0.9% pregabalin eye drops to D2, BXD48 (D allele), B6, and BXD14 (B allele) mice Pharmacodynamic parameters (mean ± SEM) Mouse strain D2 BXD48 B6 BXD14 Baseline IOP (mmHg) 17.2 ± 0.3 16.8 ± 0.5 15.3 ± 0.1 17.8 ± 0.6 IOP at Tmax (mmHg) — 14.4 ± 0.7 12.7 ± 0.8 12.7 ± 0.5 ΔIOP (mmHg) — −2.4 ± 0.6 −2.7 ± 0.8 −5.1 ± 0.8 % Reduction in IOP — 14.3 ± 3.9 17.3 ± 5.4 28.6 ± 3.5 Tmax (h) — 1.5 ± 0.2 2.8 ± 0.5 3.2 ± 0.3 Tend (h) — 3.0 ± 0.5 6.0 ± 0.4 6.5 ± 0.3 AUCtotal (% h) — 26.6 ± 4.5 67.1 ± 19.1 88.7 ± 12.6 Statistical comparisons among different strains of mice after application of 0.9% pregabalin eye drops Pharmacodynamic parameters (mean ± SEM) Overall P-value D2 vs. BXD48 D2 vs. B6 D2 vs. BXD14 BXD48 vs. B6 BXD48 vs. BXD14 B6 vs. BXD14 % Reduction in IOP 0.001 >0.05 >0.05 <0.001 >0.05 <0.05 >0.05 Tmax (h) <0.0001 <0.05 <0.001 <0.0001 <0.05 <0.01 >0.05 Tend (h) <0.0001 <0.001 <0.0001 <0.0001 <0.001 <0.0001 >0.05 AUCtotal (% h) 0.0002 >0.05 <0.01 <0.001 >0.05 <0.01 >0.05 Overall P-value represents the outcome of the one-way ANOVA analysis. Individual AUC-values represent the outcome of Tukey–Kramer multiple comparisons tests. Discussion b AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours Table 2 Pharmacodynamic parameters after application of 0.9% pregabalin eye drops to D2, BXD48 (D allele), B6, and BXD14 (B allele) mice Pharmacodynamic parameters (mean ± SEM) Mouse strain D2 BXD48 B6 BXD14 Baseline IOP (mmHg) 17.2 ± 0.3 16.8 ± 0.5 15.3 ± 0.1 17.8 ± 0.6 IOP at Tmax (mmHg) — 14.4 ± 0.7 12.7 ± 0.8 12.7 ± 0.5 ΔIOP (mmHg) — −2.4 ± 0.6 −2.7 ± 0.8 −5.1 ± 0.8 % Reduction in IOP — 14.3 ± 3.9 17.3 ± 5.4 28.6 ± 3.5 Tmax (h) — 1.5 ± 0.2 2.8 ± 0.5 3.2 ± 0.3 Tend (h) — 3.0 ± 0.5 6.0 ± 0.4 6.5 ± 0.3 AUCtotal (% h) — 26.6 ± 4.5 67.1 ± 19.1 88.7 ± 12.6 Table 2 Pharmacodynamic parameters after application of 0.9% pregabalin eye drops to D2, BXD48 (B allele) mice Statistical comparisons among different strains of mice after application of 0.9% pregabalin eye drops Statistical comparisons among different strains of mice after application of 0.9% pregabalin eye drops Pharmacodynamic parameters (mean ± SEM) Overall P-value D2 vs. BXD48 D2 vs. B6 D2 vs. BXD14 BXD48 vs. B6 BXD48 vs. BXD14 B6 vs. BXD14 % Reduction in IOP 0.001 >0.05 >0.05 <0.001 >0.05 <0.05 >0.05 Tmax (h) <0.0001 <0.05 <0.001 <0.0001 <0.05 <0.01 >0.05 Tend (h) <0.0001 <0.001 <0.0001 <0.0001 <0.001 <0.0001 >0.05 AUCtotal (% h) 0.0002 >0.05 <0.01 <0.001 >0.05 <0.01 >0.05 Overall P-value represents the outcome of the one-way ANOVA analysis. Individual AUC-values represent the outcome of Tukey–Kramer multiple comparisons tests. AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours Pharmacodynamic parameters (mean ± SEM) Overall P-value D2 vs. BXD48 D2 vs. B6 D2 vs. BXD14 BXD48 vs. B6 transfer, and contraction of smooth muscle cells. The net result of pregabalin binding to CACNA2D1 could be threefold: a reduc- tion in the production of AH by the CB; a relaxation of the TM and resultant facilitation of the outﬂow of AH; and relaxation of the CM, which would have an antagonistic effect on TM relaxation. Methods i l Animals. A total of 65 BXD strains, parents, and F1 crosses (548 mice at 9.1–13 months age) were used for QTL mapping of IOP in mice aged 9.1–13 months. The gender distribution was relatively equal (313 females and 235 males). IOP measured from 64 BXD strains aged 1–2.1 months (data available on GeneNetwork) were used for comparison. The parental strains C57BL/6J (n = 6; aged 3–5 months), DBA/2J (n = 4; aged 3–5 months), and two BXD strains— BXD14 and BXD48 (n = 6; aged 3–5 months)—were used to evaluate the IOP- lowering effects of pregabalin. C57BL/6 J and DBA/2 J were used as controls. BXD14 and BXD48 strains were selected based on presence of the B and D hap- lotype of Cacna2d1, respectively. Dutch belted rabbits (n = 5; procured from Robinson Services, Mocksville, NC, USA) weighing ~2.0–2.5 kg were used to test IOP-lowering effects of pregabalin. All procedures involving mice and rabbits were approved by the Animal Care and Use review board of the University of Tennessee Health Science Center (UTHSC) and followed the Association of Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research in addition to the guidelines for laboratory animal experiments (Institute of Laboratory Animal Resources, Public Health Service Policy on Humane Care and Use of Laboratory Animals). Animals were housed under cyclic light (12 h on:12 h off) with 35% humidity in a speciﬁc pathogen-free facility at UTHSC, and were allowed free access to water and food. Gene set enrichment analysis. GSEA was performed as described previously using version 2.2.0 (http://www.broadinstitute.org/gsea35, 62). Brieﬂy, all nominally signiﬁcant correlates (uncorrected P ≤0.05, 1603 genes) of Cacna2d1 were extracted from the whole-eye transcript data and pre-ranked based on their cor- relation coefﬁcients (Supplementary Data 1). Molecular Signatures Database (MSigDB) version 5.0 was used as the database to perform GSEA. GSEA was performed on these pre-ranked genes using 1000 permutations. This was followed by building modules of related pathways based on at least 25% gene overlap between pathways using the enrichment map strategy. The enrichment score (ES) was calculated as cumulative score from 0. The ES was normalized to account for the size of the gene set being used. To control the expected proportions of false positives, the FDR for P-values was calculated using the Benjamini and Hochberg method implemented in LIMMA. Human donor tissue. NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Epithelial secretory activity AH production (ciliary body) IOP AH outflow (trabecular meshwork and/or ciliary muscle) Water transfer Smooth muscle relaxation Cytoskeletal remodeling Ca2+ homeostasis Extracellular Pregabalin Intracellular α2 Ca2+ Ca2+ s-s δ γ α1 β Fig. 6 Proposed mechanism of IOP reduction by pregabalin. We predict that pregabalin binds to CACNA2D1 localized to CB, TM, and/or CM cells. Pregabalin binding causes the α1 pore to close, leading to a decrease in Ca2+ inﬂux into cells and a resultant decrease in free cytosolic Ca2+. Consequently, the cells relax and aqueous humor inﬂow may be reduced and/or outﬂow may be increased, leading to a reduction in IOP AH production (ciliary body) AH outflow (trabecular meshwork and/or ciliary muscle) Cytoskeletal remodeling Fig. 6 Proposed mechanism of IOP reduction by pregabalin. We predict that pregabalin binds to CACNA2D1 localized to CB, TM, and/or CM cells. Pregabalin binding causes the α1 pore to close, leading to a decrease in Ca2+ inﬂux into cells and a resultant decrease in free cytosolic Ca2+. Consequently, the cells relax and aqueous humor inﬂow may be reduced and/or outﬂow may be increased, leading to a reduction in IOP Fig. 6 Proposed mechanism of IOP reduction by pregabalin. We predict that pregabalin binds to CACNA2D1 localized to CB, TM, and/or CM cells. Pregabalin binding causes the α1 pore to close, leading to a decrease in Ca2+ inﬂux into cells and a resultant decrease in free cytosolic Ca2+. Consequently, the cells relax and aqueous humor inﬂow may be reduced and/or outﬂow may be increased, leading to a reduction in IOP offer a potential therapeutic target that could preserve vision by reducing IOP. In the future, a similar approach could provide a bridge between systems biology and effective drug target identiﬁcation. Evaluation of CACNA2D1 in GWAS repositories. SNPs located within the CACNA2D1 human genomic region were evaluated in the NEIGHBORHOOD consortium metadata16. To capture regulatory regions, the genomic locus was deﬁned as all coding sequences + 50 base pairs on either side of the ﬁrst/last exon. eQTL analysis of IOP candidate genes. IOP data from BXD mice aged 1–2.1 months (64 strains) and 9.1–13 months (65 strains) are publically available as BXD published phenotypes record IDs 16337 (http://www.gn2.genenetwork.org/ show_trait?trait_id=16337&dataset=BXDPublish) and 16340 (http://www.gn2. genenetwork.org/show_trait?trait_id=16340&dataset=BXDPublish), respectively. Candidate genes were identiﬁed using the tools available on the GeneNetwork/ UCSC Genome browser. Discussion b Given the marked effect that pregabalin has on base- line IOP in BXD mice carrying the B haplotype of Cacna2d1, we predict that the balance between the effect of the drug on TM cells and also CM cells would be in favor of increased outﬂow through the TM. If our hypothesis is valid, pregabalin binding to both CB and TM, could both reduce AH production and increase AH outﬂow. Table 3 Pharmacodynamic parameters after application of 0.9% pregabalin eye drops to Dutch belted rabbits Pharmacodynamic parameters (Mean ± SEM) Baseline IOP (mmHg) 20.7 ± 0.5 IOP at Tmax (mmHg) 16.1 ± 1.4 ΔIOP (mmHg) −4.6 ± 0.6 % Reduction in IOP 22.1 ± 2.8 Tmax (h) 2.4 ± 0.6 Tend (h) 6.0 ± 0.0 AUCtotal (% h) 78.2 ± 4.2 AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours Table 3 Pharmacodynamic parameters after application of 0.9% pregabalin eye drops to Dutch belted rabbits Pharmacodynamic parameters (Mean ± SEM) Baseline IOP (mmHg) 20.7 ± 0.5 IOP at Tmax (mmHg) 16.1 ± 1.4 ΔIOP (mmHg) −4.6 ± 0.6 % Reduction in IOP 22.1 ± 2.8 Tmax (h) 2.4 ± 0.6 Tend (h) 6.0 ± 0.0 AUCtotal (% h) 78.2 ± 4.2 AUCtotal (% h) area under the curve in percent IOP reduction x hours, Tend (h) time to end of response in hours, Tmax (h) time to maximum response in hours In summary, our study identiﬁed Cacna2d1 as a modiﬁer of IOP. We offer a hypothesis regarding the modulation of IOP and NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 8 ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 Gene expression was used as a microtrait to map reg- ulatory eQTLs for the differences in mRNA expression levels over the panel of BXD lines because it is a heritable trait. Whole-eye transcript data from BXD strains available on GeneNetwork as Eye M430v2 (Sep08) RMA (http://www. genenetwork.org/webqtl/main.py?FormID=sharinginfo&GN_AccessionId=207) were used to identify cis (locally)-regulated genes from within intervals of inter- est34. The same transcript data were used to identify genes whose expression correlated with IOP at 9.1–13 months of age. Only probes that did not bind to regions containing SNPs were used to avoid hybridization artifacts that may cause differences in expression due to technical error rather than biological variance. All probes were veriﬁed by BLAST-like alignment tool (BLAT) (University of California Santa Cruz (UCSC) Genome Browser). Correlation analyses for initial candidate gene selection used Pearson correlation coefﬁcients. Genes were con- sidered as being cis-regulated if the associated marker was localized within a 10 Mb distance of the gene position34. Methods i l Our research and the protocol for collection of human retinas were conducted in accordance with the tenets of the Declaration of Helsinki and approved by University of Tennessee Health Science Center Institutional Review Board. Informed consent was received from donors or donor family members. Parafﬁn sections from a human donor eye (70-year-old, male) with no diagnosis of ocular disease were obtained from National Disease Research Inter- change (Philadelphia, PA). No donor details apart from age, sex, and cause of death were provided. IOP measurement. IOP of mice was measured using an induction-impact ton- ometer (Tonolab tonometer, Colonial Medical Supply, Franconia, NH) for rodents according to the manufacturer’s recommended procedures. The IOP of rabbits was measured using a Tono-Pen AVIA (Reichert Technologies, Depew, NY) while rabbits were held in a rabbit restrainer. Six consecutive IOP readings were aver- aged. For each mouse, the tested ophthalmic formulation (10 μl) was applied to the right eye, while the left eye received the blank and serve as a control29, 60. One hundred microliters of 0.9% pregabalin eye drops were applied into the inferior conjunctival sac of the right eye of the Dutch belted rabbits, while the left eye received the blank eye drops61. All results were expressed as the mean percentage reduction in IOP from baseline (mean% reduction ± SEM). Immunohistochemistry. Our standard methods were used to immunolocalize CACNA2D1 in mouse and human eye sections30, 63, 64. An anti-CACNA2D1 antibody (Bioss Antibodies, Woburn, MA; catalog #bs-11981R; 1:100) was used in these studies along with goat anti-rabbit Alexa ﬂuor 488 secondary antibody (Invitrogen, Waltham, MA; catalog #A11034; 1:200), and TO-PRO-3 iodide (Invitrogen, Waltham, MA; catalog #T3605) as a nuclear counterstain. Sections were viewed and images were obtained using a Nikon C1 confocal microscope (Nikon, NY). Identiﬁcation of IOP-modulating loci in mice. IOP data generated from 9.1–13- month-old BXD strains was integrated into the GeneNetwork database. The identiﬁcation and mapping of phenotypic QTL was performed by linking trait data to genotypes at known genetic marker loci using the WebQTL module on Gene- Network (http://www.genenetwork.org/)26, 29, 30, 32. Topical pregabalin ophthalmic formulations. Blank ophthalmic eye drops were prepared by dissolving 2% hydroxypropylmethylcellulose (Sigma-Aldrich, St. Louis, MO) in PBS pH 7.4 to use it as a vehicle (80%) in combination with propylene glycol (10%; Sigma-Aldrich) and PEG 300 (10%; Sigma-Aldrich). Pregabalin (Sigma-Aldrich, St. ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 13. Nag, A. et al. A genome-wide association study of intra-ocular pressure suggests a novel association in the gene FAM125B in the TwinsUK cohort. Hum. Mol. Genet. 23, 3343–3348 (2014). using four different concentrations of the drug (0.3, 0.6, 0.9, and 1.2%) by dis- solving pregabalin in the previously prepared blank ophthalmic eye drops. These studies were conducted using a single dose–response design. Ten microliters of the pregabalin formulations or the blank ophthalmic eye drops were used in the mouse studies60 (n = 4–6). One hundred microliters of either formulation were used in the rabbit study (n = 5). IOP was measured prior to treatment and hourly until the IOP returned to baseline levels. 14. Li, Z. et al. A common variant near TGFBR3 is associated with primary open angle glaucoma. Hum. Mol. Genet. 24, 3880–3892 (2015). 15. Springelkamp, H. et al. ARHGEF12 inﬂuences the risk of glaucoma by increasing intraocular pressure. Hum. Mol. Genet. 24, 2689–2699 (2015). Several pharmacodynamics parameters were used to evaluate pregabalin including maximum reduction in IOP (% reduction in IOP), time required to reach maximum decrease of IOP (Tmax), time required for IOP to return to baseline (i.e., end of drug effect; Tend), and total area under the percent reduction in IOP-vs-time curve (AUCtotal). Values of AUCtotal were determined using a linear trapezoidal method65. Results were reported as mean ± SEM. 16. Bailey, J. N. et al. Genome-wide association analysis identiﬁes TXNRD2, ATXN2 and FOXC1 as susceptibility loci for primary open-angle glaucoma. Nat. Genet. 48, 189–94 (2016). 17. Wang, X. et al. Joint mouse-human phenome-wide association to test gene function and disease risk. Nat. Commun. 7, 10464 (2016). 18. Houtkooper, R. H. et al. Mitonuclear protein imbalance as a conserved longevity mechanism. Nature 497, 451–7 (2013). 19. Hager, R., Lu, L., Rosen, G. D. & Williams, R. W. Genetic architecture supports mosaic brain evolution and independent brain-body size regulation. Nat. Commun. 3, 1079 (2012). Statistics. To determine how much of the variation in candidate gene across the cohort was due to genetic effects, we calculated the heritability of the candidate gene using the formula of Hegmann and Possidente26, 32, 33. Pearson product- moment correlations were calculated using the Correlation matrix tool available in GeneNetwork. References 1. Kwon, Y. H., Fingert, J. H., Kuehn, M. H. & Alward, W. L. M. Primary open- angle glaucoma. N. Engl. J. Med. 360, 1113–1124 (2009). 32. Lu, H. et al. Genetic dissection of the Gpnmb network in the eye. Invest. Ophthalmol. Vis. Sci. 52, 4132–42 (2011). g g g 2. Goel, M., Picciani, R. G., Lee, R. K. & Bhattacharya, S. K. Aqueous humor dynamics: a review. Open Ophthalmol. J. 4, 52–9 (2010). 33. Swaminathan, S., Lu, H., Williams, R. W., Lu, L. & Jablonski, M. M. Genetic modulation of the iris transillumination defect: a systems genetics analysis using the expanded family of BXD glaucoma strains. Pigment Cell Melanoma Res. 26, 487–98 (2013). 3. Tamm, E. The trabecular meshwork outﬂow pathways: structural and functional aspects. Exp. Eye. Res. 88, 648–655 (2009). 4. Llobet, A., Gasull, X. & Gual, A. Understanding trabecular meshwork physiology: a key to the control of intraocular pressure? News Physiol. Sci. 18, 205–9 (2003). 34. Freeman, N. E. et al. Genetic networks in the mouse retina: growth associated protein 43 and phosphatase tensin homolog network. Mol. Vis. 17, 1355–72 (2011). 5. van Koolwijk, L. M. et al. Genetic contributions to glaucoma: heritability of intraocular pressure, retinal nerve ﬁber layer thickness, and optic disc morphology. Invest. Ophthalmol. Vis. Sci. 48, 3669–76 (2007). 35. Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression proﬁles. Proc. Natl Acad. Sci. USA 102, 15545–50 (2005). 6. Carbonaro, F., Andrew, T., Mackey, D. A., Spector, T. D. & Hammond, C. J. Heritability of intraocular pressure: a classical twin study. Br. J. Ophthalmol. 92, 1125–8 (2008). 36. Schleithoff, L., Mehrke, G., Reutlinger, B. & Lehmann-Horn, F. Genomic structure and functional expression of a human alpha(2)/delta calcium channel subunit gene (CACNA2). Genomics 61, 201–9 (1999). 7. Thorleifsson, G. et al. Common variants near CAV1 and CAV2 are associated with primary open-angle glaucoma. Nat. Genet. 42, 906–9 (2010). 37. Risgaard, B. et al. High prevalence of genetic variants previously associated with Brugada syndrome in new exome data. Clin. Genet. 84, 489–95 (2013). 8. Wiggs, J. et al. Common variants at 9p21 and 8q22 are associated with increased susceptibility to optic nerve degeneration in glaucoma. PLoS Genet. 8, e1002654 (2012). 38. Templin, C. et al. Identiﬁcation of a novel loss-of-function calcium channel gene mutation in short QT syndrome (SQTS6). Eur. Heart J. 32, 1077–1088 (2011). 9. Hysi, P. et al. Received: 1 December 2016 Accepted: 28 June 2017 30. Chintalapudi, S. R., Morales-Tirado, V. M., Williams, R. W. & Jablonski, M. M. Multipronged approach to identify and validate a novel upstream regulator of Sncg in mouse retinal ganglion cells. FEBS J. 283, 678–93 (2016). 31. Anderson, M. G. et al. Mutations in genes encoding melanosomal proteins cause pigmentary glaucoma in DBA/2J mice. Nat. Genet. 30, 81–5 (2002). Methods i l Louis, MO) ophthalmic eye drops were prepared 9 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications ARTICLE Innate immune network in the retina nerve crush. Invest. Ophthalmol. Vis. Sci. 54, 2599–606 (201 29. Lu, H., Lu, L., Williams, R. & Jablonski, M. Iris transillumination defect and its gene modulators do not correlate with intraocular pressure in the BXD family of mice. Mol. Vis. 22, 224–233 (2016). Received: 1 December 2016 Accepted: 28 June 2017 ARTICLE Pearson correlation, denoted by r, measures the strength of a linear association between two variables, which in our case are IOP (mmHg) and Cac- na2d1 mRNA expression. Partial correlation within-strain and between-strain variances, genotypes vs. IOP, and genotype vs. gene expression were calculated with analysis of variance (ANOVA) using SAS 6.0. (Cary, NC) or GraphPad Prism- 7 software (La Jolla, CA). Statistical analyses of the pharmacodynamics results for pregabalin were performed using one-way ANOVA with Tukey–Kramer multiple comparisons test. All pharmacodynamic parameters as well as the statistical ana- lysis of the results were also calculated using GraphPad Prism-7 software (La Jolla, CA). 20. Andreux, P. A. et al. Systems genetics of metabolism: the use of the BXD murine reference panel for multiscalar integration of traits. Cell 150, 1287–99 (2012). 21. Peirce, J. L., Lu, L., Gu, J., Silver, L. M. & Williams, R. W. A new set of BXD recombinant inbred lines from advanced intercross populations in mice. BMC Genet. 5, 7 (2004). 22. Taylor, B. A. et al. Genotyping new BXD recombinant inbred mouse strains and comparison of BXD and consensus maps. Mamm. Genome 10, 335–48 (1999). 23. Chesler, E. J. et al. Complex trait analysis of gene expression uncovers polygenic and pleiotropic networks that modulate nervous system function. Nat. Genet. 37, 233–42 (2005). 24. Ashbrook, D. G., Gini, B. & Hager, R. Genetic variation in offspring indirectly inﬂuences the quality of maternal behaviour in mice. eLife 4, e11814 (2015). Data availability. The data sets generated during and/or analyzed during the current study are available in the GeneNetwork repository as follows: IOP data from BXD mice aged 1–2.1 months and 9.1–13 months are publically available as BXD published phenotypes record IDs 16337 (http://www.gn2.genenetwork.org/ show_trait?trait_id=16337&dataset=BXDPublish) and 16340, (http://www.gn2. genenetwork.org/show_trait?trait_id=16340&dataset=BXDPublish), respectively. Cacna2d1 expression data is available as Trait ID 1446324_at and 1449999_a_at in the Eye M430v2 (Sep08) RMA database (http://www.genenetwork.org/webqtl/ main.py?FormID=sharinginfo&GN_AccessionId=207). 25. Williams, E. G. et al. Systems proteomics of liver mitochondria function. Science 352, aad0189 (2016). 26. Lu, H. et al. Complex interactions of Tyrp1 in the eye. Mol. Vis. 17, 2455–2468 (2011). 27. Jablonski, M. M. et al. Genetic pathways regulating glutamate levels in retinal Muller cells. Neurochem. Res. 36, 594–603 (2011). 28. Templeton, J. P. et al. Innate immune network in the retina activated by optic nerve crush. Invest. Ophthalmol. Vis. Sci. 54, 2599–606 (2013). 28. Templeton, J. P. et al. Acknowledgements 19, 271–295 (2000). R01EY021200, R01EY022305, and P30EY013080; NIAAA Grant U01AA01666; and a Stein Innovation Award and an unrestricted grant from Research to Prevent Blindness, Inc. 48. Siegner, S., Netland, P., Schroeder, A. & Erickson, K. Effect of calcium channel blockers alone and in combination with antiglaucoma medications on intraocular pressure in the primate eye. J. Glaucoma 9, 334–339 (2000). p p y 49. Luksch, A. et al. Effect of nimodipine on ocula rblood ﬂow and colour contrast sensitivity in patients with normal tension glaucoma. Br. J. Ophthalmol. 89, 21–25 (2005). Author contributions S.R.C. and M.M.J. conceived the experiments. S.R.C., D.M., and X.W. conducted experiments. S.R.C., D.M., X.W., P.G.H., J.N.C.B., R.W.W., J.L.W., and M.M.J. partici- pated in data interpretation and discussion. M.M.J. conceptualized the project and supervised the experiments. R.W.W. and M.M.J. provided resources/materials/analysis tools for completion of the study. S.R.C. and M.M.J. wrote the manuscript. All authors reviewed and contributed intellectually to the article. S.R.C. and M.M.J. conceived the experiments. S.R.C., D.M., and X.W. conducted experiments. S.R.C., D.M., X.W., P.G.H., J.N.C.B., R.W.W., J.L.W., and M.M.J. partici- pated in data interpretation and discussion. M.M.J. conceptualized the project and supervised the experiments. R.W.W. and M.M.J. provided resources/materials/analysis tools for completion of the study. S.R.C. and M.M.J. wrote the manuscript. All authors reviewed and contributed intellectually to the article. 50. Jani, A., Goyal, R., Shah, G. & Mehta, A. Effect of calcium channel blockers on intraocular pressure in rabbits. Iranian J. Pharmacol. Ther. 4, 95–99 (2005). 51. Araie, M. & Mayama, C. Use of calcium channel blockers for glaucoma. Prog. Retin. Eye Res. 30, 54–71 (2011). y 52. Mayama, C. Calcium channels and their blockers in intraocular pressure and glaucoma. European J Pharm 739, 96–105 (2014). g p 53. Ganekal, S., Dorairaj, S., Jhanji, V. & Kudlu, K. Effect of topical calcium channel blockers on intraocular pressure in steroid-induced glaucoma. J. Curr. Glaucoma Pract. 8, 15–19 (2014). ARTICLE ARTICLE NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 42. Hugosson, T. et al. Phenotype associated with mutation in the recently identiﬁed autosomal dominant retinitis pigmentosa KLHL7 gene. Arch. Ophthalmol. 128, 772–8 (2010). 64. Chintalapudi, S. R. et al. Isolation and molecular proﬁling of primary mouse retinal ganglion cells: comparison of phenotypes from healthy and glaucomatous retinas. Front. Aging Neurosci. 8, 93 (2016). p 43. Reis, L. M. et al. Whole exome sequencing in dominant cataract identiﬁes a new causative factor, CRYBA2, and a variety of novel alleles in known genes. Hum. Genet. 132, 761–770 (2013). g g g 65. Aburahma, M. H. & Mahmoud, A. A. Biodegradable ocular inserts for sustained delivery of brimonidine tartarate: preparation and in vitro/in vivo evaluation. AAPS PharmSciTech. 12, 1335–47 (2011). 44. McElnea, E. M. et al. Oxidative stress, mitochondrial dysfunction and calcium overload in human lamina cribrosa cells from glaucoma donors. Mol. Vis. 17, 1182–1191 (2011). 9Department of Ophthalmology, Duke University Eye Center, Durham, NC 27705, USA. 10University of Wisconsin Institute for Clinical and Translational Research, Madison, WI 53705, USA. 11Department of Ophthalmology, UNC School of Medicine, Chapel Hill, NC 27517, USA. 12Department of Ophthalmology and Visual Sciences, University of Iowa, Iowa City, IA 52242, USA. 13Eye Doctors of Washington DC, Washington, Acknowledgements We thank Dr. E. Geisert, Dr. L. Lu, and Mr. B. Orr for their assistance in generating the BXD microarray data sets that were used in these analyses. We also thank Dr. H. Lu for acquiring the IOP data sets and Dr. S. Surbhi for assistance with statistical analyses. We further thank S. Ganguli and A. Shepherd for discussion and technical assistance. We acknowledge the ﬁnancial support from: the Center for Integrative and Translational Genomics at the University of Tennessee Health Science Center; NEI Grants We thank Dr. E. Geisert, Dr. L. Lu, and Mr. B. Orr for their assistance in generating the BXD microarray data sets that were used in these analyses. We also thank Dr. H. Lu for acquiring the IOP data sets and Dr. S. Surbhi for assistance with statistical analyses. We further thank S. Ganguli and A. Shepherd for discussion and technical assistance. We acknowledge the ﬁnancial support from: the Center for Integrative and Translational Genomics at the University of Tennessee Health Science Center; NEI Grants ( ) 45. Crish, S. & Calkins, D. Neurodegeneration in glaucoma: progression and calcium-dependent intracellular mechanisms. Neuroscience 176, 1–11 (2011). 46. Payne, L., Slagle, T., Cheeks, L. & Green, K. Effect of calcium channel blockers on intraocular pressure in rabbits. Ophthalmic Res. 22, 337–341 (1990). 47. Wiederholt, M., Thieme, H. & Stumpff, F. The regulation of trabecular meshwork and ciliary muscle contractility. Prog. Retin. Eye Res. 19, 271–295 (2000). 45. Crish, S. & Calkins, D. Neurodegeneration in glaucoma: progression and calcium-dependent intracellular mechanisms. Neuroscience 176, 1–11 (2011). 45. Crish, S. & Calkins, D. Neurodegeneration in glaucoma: progression and calcium-dependent intracellular mechanisms. Neuroscience 176, 1–11 (2011). 46. Payne, L., Slagle, T., Cheeks, L. & Green, K. Effect of calcium channel blockers on intraocular pressure in rabbits. Ophthalmic Res. 22, 337–341 (1990). p 46. Payne, L., Slagle, T., Cheeks, L. & Green, K. Effect of calcium channel blockers on intraocular pressure in rabbits. Ophthalmic Res. 22, 337–341 (1990). 46. Payne, L., Slagle, T., Cheeks, L. & Green, K. Effect of calcium channel blockers on intraocular pressure in rabbits. Ophthalmic Res. 22, 337–341 (1990). 47. Wiederholt, M., Thieme, H. & Stumpff, F. The regulation of trabecular meshwork and ciliary muscle contractility. Prog. Retin. Eye Res. 19, 271–295 (2000). p p ( ) 47. Wiederholt, M., Thieme, H. & Stumpff, F. The regulation of trabecular meshwork and ciliary muscle contractility. Prog. Retin. Eye Res. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. y p g 59. Yuan, Z. & Xu, S. Novel SNPs of the bovine CACNA2D1 gene and their association with carcass and meat quality traits. Mol. Biol. Rep. 38, 365–370 (2011). 60. Ibrahim, M., Abd-Elgawad, A., Soliman, O. & Jablonski, M. Natural bioadhesive biodegradable nanoparticle-based topical ophthalmic formulations for management of glaucoma. Transl. Vis. Sci. Technol. 4, 12 (2015). 61. Swaminathan, S. et al. Novel endogenous glycan therapy for retinal diseases: safety, in vitro stability, ocular pharmacokinetic modeling and bio-distribution. AAPS J. 16, 311–312 (2014). 62. Mootha, V. K. et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat. Genet. 34, 267–73 (2003). 63. Nookala, S. et al. In search of the identity of the XAP-1 antigen: a protein localized to cone outer segments. Invest. Ophthalmol. Vis. Sci. 51, 2736–43 (2010). © The Author(s) 2017 Additional information Supplementary Information accompanies this paper at doi:10.1038/s41467-017-00837-5. 54. Andersen, J. S. et al. A gene responsible for the pigment dispersion syndrome maps to chromosome 7q35-q36. Arch. Ophthalmol. 115, 384–8 (1997). Competing interests: The authors declare no competing ﬁnancial interests. 55. Wirtz, M. K. et al. GLC1F, a new primary open-angle glaucoma locus, maps to 7q35-q36. Arch. Ophthalmol. 117, 237–41 (1999). Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ 56. Klein, B. E. et al. Prevalence of glaucoma. The Beaver Dam Eye Study. Ophthalmology 99, 1499–504 (1992). Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 57. Zhang, W., Liu, W., Poradosu, E. & Ratain, M. Genome-wide identiﬁcation of genetic determinants for the cytotoxicity of perifosine. Hum. Genomics 3, 53–70 (2008). 58. Neely, G. et al. A genome-wide Drosophila screen for heat nociception identiﬁes α2δ3 as an evolutionarily conserved pain gene. Cell 143, 628–638 (2010). References Genome-wide analysis of multi-ancestry cohorts identiﬁes new loci inﬂuencing intraocular pressure and susceptibility to glaucoma. Nat. Genet. 46, 1126–1130 (2014). 39. Lei, Y., Zhang, X., Song, M., Wu, J. & Sun, X. Aqueous Humor Outﬂow Physiology in NOS3 Knockout Mice. Invest. Ophthalmol. Vis. Sci. 56, 4891–8 (2015). 10. Gharahkhani, P. et al. Common variants near ABCA1, AFAP1 and GMDS confer risk of primary open-angle glaucoma. Nat. Genet. 46, 1120–1125 (2014). 40. Kang, J. H., Wiggs, J. L., Haines, J., Abdrabou, W. & Pasquale, L. R. Reproductive factors and NOS3 variant interactions in primary open-angle glaucoma. Mol. Vis. 17, 2544–51 (2011). 11. Burdon, K. et al. Genome-wide association study identiﬁes susceptibility loci for open angle glaucoma at TMCO1 and CDKN2B-AS1. Nat. Genet. 46, 574–578 (2011). 41. Smit-McBride, Z. et al. In vivo gene expression proﬁling of retina postintravitreal injections of dexamethasone and triamcinolone at clinically relevant time points for patient care. Invest. Ophthalmol. Vis. Sci. 52, 8965–78 (2011). 12. Chen, Y. et al. Common variants near ABCA1 and in PMM2 are associated with primary open-angle glaucoma. Nat. Genet. 46, 1115–1119 (2014). 10 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications © The Author(s) 2017 30Singapore Eye Research Institute, National University of Singapore, Singapore, 168751, Singapore. 31Department of Epidemiology, Erasmus Medical Center, Rotterdam, GE 3015, The Netherlands. 32Department of Ophthalmology, Flinders University, Adelaide, SA 5042, Australia. 33Department of Epidemiology, Erasmus Medical Center, Rotterdam, GE 3015, The Netherlands. 34Statistical Genetics, QIMR Berghofer Medical Research Institute, Brisbane, QLD 4006, Australia. 35Department of Epidemiology, University Medical Center of Rotterdam, Rotterdam, GE 3015, The Netherlands. 36Department of Ophthalmology, King’s College London, London, SE5 9RS, UK. 37Department of Ophthlmology, Menzies Institute for Medical Research, University of Tasmania, Tasmania, 7000, Australia. 38Ophthalmology, Inselspital, University Hospital Bern, University of Bern, Bern, 3010, Switzerland. 39Faculty of Medicine, University of Iceland, Reykjavik, 101, Iceland. 40Department of Public Health and Primary Care, University of Cambridge, Cambridge, CB2 1TN, England. 41Department of Biochemistry, National University of Singapore, Singapore, 117596, Singapore. 42Department of Ophthalmology, Erasmus Medical Center, Rotterdam, GE 3015, The Netherlands. 43Clinical and Experimental Sciences, Faculty of Medicine, University of Southampton, Southampton, SO17 1BJ, UK. 44Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, VIC 3002, Australia. 45Statistical Genetics, QIMR Berghofer Medical Research Institute Royal Brisbane Hospital, Brisbane, QLD 4006, Australia. 46Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, Hong Kong. 47Institute of Human Genetics, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, 91054, Germany. 48deCODE genetics/Amgen, Inc., Reykjavik, IS-101, Iceland. 49MRC Human Genetics Unit, The University of Edinburgh, Edinburgh, EH4 2XU, Scotland. 50McPherson Eye Research Institute, University of Wisconsin-Madison, Madison, WI 53705, USA. 51Clinic for General and Interventional Cardiology, University Heart Center Hamburg, Hamburg, 20246, Germany NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 NATURE COMMUNICATIONS \| DOI: 10.1038/s41467-017-00837-5 DC 20036, USA. 14Glaucoma Research Center, Wills Eye Hospital, Philadelphia, PA 19107, USA. 15Channing Division of Network Medicine, Brigham and Women’s Hospital, Boston, MA 02115, USA. 16Harvard T.H. Chan School of Public Health, Boston, MA 02115, USA. 17Department of Ophthalmology, Bascom Palmer Eye Institute, University of Miami, Miami, FL 33136, USA. 18Department of Ophthalmology, Kellogg Eye Center, University of Michigan, Ann Arbor, MI 48105, USA. 19Department of Cellular Biology and Anatomy, Augusta University, Augusta, GA 30901, USA. 20John P. Hussman Institute for Human Genomics, University of Miami, Miami, FL 33136, USA. 21Department of Ophthalmology, West Virginia University, Morgantown, WV 26506, USA. 22Department of Ophthalmology and Visual Sciences, Case Western Reserve University, Cleveland, OH 44106, USA. 23Department of Ophthalmology, Einhorn Clinical Research Center, New York Eye and Ear Inﬁrmary, New York, NY 10003, USA. 24Department of Ophthalmology, NYU Langone Medical Center, New York, NY 10016, USA. 25Dr. John T. Macdonald Foundation Department of Human Genetics, John P. Hussman Institute for Human Genomics, University of Miami Health System, Miami, FL 33136, USA. 26Department of Ophthalmology, Stanford University Medical Center, Stanford, CA 94303, USA. 27Department of Ophthalmology, Mayo Clinic, Rochester, MN 85259, USA. 28Department of Genetics, Stanford University Medical Center, Stanford, CA 94305, USA. 29Department of Ophthalmology, Wilmer Eye Institute, The Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA NEIGHBORHOOD consortium Rand Allingham9, Murray Brilliant10, Don Budenz11, John Fingert12, Douglas Gaasterland13, Teresa Gaasterland13, Jonathan L. Haines4, Lisa Hark14, Michael Hauser9, Rob Igo4, Jae Hee Kang15, Peter Kraft16, Richard Lee17, Paul Lichter18, Yutao Liu19, Syoko Moroi18, Louis R. Pasquale6,15, Margaret Pericak-Vance20, Anthony Realini21, Doug Rhee22, Julia R. Richards18, Robert Ritch23, Joel Schuman24, William K. Scott25, Kuldev Singh26, Arthur Sit27, Douglas Vollrath28, Gadi Wollstein24 & Don Zack29 9Department of Ophthalmology, Duke University Eye Center, Durham, NC 27705, USA. 10University of Wisconsin Institute for Clinical and Translational Research, Madison, WI 53705, USA. 11Department of Ophthalmology, UNC School of Medicine, Chapel Hill, NC 27517, USA. 12Department of Ophthalmology and Visual Sciences, University of Iowa, Iowa City, IA 52242, USA. 13Eye Doctors of Washington DC, Washington, NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 11 ARTICLE 30Singapore Eye Research Institute, National University of Singapore, Singapore, 168751, Singapore. 31Departme Medical Center, Rotterdam, GE 3015, The Netherlands. 32Department of Ophthalmology, Flinders University, A International Glaucoma Genetics consortium Tin Aung30, Peter Bonnemaijer31, Cheng-Yu Cheng31, Jamie Craig32, Cornelia van Duijn33, Puya Gharahkhani34, Adriana Iglesias Gonzalez35, Christopher J. Hammond36, Alex Hewitt37, Rene Hoehn38, Fridbert Jonansson39, Anthony Khawaja40, Chiea Chuen Khor41, Caroline C.W. Klaver42, Andrew Lotery43, David Mackey44, Stuart MacGregor45, Calvin Pang46, Francesca Pasutto47, Kári Stefansson48, Gudmar Thorleifsson48, Unnar Thorsteinsdottir48, Veronique Vitart49, Eranga Vithana30, Terri Young50 & Tanja Zeller51 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications 12 NATURE COMMUNICATIONS\| 8: 1755 \| DOI: 10.1038/s41467-017-00837-5\| www.nature.com/naturecommunications
https://openalex.org/W2911736905	https://gmd.copernicus.org/preprints/gmd-2018-300/gmd-2018-300.pdf	English	null	On fluctuating air-sea-interaction in local models: linear theory	null	2,019	cc-by	29,244	Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. On ﬂuctuating air-sea-interaction in local models: lin Achim Wirth1 Abstract. The dynamics of three local linear models of air sea-interation commonly employed in climate or ocean simulations is compared. The models differ by whether or not the ocean velocity is included in the shear calculation applied to the ocean and the atmosphere. Analytic calculations for the models with deteministic and random forcing (white and colored) are presented. The short term behavior is similar in all models, which only small quantitative differences, while the longterm behavior differs qualitatively between the models. The ﬂuctuation-dissipation-relation, which connects the fast excitation to the slow dissipa- 5 tion, is establised for all models with random forcing. The ﬂuctuation-dissipation-theorem, which compares the response to an external forcing to internal ﬂuctuations is established for a white-noise forcing and a colored forcing when the phase space is augmented by the forcing variable. Using results from numerical integrations of stochastic differential equations shows that the ﬂuctuation-theorem, which compares the probability of positive to negative ﬂuxes of the same magnitude, averaged over time-intervals of varying length, holds for the energy gained by the ocean from the atmosphere. 10 5 time-intervals of varying length, holds for the energy gained by the ocean from the atmosphere. 10 10 Copyright statement. Copyright statement. 1 Introduction The exchange of momentum, heat, water and chemical ﬂuxes at the atmosphere-ocean interface is key to understanding the dynamics of the atmosphere, the ocean and the climate, as well as their response to changes in the forcing of the climate dynamics of the atmosphere, the ocean and the climate, as well as their response to changes in the forcing of the climate system (Stocker et al. (2013), Csanady (2001)). In the present paper the exchange of momentum, only, is considered. The 15 exchange is usually parameterized by local models (so called bulk formulas), which imitate the action of the non-explicitly resolved dynamics on the explicitely resolved dynamics in atmosphere, ocean and climate models. In the present paper the behavior of the most widely used local models subject to deterministic and random forcing are discussed. Only the local exchange between the atmosphere and the ocean is considered, neglecting the horizontal interaction within the atmosphere 15 and the ocean. Mathematically speaking the models are zero-dimensional one-component (0D1C) (see section 2). The linear 20 models used here allow for analytic solutions. When within a hierarchy of models a systematic liaison of the more involved models to models that allow for an analytic solution can be established, the scientiﬁc understanding of the process studied is increased (Wirth (2010)). and the ocean. Mathematically speaking the models are zero-dimensional one-component (0D1C) (see section 2). The linear 20 models used here allow for analytic solutions. When within a hierarchy of models a systematic liaison of the more involved models to models that allow for an analytic solution can be established, the scientiﬁc understanding of the process studied is increased (Wirth (2010)). 1 1 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. The conspicuous feature of the atmosphere ocean system is the strong difference in mass (and also heat capacity, CO2 absorption) of the two media, leading to strong difference in the characteristic time scales for inertia (and also heat, CO2 storage). In this respect there is a strong analogy to Brownian motion, with light and fast molecules colliding with heavy and slow Brownian particles. The Fluctuation-Dissipation-Relation (FDR) developed by Einstein (1906) represents the framework to describe such motion. 1 Introduction He noted that a Brownian particle in a ﬂuid is subject to two processes, a macroscopic friction and 5 microscopic ﬂuctuations, which are related as they are both due the surrounding ﬂuid (see Einstein (1906, 1956), Perrin (2014)). The FDR describes the relation between the two processes (Barrat and Hansen (2003)). The FDR is applied to a large variety of linear and non-linear problems in the ﬁeld of non-equilibrium statistical mechanics, also when the “Brownian particle” is some “slow” property of a system. In air-sea interaction the friction at the interface dissipates energy and introduces ﬂuctuations in The conspicuous feature of the atmosphere ocean system is the strong difference in mass (and also heat capacity, CO2 absorption) of the two media, leading to strong difference in the characteristic time scales for inertia (and also heat, CO2 storage). In this respect there is a strong analogy to Brownian motion, with light and fast molecules colliding with heavy and slow Brownian particles. The Fluctuation-Dissipation-Relation (FDR) developed by Einstein (1906) represents the framework to describe such motion. He noted that a Brownian particle in a ﬂuid is subject to two processes, a macroscopic friction and 5 microscopic ﬂuctuations, which are related as they are both due the surrounding ﬂuid (see Einstein (1906, 1956), Perrin (2014)). The FDR describes the relation between the two processes (Barrat and Hansen (2003)). The FDR is applied to a large variety of linear and non-linear problems in the ﬁeld of non-equilibrium statistical mechanics, also when the “Brownian particle” is some “slow” property of a system. In air-sea interaction the friction at the interface dissipates energy and introduces ﬂuctuations in 5 both media. Also in this case, dissipation and ﬂuctuations are due to the same process and a relation between the two has to 10 exist. This relation, the FDR, is established for the three models, subject to white and colored random forcing, in the present work. The major difference between Brownian motion and air-sea interaction is that the former system is conservative while the latter is dissipative and forced from the exterior. Mathematically speaking: in the former, the dynamics conserves the phase both media. Also in this case, dissipation and ﬂuctuations are due to the same process and a relation between the two has to 10 exist. This relation, the FDR, is established for the three models, subject to white and colored random forcing, in the present work. 1 Introduction In the context of a purely 2D dynamics the energy dissipation within the atmospheric and the oceanic layer, due to horizontal friction processes, decreases with an increasing Reynolds number, due to the inverse cascade of energy in two-dimensional by restoring to a constant velocity proﬁle. That is, there was no separation between the time scale of the forcing and the 25 atmospheric dynamics in the 2D numerical simulations. The white-in-time FDR model, while successful for describing the ocean response, was therefore inappropriate to describe the atmospheric dynamics. In the context of a purely 2D dynamics the energy dissipation within the atmospheric and the oceanic layer, due to horizontal friction processes, decreases with an increasing Reynolds number, due to the inverse cascade of energy in two-dimensional atmospheric dynamics in the 2D numerical simulations. The white-in-time FDR model, while successful for describing the ocean response, was therefore inappropriate to describe the atmospheric dynamics. In the context of a purely 2D dynamics the energy dissipation within the atmospheric and the oceanic layer, due to horizontal friction processes, decreases with an increasing Reynolds number, due to the inverse cascade of energy in two-dimensional turbulence (Boffetta and Ecke (2012)). In the real ocean internal energy dissipation depends on a variety of processes as frontal 30 dynamics, tidal mixing, stratiﬁcation and bottom friction (Ferrari and Wunsch (2009)). In this work three different mathematical models for parameterizing air-sea interaction which are all used in numerical models of climate dynamics are discussed. They differ by to what extend the ocean velocity is considered in the calculation of the shear force at the air-sea interface. In the ﬁrst, the ocean velocity is ignored and shear is calculated based on the atmospheric velocity only. Historically this was done in all climate models and is justiﬁed by the fact that usually atmospheric 35 turbulence (Boffetta and Ecke (2012)). In the real ocean internal energy dissipation depends on a variety of processes as frontal 30 dynamics, tidal mixing, stratiﬁcation and bottom friction (Ferrari and Wunsch (2009)). In this work three different mathematical models for parameterizing air-sea interaction which are all used in numerical models of climate dynamics are discussed. They differ by to what extend the ocean velocity is considered in the calculation of the shear force at the air-sea interface. In the ﬁrst, the ocean velocity is ignored and shear is calculated based on the atmospheric velocity only. 1 Introduction The major difference between Brownian motion and air-sea interaction is that the former system is conservative while the latter is dissipative and forced from the exterior Mathematically speaking: in the former the dynamics conserves the phase latter is dissipative and forced from the exterior. Mathematically speaking: in the former, the dynamics conserves the phase space volume, while in the latter it contracts and the dynamics takes place on a strange attractor of vanishing phase space 15 volume. A key feature of Brownian motion is the equipartition of energy between a Brownian particle and a molecule (Einstein (1906)), in the case of air-sea interaction equipartition does not hold. Although there are fundamental differences between conservative and dissipative-and-forced dynamics, many of the mathematical concepts developed can be extended from the former to the latter (Marconi et al. (2008)). In a previous publication (Wirth (2018)) the FDR was derived for one model of the atmosphere-ocean system, where the forcing on the atmosphere was white-in-time. This FDR was then compared to a two 20 dimensional numerical simulation of air-sea interaction, giving rise to turbulent dynamics. The two-dimensional numerical simulation was forced by maintaining one Fourier-mode of the dynamics at a ﬁxed value. The application of the FDR to 2D numerical simulations of turbulent dynamics succeeded for the ocean dynamics but failed for the atmosphere, as the former evolves on a time scale much slower than the forcing, while the time scale of the latter is equal to the forcing, which acts the atmosphere-ocean system, where the forcing on the atmosphere was white-in-time. This FDR was then compared to a two 20 dimensional numerical simulation of air-sea interaction, giving rise to turbulent dynamics. The two-dimensional numerical simulation was forced by maintaining one Fourier-mode of the dynamics at a ﬁxed value. The application of the FDR to 2D numerical simulations of turbulent dynamics succeeded for the ocean dynamics but failed for the atmosphere, as the former evolves on a time scale much slower than the forcing, while the time scale of the latter is equal to the forcing, which acts by restoring to a constant velocity proﬁle. That is, there was no separation between the time scale of the forcing and the 25 atmospheric dynamics in the 2D numerical simulations. The white-in-time FDR model, while successful for describing the ocean response, was therefore inappropriate to describe the atmospheric dynamics. 1 Introduction The results of the FDR are used to establish in section 5 The present work compares the three different parameterizations of air-sea transfer of momentum discussed above with 20 a linear drag law having a constant or periodic (deterministic) forcing or a white or colored random-forcing. This leads to 3 · (2 + 2) = 12 different local conﬁgurations which are discussed here. The models are introduced in section 2 and the solutions with different forcing are discussed in section 3. The resulting The present work compares the three different parameterizations of air-sea transfer of momentum discussed above with 20 a linear drag law having a constant or periodic (deterministic) forcing or a white or colored random-forcing. This leads to 3 · (2 + 2) = 12 different local conﬁgurations which are discussed here. The models are introduced in section 2 and the solutions with different forcing are discussed in section 3. The resulting FDRs for linear models with stochastic forcing are given in section 4. The results of the FDR are used to establish in section 5 the work done on the air-sea system, the energy ﬂuxes and energy dissipation in the different models. A special emphasis 25 is directed towards the consistency of the different models and their differences, quantitative and qualitative, for short- and long-term integrations. Th Fl i Di i i Th (FDT) ( i K b (1966) B d H (2003) M i l (2008)) i the work done on the air-sea system, the energy ﬂuxes and energy dissipation in the different models. A special emphasis 25 is directed towards the consistency of the different models and their differences, quantitative and qualitative, for short- and long-term integrations. The Fluctuation-Dissipation-Theorem (FDT) (see i.e. Kubo (1966), Barrat and Hansen (2003), Marconi et al. (2008)) is discussed in section 6. It considers the relation of the internal ﬂuctuations of a system to the response of an external force. the work done on the air-sea system, the energy ﬂuxes and energy dissipation in the different models. A special emphasis 25 is directed towards the consistency of the different models and their differences, quantitative and qualitative, for short- and long-term integrations. The Fluctuation-Dissipation-Theorem (FDT) (see i.e. Kubo (1966), Barrat and Hansen (2003), Marconi et al. (2008)) is discussed in section 6. It considers the relation of the internal ﬂuctuations of a system to the response of an external force. 1 Introduction Only the third model is mechanically consistent as the shear force, 5 We demonstrated in recent work using a two-dimensional model, that the third parameterization together with a quadratic drag law can lead to a generation of instability (Moulin and Wirth (2014)) and new dynamical behavior (Moulin and Wirth (2016)), that is a co-organization between the atmospheric and oceanic variables that resembles a glass-transition in condensed 10 matter physics. It is shown in Wirth (2018), by solving the Fokker-Planck equation, that the second order moments of the non-linear model can be reproduced by a linear model using an eddy-friction approach with an eddy coefﬁcient that is obtained analytically. Here, only linear models are considered, as the focus is on the analytic theory (where possible). The analytic solution of (2016)), that is a co-organization between the atmospheric and oceanic variables that resembles a glass-transition in condensed 10 matter physics. It is shown in Wirth (2018), by solving the Fokker-Planck equation, that the second order moments of the non-linear model can be reproduced by a linear model using an eddy-friction approach with an eddy coefﬁcient that is obtained analytically. Here, only linear models are considered, as the focus is on the analytic theory (where possible). The analytic solution of Here, only linear models are considered, as the focus is on the analytic theory (where possible). The analytic solution of a linear model gives the dependence on all parameters in a nonlinear model parameter dependence has to be numerically 15 evaluated for each parameter. Furthermore, in the linear models, solutions with different forcing can be simply added up, in their non-linear counterpart this is no-longer true. The prolongation to non-linear models and their numerical solutions, will be discussed elsewhere. It is furthermore important to note that the major differences between the three models already emerge in their linear versions. The present work compares the three different parameterizations of air-sea transfer of momentum discussed above with 20 a linear drag law having a constant or periodic (deterministic) forcing or a white or colored random-forcing. This leads to 3 · (2 + 2) = 12 different local conﬁgurations which are discussed here. The models are introduced in section 2 and the solutions with different forcing are discussed in section 3. The resulting FDRs for linear models with stochastic forcing are given in section 4. 1 Introduction Historically this was done in all climate models and is justiﬁed by the fact that usually atmospheric 35 2 2 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. winds have higher speeds than ocean currents. In the second, the ocean velocity is considered when the shear-force applied to the ocean is calculated, but not to the atmosphere. These two models are called “one-way” as the ocean dynamic does not act on the atmosphere, they are used, for example, whenever the atmospheric forcing is known prior to the integration of the ocean model, when an ocean only simulation is performed. Only the third model is mechanically consistent as the shear force, applied to the ocean and the atmosphere, is calculated based on the difference between the atmospheric and the oceanic velocity 5 vectors and respects Newton’s laws. This model is called “two-way”. In the present work which is mostly focused on analytic theory, the calculation of the shear is performed using a linear (Rayleigh) law. winds have higher speeds than ocean currents. In the second, the ocean velocity is considered when the shear-force applied to the ocean is calculated, but not to the atmosphere. These two models are called “one-way” as the ocean dynamic does not act on the atmosphere, they are used, for example, whenever the atmospheric forcing is known prior to the integration of the ocean model, when an ocean only simulation is performed. Only the third model is mechanically consistent as the shear force, winds have higher speeds than ocean currents. In the second, the ocean velocity is considered when the shear-force applied to the ocean is calculated, but not to the atmosphere. These two models are called “one-way” as the ocean dynamic does not act on the atmosphere, they are used, for example, whenever the atmospheric forcing is known prior to the integration of the ocean model, when an ocean only simulation is performed. 1 Introduction When it holds the average response of the system to an external force can be obtained by observing the average relaxation of 30 spontaneous ﬂuctuations. The FDT is today used in a large variety of statistical and dynamical systems as i.e. climate dynamics. If it holds for our climate system it allows for determining the response to perturbations, antropogenic or others, by studying its natural variability (see e.g. Cooper and Haynes (2011)). In the case of the linear models discussed here the FDT can be established analytically using matrix calculus. Studying the FDT in air-sea interaction is also interesting from a conceptual When it holds the average response of the system to an external force can be obtained by observing the average relaxation of 30 spontaneous ﬂuctuations. The FDT is today used in a large variety of statistical and dynamical systems as i.e. climate dynamics. If it holds for our climate system it allows for determining the response to perturbations, antropogenic or others, by studying its natural variability (see e.g. Cooper and Haynes (2011)). In the case of the linear models discussed here the FDT can be established analytically using matrix calculus. Studying the FDT in air-sea interaction is also interesting from a conceptual When it holds the average response of the system to an external force can be obtained by observing the average relaxation of 30 spontaneous ﬂuctuations. The FDT is today used in a large variety of statistical and dynamical systems as i.e. climate dynamics. If it holds for our climate system it allows for determining the response to perturbations, antropogenic or others, by studying its natural variability (see e.g. Cooper and Haynes (2011)). In the case of the linear models discussed here the FDT can be established analytically using matrix calculus. Studying the FDT in air-sea interaction is also interesting from a conceptual 3 3 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. point of view as air-sea interaction posses a dynamics on dissimilar and interacting time scales, a property found in many natural applications and processes of the climate system. 1 Introduction point of view as air-sea interaction posses a dynamics on dissimilar and interacting time scales, a property found in many natural applications and processes of the climate system. The Fluctuation Theorems (FTs) (Gallavotti and Cohen (1995a), Gallavotti and Cohen (1995b), Ciliberto et al. (2004)) for the energy exchange between the atmosphere and the ocean are numerically evaluated in section 7. When the atmosphere is force by a white or colored noise, the velocity ﬂuctuations in the atmosphere have a larger amplitude than in the ocean and 5 on average the atmosphere does work on the ocean and the ocean receives energy form the atmosphere. Instantaneous energy ﬂuxes can however also go in the opposite direction. The probability of positive versus negative ﬂuxes, averaged over ﬁnite time intervals of varying length, are subject of the FT. The results are discussed in section 8 and conclusions are presented in section 9 5 The analytic calculations are found in the vast appendixes, which form the major part of the publication. They present a 10 register for calculations concerning local models of air-sea interaction. These calculations are important as they expose the strength and weaknesses of and the differences between the models. The analytic calculations are found in the vast appendixes, which form the major part of the publication. They present a 10 register for calculations concerning local models of air-sea interaction. These calculations are important as they expose the strength and weaknesses of and the differences between the models. 10 2 Local models The turbulent friction at the atmosphere-ocean interface is commonly modeled by a quadratic friction law, where the friction force is a constant times the product of the shear speed and the shear velocity (see e.g. Stull (2012)). The linear version with 15 a constant eddy-coefﬁcient allows for analytic solutions. It is also sometimes used in numerical simulations of the climate dynamics. The friction coefﬁcient represents an average (in time and space) mimicking the real friction process. 15 The mathematical models discussed here are non-dimensionalized. The mass of the atmosphere per unit area is set to unity. The mass of the ocean per unit area is m times the mass of the atmosphere, the total mass per unit area is M = m + 1. When The mathematical models discussed here are non-dimensionalized. The mass of the atmosphere per unit area is set to unity. The mass of the ocean per unit area is m times the mass of the atmosphere, the total mass per unit area is M = m + 1. When the interaction of the atmospheric and oceanic mixed-layers are considered m ≈O(102). Three linear models commonly used 20 are discussed. These models give rise to different conﬁgurations which differ by the forcing. In the following equations S is the inverse of the friction time in the ocean. When a linear model is used (S=const) both horizontal directions are un-coupled and the problem can be considered independently for each direction and reduces to a one dimensional problem. Scalar variables are therefore employed. Newton’s third law sets the inverse friction time for the atmosphere to Sm. The forcing of the system the interaction of the atmospheric and oceanic mixed-layers are considered m ≈O(102). Three linear models commonly used 20 are discussed. These models give rise to different conﬁgurations which differ by the forcing. In the following equations S is the inverse of the friction time in the ocean. When a linear model is used (S=const) both horizontal directions are un-coupled and the problem can be considered independently for each direction and reduces to a one dimensional problem. Scalar variables are therefore employed. Newton’s third law sets the inverse friction time for the atmosphere to Sm. The forcing of the system is denoted ˜F. 2 Local models In the ﬁrst model, L1, the ocean velocity is not considered in the calculation of the shear force at the interface, 25 neither in the dynamics of the atmosphere nor the ocean: (1) (2) ∂tuL1 a = −SmuL1 a + ˜F (1) ∂tuL1 o = S uL1 a . (2) ∂tuL1 a = −SmuL1 a + ˜F ∂tuL1 o = S uL1 a . (1) (1) (2) (2) Note that this model is inconsistent as the ocean is accelerated in the direction of the atmospher Note that this model is inconsistent as the ocean is accelerated in the direction of the atmospheric velocity even when the ocean is moving in the same direction with a higher speed. The model is justiﬁed by the observation that atmospheric wind-speeds 30 are often much larger than ocean currents and when the difference of the two is considered the latter can be neglected to ﬁrst order in uo/ua. The L1 represents the classical approach to implementing air-sea interaction. is moving in the same direction with a higher speed. The model is justiﬁed by the observation that atmospheric wind-speeds 30 are often much larger than ocean currents and when the difference of the two is considered the latter can be neglected to ﬁrst order in uo/ua. The L1 represents the classical approach to implementing air-sea interaction. 4 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. In the second model, L2, the ocean velocity is considered in the calculation of the shear force at the interface when the ocean dynamics is considered, but not for the atmospheric velocity: ∂tuL2 a = −SmuL2 a + ˜F (3) ∂tuL2 o = S (uL2 a −uL2 o ). (4) ∂tuL2 a = −SmuL2 a + ˜F (3) ∂tuL2 o = S (uL2 a −uL2 o ). (4) ∂tuL2 a = −SmuL2 a + ˜F ∂tuL2 o = S (uL2 a −uL2 o ). (3) (4) (3) (3) ( ) (4) (4) Note that this model is inconsistent as interfacial friction does not conserve total (atmosphere plus ocean) momentum. 2 Local models This 5 model neglects the action of ocean currents on the atmospheric dynamics, it is used, for example, in ocean modeling when ocean models are forced by winds which are predeﬁned, available previous to the ocean simulation. The L1 and L2 models are called “one-way”. Note that this model is inconsistent as interfacial friction does not conserve total (atmosphere plus ocean) momentum. This 5 model neglects the action of ocean currents on the atmospheric dynamics, it is used, for example, in ocean modeling when ocean models are forced by winds which are predeﬁned, available previous to the ocean simulation. The L1 and L2 models are called “one-way”. In L3 the ocean velocity is considered in the calculation of the shear force at the interface, when the atmosphere and ocean dynamics is considered: 10 In L3 the ocean velocity is considered in the calculation of the shear force at the interface, when the atmosphere and ocean dynamics is considered: 10 ∂tuL3 a = −Sm(uL3 a −uL3 o ) + ˜F (5) ∂tuL3 o = S (uL3 a −uL3 o ). (6) ∂tuL3 a = −Sm(uL3 a −uL3 o ) + ˜F (5) ∂tuL3 o = S (uL3 a −uL3 o ). (6) ∂tuL3 a = −Sm(uL3 a −uL3 o ) + ˜F ∂tuL3 o = S (uL3 a −uL3 o ). (5) (6) This model obeys Newton’s laws. Including the ocean velocity on the r.h.s. of eqs. (4) and (6) damps the ocean velocity and is recently referred to as the “eddy killing” term. It is found to have a considerable effect on the ocean dynamics (see i.e. Renault et al. (2017)). 15 This model obeys Newton’s laws. Including the ocean velocity on the r.h.s. of eqs. (4) and (6) damps the ocean velocity and is recently referred to as the “eddy killing” term. It is found to have a considerable effect on the ocean dynamics (see i.e. Renault et al (2017)) 5 For each of the linear models four different kinds of forcing are distinguihed. The ﬁrst is constant forcing starting at a time t = t0. The solutions are discussed in the next section and given in the appendix B1. These conﬁgurations are denoted LxK (x = 1,2,3) for the three different models mentioned above. The second is periodic forcing ˜F = cos(ωt). These models are denoted LxP (x = 1,2,3). x = 1,2,3) the atmosphere is directly forced by a the white noise, ˜F = F. urations LxW (x = 1,2,3) the atmosphere is directly forced by a the white noise, ˜F = F. 2 Local models Augmenting the phase space dimension to render a non-Markovian process Markovian is a standard procedure. 2 Local models The solutions for the atmospheric and the oceanic velocities as well as the second-order moments, t = t0. The solutions are discussed in the next section and given in the appendix B1. These conﬁgurations are denoted LxK (x = 1,2,3) for the three different models mentioned above. The second is periodic forcing ˜F = cos(ωt). These models are denoted LxP (x = 1,2,3). The solutions for the atmospheric and the oceanic velocities as well as the second-order moments, obtained by averaging over one period of τ = 2π/ω are discussed in the next section and given in the appendix B5. 20 White-in-time random forcing F is also considered. The parameter R measures the strength of the delta-correlated ﬂuctuat- ing force, it is: White-in-time random forcing F is also considered. The parameter R measures the strength of the delta-correlated ﬂuctuat- ing force, it is: R = ∞ Z 0 ⟨F(0)F(t′)⟩Ωdt′. (7) R = ∞ Z 0 ⟨F(0)F(t′)⟩Ωdt′. R = ∞ Z 0 ⟨F(0)F(t′)⟩Ωdt′. (7) In conﬁgurations LxW (x = 1,2,3) the atmosphere is directly forced by a the white noise, ˜F = In conﬁgurations LxW (x = 1,2,3) the atmosphere is directly forced by a the white noise, ˜F = F. In the fourth series of conﬁgurations, LxC (x = 1,2,3), the same models are forced using a colored noise, which is itself a 25 solution of a Langevin equation: 25 ∂t ˜F = −µ ˜F + F (8) ∂t ˜F = −µ ˜F + F (8) It is sometimes stated that without a clear-cut separation between the relaxation time-scale and the noise correlation-time the process is non-Markovian. This problem is avoided here by using a colored noise ( ˜F has the ﬁnite correlation-time µ−1), which is itself generated by a white-noise through a linear Langevin equation. Indeed, when F is white in time, the variable ˜F 30 5 5 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. is an Ornstein-Uhlenbeck process. When the atmosphere-ocean system is forced by ˜F and the three variable system ( ˜F,ua,uo) is considered, the system is a Markov-process, while the problem is non-Markovain in the two variable system (ua,uo). 3 Solutions of local models In the local linear models all solutions are analytic, for all types of forcing considered. These models are a ﬁrm testing ground 5 for all theories and statements on air-sea interaction. In the local linear models all solutions are analytic, for all types of forcing considered. These models are a ﬁrm testing ground 5 for all theories and statements on air-sea interaction. In the local linear models all solutions are analytic, for all types of forcing considered. These models are a ﬁrm testing ground 5 for all theories and statements on air-sea interaction. First, the unforced evolution of an initial state in the three models is compared. For the consistent model L3 eq. (A26) shows, that the total inertia in the system (ua+muo) is conserved and the shear between the atmosphere and the ocean \|ua−uo\| decays with a characteristic time-scale of (SM)−1. For the L1 model eq. (A12) shows that the total inertia is conserved and every perturbation in the atmosphere decays with a characteristic time scale of (Sm)−1 and adds to the the ocean m−1 times the 10 initial atmospheric perturbation at the same characteristic time scale. Ultimately all the inertia is in the ocean which has no i ﬂ th t h I th L2 d l t b ti i th t h d th d d i l ti that the total inertia in the system (ua+muo) is conserved and the shear between the atmosphere and the ocean \|ua−uo\| decays with a characteristic time-scale of (SM)−1. For the L1 model eq. (A12) shows that the total inertia is conserved and every perturbation in the atmosphere decays with a characteristic time scale of (Sm)−1 and adds to the the ocean m−1 times the 10 initial atmospheric perturbation at the same characteristic time scale. Ultimately all the inertia is in the ocean which has no inﬂuence on the atmosphere. In the L2 model perturbations in the atmosphere and the ocean decay and a spurious slow time- scale S−1 appears in the ocean dynamics, as can be seen from eq. (A19) . Replacing the L1 model by the L2 model is not always an improvement as it leads to a decay of all motion and introduces an artiﬁcial time-scale. 3 Solutions of local models perturbation in the atmosphere decays with a characteristic time scale of (Sm)−1 and adds to the the ocean m−1 times the 10 initial atmospheric perturbation at the same characteristic time scale. Ultimately all the inertia is in the ocean which has no inﬂuence on the atmosphere. In the L2 model perturbations in the atmosphere and the ocean decay and a spurious slow time- scale S−1 appears in the ocean dynamics, as can be seen from eq. (A19) . Replacing the L1 model by the L2 model is not always an improvement as it leads to a decay of all motion and introduces an artiﬁcial time-scale. 15 Second, the solutions of the different models, subject to the same forcing, are compared. Only the atmosphere is subject to 15 an external forcing. Two extreme cases can be distinguished. The ﬁrst is the short term response and the second is the long term evolution. To consider the ﬁrst question only the LxK conﬁgurations (see appendix B1), in which a constant forcing is turned on at t = t0, have to be considered. Every forcing can be decomposed into a sum (integral) over (possibly inﬁnitesimal) step functions. As the model is linear its dynamics is a sum (integral) over a (ﬁnite or inﬁnite number) of solutions with a single Second, the solutions of the different models, subject to the same forcing, are compared. Only the atmosphere is subject to 15 an external forcing. Two extreme cases can be distinguished. The ﬁrst is the short term response and the second is the long term evolution. To consider the ﬁrst question only the LxK conﬁgurations (see appendix B1), in which a constant forcing is turned on at t = t0, have to be considered. Every forcing can be decomposed into a sum (integral) over (possibly inﬁnitesimal) step functions. As the model is linear its dynamics is a sum (integral) over a (ﬁnite or inﬁnite number) of solutions with a single (possibly inﬁnitesimal) step. From the Taylor-series expansion of the LxK conﬁgurations in appendices B1 it emerges that the 20 short time response of the L1 and the L2 model is similar to the L3 model as the ﬁrst terms agree, for the atmosphere and the ocean and the two consecutive terms have coefﬁcients which differ by at most a factor of the order of m/M ≈1. 3 Solutions of local models In the L3P model Ξ = S2/(S2 + κ2) is always smaller than one and 5 oceanic velocities approach the atmospheric velocities, when the characteristic forcing time scale increases. The normalized correlation Θ = S/ √ S2 + κ2 = √ Ξ shows that the slower the forcing the higher is the correlation between the atmospheric and oceanic velocities. For the L1P model, Ξ = S2/κ2 which approaches the consistent L3P model for a high frequency forcing, only. Values larger than unity are however non-physical and so a forcing in which the oceanic forcing-time is larger than the oceanic friction-time can not be considered with this model. This is worrisome as a forcing of the atmospheric system contains 10 components of arbitrary long time-scales. Furthermore, Θ = 0, which means that the phase shift between the atmosphere and ocean is π/2. For the L2P model Ξ and Θ are identical to the L3P model and it is clearly a better choice than the L1P model. Some of the models with random forcing have a dynamics which is not statistically stationary and time averages depend on the length of the averaging interval. Time averages are therefore replaced by ensemble averages, taken over an ensemble (ω ∈Ω) of realizations of forcing functions Fω, they are denoted by ⟨.⟩Ω, where ω is one realization of the ensemble space Ω. 15 The dynamics starts from rest at t0 = 0, for convenience. When the forcing is Gaussian in the LxW and LxC conﬁgurations the probability density functions (pdfs) of the variables ˜F,ua,uo are centered Gaussians, ﬁrst order moments vanish and the pdfs are determined by their second-order moments. Using stochastic calculus (see appendix B9 and Wirth (2018)) the second-order moments are obtained analytically. The differences in the results of the models are, again, not only quantitative but qualitative. The total momentum, atmosphere 20 plus ocean, performs a random-walk (see appendix B9B10) in the L3W and L3C models, as it is not subject to damping. Superposed to this motion is the shear mode, which performs an Ornstein-Uhlenbeck process (see appendix B9B10). The L1 model has an atmospheric mode which performs an Ornstein-Uhlenbeck process and the oceanic mode that performs a random walk which is forced by the atmospheric dynamics on which it does not retro-act. 3 Solutions of local models Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. When the forcing applied to the atmosphere is periodic (Fa = cos(κt), see appendix B5) all solutions are periodic. The ratio of the square amplitude of the ocean and the atmospheric velocities and their normalized correlations are: When the forcing applied to the atmosphere is periodic (Fa = cos(κt), see appendix B5) all solutions are periodic. The ratio of the square amplitude of the ocean and the atmospheric velocities and their normalized correlations are: Ξ = ⟨u2 o⟩τ ⟨u2a⟩τ (9) Θ = ⟨uauo⟩τ p ⟨u2a⟩τ⟨u2o⟩τ , (10) (9) (10) where averages are taken over one period τ = 2π/κ. In the L3P model Ξ = S2/(S2 + κ2) is always smaller than one and 5 oceanic velocities approach the atmospheric velocities, when the characteristic forcing time scale increases. The normalized correlation Θ = S/ √ S2 + κ2 = √ Ξ shows that the slower the forcing the higher is the correlation between the atmospheric and oceanic velocities. For the L1P model, Ξ = S2/κ2 which approaches the consistent L3P model for a high frequency forcing, only. Values larger than unity are however non-physical and so a forcing in which the oceanic forcing-time is larger than the where averages are taken over one period τ = 2π/κ. In the L3P model Ξ = S2/(S2 + κ2) is always smaller than one and 5 oceanic velocities approach the atmospheric velocities, when the characteristic forcing time scale increases. The normalized correlation Θ = S/ √ S2 + κ2 = √ Ξ shows that the slower the forcing the higher is the correlation between the atmospheric and oceanic velocities. For the L1P model, Ξ = S2/κ2 which approaches the consistent L3P model for a high frequency forcing, only. Values larger than unity are however non-physical and so a forcing in which the oceanic forcing-time is larger than the where averages are taken over one period τ = 2π/κ. 3 Solutions of local models The long-term behaviors with constant forcing of the atmosphere are however completely different (see appendix B1). For the L3 model, both the atmospheric and oceanic velocity are unbounded and increase at the same rate. For the L1 model, the (possibly inﬁnitesimal) step. From the Taylor-series expansion of the LxK conﬁgurations in appendices B1 it emerges that the 20 short time response of the L1 and the L2 model is similar to the L3 model as the ﬁrst terms agree, for the atmosphere and the ocean and the two consecutive terms have coefﬁcients which differ by at most a factor of the order of m/M ≈1. The long-term behaviors with constant forcing of the atmosphere are however completely different (see appendix B1). For the L3 model, both the atmospheric and oceanic velocity are unbounded and increase at the same rate. For the L1 model, the the L3 model, both the atmospheric and oceanic velocity are unbounded and increase at the same rate. For the L1 model, the atmospheric velocity is bounded while the oceanic velocity is unbounded and increases at a rate that is M/m ≈1 higher as 25 compared to the L3 model. For the L2 model, both the atmospheric velocity and oceanic velocity are bounded. So differences are not only quantitative but also qualitative and the L1 performs better in a coupled simulation, when the ocean dynamics is considered. This is important to note, as it is the ocean dynamics community that favors a passage from the L1 to the L2 parameterization. atmospheric velocity is bounded while the oceanic velocity is unbounded and increases at a rate that is M/m ≈1 higher as 25 compared to the L3 model. For the L2 model, both the atmospheric velocity and oceanic velocity are bounded. So differences are not only quantitative but also qualitative and the L1 performs better in a coupled simulation, when the ocean dynamics is considered. This is important to note, as it is the ocean dynamics community that favors a passage from the L1 to the L2 parameterization. 6 When the forcing applied to the atmosphere is periodic (Fa = cos(κt), see appendix B5) all solutions are periodic. The ratio of the square amplitude of the ocean and the atmospheric velocities and their normalized correlations are: Ξ ⟨u2 o⟩τ (9) Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. 4 Fluctuation Dissipation Relation At the interface the ocean (eqs. (2), (4) and (6)) is subject to forcing by the atmosphere, given by Sua, and (in the L2 and At the interface the ocean (eqs. (2), (4) and (6)) is subject to forcing by the atmosphere, given by Sua, and (in the L2 and L3 models) dissipation, given by −Suo. They both are due to the same process and must be therefore related. In analogy 5 to Brownian motion I call this relation, when expressed by the second-order moments, the oceanic Fluctuation Dissipation Relation (FDR). The atmosphere (eq. (1), (3) and (5)) is subject to outside forcing (given by F), dissipation given by −Smua and, in the L3 model, to ﬂuctuations by the ocean, given by Smuo. The atmospheric FDR consist in the relation of the three processes in the equation of the second-order moments. Furthermore, the total momentum, atmosphere plus ocean, is subject L3 models) dissipation, given by −Suo. They both are due to the same process and must be therefore related. In analogy 5 to Brownian motion I call this relation, when expressed by the second-order moments, the oceanic Fluctuation Dissipation Relation (FDR). The atmosphere (eq. (1), (3) and (5)) is subject to outside forcing (given by F), dissipation given by −Smua and, in the L3 model, to ﬂuctuations by the ocean, given by Smuo. The atmospheric FDR consist in the relation of the three processes in the equation of the second-order moments. Furthermore, the total momentum, atmosphere plus ocean, is subject to external forcing but not to internal dissipation in the case of the L1 and L3 model. In the L2 model, the total momentum is 10 inﬂuenced by the forcing and the ocean velocity. The latter is nonphysical. 10 to external forcing but not to internal dissipation in the case of the L1 and L3 model. In the L2 model, the total momentum is 10 inﬂuenced by the forcing and the ocean velocity. The latter is nonphysical. As an example the FDR in the conﬁguration L3W is considered after the initial spin up, that is for t ≪(SM)−1. The FDR is obtained by multiplying eq. (5) by ua, eq. (6) by uo and ensemble averaging. 3 Solutions of local models In the L2 model a damping is added to the ocean mode as compared to the L1 model which leads to an Ornstein-Uhlenbeck process also in the ocean. In the L3 model 25 there is a constant and equal growth-rate of all second-order moments (see appendix B11). In the L1 model the linear growth is only present in the ocean with a growth rate which is by a factor (M/m)2 ≈1 higher than for the L3 model. In the L2 model all second order moments are bounded. All the results from analytical calculations are given in appendix B11 and tables 1 and 2. The differences between the models with white and colored noise (LxW and LxC) are only quantitative and tables 1 and 2 show that for short correlations times (µ ≪SM) the dynamics of the colored-noise cases converges to the corresponding 30 white-noise cases. It is important to note that some of these models do not lead to a (statistically) stationary state, but their ensemble aver- ages evolve in time. All the processes are, however, of stationary increment, that is, the time increments of random variables 7 7 (ua,uo,F) and linear combinations thereof are stationary (⟨x(t) −x(t + ∆t)⟩Ωdepends on ∆t but not on t). The dynamics is a sum of Ornstein-Uhlenbeck processes and random walks, which are all of statistically stationary increment. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-3 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. (ua,uo,F) and linear combinations thereof are stationary (⟨x(t) −x(t + ∆t)⟩Ωdepends on ∆t but not on t). The dynamics is a sum of Ornstein-Uhlenbeck processes and random walks, which are all of statistically stationary increment. (ua,uo,F) and linear combinations thereof are stationary (⟨x(t) −x(t + ∆t)⟩Ωdepends on ∆t but not on t). The dynamics is a sum of Ornstein-Uhlenbeck processes and random walks, which are all of statistically stationary increment. 4 Fluctuation Dissipation Relation The second-order moments are given by the correlation matrix (B41), which leads to: 1 2∂t⟨u2 a⟩Ω = SmR M 2    2t + m −2 SM \| {z } ﬂuctuation −2t −m2 + 4m SM \| {z } dissipation    + R \|{z} forcing = R M 2 (11) 15 1 2∂t⟨u2 o⟩Ω = SR M 2    2t + m −2 SM \| {z } ﬂuctuation −2t + 3 SM \| {z } dissipation    = R M 2 (12) 1 2∂t⟨u2 a + mu2 o⟩Ω = SmR M 2     −M S \|{z} dissipation    + R \|{z} forcing = R M . (13) 15 15 (11) (12) R 2     −M S \|{z} dissipation    + R \|{z} forcing = R M . (13) (13) For the atmosphere (eq. (11)) the ﬂuctuation terms are due to the atmosphere-ocean correlation, on average they drive the atmospheric dynamics, as the ocean dynamics reduces the friction, on average. The dissipation terms are due to the atmospheric auto-correlation. The forcing term drives the atmospheric dynamics and the whole system. The sum of the ﬂuctuating, the 0 di i ti d th f l d t t t i ti i f th l it i th t h For the atmosphere (eq. (11)) the ﬂuctuation terms are due to the atmosphere-ocean correlation, on average they drive the atmospheric dynamics, as the ocean dynamics reduces the friction, on average. The dissipation terms are due to the atmospheric For the atmosphere (eq. (11)) the ﬂuctuation terms are due to the atmosphere-ocean correlation, on average they drive the atmospheric dynamics, as the ocean dynamics reduces the friction, on average. The dissipation terms are due to the atmospheric auto-correlation. The forcing term drives the atmospheric dynamics and the whole system. The sum of the ﬂuctuating, the 20 dissipation and the force leads to a constant-in-time increase of the square velocity in the atmosphere. atmospheric dynamics, as the ocean dynamics reduces the friction, on average. The dissipation terms are due to the atmospheric auto-correlation. The forcing term drives the atmospheric dynamics and the whole system. The sum of the ﬂuctuating, the 20 dissipation and the force leads to a constant-in-time increase of the square velocity in the atmosphere. 20 auto-correlation. 4 Fluctuation Dissipation Relation The latter compares the atmosphere- and ocean-variance, while the former is based on the difference of the ocean-variance and the ocean-atmosphere correlation, only. 25 4 Fluctuation Dissipation Relation 15 It is essential to note that in the linear models discussed in this subsection the forcing can be a linear combination of different forcing proposed with different periods and correlation times. The second-order moments are the sum of the individual second- order moments, that is cross-correlations of variables with a different forcing vanish. ) y q y p g g y It is essential to note that in the linear models discussed in this subsection the forcing can be a linear combination of different forcing proposed with different periods and correlation times. The second-order moments are the sum of the individual second- order moments, that is cross-correlations of variables with a different forcing vanish. It is essential to note that in the linear models discussed in this subsection the forcing can be a linear combination of different forcing proposed with different periods and correlation times. The second-order moments are the sum of the individual second- order moments, that is cross-correlations of variables with a different forcing vanish. When the forcing is a combination of a random and a periodic forcing it is important to note that the periodic part does not contribute to the (linear) growth-rate and that it also does not contribute to the difference in the correlation between the ocean- 20 variance and the ocean-atmosphere correlation, both facts are related. The periodic part is however important when it comes to evaluating the difference between the atmosphere- and ocean-variance. This is a possible explanation, why the estimation of the friction parameter was successful in Wirth (2018) but not the mass-ratio between the atmosphere and the ocean. The latter compares the atmosphere- and ocean-variance, while the former is based on the difference of the ocean-variance and the ocean-atmosphere correlation, only. 25 When the forcing is a combination of a random and a periodic forcing it is important to note that the periodic part does not contribute to the (linear) growth-rate and that it also does not contribute to the difference in the correlation between the ocean- 20 variance and the ocean-atmosphere correlation, both facts are related. The periodic part is however important when it comes to evaluating the difference between the atmosphere- and ocean-variance. This is a possible explanation, why the estimation of the friction parameter was successful in Wirth (2018) but not the mass-ratio between the atmosphere and the ocean. 4 Fluctuation Dissipation Relation This is a double ﬂuctuation-dissipation relation: the dissipation and the ﬂuctuation are related, ﬁrstly, by the equal growth rate of their squares and secondly by the difference between them. Brownian motion leads to an equipartition of energy between molecules and Brownian particles which in air-sea interaction is substituted by the equal growth rate of the square velocities of the Equations (11) and (12) show the equilibrium between the ﬂuctuation-dissipation, the forcing and the energy growth. This is a double ﬂuctuation-dissipation relation: the dissipation and the ﬂuctuation are related, ﬁrstly, by the equal growth rate of their 5 squares and secondly by the difference between them. Brownian motion leads to an equipartition of energy between molecules 5 and Brownian particles, which, in air-sea interaction is substituted by the equal growth rate of the square velocities of the atmosphere, the ocean and their covariance. When observing the second-order moments the parameters in the linear model are given by: When observing the second-order moments the parameters in the linear model are given by: S = ∂t⟨u2 o⟩Ω 2⟨uauo −u2o⟩Ω (14) M = ⟨u2 a −u2 o⟩Ω ⟨uauo −u2o⟩Ω . (15) 0 S = ∂t⟨u2 o⟩Ω 2⟨uauo −u2o⟩Ω M = ⟨u2 a −u2 o⟩Ω ⟨uauo −u2o⟩Ω . 10 (14) (15) It is straightforward to determine the FDR for LW1 and LW2 using results from the appendix B11. In the L1 and L2 model the ﬂuctuation is neglected in the atmosphere and only the forcing term and the dissipation term are present. In the L1 model the dissipation is neglected in the ocean and only the ﬂuctuation is present and the ocean performs a random-walk. A FDR can be established for the conﬁgurations with a forcing by a colored noise (LxC), in the same way (see appendix B11). They show qualitatively the same behavior than the corresponding conﬁgurations forced by a white noise. 15 It is essential to note that in the linear models discussed in this subsection the forcing can be a linear combination of different forcing proposed with different periods and correlation times. The second-order moments are the sum of the individual second- order moments, that is cross-correlations of variables with a different forcing vanish. A FDR can be established for the conﬁgurations with a forcing by a colored noise (LxC), in the same way (see appendix B11). They show qualitatively the same behavior than the corresponding conﬁgurations forced by a white noise. 4 Fluctuation Dissipation Relation The forcing term drives the atmospheric dynamics and the whole system. The sum of the ﬂuctuating, the 20 dissipation and the force leads to a constant-in-time increase of the square velocity in the atmosphere. 20 Concerning the ocean (eq. (12)) the ﬂuctuation terms originate from the atmosphere-ocean correlation, they are driving the ocean dynamics, on average. The oceanic auto-correlation leads to the dissipation terms. The ﬁrst and third term, due to the total-inertia mode in the atmosphere and the ocean cancel, as the total inertia mode performs a random walk (Wiener-process) with no shear associated to it. The second and forth term due to the shear-mode in the atmosphere and the ocean, respectively, 25 lead to a statistically stationary dynamics, an Ornstein-Uhlenbeck process. The sum the ﬂuctuating force and the dissipation leads to a constant-in-time increase of the square velocity in the ocean which is equal to the increase in the atmosphere. The with no shear associated to it. The second and forth term due to the shear-mode in the atmosphere and the ocean, respectively, 25 lead to a statistically stationary dynamics, an Ornstein-Uhlenbeck process. The sum the ﬂuctuating force and the dissipation leads to a constant-in-time increase of the square velocity in the ocean which is equal to the increase in the atmosphere. The 8 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. atmospheric ﬂuctuations are the only forcing acting on the ocean The total energy is forced by the exterior and dissipated due to the internal shear (eq. (13)). atmospheric ﬂuctuations are the only forcing acting on the ocean The total energy is forced by the exterior and dissipated due to the internal shear (eq. (13)). Equations (11) and (12) show the equilibrium between the ﬂuctuation-dissipation, the forcing and the energy growth. This is a double ﬂuctuation-dissipation relation: the dissipation and the ﬂuctuation are related, ﬁrstly, by the equal growth rate of their squares and secondly by the difference between them. Brownian motion leads to an equipartition of energy between molecules and Brownian particles, which, in air-sea interaction is substituted by the equal growth rate of the square velocities of the atmosphere, the ocean and their covariance. Equations (11) and (12) show the equilibrium between the ﬂuctuation-dissipation, the forcing and the energy growth. 5 Energetics The ﬂuxes of kinetic energy in the system, are detailed in ﬁg. 1. The forcing injects energy in the atmosphere, part of it leads to an increase of the energy of the atmosphere Pa and part of it is transferred to the atmosphere-ocean interface Pa→i. At the interface energy is transferred between the atmosphere and the ocean and also dissipated Pdissip. The energy ﬂux to the ocean Pi→o leads to an energy increase Po. The dynamics in the system does not, for all models, converge to a stationary state, in 30 The ﬂuxes of kinetic energy in the system, are detailed in ﬁg. 1. The forcing injects energy in the atmosphere, part of it leads to an increase of the energy of the atmosphere Pa and part of it is transferred to the atmosphere-ocean interface Pa→i. At the interface energy is transferred between the atmosphere and the ocean and also dissipated Pdissip. The energy ﬂux to the ocean Pi→o leads to an energy increase Po. The dynamics in the system does not, for all models, converge to a stationary state, in 30 which Pa = Po = 0. Pi→o leads to an energy increase Po. The dynamics in the system does not, for all models, converge to a stationary state, in 30 which Pa = Po = 0. 30 9 9 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. atmosphere: Pa PF interface: Pdissip ocean: Po Pi→o Pa→i Figure 1. Schematic of energy ﬂuxes in the atmosphere ocean system Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. atmosphere: Pa PF interface: Pdissip ocean: Po Pi→o Pa→i Figure 1. Schematic of energy ﬂuxes in the atmosphere ocean system ocean: Po Figure 1. Schematic of energy ﬂuxes in the atmosphere ocean system Figure 1. 5 Energetics Schematic of energy ﬂuxes in the atmosphere ocean system When the forcing is periodic averages over one period are taken and when the forcing is stochastic ensemble averages are performed. For convenience the same symbol ⟨.⟩denotes the averages, when the forcing is periodic ⟨.⟩= ⟨.⟩τ and when the forcing is stochastic ⟨.⟩= ⟨.⟩Ω. The power injected in the system is PF = ⟨ua ˜F⟩. The increase of energy in the atmosphere Pa = 1 2∂t⟨u2 a⟩and the ocean Po = m 2 ∂t⟨u2 o⟩are obtained by ﬁrst multiplying eqs. (1, 3, 5) by ua and eqs. (2, 4, 6) by uo and then averaging. In the consistent model L3, the auto-correlation of the atmosphere and the ocean dissipate the energy in the 5 atmosphere and the ocean, respectively. In the L1 model the dissipation in the ocean is omitted. By looking at the equations the important role of the correlation between the atmospheric and oceanic velocity (the ﬂuctuation term) ⟨uauo⟩becomes clear. On average it is non-negative in all models and conﬁgurations. In the consistent model L3 it has two effects it reduces the energy dissipation in the atmosphere and injects the energy in the ocean, both are equal, not only in magnitude but also in sign then averaging. In the consistent model L3, the auto-correlation of the atmosphere and the ocean dissipate the energy in the 5 atmosphere and the ocean, respectively. In the L1 model the dissipation in the ocean is omitted. By looking at the equations the important role of the correlation between the atmospheric and oceanic velocity (the ﬂuctuation term) ⟨uauo⟩becomes clear. On average it is non-negative in all models and conﬁgurations. In the consistent model L3 it has two effects it reduces the energy dissipation in the atmosphere and injects the energy in the ocean, both are equal, not only in magnitude but also in sign (+Sm⟨uauo⟩). In the L1 and L2 models the reduction of the energy dissipation in the atmosphere, due to ocean velocities, 10 is omitted. Therefore, the L1 model suffers from an increased energy dissipation in the atmosphere and a reduced energy dissipation in the ocean, while in the L2 model only the ﬁrst is present. (+Sm⟨uauo⟩). In the L1 and L2 models the reduction of the energy dissipation in the atmosphere, due to ocean velocities, 10 is omitted. 5 Energetics Therefore, the L1 model suffers from an increased energy dissipation in the atmosphere and a reduced energy dissipation in the ocean, while in the L2 model only the ﬁrst is present. Pa→i = PF −Pa (16) Pi→o = Po (17) 15 Pdissip = Pa→i −Pi→o = PF −Pa −Po. (18) Pa→i = PF −Pa (16) Pi→o = Po (17) 15 Pdissip = Pa→i −Pi→o = PF −Pa −Po. (18) (16) (17) (18) (17) (18) If the ﬂuxes are time independent, or long-time averages are taken, they are all non-negative for the L2 and L3 model (see also the discussion of the Fluctuation Theorem in section 7). In the L1 model Pdissip < 0 when uoua > u2 a, which is ofcourse unphysical. 10 Exp. PF Pa/PF Pa→i/PF Pdissip PF Po/PF Pa+Po PF Pa/Pa→i Pa+Po Pdissip η L1P Sm κ2+(Sm)2 0 Sm κ2+(Sm)2 Sm κ2+(Sm)2 0 0 0 0 0 L2P Sm κ2+(Sm)2 0 Sm κ2+(Sm)2 Sm κ2+(Sm)2 0 0 0 0 0 L3P Sm κ2+(SM)2 0 Sm κ2+(SM)2 Sm κ2+(SM)2 0 0 0 0 0 L1W R 0 1 m2−1 m2 1 m2 1 m2 0 1 m2−1 1 m2 L2W R 0 1 1 0 0 0 0 0 L3W R 1 M2 M2−1 M2 m M m M2 1 M 1 M2−1 1 m 1 M+1 L1C R µ(µ+Sm) 0 1 µ(m2−1)−Sm µm2 µ+Sm µm2 µ+Sm µm2 0 µ+Sm µ(m2−1)−Sm µ−Sm µm2 L2C R µ(µ+SM) 0 1 1 0 0 0 0 0 L3C R(µ+S) µ2(µ+SM) µ+SM M2(µ+S) µ(M2−1)+SMm M2(µ+S) µm M(µ+S) m(µ+SM) M2(µ+S) µ+SM M(µ+S) µ+SM µ(M2−1)+SMm µ+SM µm µ+SM µ(M+1)+SM Table 1. Energy ﬂuxes for t ≫(SM)−1,µ−1. The last column is the efﬁciency in the system as it compares the energy growth in the system to the energy injection. Note that for µ ≫SM, LCx converges to LWx if R →Rµ2. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Exp. 5 Energetics PF Pa/PF Pa→i/PF Pdissip PF Po/PF Pa+Po PF Pa/Pa→i Pa+Po Pdissip η L1P Sm κ2+(Sm)2 0 Sm κ2+(Sm)2 Sm κ2+(Sm)2 0 0 0 0 0 L2P Sm κ2+(Sm)2 0 Sm κ2+(Sm)2 Sm κ2+(Sm)2 0 0 0 0 0 L3P Sm κ2+(SM)2 0 Sm κ2+(SM)2 Sm κ2+(SM)2 0 0 0 0 0 L1W R 0 1 m2−1 m2 1 m2 1 m2 0 1 m2−1 1 m2 L2W R 0 1 1 0 0 0 0 0 L3W R 1 M2 M2−1 M2 m M m M2 1 M 1 M2−1 1 m 1 M+1 L1C R µ(µ+Sm) 0 1 µ(m2−1)−Sm µm2 µ+Sm µm2 µ+Sm µm2 0 µ+Sm µ(m2−1)−Sm µ−Sm µm2 L2C R µ(µ+SM) 0 1 1 0 0 0 0 0 L3C R(µ+S) µ2(µ+SM) µ+SM M2(µ+S) µ(M2−1)+SMm M2(µ+S) µm M(µ+S) m(µ+SM) M2(µ+S) µ+SM M(µ+S) µ+SM µ(M2−1)+SMm µ+SM µm µ+SM µ(M+1)+SM Table 1. Energy ﬂuxes for t ≫(SM)−1,µ−1. The last column is the efﬁciency in the system as it compares the energy growth in the system to the energy injection. Note that for µ ≫SM, LCx converges to LWx if R →Rµ2. to the energy injection. Note that for µ ≫SM, LCx converges to LWx if R →Rµ2. Using eqs. (16) - (18) allows to calculate the remaining energy ﬂuxes. Note that in L3, Pdissip = Sm⟨(ua −uo)2⟩. Detailed results for all the energy ﬂuxes are given in table 1. Using eqs. (16) - (18) allows to calculate the remaining energy ﬂuxes. Note that in L3, Pdissip = Sm⟨(ua −uo)2⟩. Detailed results for all the energy ﬂuxes are given in table 1. The efﬁciency of the power transfer, is the power gained by the ocean at the interface divided by the power lost by the atmosphere at the interface: The efﬁciency of the power transfer, is the power gained by the ocean at the interface divided by the power lost by the atmosphere at the interface: η = Pi→o Pa→i = 1 −Pdissip Pa→i . 5 η = Pi→o Pa→i = 1 −Pdissip Pa→i . (19) 5 (19) The important question of the efﬁciency of the power transfer in the air-sea system, its dependence on the parameters and its representation in different models, has, to the best of my knowledge, never been addressed. 6 Fluctuation Dissipation Theorem : Response Theory The Fluctuation Dissipation Theorem (FDT) compares the response of the system subject to an external perturbation to the 5 internal ﬂuctuations of the system. This is related to the Onsager’s principle which states that the system relaxes from a forced state to the unforced dynamics, in the same manner as if the forced state were due to an internal ﬂuctuation of the system. The expressions FDT, Onsager’s principle and also response theory are often interchanged in applications precise deﬁnitions are given in appendixes C1 and C2. Mathematically speaking: if x(t) is the state vector of the system the correlation matrix is C(t ∆t) = ⟨x(t)xt(t + ∆t)⟩and the normalized correlation matrix is C(t ∆t)C(t 0)−1 The average decay of a perturbation 10 The Fluctuation Dissipation Theorem (FDT) compares the response of the system subject to an external perturbation to the 5 internal ﬂuctuations of the system. This is related to the Onsager’s principle which states that the system relaxes from a forced state to the unforced dynamics, in the same manner as if the forced state were due to an internal ﬂuctuation of the system. The expressions FDT, Onsager’s principle and also response theory are often interchanged in applications precise deﬁnitions are given in appendixes C1 and C2. Mathematically speaking: if x(t) is the state vector of the system the correlation matrix is C(t,∆t) = ⟨x(t)xt(t + ∆t)⟩and the normalized correlation matrix is C(t,∆t)C(t,0)−1. The average decay of a perturbation 10 ¯x is given by the perturbation matrix, ⟨x(t + ∆t)⟩= χ(t,∆t)¯x. The FDT holds if: C(t,∆t)C(t,0)−1 = χ(t,∆t). (20) The processes considered here are of stationary increment and the perturbation matrices are independent of the actual time t and so are the normalized correlation matrices (see also appendices A1 - A3 and B11). The application of the FDT in the case The processes considered here are of stationary increment and the perturbation matrices are independent of the actual time t and so are the normalized correlation matrices (see also appendices A1 - A3 and B11). The application of the FDT in the case of simple Langevin equations is discussed in appendix C1 for the white-noise forcing and in appendix C2, for forcing with a 15 colored-noise. of simple Langevin equations is discussed in appendix C1 for the white-noise forcing and in appendix C2, for forcing with a 15 colored-noise. 5 Energetics ⟨u2 a −uauo⟩ ⟨uauo −u2 o⟩ ⟨u2 a −u2 o⟩ ⟨(ua −uo)2⟩ L1P 1 −S2 κ2 1 −S2 κ2 1 −S2 κ2 L2P κ2 S2+κ2 0 κ2 S2+κ2 κ2 S2+κ2 L3P 1 0 1 1 L1W m−1 Sm2 t-dep. t-dep. t-dep. L2W m SM2 0 m SM2 m SM2 L3W M+1 SM2 1 SM2 M+2 SM2 M SM2 L1C (m−1)µ−Sm Sm2(µ+SM) t-dep. t-dep. t-dep. L2C µ2 SM(S+µ)(Sm+µ) 0 µ2 SM(S+µ)(Sm+µ) µ2 SM(S+µ)(Sm+µ) L3C SM+(M+1)µ SM2(SM+µ) 1 SM2 2SM+(M+2)µ SM2(SM+µ) Mµ SM2(SM+µ) Table 2. Differences of second-order-moments of the velocity (normalized by 2(κ2 +(Sm)2) for L1P and L2P, by 2(κ2 +(SM)2) for L3P, by R for LxW and by R/µ2 for LxC). Note that for µ ≫SM, LxC converges to LxW the symmetry of the L3 model it is evident that a necessary condition for the atmosphere to receive energy at the interface is u2 o > u2 a. It is important to note that a less energetic forced atmosphere can do work on the ocean (when m > 1) . In the L1 model the ocean receives energy whenever ua · uo > 0. the symmetry of the L3 model it is evident that a necessary condition for the atmosphere to receive energy at the interface is u2 o > u2 a. It is important to note that a less energetic forced atmosphere can do work on the ocean (when m > 1) . In the L1 model the ocean receives energy whenever ua · uo > 0. 5 Energetics Note that when no averaging is performed, the instantaneous ﬂuxes in a single experiment is considered, η = uo/ua in the L1 and L3 model, while η = (1 − uo/ua)uo/ua for L2. When η > 1 the ocean is providing energy to the atmosphere. When the constant forcing is considered, the initial behavior of the efﬁciency is identical, to leading order, in the three model. The long-term behaviors differ: in L1 η 10 grows linearly to inﬁnity, in L2 it converges to η = 0 and in L3 int converges to η = 1. A striking feature, shown in table 1, is that for the different models the efﬁciency is of different order in the mass ratio m, when random forcing, white or colored, is applied. So again, the differences are not only quantitative, expressed by different prefactors, but they are clearly qualitative. For the L3 model, the only model that respect Newton’s laws, all second order moments have the same constant growth rate and so the differences of these second order moments are constant in time, they are given in table 2. 15 In a perfect gas in equilibrium with atoms of different mass the kinetic energy of each atom, measured by the temperature, is equal on average and heat ﬂows on average from the hotter to the colder substance (second law of thermodynamics). For the forced and dissipative air-sea interaction of the L2 and L3 models, the energetic inﬂuence of the interface on the ocean is: Sm(uauo −u2 o) which shows that a necessary condition for the ocean to receive energy at the interface is: u2 a > u2 o. This is also true when averages are taken ⟨u2 a⟩> ⟨u2 o⟩, a consequence of the Cauchy-Schwarz inequality. This is reﬂected in the 20 results presented in table 2, as all entries of the third column, giving ⟨u2 a −u2 o⟩for the L2 and L3 model are positive. Due to 11 Exp. ⟨u2 a −uauo⟩ ⟨uauo −u2 o⟩ ⟨u2 a −u2 o⟩ ⟨(ua −uo)2⟩ L1P 1 −S2 κ2 1 −S2 κ2 1 −S2 κ2 L2P κ2 S2+κ2 0 κ2 S2+κ2 κ2 S2+κ2 L3P 1 0 1 1 L1W m−1 Sm2 t-dep. t-dep. t-dep. L2W m SM2 0 m SM2 m SM2 L3W M+1 SM2 1 SM2 M+2 SM2 M SM2 L1C (m−1)µ−Sm Sm2(µ+SM) t-dep. t-dep. t-dep. 5 Energetics L2C µ2 SM(S+µ)(Sm+µ) 0 µ2 SM(S+µ)(Sm+µ) µ2 SM(S+µ)(Sm+µ) L3C SM+(M+1)µ SM2(SM+µ) 1 SM2 2SM+(M+2)µ SM2(SM+µ) Mµ SM2(SM+µ) Table 2. Differences of second-order-moments of the velocity (normalized by 2(κ2 +(Sm)2) for L1P and L2P, by 2(κ2 +(SM)2) for L3P, by R for LxW and by R/µ2 for LxC). Note that for µ ≫SM, LxC converges to LxW Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Exp. ⟨u2 a −uauo⟩ ⟨uauo −u2 o⟩ ⟨u2 a −u2 o⟩ ⟨(ua −uo)2⟩ L1P 1 −S2 κ2 1 −S2 κ2 1 −S2 κ2 L2P κ2 S2+κ2 0 κ2 S2+κ2 κ2 S2+κ2 L3P 1 0 1 1 L1W m−1 Sm2 t-dep. t-dep. t-dep. L2W m SM2 0 m SM2 m SM2 L3W M+1 SM2 1 SM2 M+2 SM2 M SM2 L1C (m−1)µ−Sm Sm2(µ+SM) t-dep. t-dep. t-dep. L2C µ2 SM(S+µ)(Sm+µ) 0 µ2 SM(S+µ)(Sm+µ) µ2 SM(S+µ)(Sm+µ) L3C SM+(M+1)µ SM2(SM+µ) 1 SM2 2SM+(M+2)µ SM2(SM+µ) Mµ SM2(SM+µ) Table 2. Differences of second-order-moments of the velocity (normalized by 2(κ2 +(Sm)2) for L1P and L2P, by 2(κ2 +(SM)2) for L3P, by R for LxW and by R/µ2 for LxC). Note that for µ ≫SM, LxC converges to LxW Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Exp. ⟨u2 a −uauo⟩ ⟨uauo −u2 o⟩ ⟨u2 a −u2 o⟩ ⟨(ua −uo)2⟩ L1P 1 −S2 κ2 1 −S2 κ2 1 −S2 κ2 L2P κ2 S2+κ2 0 κ2 S2+κ2 κ2 S2+κ2 L3P 1 0 1 1 L1W m−1 Sm2 t-dep. t-dep. t-dep. L2W m SM2 0 m SM2 m SM2 L3W M+1 SM2 1 SM2 M+2 SM2 M SM2 L1C (m−1)µ−Sm Sm2(µ+SM) t-dep. t-dep. t-dep. L2C µ2 SM(S+µ)(Sm+µ) 0 µ2 SM(S+µ)(Sm+µ) µ2 SM(S+µ)(Sm+µ) L3C SM+(M+1)µ SM2(SM+µ) 1 SM2 2SM+(M+2)µ SM2(SM+µ) Mµ SM2(SM+µ) Table 2. Differences of second-order-moments of the velocity (normalized by 2(κ2 +(Sm)2) for L1P and L2P, by 2(κ2 +(SM)2) for L3P, by R for LxW and by R/µ2 for LxC). Note that for µ ≫SM, LxC converges to LxW Exp. 6 Fluctuation Dissipation Theorem : Response Theory Even in the case of the random walk, where the perturbation (forced or internal) does not decay, the FDT applies, as eq. (20) is veriﬁed. 5 For the colored-noise forcing the perturbation matrix in the augmented phase space is given for the L1C model in eq. (B53), for the L2C model in eq. (B64) and for the L3C model in eq. (B76). The failure of the FDT for a colored-noise forcing in a phase 10 space consisting only of the velocities is due to the fact that the forcing has a ﬁnite correlation time and so the future forcing is correlated to the actual velocities and the correlations of the velocities is not the same as described by the perturbation matrix. The decay of a perturbation is therefore dependent on how the perturbation was reached, that is the system is not Markovian. This is shown in appendix C2 and discussed in detail by Balakrishnan (1979). for the L2C model in eq. (B64) and for the L3C model in eq. (B76). The failure of the FDT for a colored-noise forcing in a phase 10 space consisting only of the velocities is due to the fact that the forcing has a ﬁnite correlation time and so the future forcing is correlated to the actual velocities and the correlations of the velocities is not the same as described by the perturbation matrix. The decay of a perturbation is therefore dependent on how the perturbation was reached, that is the system is not Markovian. This is shown in appendix C2 and discussed in detail by Balakrishnan (1979). The FDT relies strongly on Gaussian statistics (see e.g. Cooper and Haynes (2011)) of the variables, which is assured 15 in linear-models but the statistics in non-linear models is clearly non-Gaussian. As, to the best of my knowledge, no analytic solution exists for the non-linear versions of the air-sea interaction models the FDT has to be explored by numerical experiment. This will be done elsewhere. The FDT relies strongly on Gaussian statistics (see e.g. Cooper and Haynes (2011)) of the variables, which is assured 15 in linear-models but the statistics in non-linear models is clearly non-Gaussian. As, to the best of my knowledge, no analytic solution exists for the non-linear versions of the air-sea interaction models the FDT has to be explored by numerical experiment. This will be done elsewhere. 7 Fluctuation Theorem The average states and ﬂuxes in the different models investigated as a function of their parameters are given in section 5. 20 The probability density functions of the variables representing the atmospheric and ocean velocities are centered Gaussian variables and they are therefore completely described by their variance. Fluxes are products of Gaussian random-variables and are not Gaussian. The present section discusses instantaneous deviations from the average values, the ﬂuctuations of the system and their persistence in time. To this end the Fluctuation Theorem (FT) (Gallavotti and Cohen (1995a), Gallavotti and Cohen The average states and ﬂuxes in the different models investigated as a function of their parameters are given in section 5. 20 The probability density functions of the variables representing the atmospheric and ocean velocities are centered Gaussian variables and they are therefore completely described by their variance. Fluxes are products of Gaussian random-variables and are not Gaussian. The present section discusses instantaneous deviations from the average values, the ﬂuctuations of the system and their persistence in time. To this end the Fluctuation Theorem (FT) (Gallavotti and Cohen (1995a), Gallavotti and Cohen The average states and ﬂuxes in the different models investigated as a function of their parameters are given in section 5. 20 The probability density functions of the variables representing the atmospheric and ocean velocities are centered Gaussian variables and they are therefore completely described by their variance. Fluxes are products of Gaussian random-variables and are not Gaussian. The present section discusses instantaneous deviations from the average values, the ﬂuctuations of the system and their persistence in time. To this end the Fluctuation Theorem (FT) (Gallavotti and Cohen (1995a), Gallavotti and Cohen (1995b), Ciliberto et al. (2004)) is discussed for the ﬂuxes Pa→i;ω and Pi→o;ω. The FT expresses properties of these quantities 25 evaluated along ﬂuctuating trajectories (indexed by ω ∈Ω). The second law of thermodynamics states that heat always ﬂows spontaneously from hotter to colder bodies. The FT speciﬁes that this property is true on average but locally in-time-and-space counter ﬂuxes are present. The relation of the probability of positive versus negative ﬂuxes of a given magnitude and their persistence in time is subject of the FT. The concepts of the FT are applied to a variety of problems and quantities and is also extended to deterministic dynamical 30 systems. In the present work the analysis of Ciliberto et al. 6 Fluctuation Dissipation Theorem : Response Theory 15 The calculations concerning the application of the FDT for the models considered is given in appendix B9 eq. (B38), which show that the FDT applies in the models with the white-noise forcing and in the model with the colored-noise forcing, when the phase space is augmented by the variable ˜F representing the colored-noise forcing. Indeed, the FDT can be veriﬁed in these 12 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. models by explicitly calculating and comparing the matrices in eq. (20). The normalized correlation matrix with a white noise forcing, which is equal to the perturbation matrix, is given for the L1 model in eq. (A12), for the L2 model in eq. (A19) and for the L3 model in eq. (A26). The decay of an initial perturbation in the models was discussed in detail at the beginning of section 3. In the L1 model the ocean dynamics is undamped (see eq. (A12)) and performs a random-walk. In the L3 model the total momentum is undamped (see eq. (A26)) and performs a random-walk. The random walk, has the martingale property, that is 5 the expectation for future values is equal to the actual value. This leads to an inﬁnite memory in the process and inﬁnite long correlations ⟨utotal(t)utotal(t + ∆t)⟩= ⟨utotal(t)2⟩∀∆t > 0. Even in the case of the random walk, where the perturbation (forced or internal) does not decay, the FDT applies, as eq. (20) is veriﬁed. models by explicitly calculating and comparing the matrices in eq. (20). The normalized correlation matrix with a white noise forcing, which is equal to the perturbation matrix, is given for the L1 model in eq. (A12), for the L2 model in eq. (A19) and for the L3 model in eq. (A26). The decay of an initial perturbation in the models was discussed in detail at the beginning of section 3. In the L1 model the ocean dynamics is undamped (see eq. (A12)) and performs a random-walk. In the L3 model the total momentum is undamped (see eq. (A26)) and performs a random-walk. The random walk, has the martingale property, that is 5 the expectation for future values is equal to the actual value. This leads to an inﬁnite memory in the process and inﬁnite long correlations ⟨utotal(t)utotal(t + ∆t)⟩= ⟨utotal(t)2⟩∀∆t > 0. 7 Fluctuation Theorem (2004) who tested the FT in two examples of turbulent ﬂows in laboratory experiments is applied. Section 5 showed that on average the atmosphere gains energy by the random forcing and looses energy at the interface and the ocean gains energy at the interface. Also in the case of air-sea interaction instantaneous The concepts of the FT are applied to a variety of problems and quantities and is also extended to deterministic dynamical 30 systems. In the present work the analysis of Ciliberto et al. (2004) who tested the FT in two examples of turbulent ﬂows in laboratory experiments is applied. Section 5 showed that on average the atmosphere gains energy by the random forcing and looses energy at the interface and the ocean gains energy at the interface. Also in the case of air-sea interaction instantaneous The concepts of the FT are applied to a variety of problems and quantities and is also extended to deterministic dynamical 30 systems. In the present work the analysis of Ciliberto et al. (2004) who tested the FT in two examples of turbulent ﬂows in laboratory experiments is applied. Section 5 showed that on average the atmosphere gains energy by the random forcing and looses energy at the interface and the ocean gains energy at the interface. Also in the case of air-sea interaction instantaneous 13 ﬂuxes can go in the opposite direction (Moulin and Wirth (2016)). The FT quantiﬁes the asymmetry of the pdf of averages of the ﬂuxes with respect to zero. It compares the probability of having a positive event to the probability of having a negative event of the same magnitude for averages of the ﬂuxes over a time interval τ. Do the symmetries implied by the Fluctuation Theorem (FT) apply to the momentum ﬂuxes Pa→i;ω and Pi→o;ω? The ﬂuxes are quadratic quantities and their statistics is therefore not Gaussian. Recently a closed form of the probability density function f(Z) of the product of two correlated Gaussian variables 5 Z = XY with vanishing means and variances σ2 x and σ2 y and correlation ρ has been obtained (Nadarajah and Pogány (2016) and Gaunt (2018)) in terms of a modiﬁed Bessel function of the second kind of order zero K0(z) = R ∞ 0 cos(z sinht)dt: ﬂuxes can go in the opposite direction (Moulin and Wirth (2016)). 7 Fluctuation Theorem The normalized time average over an interval τ is denoted by: Z(t) τ = 1 τ⟨Z(t)⟩ τ Z 0 Z(t + τ ′)dτ ′. (23) Z(t) τ = 1 τ⟨Z(t)⟩ τ Z 0 Z(t + τ ′)dτ ′. (23) When the interval τ is lager than the characteristic time of the system, the FT holds when: When the interval τ is lager than the characteristic time of the system, the FT holds when: SZ τ (z) = στz, (24) 15 SZ τ (z) = στz, 15 (24) e variable σ is called the contraction rate (see Ciliberto et al. (2004)), it depends on the problem con The power the atmosphere looses at the interface Pa→i;ω and the power the ocean receives from the atmosphere at the interface Pi→o;ω along a trajectory ω ∈Ωis investigated. Both differ by the work dissipated at the interface (see section 5). The ensemble averages of all these quantities are positive, but negative ﬂuxes exist, even when temporal averages over time The power the atmosphere looses at the interface Pa→i;ω and the power the ocean receives from the atmosphere at the interface Pi→o;ω along a trajectory ω ∈Ωis investigated. Both differ by the work dissipated at the interface (see section 5). The ensemble averages of all these quantities are positive, but negative ﬂuxes exist, even when temporal averages over time intervals of length τ are considered. The FT states that the probability of ﬁnding a positive ﬂux of magnitude z divided by the 20 probability of a negative ﬂux with the same magnitude increases exponentially with the value z and the averaging period τ The power the atmosphere looses at the interface Pa→i;ω and the power the ocean receives from the atmosphere at the interface Pi→o;ω along a trajectory ω ∈Ωis investigated. Both differ by the work dissipated at the interface (see section 5). The ensemble averages of all these quantities are positive, but negative ﬂuxes exist, even when temporal averages over time intervals of length τ are considered. The FT states that the probability of ﬁnding a positive ﬂux of magnitude z divided by the 20 probability of a negative ﬂux with the same magnitude increases exponentially with the value z and the averaging period τ. intervals of length τ are considered. 7 Fluctuation Theorem The FT quantiﬁes the asymmetry of the pdf of averages of the ﬂuxes with respect to zero. It compares the probability of having a positive event to the probability of having a negative event of the same magnitude for averages of the ﬂuxes over a time interval τ. Do the symmetries implied by the Fluctuation Theorem (FT) apply to the momentum ﬂuxes Pa→i;ω and Pi→o;ω? The ﬂuxes are quadratic quantities and their statistics is therefore not Gaussian. Recently a closed form of the probability density function f(Z) of the product of two correlated Gaussian variables 5 Z = XY with vanishing means and variances σ2 x and σ2 y and correlation ρ has been obtained (Nadarajah and Pogány (2016) and Gaunt (2018)) in terms of a modiﬁed Bessel function of the second kind of order zero K0(z) = R ∞ 0 cos(z sinht)dt: Gaussian. Recently a closed form of the probability density function f(Z) of the product of two correlated Gaussian variables 5 Z = XY with vanishing means and variances σ2 x and σ2 y and correlation ρ has been obtained (Nadarajah and Pogány (2016) and Gaunt (2018)) in terms of a modiﬁed Bessel function of the second kind of order zero K0(z) = R ∞ 0 cos(z sinht)dt: f(z) = 1 πσxσy p 1 −ρ2 exp ρz σxσy(1 −ρ2) K0 \|z\| σxσy(1 −ρ2) . f(z) = 1 πσxσy p 1 −ρ2 exp ρz σxσy(1 −ρ2) K0 \|z\| σxσy(1 −ρ2) . (21) ρz σxσy(1 −ρ2) K0 \|z\| σxσy(1 −ρ2) . (21) (21) The corresponding symmetry function is: The corresponding symmetry function is: The corresponding symmetry function is: SZ(z) = ln f(z) f(−z) = βz, 10 SZ(z) = ln f(z) f(−z) = βz, (22) 0 SZ(z) = ln f(z) f(−z) = βz, (22) 10 (22) which is linear in the variable z and the prefactor is β = 2ρ/((1−ρ2)σxσy). The normalized time average over an interval τ is denoted by: which is linear in the variable z and the prefactor is β = 2ρ/((1−ρ2)σxσy). The normalized time average over an interval τ is denoted by: which is linear in the variable z and the prefactor is β = 2ρ/((1−ρ2)σxσy). 7 Fluctuation Theorem Th t d i th i l l l ti S 10 3 d 100 D t il l i i ll th t h ti Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Figure 2. Probability density function of Za/o;ω(t) τ at t = 300 and τ = 0, atmosphere full-line, ocean dashed-line and for τ = 100, atmo- sphere dotted-line, ocean dashed-dotted-line. All are clearly non-Gaussian Figure 2. Probability density function of Za/o;ω(t) τ at t = 300 and τ = 0, atmosphere full-line, ocean dashed-line and for τ = 100, atmo- sphere dotted-line, ocean dashed-dotted-line. All are clearly non-Gaussian non-stationary processes with a constant-in-time average value. More precisely: the pdfs of the variables Za;ω τ and Zo;ω(t) τ depend on t and τ, but ⟨Za;ω(t) τ⟩Ωand ⟨Zo;ω(t) τ⟩Ωare independent of the time t and the averaging period τ. The parameters used in the numerical calculations are S = 10−3 and m = 100. Details on solving numerically the stochastic differential equations are given in Wirth (2018). The numerical results presented in ﬁg. 2 show that the pdfs of Za;ω(t) τ and Zo;ω(t) τ for t = 300, and τ = 0,100 are non Gaussian. The exponential scaling of the symmetry function for the ocean 5 SZo τ (z) for t = 300, and τ = 0,100,200 is clearly present in ﬁg. 3. This also means that zero is a special value, which is already conspicuous in ﬁg 2. The scaling exponents for t = 10,20,30 and 50 as a function of τ are given in ﬁg. 4, it can be veriﬁed that eq. (24) holds, when the absolute and the averaging time exceeds the characteristic time t,τ > (SM)−1 meaning that the FT applies asymptotically, as in Ciliberto et al. (2004). The change of slope for the different values of t is well ﬁtted 5 by a σ ∝t−1 law. 10 For the atmosphere the probability of having a negative ﬂux Pa→i(t), that is the atmosphere receives energy at the interface due to the ocean dynamics is small even in instantaneous pdfs. 7 Fluctuation Theorem The FT states that the probability of ﬁnding a positive ﬂux of magnitude z divided by the 20 probability of a negative ﬂux with the same magnitude increases exponentially with the value z and the averaging period τ. In the problem considered here the variable Z(t) τ is either the time-averaged energy the atmosphere does on the ocean Pa→i;ω τ, divided by its ensemble average, or the time-averaged work the ocean receives from the atmosphere Pi→o;ω τ, divided by its ensemble average. More precisely, the random variables: In the problem considered here the variable Z(t) τ is either the time-averaged energy the atmosphere does on the ocean Pa→i;ω τ, divided by its ensemble average, or the time-averaged work the ocean receives from the atmosphere Pi→o;ω τ, divided by its ensemble average. More precisely, the random variables: Za;ω(t) τ = Pa→i;ω(t) τ ⟨Pa→i;ω(t)⟩Ω and Zo;ω(t) τ = Pi→o;ω(t) τ ⟨Pi→o;ω(t)⟩Ω , (25) 25 Za;ω(t) τ = Pa→i;ω(t) τ ⟨Pa→i;ω(t)⟩Ω and Zo;ω(t) τ = Pi→o;ω(t) τ ⟨Pi→o;ω(t)⟩Ω , (25) 25 Za;ω(t) τ = Pa→i;ω(t) τ ⟨Pa→i;ω(t)⟩Ω 25 (25) for all the models are considered. Ensemble averages of the ﬂuxes can be obtained analytically for the linear models, but I do not know their pdfs. These investigations are therefore numerical even for the linear models considered here. for all the models are considered. Ensemble averages of the ﬂuxes can be obtained analytically for the linear models, but I do not know their pdfs. These investigations are therefore numerical even for the linear models considered here. First, the L3 model is discussed. It is important to note that although ⟨Pa→i;ω(t)⟩Ωand ⟨Pi→o;ω(t)⟩Ωare constant in time (after an initial spin-up of O((SM)−1), see Wirth (2018)) the pdfs are not (see Fig 2). Which means that the energy transfers are 14 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Figure 2. Probability density function of Za/o;ω(t) τ at t = 300 and τ = 0, atmosphere full-line, ocean dashed-line and for τ = 100, atmo- sphere dotted-line, ocean dashed-dotted-line. All are clearly non-Gaussian non-stationary processes with a constant-in-time average value. More precisely: the pdfs of the variables Za;ω τ and Zo;ω(t) τ depend on t and τ, but ⟨Za;ω(t) τ⟩Ωand ⟨Zo;ω(t) τ⟩Ωare independent of the time t and the averaging period τ. 7 Fluctuation Theorem Negative events in an ensemble sizes of 3·107 were so few that the symmetry function could not be obtained with a sufﬁcient accuracy. For the atmosphere the probability of having a negative ﬂux Pa→i(t), that is the atmosphere receives energy at the interface due to the ocean dynamics is small even in instantaneous pdfs. Negative events in an ensemble sizes of 3·107 were so few that the symmetry function could not be obtained with a sufﬁcient accuracy. 15 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Figure 3. Lin-log plot of (SZo τ )−1 for t = 300, τ = 0 full-line, 10 dotted-line, 20 dashed-line Figure 3. Lin-log plot of (SZo τ )−1 for t = 300, τ = 0 full-line, 10 dotted-line, 20 dashed-line For calculations with the colored noise model, the same parameters as in the white-noise calculations are used and µ = 10−2. In this case the forcing time-scale µ−1 is actually slower than the atmospheric friction time-scale (Sm)−1 but faster than the oceanic friction time-scale S−1. This mimics the fact, that the fast motion in the atmospheric boundary-layer is forced by the slower synoptic dynamics above. The forcing time of the oceanic mixed layer is then determined by the mass ratio m between the oceanic and the atmospheric layers, it is the slowest time-scale. Results (not shown) from the models with colored noise 5 agree qualitatively with the white-noise forcing, that is, they indicate that the energy ﬂux to the ocean obeys a FT. 5 the oceanic and the atmospheric layers, it is the slowest time-scale. Results (not shown) from the models with colored noise 5 agree qualitatively with the white-noise forcing, that is, they indicate that the energy ﬂux to the ocean obeys a FT. Numerical integration of the L1W, L2W, L1C and L2C models show that Pi→o;ω obeys the FT as in the L3 model. The atmospheric ﬂux Pa→i(t) is always positive in the L1 and L2 models so the FT can not be considered. 8 Discussion When ocean velocities are not considered in models of air sea-interaction the atmosphere loses, on average, more energy and 10 the ocean gains more energy, as when the ocean velocities are taken into account. Previous publications on the comparison of different models of air-sea interaction focus on quantitative differences. This is justiﬁed when the short-term dynamics is considered, as shown above. At longer time-scales the differences are not only 16 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Figure 4. Scaling exponents for t = 10 full-line, 20 dotted-line, 30 dashed-line and 50 dashed-dotted-line, as a function of τ and their linear ﬁt (thin full lines), the slope is σ. Figure 4. Scaling exponents for t = 10 full-line, 20 dotted-line, 30 dashed-line and 50 dashed-dotted-line, as a function of τ and their linear ﬁt (thin full lines), the slope is σ. quantitative but qualitative, as for some models stationary states in the ocean the atmosphere or in both are reached while in others this is not the case. An example is the “eddy-killing” term (see Renault et al. (2017)), that includes the ocean velocity in the shear calculation. In the short term its impact is small and can be parameterized by changing the friction coefﬁcient. In the long-term it imposes a convergence to a stationary state, when only implemented in the ocean, when implemented in the quantitative but qualitative, as for some models stationary states in the ocean the atmosphere or in both are reached while in others this is not the case. An example is the “eddy-killing” term (see Renault et al. (2017)), that includes the ocean velocity in the shear calculation. In the short term its impact is small and can be parameterized by changing the friction coefﬁcient. In the long-term it imposes a convergence to a stationary state, when only implemented in the ocean, when implemented in the atmosphere and the ocean it leads to a divergence in both layers, whereas neglecting it totally leads to a divergence in the ocean 5 only. In more involved models, divergence is avoided by other processes as non-linear interactions or data assimilation, who assure it differently. 8 Discussion When these processes supplante an incomplete representation of “eddy killing” the converged state will differ and so will the energy balance. The small differences in the short term behavior and the qualitative differences in the long-term behavior between the models, indicate that the choice of the model does not matter when weather or ocean forecasts atmosphere and the ocean it leads to a divergence in both layers, whereas neglecting it totally leads to a divergence in the ocean 5 only. In more involved models, divergence is avoided by other processes as non-linear interactions or data assimilation, who assure it differently. When these processes supplante an incomplete representation of “eddy killing” the converged state will differ and so will the energy balance. The small differences in the short term behavior and the qualitative differences in the long-term behavior between the models, indicate that the choice of the model does not matter when weather or ocean forecasts are performed, but it might be crucial in climate simulations. 10 5 10 The discussion of the FDT establishes when the response to an external perturbation can be obtained from internal ﬂuctu- ations of the system. In the simple system discussed here we can see analytically when it is veriﬁed and fails and how the failure can be removed by extending the phase space. To determine the response to a sudden change in the external forcing is key in many applications, as for example the response of the atmospheric and oceanic planetary-boundary-layer dynamics to 17 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. a change in the synoptic weather condition. The presented calculations can also be used to guide applications of the FDT to systems with large, but not inﬁnite, time separation. a change in the synoptic weather condition. The presented calculations can also be used to guide applications of the FDT to systems with large, but not inﬁnite, time separation. The FT concerns the transfer of energy between the atmosphere and the interface and the interface and the ocean on different time scales. The temporal down-scaling is solved when we can obtain the pdf of short term averages from the pdf of longer term averages. The temporal up-scaling is solved when we can obtain the pdf of long term averages from the pdf of short term 5 averages. 8 Discussion If The major difference between a two-dimensional and a local model is that the former contains horizontal advection of momentum while the latter does not. It is thus not clear which variable of the two-dimensional model has to be considered using the insight from the local models; is it the local velocity, the velocity advected by the total inertia mode or by the ocean dynamics. Or do we have to consider coarse grained variables for which the importance of horizontal advection is reduced? If this is the case we have to deﬁne a coarse graining scale that is sufﬁcient or optimal in some sense. 20 this is the case we have to deﬁne a coarse graining scale that is sufﬁcient or optimal in some 20 this is the case we have to deﬁne a coarse graining scale that is sufﬁcient or optimal in some sense. 20 8 Discussion The FT relates temporal averages over different time scales and puts a large constraint on the probability density functions (pdfs) of the averaged energy transfers over different time scales. The FT is key to understanding and modeling the climate dynamics as in all observations and models some time-and-space averaging is present. It is not even clear what is the averaging period associated to a variable in a model. The FT gives us a hint of what to expect when passing, for example, from thl d i t ti / f i t d il h l 10 5 monthly-averaged interaction / forcing to daily or hourly averages. 10 In Renault et al. (2017) it was suggested, using satellite observations, that to leading order, the effect of mesoscale ocean currents on the surface stress (the "eddy killing") can be parameterized as a linear function of the wind. Throughout the above presented results we see that differences between the three models considered, are not only quantitative but are qualitative, when long-term behavior is considered. This shows that we can not improve the L1 or L2 model by adjusting the friction parameter to obtain the behavior of the consistent L3 model. 15 y g g y y g In Renault et al. (2017) it was suggested, using satellite observations, that to leading order, the effect of mesoscale ocean currents on the surface stress (the "eddy killing") can be parameterized as a linear function of the wind. Throughout the above presented results we see that differences between the three models considered, are not only quantitative but are qualitative, when long-term behavior is considered. This shows that we can not improve the L1 or L2 model by adjusting the friction parameter to obtain the behavior of the consistent L3 model. 15 parameter to obtain the behavior of the consistent L3 model. 15 The major difference between a two-dimensional and a local model is that the former contains horizontal advection of momentum while the latter does not. It is thus not clear which variable of the two-dimensional model has to be considered using the insight from the local models; is it the local velocity, the velocity advected by the total inertia mode or by the ocean dynamics. Or do we have to consider coarse grained variables for which the importance of horizontal advection is reduced? Appendix A: The Models In this section the solution of the linear models L1, L2 and L3 are solved using linear algebra. The linear differential equation 10 In this section the solution of the linear models L1, L2 and L3 are solved using linear algebra. The linear differential equation 10 ∂tu = −αu + F(t) (A1) In this section the solution of the linear models L1, L2 and L3 are solved using linear algeb 10 ∂tu = −αu + F(t) (A1) ∂tu = −αu + F(t) (A1) with initial condition u(t0), has solutions: with initial condition u(t0), has solutions: u(t) = It t0(α) + u(t0)eα(t0−t), (A2) with the symbol: It t0(α) = t Z t0 eα(t′−t)F(t′)dt′. 15 (A3) In the multidimensional case we have: In the multidimensional case we have: ∂tu = −Pu + F F(t) = −ADA−1 + F F(t) (A4) ∂tu = −Pu + F F(t) = −ADA−1 + F F(t) ∂tu = −Pu + F F(t) = −ADA−1 + F F(t) (A4) where D is a diagonal (or Jacobi normal) matrix, F a constant coefﬁcient vector and F(t) a time dependent scalar forcing. the solution with initial condition u(t0) is: u(t) = AIt t0(D)A−1F + AeD(t0−t)A−1u(t0). 20 (A5) 20 Note that in all our applications α and the eigenvalues of D are positive or zero. 9 Conclusions In the present work deterministic and stochastic calculus to established models of momentum transfer at the air-sea interface are considered. The results of these idealized models have implications for present day simulations of climate dynamics and are important to understand the mechanical energy transfer between the ocean and the atmosphere. It is furthermore interesting to see the concepts of non-equilibrium statistical dynamics applied to a ﬁeld of climate science 25 which is sufﬁciently simple so that most results can be obtained analytically. This, together with the detailed calculations given in the appendixes, will further the advancement of this concepts in the ﬁeld. It is furthermore interesting to see the concepts of non-equilibrium statistical dynamics applied to a ﬁeld of climate science 25 which is sufﬁciently simple so that most results can be obtained analytically. This, together with the detailed calculations given in the appendixes, will further the advancement of this concepts in the ﬁeld. After having here considered linear 0D models, which allow for analytic calculations to a large extent, future work will be dedicated to extend these work to more involved, non-linear, 1D, 2D and 3D numerical models and to consider the FDR, FDT After having here considered linear 0D models, which allow for analytic calculations to a large extent, future work will be dedicated to extend these work to more involved, non-linear, 1D, 2D and 3D numerical models and to consider the FDR, FDT and FT in observations. 30 and FT in observations. 30 The here presented theory is not restricted to momentum transfer, but can also be employed to study heat exchange between the atmosphere and the ocean, or to other processes in the climate system with diverse characteristic time scales. The here presented theory is not restricted to momentum transfer, but can also be employed to study heat exchange between the atmosphere and the ocean, or to other processes in the climate system with diverse characteristic time scales. 18 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. In this work basic models of air-sea interaction are investigated using modern concepts of non-equilibrium statistical dynam- ics. In the sense of Kuhn (1963) this work is “normal science”, puzzle solving within an established framework. 9 Conclusions It is important to be familiar with these calculations and and their results because they demonstrate how far analytic or semi-analytic calculus brings us and they also help us to understand observations. If discrepancy between the former and the latter arises, which can not be solved, they possibly lead to a paradigm shift. But for this to happen the science within a certain paradigm has to be 5 scrutinized (see again Kuhn (1963)). Only after this is accomplished deviations from the theory, as a glass transition in the air-sea interaction framework discovered in Moulin and Wirth (2016) can be studied. In this work basic models of air-sea interaction are investigated using modern concepts of non-equilibrium statistical dynam- ics. In the sense of Kuhn (1963) this work is “normal science”, puzzle solving within an established framework. It is important to be familiar with these calculations and and their results because they demonstrate how far analytic or semi-analytic calculus brings us and they also help us to understand observations. If discrepancy between the former and the latter arises, which can not be solved they possibly lead to a paradigm shift But for this to happen the science within a certain paradigm has to be 5 5 Code availability. The data used in section 7 was produced by the FORTRAN code available under open acces under: https://zenodo.org/record/2530007 Code availability. The data used in section 7 was produced by the FORTRAN code available under open acces Code availability. The data used in section 7 was produced by the FORTRAN code available under open acces under: https://zenodo.org/record/253 A1 Model L1 (A11) P = ADA−1 =  −Sm 0 S 0  =  m 0 −1 1    −Sm 0 0 0    m−1 0 m−1 1  . (A10) The solution is: The solution is:  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo  = A  It t0(Sm) 0 0 It t0(0)  A−1  Fa Fo   =   It t0(Sm) 0 −It t0(Sm) + It t0(0) /m It t0(0)    Fa Fo  . 5 5 (A11) In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: ation at ∆t = 0, (¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as:  uP a (∆t) uP o (∆t)  =  e−Sm∆t 0 1−e−Sm∆t m 1    ¯ua ¯uo  .  uP a (∆t) uP o (∆t)  =  e−Sm∆t 0 1−e−Sm∆t m 1    ¯ua ¯uo  . (A12) (A12) Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. A1 Model L1 The system is forced and damped and the atmospheric dynamics acts on the ocean without considering the ocean velocity. A coupling which is still used in some climate models. ∂tua = −Smua + Fa 25 (A7) ∂tuo = Sua + Fo 19 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. ∂t  ua uo  =  −Sm 0 S 0    ua uo  +  Fa Fo   (A8) λ1 = −Sm, e1 =  m −1  ; λ2 = 0, e2 =  0 1   (A9) ∂t  ua uo  =  −Sm 0 S 0    ua uo  +  Fa Fo   (A8)        λ1 = −Sm, e1 =  m −1  ; λ2 = 0, e2 =  0 1   (A9) λ1 = −Sm, e1 =  m −1  ; λ2 = 0, e2 =  0 1   (A9) P = ADA−1 =  −Sm 0 S 0  =  m 0 −1 1    −Sm 0 0 0    m−1 0 m−1 1  . (A10) P = ADA−1 =  −Sm 0 S 0  =  m 0 −1 1    −Sm 0 0 0    m−1 0 m−1 1  . (A10) The solution is:  ua(t)  AIt (D)A−1  Fa  A  It t0(Sm) 0  A−1  Fa  P = ADA−1 =  −Sm 0 S 0  =  m 0 −1 1    −Sm 0 0 0    m−1 0 m−1 1  . (A10) The solution is:  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo  = A  It t0(Sm) 0 0 It t0(0)  A−1  Fa Fo   =   It t0(Sm) 0 −It t0(Sm) + It t0(0) /m It t0(0)    Fa Fo  . A2 Model L2 10 (A18) 5  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo    ua(t) uo(t)  = A  It t0(Sm) 0 0 It t0(S)  A−1  Fa Fo     It t0(Sm) 0 −It t0(Sm) + It t0(S) /(m −1) It t0(S)    Fa Fo  . (A18) 5 In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as:  uP a (∆t) uP o (∆t)  =   e−Sm∆t 0 e−S∆t−e−Sm∆t m−1 e−S∆t    ¯ua ¯uo  .  uP a (∆t) uP o (∆t)  =   e−Sm∆t 0 e−S∆t−e−Sm∆t m−1 e−S∆t    ¯ua ¯uo  . (A19) (A19) Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. A2 Model L2 10 The system is forced with damping and the atmospheric dynamics forces the ocean, ocean velocity is taken into account for the ocean dynamics but not in the atmospheric dynamics (Newton’s third law is not respected). ∂tua = −Smua + Fa (A13) ∂tuo = S(ua −uo) + Fo (A14) ∂tua = −Smua + Fa ∂tuo = S(ua −uo) + Fo 15 15 ∂t  ua uo  =  −Sm 0 S −S    ua uo  +  Fa Fo   (A15) λ1 = −Sm, e1 =  m −1 −1  ; λ2 = −S, e2 =  0 1   (A16) ∂t  ua uo  =  −Sm 0 S −S    ua uo  +  Fa Fo   (A15) ∂t  ua uo  =  −Sm 0 S −S    ua uo  +  Fa Fo   (A15) λ1 = −Sm, e1 =  m −1 −1  ; λ2 = −S, e2 =  0 1   20 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. P = ADA−1 =  −Sm 0 S −S  =  m −1 0 −1 1    −Sm 0 0 −S    (m −1)−1 0 (m −1)−1 1  . (A17) (A17) The solution is: The solution is:  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo    ua(t) uo(t)  = A  It t0(Sm) 0 0 It t0(S)  A−1  Fa Fo   =   It t0(Sm) 0 −It t0(Sm) + It t0(S) /(m −1) It t0(S)    Fa Fo  . A3 Model L3 (A25) n is:  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo    ua(t) uo(t)  = A  It t0(SM) 0 0 It t0(0)  A−1  Fa Fo   (SM) + It t (0) −mIt t (SM) + mIt t (0)  Fa  (A25) In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: ial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: 5 In the absence of forcing an initial perturbation at ∆t = 0, (¯ua, ¯uo) evolves as: 5 5  uP a (∆t) uP o (∆t)  = 1 M  1 + me−SM∆t m(1 −e−SM∆t) 1 −e−SM∆t m + e−SM∆t    ¯ua ¯uo  . (A26)  uP a (∆t) uP o (∆t)  = 1 M  1 + me−SM∆t m(1 −e−SM∆t) 1 −e−SM∆t m + e−SM∆t    ¯ua ¯uo  . (A26) (A26) Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t Note that the perturbation matrix above: χ(∆t) = Aexp(Dt)A−1, is independent of the time t. B1 Constant Forcing In this appendix Fa = 1 and Fo = 0. Put into eq. (A3) this leads to: (B1) (B2) It t0(0) = t −t0 15 It t0(α) = 1 α(1 −exp(−α(t −t0))), if α ̸= 0. It t0(0) = t −t0 15 (B1) It t0(α) = 1 α(1 −exp(−α(t −t0))), if α ̸= 0. (B2) The solutions for the different models are: Appendix B: Experiments In all experiments only the atmosphere is forced, Fo = 0. Note that in the L3 model the atmosphere and the ocean are treated similarly and they only differ by the mass ratio m. The dynamics due to a forcing of the ocean can be represented by choosing 10 m < 1 and interchanging the subscripts. The dynamics of a forcing of the ocean and atmosphere can be obtained by adding a model forced by the ocean and the same model forced by the atmosphere due to the linearity of the model. A3 Model L3 The atmosphere is forced with damping and the atmospheric dynamics forces the ocean, the ocean velocity is taken into 10 account. ∂tua = Sm(uo −ua) + Fa (A20) ∂tuo = S(ua −uo) + Fo (A21) ∂tua = Sm(uo −ua) + Fa ∂tuo = S(ua −uo) + Fo ∂tuo = S(ua −uo) + Fo (A21) ∂t  ua uo  =  −Sm Sm S −S    ua uo  +  Fa Fo   15 (A22) 15 λ1 = −SM, e1 =  m −1  ; λ2 = 0, e2 =  1 1  . (A23) λ2 = 0, e2 =  1 1  . (A23) (A23) The ﬁrst eigenvector corresponds to the shear mode and the second to the total-inertia mode. The ﬁrst eigenvector corresponds to the shear mode and the second to the total-inertia mode. P = ADA−1 =  −Sm Sm S −S  =  m 1 −1 1    −SM 0 0 0  1 M  1 −1 1 m  . P = ADA−1 =  −Sm Sm S −S  =  m 1 −1 1    −SM 0 0 0  1 M  1 −1 1 m  . (A24) P = ADA−1 =  −Sm Sm S −S  =  m 1 −1 1    −SM 0 0 0  1 M  1 −1 1 m  . (A24) m 1 1 1    −SM 0 0 0  1 M  1 −1 1 m  . (A24) (A24) 21 21 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. The solution is: The solution is:  ua(t) uo(t)  = AIt t0(D)A−1  Fa Fo    ua(t) uo(t)  = A  It t0(SM) 0 0 It t0(0)  A−1  Fa Fo   = 1 M  mIt t0(SM) + It t0(0) −mIt t0(SM) + mIt t0(0) −It t0(SM) + It t0(0) It t0(SM) + mIt t0(0)    Fa Fo  . B4 Experiment L3K  ua(t) uo(t)  = 1 M  t −t0 + (1 −exp(−SM(t −t0)))m/(SM) t −t0 −(1 −exp(−SM(t −t0)))/(SM)   0 (B9) 10 10 The Taylor series expansion for small times (t −t0) ≪(Sm)−1 is: The Taylor series expansion for small times (t −t0) ≪(Sm)−1 is:  ua(t) uo(t)  =   t −t0 −Sm 2 (t −t0)2 + S2mM 6 (t −t0)3 −... S 2 (t −t0)2 −S2M 6 (t −t0)3 + S3M 2 24 (t −t0)4 −...   (B10) Asymptotics for large times (t −t0) ≫(SM)−1: Asymptotics for large times (t −t0) ≫(SM)−1:  ua(t) uo(t)  =   t−t0 M + m SM2 t−t0 M − 1 SM2    ua(t) uo(t)  =   t−t0 M + m SM2 t−t0 M − 1 SM2   (B11) (B11) B2 Experiment L1K  ua(t) uo(t)  = 1 Sm   1 −exp(−Sm(t −t0)) [Sm(t −t0) −1 + exp(−Sm(t −t0))]/m  . (B3) The Taylor series expansion for small times (t −t0) ≪(Sm)−1 is: 20  ua(t) uo(t)  =   t −t0 −Sm 2 (t −t0)2 + (Sm)2 6 (t −t0)3 −... S 2 (t −t0)2 −S2m 6 (t −t0)3 + S3m2 24 (t −t0)4 −...  . (B4) 22 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Asymptotics for large times (t −t0) ≫(Sm)−1: Asymptotics for large times (t −t0) ≫(Sm)−1:  ua(t) uo(t)  =   1 Sm t−t0 m − 1 Sm2    ua(t) uo(t)  =   1 Sm t−t0 m − 1 Sm2   (B5) B3 Experiment L2K  ua(t) uo(t)  =   [1 −exp(−Sm(t −t0))]/(Sm) [1 −exp(−S(t −t0)) −(1 −exp(−Sm(t −t0))/m]/(S(m −1))  . (B6) The Taylor series expansion for small times (t −t0) ≪(Sm)−1 is: 5 The Taylor series expansion for small times (t −t0) ≪(Sm)−1 is: 5  ua(t) uo(t)  =   t −t0 −Sm 2 (t −t0)2 + (Sm)2 6 (t −t0)3 −... S 2 (t −t0)2 −S2M 6 (t −t0)3 + S3(m2+m+1) 24 (t −t0)4 −...  . (B7) Asymptotics for large times (t −t0) ≫S−1: Asymptotics for large times (t −t0) ≫S−1: Asymptotics for large times (t −t0) ≫S−1:  ua(t) uo(t)  =   1 Sm 1 Sm  . (B8) B4 Experiment L3K B5 Periodic Forcing 15 In this appendix Fa = cos(κt). Put into eq. (A3) and starting the integration at t0 = −∞(ignoring transients), leads to: It −∞(α) = 1 α2 + κ2 (κsin(κt) + αcos(κt)). (B1 In this appendix Fa = cos(κt). Put into eq. (A3) and starting the integration at t0 = −∞(ignoring transients), leads to: It −∞(α) = 1 α2 + κ2 (κsin(κt) + αcos(κt)). (B1 In this appendix Fa = cos(κt). Put into eq. (A3) and starting the integration at t0 = −∞(ignoring transients), leads to: It −∞(α) = 1 α2 + κ2 (κsin(κt) + αcos(κt)). (B1 (B12) 23 23 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. B6 Experiment L1P Solution Solution  ua(t) uo(t)  = 1 (Sm)2 + κ2   (κsin(κt) + Smcos(κt)) S/κ(Smsin(κt) −κcos(κt))  . (B13)  ua(t) uo(t)  = 1 (Sm)2 + κ2   (κsin(κt) + Smcos(κt)) S/κ(Smsin(κt) −κcos(κt))  . (B13) The averages over one period τ = 2π/κ are denoted by ⟨.⟩τ. First order moments all vanish, for the second order moments we t The averages over one period τ = 2π/κ are denoted by ⟨.⟩τ. First order moments all vanish, for the second order moments we get: 5 get: 5 get: 2⟨u2 a⟩τ = 1 (Sm)2 + κ2 2⟨u2 o⟩τ = S2 κ2((Sm)2 + κ2) ⟨u2 o⟩τ ⟨u2a⟩τ = S2 κ2 ⟨uauo⟩τ = 0. (B14) 2⟨u2 a⟩τ = 1 (Sm)2 + κ2 2⟨u2 o⟩τ = S2 κ2((Sm)2 + κ2) ⟨u2 o⟩τ ⟨u2a⟩τ = S2 κ2 ⟨uauo⟩τ = 0. (B14) B7 Experiment L2P 10 B7 Experiment L2P 10 ⟨u2 o⟩τ ⟨u2a⟩τ = S2 κ2 + S2 . (B19) (B20) (B21) B9 Stochastic calculus 5 B7 Experiment L2P 10 B7 Experiment L2P 10 ua = 1 (Sm)2 + κ2 [κsin(κt) + Smcos(κt)] (B15) uo = S ((Sm)2 + κ2)(S2 + κ2)[SMκsin(κt) + (S2m −κ2)cos(κt)] (B16) 2⟨u2 a⟩τ = 1 (Sm)2 + κ2 2⟨u2 o⟩τ = 2⟨uauo⟩τ = S2 ((Sm)2 + κ2)(S2 + κ2) 15 ⟨u2 o⟩τ ⟨u2a⟩τ = S2 S2 + κ2 B8 Experiment L3P ut(t) = sin(κt) κ us(t) = κsin(κt) + SM cos(κt) κ2 + (SM)2 (B17) 20 ua(t) = 1 M (ut + mus) = 1 κ2 + (SM)2 κ2 + S2M κ sin(κt) + Smcos(κt) uo(t) = 1 M (ut −us) = S κ(κ2 + (SM)2) [SM sin(κt) −κcos(κt)] (B18) ua = 1 (Sm)2 + κ2 [κsin(κt) + Smcos(κt)] uo = S ((Sm)2 + κ2)(S2 + κ2)[SMκsin(κt) + (S2m −κ2)cos(κt)] 2⟨u2 a⟩τ = 1 (Sm)2 + κ2 2⟨u2 o⟩τ = 2⟨uauo⟩τ = S2 ((Sm)2 + κ2)(S2 + κ2) 15 ⟨u2 o⟩τ ⟨u2a⟩τ = S2 S2 + κ2 B8 Experiment L3P ua = 1 (Sm)2 + κ2 [κsin(κt) + Smcos(κt)] uo = S ((Sm)2 + κ2)(S2 + κ2)[SMκsin(κt) + (S2m −κ2)cos(κt)] 2⟨u2 a⟩τ = 1 (Sm)2 + κ2 2⟨u2 o⟩τ = 2⟨uauo⟩τ = S2 ((Sm)2 + κ2)(S2 + κ2) 15 ⟨u2 o⟩τ ⟨u2a⟩τ = S2 S2 + κ2 (B15) (B16) B8 Experiment L3P ut(t) = sin(κt) κ us(t) = κsin(κt) + SM cos(κt) κ2 + (SM)2 0 ua(t) = 1 M (ut + mus) = 1 κ2 + (SM)2 κ2 + S2M κ sin(κt) + Smcos(κt) (B17) 20 20 ua(t) = 1 M (ut + mus) = 1 κ2 + (SM)2 κ2 + S2M κ sin(κt) + Smcos(κt) uo(t) = 1 M (ut −us) = S κ(κ2 + (SM)2) [SM sin(κt) −κcos(κt)] (B18) 24 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. which leads to: 2⟨u2 a⟩τ = 1 κ2 + (SM)2 κ2 + S2 κ2 (B19) 2⟨u2 o⟩τ = 2⟨uauo⟩t = 1 κ2 + (SM)2 S2 κ2 . (B20) ⟨u2 o⟩τ ⟨u2a⟩τ = S2 κ2 + S2 . (B21) B9 Stochastic calculus 5 which leads to: 2⟨u2 a⟩τ = 1 κ2 + (SM)2 κ2 + S2 κ2 2⟨u2 o⟩τ = 2⟨uauo⟩t = 1 κ2 + (SM)2 S2 κ2 . B10 Random walk and Ornstein-Uhlenbeck process The following identities are used: t Z −∞ δ(t −t′)dt′ = 1/2 t Z t0 t Z t0 δ(t′′ −t′)dt′′dt′ = t −t0 ⟨Fω(t)⟩Ω= 0 ⟨Fω(t′)Fω(t′′)⟩Ω= 2Rδ(t′′ −t′). (B22) (B23) 10 10 (B25) In the sequel the subscript ω is omitted. Below are the equations for a random-walk uR and a Ornstein-Uhlenbeck process uO, the solution of a Langevin equation. In the sequel the subscript ω is omitted. Below are the equations for a random-walk uR and a Ornstein-Uhlenbeck process uO, the solution of a Langevin equation. In the sequel the subscript ω is omitted. Below are the equations for a random-walk uR and a Ornstein-Uhlenbeck process uO, the solution of a Langevin equation. ∂tuR = F (B26) ∂tuO = −SuO + F. (B27) 5 ∂tuR = F ∂tuR = F ∂tuO = −SuO + F. 15 ∂tuO = −SuO + F. 15 Solutions starting from rest at t0, uR(t0) = uO(t0) = 0, are: Solutions starting from rest at t0, uR(t0) = uO(t0) = 0, are: uR(t) = R t t0 F(t′)dt′ (B28) uO(t) = R t t0 eS(t′−t)F(t′)dt′. (B29) uR(t) = R t t0 F(t′)dt′ uO(t) = R t t0 eS(t′−t)F(t′)dt′. uR(t) = R t t0 F(t′)dt′ R t0 uO(t) = R t t0 eS(t′−t)F(t′)dt′. 25 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. It follows that: ⟨uR⟩Ω= ⟨uO⟩Ω= 0. Second order moments are (note that as processes are Gaussian ﬁrst and second order moments completely determine the stochastic processes): ⟨u2 R(t)⟩Ω = t Z t0 t Z t0 ⟨F(t′)F(t′′)⟩Ωdt′′dt′ = 2R(t −t0) ⟨uR(t)uO(t)⟩Ω = t Z t0 t Z t0 eS(t′−t)⟨F(t′)F(t′′)⟩Ωdt′′dt′ 5 = 2R S (1 −eS(t0−t)) ⟨u2 O(t)⟩Ω = t Z t0 t Z t0 eS(t′+t′′−2t)⟨F(t′)F(t′′)⟩Ωdt′′dt′ = R S (1 −e2S(t0−t)). (B30) 5 (B31) (B32) It t0(α) = t Z t0 eα(t′−t)F(t′)dt′. It t0(α) = t Z t0 eα(t′−t)F(t′)dt′. (B33) It t0(α) = t Z t0 eα(t′−t)F(t′)dt′. (B33) (B33) 10 10 ⟨It t0(α)It+∆t t0 (β)⟩Ω = 2R(t −t0)e−β∆t if α + β = 0 (B34) ⟨It t0(α)It+∆t t0 (β)⟩Ω = 2R α + β (1 −exp(−(α + β)(t −t0))e−β∆t if α + β ̸= 0. B11 Experiments with stochastic forcing riment L1W B10 Random walk and Ornstein-Uhlenbeck process coordinate system spanned by the eigenvectors, in this case P = D is diagonal and A is the identity matrix, then: coordinate system spanned by the eigenvectors, in this case P = D is diagonal and A is the identity matrix, then: te system spanned by the eigenvectors, in this case P = D is diagonal and A is the identity matrix, C(t,∆t)C(t,0)−1 = A⟨It+∆t t0 (D)A−1F · F T (A−1)T It t0(D)⟩Ω (⟨It t0(D)A−1F · F T (A−1)T It t0(D)⟩Ω)−1AT = ⟨It+∆t t0 (D)F · F T It t0(D)⟩Ω(⟨It t0(D)F · F T It t0(D)⟩Ω)−1 = e−D∆t. (B38) C(t,∆t)C(t,0)−1 = A⟨It+∆t t0 (D)A−1F · F T (A−1)T It t0(D)⟩Ω (⟨It t0(D)A−1F · F T (A−1)T It t0(D)⟩Ω)−1AT = ⟨It+∆t t0 (D)F · F T It t0(D)⟩Ω(⟨It t0(D)F · F T It t0(D)⟩Ω)−1 = e−D∆t. 5 (B38) 5 To obtain the last equality above, we used eq. (B37). The normalized correlation matrix is equal to the perturbation matrix and thus proves the FDT (see section 6 and appendixes C1 and C2)). To obtain the last equality above, we used eq. (B37). The normalized correlation matrix is equal to the perturbation matrix and thus proves the FDT (see section 6 and appendixes C1 and C2)). B10 Random walk and Ornstein-Uhlenbeck process (B35) ⟨It t0(α)It+∆t t0 (β)⟩Ω = 2R(t −t0)e−β∆t if α + β = 0 ⟨It t0(α)It+∆t t0 (β)⟩Ω = 2R α + β (1 −exp(−(α + β)(t −t0))e−β∆t if α + β ̸= 0. Also note that in all cases: Also note that in all cases: Also note that in all cases: ⟨It t0(α)It+∆t t0 (β)⟩Ω= ⟨It t0(α)It t0(β)⟩Ωe−β∆t. ⟨It t0(α)It+∆t t0 (β)⟩Ω= ⟨It t0(α)It t0(β)⟩Ωe−β∆t. ⟨It t0(α)It+∆t t0 (β)⟩Ω= ⟨It t0(α)It t0(β)⟩Ωe−β∆t. (B36) (B36) The correlation matrix (t-dependence is kept to deal with non-stationary processes) is 5 The correlation matrix (t-dependence is kept to deal with non-stationary processes) is 15 The correlation matrix (t-dependence is kept to deal with non-stationary processes) is 15 15 C(t,∆t) = ⟨u(t + ∆t) · u(t)T ⟩Ω = A⟨It+∆t t0 (D)A−1F · F T (A−1)T It t0(D)⟩ΩAT = Ae−D∆t⟨It t0(D)A−1F · F T (A−1)T It t0(D)⟩ΩAT , (B37) C(t,∆t) = ⟨u(t + ∆t) · u(t)T ⟩Ω = A⟨It+∆t t0 (D)A−1F · F T (A−1)T It t0(D)⟩ΩAT = Ae−D∆t⟨It t0(D)A−1F · F T (A−1)T It t0(D)⟩ΩAT , (B37) (B37) where we used eqs. (A5) and (B36). Even so the matrix F ·F T is singular, this does not necessarily lead to a singular correlation matrix, as ⟨a⟩⟨b⟩̸= ⟨ab⟩. Calculations show that the normalized correlation matrix C(t,∆t)C(t,0)−1 is independent of F , if 20 none of the eigenvectors is orthogonal to F , that is all eigenvectors are subject to the forcing. To see this we transform to a matrix, as ⟨a⟩⟨b⟩̸= ⟨ab⟩. Calculations show that the normalized correlation matrix C(t,∆t)C(t,0)−1 is independent of F , if 20 none of the eigenvectors is orthogonal to F , that is all eigenvectors are subject to the forcing. To see this we transform to a matrix, as ⟨a⟩⟨b⟩̸= ⟨ab⟩. Calculations show that the normalized correlation matrix C(t,∆t)C(t,0)−1 is independent of F , if 20 none of the eigenvectors is orthogonal to F , that is all eigenvectors are subject to the forcing. To see this we transform to a 26 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. B13 Experiment LW2 For the stochastic forcing straightforward calculations, based on eqs. (A18), (B34) and (B35), lead to the correlation matrix: For the stochastic forcing straightforward calculations, based on eqs. (A18), (B34) and (B35), lead C(∆t) =  ⟨ua(t)ua(t + ∆t)⟩Ω ⟨uo(t)ua(t + ∆t)⟩Ω ⟨ua(t)uo(t + ∆t)⟩Ω ⟨uo(t)uo(t + ∆t)⟩Ω  = ⟨  ua(t + ∆t) uo(t + ∆t)  · (ua(t),uo(t))⟩Ω = ⟨   It+∆t t0 (Sm) 0 −It+∆t t0 (Sm)+It+∆t t0 (S) m−1 It+∆t t0 (S)    1 0 0 0    It t0(Sm) −It t0(Sm)+It t0(S) m−1 0 It t0(S)  ⟩Ω = ⟨   It t0(Sm)It+∆t t0 (Sm) It+∆t t0 (Sm) −It t0(Sm)+It t0(S) m−1 It t0(Sm) −It+∆t t0 (Sm)+It+∆t t0 (S) m−1 (−It+∆t t0 (Sm)+It+∆t t0 (S))(−It t0(Sm)+It t0(S)) (m−1)2  ⟩Ω = R   e−Sm∆t Sm e−Sm∆t SMm 1 S(m−1)( 2e−S∆t M −e−Sm∆t m ) 1 SM(m−1)(e−S∆t −e−Sm∆t m )  . (B40) Straight forward calculations show that C(∆t)C(0)−1 is equal to the perturbation matrix (A19) which profs the FDT. 5 (B40) Straight forward calculations show that C(∆t)C(0)−1 is equal to the perturbation matrix (A19) which profs the FDT. Straight forward calculations show that C(∆t)C(0)−1 is equal to the perturbation matrix (A19) w B12 Experiment L1W For the stochastic forcing straightforward calculations, based on eqs. (A11), (B34) and (B35), and supposing that t−t0 ≫S−1, 10 that is: e−S(t−t0) ≈0, leads to the correlation matrix: For the stochastic forcing straightforward calculations, based on eqs. (A11), (B34) and (B35), and supposing that t−t0 ≫S−1, 10 that is: e−S(t−t0) ≈0, leads to the correlation matrix: 10 C(∆t) =  ⟨ua(t)ua(t + ∆t)⟩Ω ⟨uo(t)ua(t + ∆t)⟩Ω ⟨ua(t)uo(t + ∆t)⟩Ω ⟨uo(t)uo(t + ∆t)⟩Ω  = ⟨  ua(t + ∆t) uo(t + ∆t)  · (ua(t),uo(t))⟩Ω = ⟨   It+∆t t0 (Sm) 0 −It+∆t t0 (Sm)+It+∆t t0 (0) m It+∆t t0 (0)    1 0 0 0    It t0(Sm) −It t0(Sm)+It t0(0) m 0 It t0(0)  ⟩Ω = ⟨   It t0(Sm)It+∆t t0 (Sm) It+∆t t0 (Sm) −It t0(Sm)+It t0(0) m It t0(Sm) −It+∆t t0 (Sm)+It+∆t t0 (0) m (−It+∆t t0 (Sm)+It+∆t t0 (0))(−It t0(Sm)+It t0(0)) m2  ⟩Ω = R   e−Sm∆t Sm e−Sm∆t Sm2 2 −Sm∆t 1 2+ −Sm∆t 2(t t )  . (B39) 5 15 (B39) 15 Straight forward calculations show that C(∆t)C(0)−1 is equal to the perturbation matrix (A12) which profs the FDT. 27 27 B13 Experiment LW2 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. B14 Experiment LW3 For the stochastic forcing straightforward calculations, based on eq. (A25), (B34) and (B35), lead to the correlation matrix: For the stochastic forcing straightforward calculations, based on eq. (A25), (B34) and (B35), lead to the correlation matrix: For the stochastic forcing straightforward calculations, based on eq. (A25), (B34) and (B35), le C(∆t) =  ⟨ua(t)ua(t + ∆t)⟩Ω ⟨uo(t)ua(t + ∆t)⟩Ω ⟨ua(t)uo(t + ∆t)⟩Ω ⟨uo(t)uo(t + ∆t)⟩Ω  = ⟨  ua(t + ∆t) uo(t + ∆t)  · (ua(t),uo(t))⟩Ω = ⟨1 M  mIt+∆t t0 (SM) + It+∆t t0 (0) −mIt+∆t t0 (SM) + mIt+∆t t0 (0) −It+∆t t0 (SM) + It+∆t t0 (0) It+∆t t0 (SM) + mIt+∆t t0 (0)    1 0 0 0   1 M   mIt t0(SM) + It t0(0) −It t0(SM) + It t0(0) −mIt t0(SM) + mIt t0(0) It t0(SM) + mIt t0(0)  ⟩Ω = e−SM∆t M 2 ⟨  mIt t0(SM)(mIt t0(SM) + It t0(0)) mIt0(SM)(−It t0(SM) + It t0(0)) −It0(SM)(mIt t0(SM) + It t0(0)) −It t0(SM)(−It t0(SM) + It t0(0))  ⟩Ω + 1 M 2 ⟨  It t0(0)(mIt t0(SM) + It t0(0)) It t0(0)(−It t0(SM) + It t0(0)) It0(0)(mIt t0(SM) + It t0(0)) It t0(0)(−It t0(SM) + It t0(0))  ⟩Ω = R M 2 (e−SM∆t SM  m(m + 2) m −m −2 −1  + 2 SM  m −1 m −1  + 2(t −t0)  1 1 1 1  ). B14 Experiment LW3 (B41) St i ht f d l l ti h th t C(∆t)C(0)−1 i l t th t b ti t i (A26) hi h f th FDT C(∆t) =  ⟨ua(t)ua(t + ∆t)⟩Ω ⟨uo(t)ua(t + ∆t)⟩Ω ⟨ua(t)uo(t + ∆t)⟩Ω ⟨uo(t)uo(t + ∆t)⟩Ω  = ⟨  ua(t + ∆t) uo(t + ∆t)  · (ua(t),uo(t))⟩Ω 10 = ⟨1 M  mIt+∆t t0 (SM) + It+∆t t0 (0) −mIt+∆t t0 (SM) + mIt+∆t t0 (0) −It+∆t t0 (SM) + It+∆t t0 (0) It+∆t t0 (SM) + mIt+∆t t0 (0)    1 0 0 0   1 M   mIt t0(SM) + It t0(0) −It t0(SM) + It t0(0) −mIt t0(SM) + mIt t0(0) It t0(SM) + mIt t0(0)  ⟩Ω   = e−SM∆t M 2 ⟨  mIt t0(SM)(mIt t0(SM) + It t0(0)) mIt0(SM)(−It t0(SM) + It t0(0)) −It0(SM)(mIt t0(SM) + It t0(0)) −It t0(SM)(−It t0(SM) + It t0(0))  ⟩Ω + 1 M 2 ⟨  It t0(0)(mIt t0(SM) + It t0(0)) It t0(0)(−It t0(SM) + It t0(0)) It0(0)(mIt t0(SM) + It t0(0)) It t0(0)(−It t0(SM) + It t0(0))  ⟩Ω = R M 2 (e−SM∆t SM  m(m + 2) m −m −2 −1  + 2 SM  m −1 m −1  + 2(t −t0)  1 1 1 1  ). (B41) (B41) 15 Straight forward calculations show that C(∆t)C(0)−1 is equal to the perturbation matrix (A26) which profs the FDT. 28 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. B15 Experiment L1C ∂t ˜F = − µ ˜F + F (B42) ∂tua = − Smua + ˜F (B43) ∂tuo = S ua (B44) ∂t ˜F = − µ ˜F + F (B42) ∂tua = − Smua + ˜F (B43) ∂tuo = S ua (B44) ∂t ˜F = − µ ˜F + F ∂tua = − Smua + ˜F ∂tuo = S ua 5 ∂t     ˜F ua uo    =     −µ 0 0 1 −Sm 0 0 S 0         ˜F ua uo    +     F 0 0     (B45) λ1 = −µ, e1 =     µ(µ −Sm) −µ S    ; λ2 = −Sm, e2 =     0 m −1    ; λ3 = 0, e3 =     0 0 1     (B46) P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S 0    =     µ(µ −Sm) 0 0 −µ m 0 S −1 1         −µ 0 0 0 −Sm 0 0 0 0         [µ(µ −Sm)]−1 0 0 [m(µ −Sm)]−1 m−1 0 [µm]−1 m−1 1    . P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S 0    =  µ(µ −Sm) 0 0  −µ 0 0  [µ(µ −Sm)]−1 0 0  P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S 0    = P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S 0    =     µ(µ −Sm) 0 0 −µ m 0 S −1 1         −µ 0 0 0 −Sm 0 0 0 0         [µ(µ −Sm)]−1 0 0 [m(µ −Sm)]−1 m−1 0 [µm]−1 m−1 1    . From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): ⟨u2 a(t)⟩Ω = R Smµ(µ + Sm) (B49) ⟨ua(t) ˜F(t)⟩Ω = R µ(µ + Sm) (B50) ⟨u2 o(t)⟩Ω = R(3Sm3 + 2Sm2µ −2Smµ2 −3µ3) µ3m3(µ2 −Sm2) + 2R(t −t0) µ2m2 (B51) ⟨ua(t)uo(t)⟩Ω = R Sµ2m2 (B52) 5 In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as:     ˜F P (t) uP a (t) uP o (t)    =     e−µt 0 0 e−Smt−e−µt µ−Sm e−Smt 0 µ(1−e−Smt)−Sm(1−e−µt) µm(µ−Sm) 1−e−Smt m 1         ¯˜F ¯ua ¯uo    . (B53) µt 0 0 −e−µt Sm e−Smt 0 −Sm(1−e−µt) −Sm) 1−e−Smt m 1         ¯˜F ¯ua ¯uo    . (B53) (B53) The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A12), a simple consequence of linearity. The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A12), a simple consequence of linearity. The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A12), a simple consequence of linearity. to eq. (A12), a simple consequence of linearity. B15 Experiment L1C (B47) The solution is: 10 10     ˜F ua(t) uo(t)    = AIt t0(D)A−1     1 0 0    dt′ = A     It t0(µ) 0 0 0 It t0(Sm) 0 0 0 It t0(0)    A−1     1 0 0     =     It t0(µ) −It t0(µ) + It t0(Sm) /(µ −Sm) SmIt t0(µ) −µIt t0(Sm) + (µ −Sm)It t0(0) /[mµ(µ −Sm)]    . (B48) 29 29 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): B16 Experiment L2C 10 (B59) The solution is: The solution is: The solution is:     ˜F ua(t) uo(t)    = AIt t0(D)A−1     1 0 0     = A     It t0(µ) 0 0 0 It t0(Sm) 0 0 0 It t0(S)    A−1     1 0 0     =     It t0(µ) It t0(µ) −It t0(Sm) /(Sm −µ) S(m −1)It t0(µ) + (S −µ)It t0(Sm) −(Sm −µ)It t0(S) /[(m −1)(Sm −µ)(S −µ)]    . (B60)     ˜F ua(t) uo(t)    = AIt t0(D)A−1     1 0 0   5 5 =     It t0(µ) It t0(µ) −It t0(Sm) /(Sm −µ) S(m −1)It t0(µ) + (S −µ)It t0(Sm) −(Sm −µ)It t0(S) /[(m −1)(Sm −µ)(S −µ)]    . (B60) =     It t0(µ) It t0(µ) −It t0(Sm) /(Sm −µ) From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): ⟨u2 a(t)⟩Ω = R µSm(Sm + µ) (B61) ⟨ua(t) ˜F(t)⟩Ω = R µ(Sm + µ) (B62) 10 ⟨u2 o(t)⟩Ω = ⟨ua(t)uo(t) = R(SM + µ) mSMµ(S + µ)(Sm + µ). (B63) (B61) + µ) o(t) = R(SM + µ) mSMµ(S + µ)(Sm + µ). (B63) SM + µ) S + µ)(Sm + µ). (B63) nitial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as:     ˜F P (t) uP a (t) uP o (t)    =     e−µt 0 0 e−Smt−e−µt µ−Sm e−Smt 0 (m−1)e−µt+(S−µ)e−Smt−(Sm−µ)e−St (m−1)(S−µ)(Sm−µ) e−St−e−Smt m−1 e−St         ¯˜F ¯ua ¯uo    . B16 Experiment L2C 10 ∂t ˜F = − µ ˜F + F (B54) ∂tua = − Smua + ˜F (B55) ∂tuo = S (ua −uo) (B56) ∂t ˜F = − µ ˜F + F ∂tua = − Smua + ˜F ∂tuo = S (ua −uo) ∂t     ˜F ua uo    =     −µ 0 0 1 −Sm 0 0 S −S         ˜F ua uo    +     F 0 0     (B57) ∂t     ˜F ua uo    =     −µ 0 0 1 −Sm 0 0 S −S         ˜F ua uo    +     F 0 0     (B57) λ1 = −µ, e1 =     (Sm −µ)(S −µ) (S −µ) S    ; λ2 = −Sm, e2 =     0 m −1 −1    ; λ3 = −S, e3 =     0 0 1     (B58) 15 (B58) 30 30 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S −S    =     (Sm −µ)(S −µ) 0 0 (S −µ) m −1 0 S −1 1         −µ 0 0 0 −Sm 0 0 0 −S         [(Sm −µ)(S −µ)]−1 0 0 −[(m −1)(Sm −µ)]−1 (m −1)−1 0 −[(m −1)(S −µ)]−1 (m −1)−1 1    . (B59) P = ADA−1 =     −µ 0 0 1 −Sm 0 0 S −S    =     (Sm −µ)(S −µ) 0 0 (S −µ) m −1 0 S −1 1         −µ 0 0 0 −Sm 0 0 0 −S         [(Sm −µ)(S −µ)]−1 0 0 −[(m −1)(Sm −µ)]−1 (m −1)−1 0 −[(m −1)(S −µ)]−1 (m −1)−1 1    . B16 Experiment L2C 10 (B64)     ˜F P (t) uP a (t) uP o (t)    =     e−µt 0 0 e−Smt−e−µt µ−Sm e−Smt 0 (m−1)e−µt+(S−µ)e−Smt−(Sm−µ)e−St (m−1)(S−µ)(Sm−µ) e−St−e−Smt m−1 e−St         ¯˜F ¯ua ¯uo    . (B64) (B64) The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A19), a simple consequence of linearity. 15 The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A19), a simple consequence of linearity. 15 to eq. (A19), a simple consequence of linearity. 15 15 31 31 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. B17 Experiment L3C Full interaction both ways. Full interaction both ways. ∂t ˜F = − µ ˜F + F (B65) ∂tua = − Sm(ua −uo) + ˜F (B66) ∂tuo = S (ua −uo) (B67) 5 ∂t ˜F = − µ ˜F + F (B65) ∂tua = − Sm(ua −uo) + ˜F (B66) ∂tuo = S (ua −uo) (B67) 5 ∂t ˜F = − µ ˜F + F ∂tua = − Sm(ua −uo) + ˜F ∂tuo = S (ua −uo) 5 5 ∂t     ˜F ua uo    =     −µ 0 0 1 −Sm Sm 0 S −S         ˜F ua uo    +     F 0 0     ∂t     ˜F ua uo    =     −µ 0 0 1 −Sm Sm 0 S −S         ˜F ua uo    +     F 0 0     (B68) ∂t     F ua uo    =     µ 0 0 1 −Sm Sm 0 S −S         F ua uo    +     F 0 0     (B68) (B68) λ1 = −µ, e1 =     µ(µ −SM) S −µ S    ; λ2 = −SM, e2 =     0 m −1    ; λ3 = 0, e3 =     0 1 1     (B69) P = ADA−1 =     −µ 0 0 1 −Sm Sm 0 S −S    =     µ(µ −SM) 0 0 S −µ m 1 S −1 1         −µ 0 0 0 −SM 0 0 0 0         [µ(µ −SM)]−1 0 0 [M(µ −SM)]−1 M −1 −M −1 (µM)−1 M −1 mM −1    . B17 Experiment L3C From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): From this follows (after dropping decaying exponentials): ⟨u2 a(t)⟩Ω = R(−3S2M 2 + (2m2 −m −3)Sµ + m(m + 4)µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B72) ⟨ua(t) ˜F(t)⟩Ω = R(µ + S) µ2(SM + µ) (B73) ⟨u2 o(t)⟩Ω = −R(3S2M 2 + 5(m + 1)Sµ + 3µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B74) ⟨ua(t)uo(t)⟩Ω = R(−3S2M 2 + (m2 −3m −4)Sµ + (m −2)µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B75) 5 ⟨u2 a(t)⟩Ω = R(−3S2M 2 + (2m2 −m −3)Sµ + m(m + 4)µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B72) ⟨ua(t) ˜F(t)⟩Ω = R(µ + S) µ2(SM + µ) (B73) ⟨u2 o(t)⟩Ω = −R(3S2M 2 + 5(m + 1)Sµ + 3µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B74) ⟨ua(t)uo(t)⟩Ω = R(−3S2M 2 + (m2 −3m −4)Sµ + (m −2)µ2) M 3Sµ3(SM + µ) + 2R(t −t0) µ2M 2 (B75) In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: In the absence of forcing an initial perturbation at t = 0, ( ¯˜F, ¯ua, ¯uo) evolves as: 0 Mt m(1−e−SMt) M Mt e−SMt+m M         ¯˜F ¯ua ¯uo    . (B76)     ˜F P (t) uP a (t) uP o (t)    =     e−µt 0 0 M(S−µ)e−µt+mµe−SMt+(µ−SM) µM(µ−SM) 1+me−SMt M m(1−e−SMt) M SMe−µt−µe−SMt−(SM−µ) µM(µ−SM) 1−e−SMt M e−SMt+m M         ¯˜F ¯ua ¯uo    . (B76) (B76) The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A26), a simple consequence of linearity. The perturbation matrix above is obtained by calculating Aexp(Dt)A−1. Note that the lower-right 2×2 sub-matrix is identical to eq. (A26), a simple consequence of linearity. Appendix C: Fluctuation Dissipation Theorem 10 The ﬂuctuation dissipation theorem applies to a system, if the system relaxes from a forced state to the unforced dynamics, in the same manner as if the forced state were due to an internal ﬂuctuation of the system. The average response of a system to an external small amplitude forcing is: ponse of a system to an external small amplitude forcing is: ⟨v(t,u)⟩Ω= ⟨u(t)⟩Ω+ t Z −∞ ˜µ(t,s)F(s)ds + O(F 2), ⟨v(t,u)⟩Ω= ⟨u(t)⟩Ω+ t Z −∞ ˜µ(t,s)F(s)ds + O(F 2), (C1) (C1) where the ﬁrst term on the r.h.s. is the unforce dynamics. The upper bound of the integral is imposed by causality. In the 15 linear case only the ﬁrst term in the Taylor expansion of the perturbation has to be considered (O(F 2) = 0) and we can put ⟨u(t)⟩Ω= 0 and v(t,u) = v(t) as the evolution does not depend on the state u. When the system is stationary we can simplify to µ(t −s) = ˜µ(t,s). Langevin equation (white noise) B17 Experiment L3C P = ADA−1 =     −µ 0 0 1 −Sm Sm 0 S −S    = P = ADA−1 =     −µ 0 0 1 −Sm Sm 0 S −S    =     µ(µ −SM) 0 0 S −µ m 1 S −1 1         −µ 0 0 0 −SM 0 0 0 0         [µ(µ −SM)]−1 0 0 [M(µ −SM)]−1 M −1 −M −1 (µM)−1 M −1 mM −1    . (B70) The solution is: 10    10     ˜F ua(t) uo(t)    = AIt t0(D)A−1     1 0 0    dt′ = A     It t0(µ) 0 0 0 It t0(SM) 0 0 0 It t0(0)dt′    A−1     1 0 0     =     It t0(µ) M(S −µ)It t0(µ) + mµIt t0(SM) + (µ −SM)It t0(0) /[Mµ(µ −SM)] SMIt t0(µ) −µIt t0(SM) + (µ −SM)It t0(0) /[Mµ(µ −SM)]    . (B71) =     M(S −µ)It t0(µ) + mµIt t0(SM) + (µ −SM)It t0(0) /[Mµ(µ −SM)] SMIt t0(µ) −µIt t0(SM) + (µ −SM)It t0(0) /[Mµ(µ −SM)]    . (B71) 32 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. C1 Example: Langevin equation (white noise) Note that for Brownian motion (S = 0) perturbations do not decay and the process has the martingale property, but also in this case the normalized correlation matrix does not depend on the absolute time t0 as the process is of stationary increment. C1 Example: Langevin equation (white noise) The Ornstein-Uhlenbeck process (S > 0) and Brownian motion (S = 0) is considered: 20 The Ornstein-Uhlenbeck process (S > 0) and Brownian motion (S = 0) is considered: 20 The Ornstein-Uhlenbeck process (S > 0) and Brownian motion (S = 0) is considered 20 ∂tuO = −SuO + F. (C2) ∂tuO = −SuO + F. (C2) The response function is, using eq. (A3): µ(∆t) = exp(−S∆t). (C3) µ(∆t) = exp(−S∆t). 33 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Straight forward calculations using eq. (B27) show that an initial perturbation decreases as: uO(t0 + ∆t) uO(t0) = χ(∆t) = exp(−S∆t). uO(t0 + ∆t) uO(t0) = χ(∆t) = exp(−S∆t). (C4) (C4) On the other hand we can show, using the same equation, that time correlation is: On the other hand we can show, using the same equation, that time correlation is: (C5) C(t0,∆t) = ⟨uO(t0)uO(t0 + ∆t)⟩Ω= exp(−S∆t) and the normalized correlation matrix is: 5 and the normalized correlation matrix is: 5 ˜C(∆t) = C(t0,∆t)C(t0,0)−1 = exp(−S∆t) (C6) This leads to: This leads to: µ(∆t) = ˜C(∆t) = χ(∆t). µ(∆t) = ˜C(∆t) = χ(∆t). (C7) Historically the ﬁrst equality is the ﬁrst FDT and the second equality is Onsager’s principle (see Barrat and Hansen (2003)). Today the second equality which matches the decay of an initial perturbation with the normalized correlation matrix is referred 10 to as the FDT. Today the second equality which matches the decay of an initial perturbation with the normalized correlation matrix is referred 10 to as the FDT. Today the second equality which matches the decay of an initial perturbation with the normalized correlation matrix is referred 10 to as the FDT. Note that for Brownian motion (S = 0) perturbations do not decay and the process has the martingale property, but also in this case the normalized correlation matrix does not depend on the absolute time t0 as the process is of stationary increment. Eq. (C11) shows that: µ(∆t) = χ(∆t). Eq. (C11) shows that: µ(∆t) = χ(∆t). The time-lagged correlation matrix is: C(t) =  ⟨˜F(t0) ˜F(t0 + ∆t)⟩Ω ⟨uOC(t0) ˜F(t0 + ∆t)⟩Ω ⟨˜F(t0)uOC(t0 + ∆t)⟩Ω ⟨uOC(t0)uOC(t0 + ∆t)⟩Ω   C(t) =  ⟨˜F(t0) ˜F(t0 + ∆t)⟩Ω ⟨uOC(t0) ˜F(t0 + ∆t)⟩Ω ⟨˜F(t0)uOC(t0 + ∆t)⟩Ω ⟨uOC(t0)uOC(t0 + ∆t)⟩Ω   5 C(t) =  ⟨˜F(t0) ˜F(t0 + ∆t)⟩Ω ⟨uOC(t0) ˜F(t0 + ∆t)⟩Ω ⟨˜F(t0)uOC(t0 + ∆t)⟩Ω ⟨uOC(t0)uOC(t0 + ∆t)⟩Ω   5 = R   e−µ∆t µ e−µ∆t µ(S+µ) 2 e−µ∆t−e−S∆t S2−µ2 + e−µ∆t µ(S+µ) Se−µ∆t−µe−S∆t µS(S2−µ2)  . (C13)   = R   e−µ∆t µ e−µ∆t µ(S+µ) 2 e−µ∆t−e−S∆t S2−µ2 + e−µ∆t µ(S+µ) Se−µ∆t−µe−S∆t µS(S2−µ2)  . (C13) µ∆t µ e−µ∆t µ(S+µ) ∆t + e−µ∆t µ(S+µ) Se−µ∆t−µe−S∆t µS(S2−µ2)  . (C13) (C13) Note that: ∂t⟨uOC(t0)uOC(t0 +t)⟩Ω\|t=0 = 0, so that contrary to the white-noise case the correlation is differentiable at t = 0. Calculations give: Note that: ∂t⟨uOC(t0)uOC(t0 +t)⟩Ω\|t=0 = 0, so that contrary to the white-noise case the correlation is differentiable at t = 0. Calculations give: Calculations give: Calculations give: C(0)−1 = R−1   S + µ −S(S + µ) −S(S + µ) S(S + µ)2   (C14) and the normalized correlation matrix: 10 and the normalized correlation matrix: 10 ˜C(∆t) = C(∆t)C(0)−1 =   e−µ∆t 0 e−µ∆t−e−S∆t S−µ e−S∆t  . (C15) As for the white noise case we get: µ(∆t) = ˜C(∆t) = χ(∆t). (C16) µ(∆t) = ˜C(∆t) = χ(∆t). µ(∆t) = ˜C(∆t) = χ(∆t). (C16) The ﬁrst equality is the (ﬁrst) FDT and the second equality is Onsager’s principle. The ﬁrst equality is the (ﬁrst) FDT and the second equality is Onsager’s principle. The ﬁrst equality is the (ﬁrst) FDT and the second equality is Onsager’s principle. 15 When only eq. (C9) forced by a colored noise is considered the FDT does not apply. Indeed, an imposed perturbation ¯uOC 15 still has the same decay of the white noise case given by eq. (C4), as the decay in the linear equation does not depend on the noise. The response function is also identical to eq. (C3), as the response in the linear equation does not depend on the noise. It follows, that µu(∆t) = χu(∆t). When only eq. (C9) forced by a colored noise is considered the FDT does not apply. C2 Example: Langevin equation with colored noise ∂t ˜F = − µ ˜F + F (C8) 5 ∂tuOC = − SuOC + ˜F. (C9) ∂t ˜F = − µ ˜F + F 15 ∂tuOC = − SuOC + ˜F. (C8) = µF + F = − SuOC + ˜F. (C9) The mathematical structure is the same as the L2 model and the solution is: The mathematical structure is the same as the L2 model and the solution is: The mathematical structure is the same as the L2 model and the solution is: ˜F(t) = t Z t0 eµ(t′−t)F(t′)dt′ + eµ(t0−t) ˜F(t0) (C10) uOC(t) = t Z t0 eS(t′−t) ˜F(t′)dt′ + eS(t0−t)uOC(t0) = 1 µ −S t Z t0 (eS(t′−t) −eµ(t′−t))F(t′)dt′ 0 + eS(t0−t)uOC(t0) + ˜F(t0) µ −S (eS(t0−t) −eµ(t0−t)). (C11) ˜F(t) = t Z t0 eµ(t′−t)F(t′)dt′ + eµ(t0−t) ˜F(t0) ˜F(t) = t Z t0 eµ(t′−t)F(t′)dt′ + eµ(t0−t) ˜F(t0) (C10) u (t) t Z eS(t′−t) ˜F(t′)dt′ + eS(t0−t)u (t ) (C10) uOC(t) = t Z t0 eS(t′−t) ˜F(t′)dt′ + eS(t0−t)uOC(t0) = 1 µ −S t Z t0 (eS(t′−t) −eµ(t′−t))F(t′)dt′ 20 + eS(t0−t)uOC(t0) + ˜F(t0) µ −S (eS(t0−t) −eµ(t0−t)). + eS(t0−t)uOC(t0) + ˜F(t0) µ −S (eS(t0−t) −eµ(t0−t)). (C11) (C11) It is important to see that the system composed of eqs. (C8) and (C9) is forced by a white noise and does obey the FDT (in 2D space), when only eq. (C9) is considered forced by a colored noise the FDT (in 1D space) is not observed. 34 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. c⃝Author(s) 2019. CC BY 4.0 License. c⃝Author(s) 2019. CC BY 4.0 License. More precisely, in 2D space: a perturbation ( ¯F, ¯uOC) (putting F = 0 in eqs. (C10) and (C11)) decreases as:  ˜F P (∆t) uP OC(∆t)  = χ(∆t)  ¯F ¯uOC  =   e−µ∆t 0 e−µ∆t−e−S∆t S−µ e−S∆t    ¯F ¯uOC  . (C12) Eq. (C11) shows that: µ(∆t) = χ(∆t). Indeed, an imposed perturbation ¯uOC 15 still has the same decay of the white noise case given by eq. (C4), as the decay in the linear equation does not depend on the noise. The response function is also identical to eq. (C3), as the response in the linear equation does not depend on the noise. It follows, that µu(∆t) = χu(∆t). Whereas the scalar calculations give: ˜Cu(∆t) = ⟨uOC(t0)uOC(t0 + ∆t)⟩Ω ⟨u2 OC(t0)⟩Ω = Se−µ∆t −µe−S∆t S −µ , (C17) 20 (t0 + ∆t)⟩Ω t0)⟩Ω = Se−µ∆t −µe−S∆t S −µ , (C1 ˜Cu(∆t) = ⟨uOC(t0)uOC(t0 + ∆t)⟩Ω ⟨u2 OC(t0)⟩Ω = Se−µ∆t −µe−S∆t S −µ , 20 (C17) ⟨uOC(t0)⟩Ω S µ which does not agree with the response function or the decay law of a perturbation and so neither the FDT nor Onsager’s principle is observed. The failure of the FDT is due to the non-vanishing auto-correlation time of ˜F. A consequence of this is that ⟨uOC(t0) ˜F(t0 + t))⟩Ω̸= 0 even if t > 0, meaning that the future forcing is correlated to the actual state. which does not agree with the response function or the decay law of a perturbation and so neither the FDT nor Onsager’s principle is observed. The failure of the FDT is due to the non-vanishing auto-correlation time of ˜F. A consequence of this is that ⟨uOC(t0) ˜F(t0 + t))⟩Ω̸= 0 even if t > 0, meaning that the future forcing is correlated to the actual state. The FDT applies only when the forcing correlation time vanishes, that is µ ≫S or a generalized Langevin equation is used, that is the friction term is presented by a memory kernel and eq. (C9) replaced by: 25 ∂tuOC(t) = − S µ t Z −∞ eµ(t′−t)uOC(t′)dt′ + ˜F. (C18) 35 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. References Balakrishnan, V.: Fluctuation-dissipation theorems from the generalised Langevin equation, Pramana, 12, 301–315, 1979. Barrat, J.-L. and Hansen, J.-P.: Basic concepts for simple and complex liquids, Cambridge University Press, 2003. Boffetta, G. and Ecke, R. E.: Two-dimensional turbulence, Annual Review of Fluid Mechanics, 44, 427–451, 2012. Fluctuation-dissipation theorems from the generalised Langevin equation, Pramana, 12, 301–315, 1979. Hansen, J.-P.: Basic concepts for simple and complex liquids, Cambridge University Press, 2003. Ecke, R. E.: Two-dimensional turbulence, Annual Review of Fluid Mechanics, 44, 427–451, 2012. Ciliberto, S., Garnier, N., Hernandez, S., Lacpatia, C., Pinton, J.-F., and Chavarria, G. R.: Experimental test of the Gallavotti–Cohen ﬂuctua- 5 tion theorem in turbulent ﬂows, Physica A: Statistical Mechanics and its Applications, 340, 240–250, 2004. Ciliberto, S., Garnier, N., Hernandez, S., Lacpatia, C., Pinton, J.-F., and Chavarria, G. R.: Experimental test of the Gallavotti–Cohen ﬂuctua- 5 tion theorem in turbulent ﬂows, Physica A: Statistical Mechanics and its Applications, 340, 240–250, 2004. Cooper, F. C. and Haynes, P. H.: Climate sensitivity via a nonparametric ﬂuctuation–dissipation theorem, Journal of the Atmospheric Sci- ences, 68, 937–953, 2011. Csanady, G. T.: Air-sea interaction: laws and mechanisms, Cambridge University Press, 2001. Ciliberto, S., Garnier, N., Hernandez, S., Lacpatia, C., Pinton, J.-F., and Chavarria, G. R.: Experimental test of the Gallavotti–Cohen ﬂuctua- 5 tion theorem in turbulent ﬂows, Physica A: Statistical Mechanics and its Applications, 340, 240–250, 2004. Cooper F C and Haynes P H : Climate sensitivity via a nonparametric ﬂuctuation dissipation theorem Journal of the Atmospheric Sci tion theorem in turbulent ﬂows, Physica A: Statistical Mechanics and its Applications, 340, 240–250, 2004. Cooper, F. C. and Haynes, P. H.: Climate sensitivity via a nonparametric ﬂuctuation–dissipation theorem, Journal of the Atmospheric Sci- ences, 68, 937–953, 2011. Cooper, F. C. and Haynes, P. H.: Climate sensitivity via a nonparametric ﬂuctuation–dissipation theorem, Journal of the Atmospheric Sci- ences, 68, 937–953, 2011. Cooper, F. C. and Haynes, P. H.: Climate sensitivity via a nonparametric ﬂuctuation–dissipation theorem, Journal of the Atmospheric Sci- ences, 68, 937–953, 2011. Csanady, G. T.: Air-sea interaction: laws and mechanisms, Cambridge University Press, 2001. anady, G. T.: Air-sea interaction: laws and mechanisms, Cambridge University Press, 2001. Einstein, A.: Zur theorie der Brownschen Bewegung, Annalen der Physik, 324, 371–381, 1906. 10 Einstein, A.: Investigations on the Theory of the Brownian Movement, Courier Corporation, 1956. Ferrari, R. Eq. (C11) shows that: µ(∆t) = χ(∆t). A more general discussion of the problem of causality due to time-correlated noise and the generalized Langevin equation are given by Balakrishnan (1979). As the system is linear the pdf’s of the variables are Gaussian. In the unperturbed system averages vanish and second-order moments are given in eq. (C13) by setting t = 0. Author contributions. AW has performed the coding, the research and the writing of the paper 5 5 Competing interests. AW has no competing interest Competing interests. AW has no competing interest Acknowledgements. This work was funded by Labex OASUG@2020 (Investissement d’avenir - ANR10 LABX56). 36 Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. Geosci. Model Dev. Discuss., https://doi.org/10.5194/gmd-2018-300 Manuscript under review for journal Geosci. Model Dev. Discussion started: 29 January 2019 c⃝Author(s) 2019. CC BY 4.0 License. References and Wunsch, C.: Ocean circulation kinetic energy: Reservoirs, sources, and sinks, Annual Review of Fluid Mechanics, 41, 2009. Gallavotti, G. and Cohen, E. G. D.: Dynamical ensembles in nonequilibrium statistical mechanics, Physical Review Letters, 74, 2694, 1995a. Gallavotti, G. and Cohen, E. G. D.: Dynamical ensembles in stationary states, Journal of Statistical Physics, 80, 931–970, 1995b. Ferrari, R. and Wunsch, C.: Ocean circulation kinetic energy: Reservoirs, sources, and sinks, Annual Review of Fluid Mechanics, 41, 2009. Gallavotti, G. and Cohen, E. G. D.: Dynamical ensembles in nonequilibrium statistical mechanics, Physical Review Letters, 74, 2694, 1995a. G ll tti G d C h E G D D i l bl i t ti t t J l f St ti ti l Ph i 80 931 970 1995b d Wunsch, C.: Ocean circulation kinetic energy: Reservoirs, sources, and sinks, Annual Review of Fluid Mech and Cohen, E. G. D.: Dynamical ensembles in nonequilibrium statistical mechanics, Physical Review Letters, Gallavotti, G. and Cohen, E. G. D.: Dynamical ensembles in stationary states, Journal of Statistical Physics, 80, 931–970, 1995b. Gaunt, R. E.: A note on the distribution of the product of zero-mean correlated normal random variables, Statistica Neerlandica, 2018. 15 Kubo, R.: The ﬂuctuation-dissipation theorem, Reports on progress in physics, 29, 255, 1966. Kuhn, T. S.: The structure of scientiﬁc revolutions, vol. 2, University of Chicago press Chicago, 1963. Marconi, U. M. B., Puglisi, A., Rondoni, L., and Vulpiani, A.: Fluctuation–dissipation: response theory in statistical physics, Physics reports, 461, 111–195, 2008. A note on the distribution of the product of zero-mean correlated normal random variables, Statistica Neerlan , p , , Kubo, R.: The ﬂuctuation-dissipation theorem, Reports on progress in physics, 29, 255, 1966. Kuhn, T. S.: The structure of scientiﬁc revolutions, vol. 2, University of Chicago press Chicago, 1963. Marconi, U. M. B., Puglisi, A., Rondoni, L., and Vulpiani, A.: Fluctuation–dissipation: response theory in statistical physics, Physics reports, e ﬂuctuation-dissipation theorem, Reports on progress in physics, 29, 255, 1966. Kubo, R.: The ﬂuctuation-dissipation theorem, Reports on progress in physics, 29, 255, 1966. Kuhn, T. S.: The structure of scientiﬁc revolutions, vol. 2, University of Chicago press Chicago, 1963. Marconi, U. M. B., Puglisi, A., Rondoni, L., and Vulpiani, A.: Fluctuation–dissipation: response theory in statistical physics, Physics reports, 461, 111–195, 2008. The structure of scientiﬁc revolutions, vol. 2, University of Chicago press Chicago, 1963. Marconi, U. M. References B., Puglisi, A., Rondoni, L., and Vulpiani, A.: Fluctuation–dissipation: response theory in statistical physics, Physics reports, 461, 111–195, 2008. nd Wirth, A.: A drag-induced barotropic instability in air–sea interaction, Journal of Physical Oceanography, 44 Moulin, A. and Wirth, A.: A drag-induced barotropic instability in air–sea interaction, Journal of Physical Oceanography, 44, 733–741, 2014. 20 Moulin, A. and Wirth, A.: Momentum Transfer Between an Atmospheric and an Oceanic Layer at the Synoptic and the Mesoscale: An Idealized Numerical Study, Boundary-Layer Meteorology, 160, 551–568, 2016. Nadarajah, S. and Pogány, T. K.: On the distribution of the product of correlated normal random variables, Comptes Rendus Mathematique, 354, 201–204, 2016. Moulin, A. and Wirth, A.: Momentum Transfer Between an Atmospheric and an Oceanic Layer at the Synoptic and the Mesoscale: An Idealized Numerical Study, Boundary-Layer Meteorology, 160, 551–568, 2016. N d j h S d P á T K O th di t ib ti f th d t f l t d l d i bl C t R d M th ti y y y gy Nadarajah, S. and Pogány, T. K.: On the distribution of the product of correlated normal random variables, Comptes Rendus Mathematique, 354, 201–204, 2016. Perrin, J.: Atomes (Les), CNRS Editions, Paris, ISBN: 978-2-271-08260-2, 2014. 25 Renault, L., McWilliams, J. C., and Masson, S.: Satellite Observations of Imprint of Oceanic Current on Wind Stress by Air-Sea Coupling, Scientiﬁc reports, 7, 17 747, 2017. Stocker, T. F., Qin, D., Plattner, G., Tignor, M., Allen, S., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P.: Climate change 2013: the physical science basis. Intergovernmental panel on climate change, working group I contribution to the IPCC ﬁfth assessment report Renault, L., McWilliams, J. C., and Masson, S.: Satellite Observations of Imprint of Oceanic Current on Wind Stress by Air-Sea Coupling, Scientiﬁc reports, 7, 17 747, 2017. S k T F Qi D Pl G Ti M All S B h J N l A Xi Y B V d Mid l P Cli h 2013 p , , , Stocker, T. F., Qin, D., Plattner, G., Tignor, M., Allen, S., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P.: Climate change 2013: the physical science basis. Intergovernmental panel on climate change, working group I contribution to the IPCC ﬁfth assessment report (AR5) N Y k 2013 30 30 (AR5), New York, 2013. 30 Stull, R. References B.: An introduction to boundary layer meteorology, vol. 13, Springer Science & Business Media, 2012. Wirth, A.: Etudes et évaluation de processus océaniques par des hiérarchies de modèles, Habilitation à Dirige des Recherches (HDR), 2010. Wirth, A.: A Fluctuation–Dissipation Relation for the Ocean Subject to Turbulent Atmospheric Forcing, Journal of Physical Oceanography, 48, 831–843, 2018. (AR5), New York, 2013. 30 Stull, R. B.: An introduction to boundary layer meteorology, vol. 13, Springer Science & Business Media, 2012. Stull, R. B.: An introduction to boundary layer meteorology, vol. 13, Springer Science & Business Media, 2012. Stull, R. B.: An introduction to boundary layer meteorology, vol. 13, Springer Science & Business Media, 2012. Wirth, A.: Etudes et évaluation de processus océaniques par des hiérarchies de modèles, Habilitation à Dirige des Recherches (HDR), 2010. Wirth, A.: A Fluctuation–Dissipation Relation for the Ocean Subject to Turbulent Atmospheric Forcing, Journal of Physical Oceanography, 48, 831–843, 2018. Wirth, A.: Etudes et évaluation de processus océaniques par des hiérarchies de modèles, Habilitation à Dirige Wirth, A.: A Fluctuation–Dissipation Relation for the Ocean Subject to Turbulent Atmospheric Forcing, Journal of Physical Oceanography, 48, 831–843, 2018. 37
https://openalex.org/W4391110368	https://www.nature.com/articles/s41598-024-52359-y.pdf	English	null	A MEK inhibitor arrests the cell cycle of human conjunctival fibroblasts and improves the outcome of glaucoma filtration surgery	Scientific reports	2,024	cc-by	8,315	A MEK inhibitor arrests the cell cycle of human conjunctival fibroblasts and improves the outcome of glaucoma filtration surgery OPEN Jinhee Lee , Megumi Honjo * & Makoto Aihara Better agents are needed to improve glaucoma filtration surgery outcomes compared to current ones. The purpose of this study is to determine whether mitogen-activated protein kinase kinase (MEK) inhibitors can effectively arrest the cell cycle of human conjunctival fibroblasts (HCFs) and inhibit the formation of fibrosis and scarring following glaucoma filtration surgery. A cell counting kit‑8 assay revealed that the MEK inhibitor PD0325901 exhibited concentration-dependent growth inhibition of HCFs. Quantitative PCR, immunocytochemistry, and western blotting demonstrated decreased expression of proliferating cell nuclear antigen (PCNA) and cyclin D1 and increased expression of p27 in HCFs treated with PD0325901. Flow cytometry indicated that PD0325901 arrested the cell cycle of HCFs in the G0/1 phase. The cell-migration assay showed that HCF migration rate was significantly suppressed by PD0325901 exposure. Rabbits were divided into PD0325901-treatment and control groups, and glaucoma filtration surgery was performed. Although intraocular pressure did not differ between PD0325901-treatment and control groups, bleb height was greater in the treatment group. Histopathological evaluation revealed that fibrotic changes were significantly attenuated in the PD0325901-treatment group compared to the control group. In conclusion, the MEK inhibitor impedes HCF proliferation via cell-cycle arrest and may be beneficial for glaucoma filtration surgery by reducing bleb scarring. Glaucoma filtration surgery is the standard procedure for glaucoma patients whose intraocular pressure (IOP) cannot be adequately controlled with medication1. Glaucoma filtration surgery facilitates the drainage of aque- ous humor under the conjunctiva. However, excessive postoperative scar formation can lead to the loss of sub- conjunctival space, resulting in surgical failure1. Currently, 5-fluorouracil (5-FU) and mitomycin-C (MMC) are applied during the perioperative period to inhibit scar formation2. These agents are commonly used for control- ling postoperative IOP, however, they can induce complications such as bleb leakage, hypotony, avascularized bleb, and endophthalmitis3. Consequently, there is a need for drugs that can safely lower IOP with fewer side effects than 5-FU and MMC, or can improve the benefits of existing anti-scarring drugs. f pi g g g Existing evidence indicates that several factors in the aqueous humor contribute to postoperative subcon- junctival fibrosis4. Tumor growth factor (TGF)-β is a major aqueous mediator that induces the contraction, proliferation, and migration of human conjunctival fibroblasts (HCFs), and has been used in in vitro glaucoma filtration surgery experiments5. www.nature.com/scientificreports www.nature.com/scientificreports Department of Ophthalmology, Graduate School of Medicine, The University of Tokyo, 7‑3‑1 Hongo Bunkyo‑ku, Tokyo 113‑8655, Japan. email: honjomegumi@gmail.com Scientific Reports \| (2024) 14:1871 Effects of PD0325901 on toxicity and proliferation f 3 59 y p PD0325901 (Mirdametinib; FUJIFILM Wako Pure Chemical, Osaka, Japan), a selective MEK inhibitor, was used. First, the 24-h toxicity of PD0325901 in HCFs was confirmed using the cell counting kit‑8 (CCK‑8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan) assay (Fig. 1). The OD450 of cells exposed to PD0325901 did not differ from that of the 0.1% dimethyl sulfoxide (DMSO) control. However, 96 h of PD0325901 exposure showed a concentration-dependent HCF growth-inhibition effect. As concentrations of 1 µM or higher showed significant growth inhibition at all time points after 48 h (P < 0.041), and there was no difference in OD450 up to 72 h at concentrations of 100 nM and 316 nM compared to DMSO (P > 0.097), we used 1 and 10 µM PD0325901 in subsequent experiments. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Mitogen-activated protein kinase kinase (MEK) inhibitors have demonstrated efficacy in both arresting the cell cycle11 and preventing fibrosis through a rapidly accelerated fibrosarcoma (Raf)/MEK/extracellular signal- regulated kinase (ERK) pathway12,13. Recently, MEK inhibitors have been applied in the treatment of melanoma14, and clinical trials using MEK inhibitors have been conducted for lung cancer15 and thyroid carcinoma16. The MEK inhibitor U0126 suppresses TGF-β1-induced myofibroblast transdifferentiation in HCFs17. However, the effects of MEK inhibitors on HCF cell-cycle arrest and their application in glaucoma filtration surgery in vivo have not been investigated.f g In this study, we investigated the effect of a MEK inhibitor on cell-cycle arrest in serum-stimulated HC and evaluated its treatment efficacy in a rabbit model of glaucoma filtration surgery. Effects of PD0325901 on mRNA expression f p Proliferating cell nuclear antigen (PCNA) plays a crucial role in the cell cycle, particularly in DNA replication and repair18. PCNA is an auxiliary protein for DNA polymerase that reaches maximal expression during the S phase of the cell cycle. Cyclin D1 plays a central role in the regulation of proliferation, linking the extracellular signaling environment to cell cycle progression19. It is involved in the transition from the G1 phase to the S phase of the cell cycle, promoting cell cycle progression by binding to and activating cyclin-dependent kinase (CDK) 4/6, which in turn phosphorylates and inactivates the retinoblastoma protein, allowing the cell to enter the S phase. p27 is a key regulator of the cell cycle, exerting multiple functions in cell cycle control, apoptosis, epigenetic modification, and transcriptional regulation20. Its inhibitory effect on the cell cycle is mediated by the inhibition of CDKs, thereby preventing the progression of the cell cycle from the G1 phase. Since the decrease in cyclin D1 and PCNA, and the increase in p27 could indicate an effect of PD0325901 on cell cycle arrest in the G1 phase, these markers were investigated in this study.i p g y Figure 2 shows the Quantitative real-time PCR (qPCR) results. PCNA expression was significantly lower in the 1- and 10-µM PD0325901 treatment groups than in the 0.1% DMSO control (P < 0.001). Compared to controls, cyclin D1 expression was significantly lower and p27 expression was significantly higher after treatment with 10 µM PD0325901 (P = 0.026, 0.013, respectively). 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 OD450 value (n = 6) a b 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 24h 48h 72h 96h 100μM 31.6μM 10μM 3.16μM 1μM 316nM 100nM DMSO 0.1% OD450 value (n = 6) Figure 1. Effects of PD0325901 on toxicity and proliferation. (a) To assess toxicity, HCFs were incubated for 24 h with PD0325901 (100 nM–100 µM or 0.1% DMSO, after which the OD450 was measured using the CCK-8 assay. No significant differences were observed among the groups. (b) To evaluate proliferation, HCFs were incubated for up to 96 h with PD0325901 or 0.1% DMSO, after which the OD450 was measured using the CCK-8 assay. The points in this figure show the average OD450 at each PD0325901 concentration or 0.1% DMSO. PD0325901 inhibited HCF proliferation in a concentration-dependent manner. Figure 1. A MEK inhibitor arrests the cell cycle of human conjunctival fibroblasts and improves the outcome of glaucoma filtration surgery OPEN However, in a previous study, administration of anti-TGF-β2 antibody during trabeculectomy did not improve surgical outcomes over a placebo6, implying that suppression of TGF-β alone is inadequate. q Fibroblasts in most tissues exist in a quiescent phase; however, surgical intervention activates tissue-repair mechanisms, leading to cell-cycle progression7. Excessive wound healing responses result in fibrotic scar formation8; therefore, strategies to regulate the cell cycle may be beneficial for successful glaucoma filtration surgery. For instance, fetal bovine serum (FBS) induces HCF proliferation and has been used in several studies of HCF cell-cycle regulation9,10. \| https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 Effects of PD0325901 on cell cycleh The percentage of HCFs in the G0/1 phase was higher, and those of the S and G2/M phases were lower, in both treatment groups versus controls (Table 1, Fig. 5, P < 0.008). Effects of PD0325901 on migration assay f g y Figure 6a shows typical images of the cell-migration assay. There were no differences in scratch width between the DMSO and PD0325901 groups at up to 24 h. However, at 48 h, the scratch disappeared in all controls but remained in all PD0325901 treatment samples; the scratch widths significantly differed between the treatment groups and controls (Fig. 6b, P < 0.001). Effects of PD0325901 on a glaucoma filtration model in rabbitsti Effects of PD0325901 on mRNA expression Effects of PD0325901 on toxicity and proliferation. (a) To assess toxicity, HCFs were incubated for 24 h with PD0325901 (100 nM–100 µM or 0.1% DMSO, after which the OD450 was measured using the CCK-8 assay. No significant differences were observed among the groups. (b) To evaluate proliferation, HCFs were incubated for up to 96 h with PD0325901 or 0.1% DMSO, after which the OD450 was measured using the CCK-8 assay. The points in this figure show the average OD450 at each PD0325901 concentration or 0.1% DMSO. PD0325901 inhibited HCF proliferation in a concentration-dependent manner. Figure 1. Effects of PD0325901 on toxicity and proliferation. (a) To assess toxicity, HCFs were incubated f 24 h i h PD0325901 (100 M 100 M 0 1% DMSO ft hi h h OD450 d i h https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| www.nature.com/scientificreports/ 0 0.2 0.4 0.6 0.8 1 1.2 0µM 1µM 10µM PCNA 0 0.2 0.4 0.6 0.8 1 0µM 1µM 10µM cyclin D1 0 0.5 1 1.5 2 0µM 1µM 10µM p27 (n = 5) * * * Relative expression to GAPDH Relative expression to GAPDH Relative expression to GAPDH Figure 2. Effects of PD0325901 on mRNA expression. PCNA expression decreased after exposure to 1 and 10 μM PD0325901 compared to 0.1% DMSO (P < 0.001). Compared to controls, cyclin D1 expression decreased and p27 expression increased after exposure to 10 μM PD0325901. P < 0.05 vs. 0.1% DMSO control. 0 0.2 0.4 0.6 0.8 1 1.2 0µM 1µM 10µM PCNA 0 0.2 0.4 0.6 0.8 1 0µM 1µM 10µM cyclin D1 0 0.5 1 1.5 2 0µM 1µM 10µM p27 (n = 5) * * * Relative expression to GAPDH Relative expression to GAPDH Relative expression to GAPDH Figure 2. Effects of PD0325901 on mRNA expression. PCNA expression decreased after exposure to 1 and 10 μM PD0325901 compared to 0.1% DMSO (P < 0.001). Compared to controls, cyclin D1 expression decreased and p27 expression increased after exposure to 10 μM PD0325901. P < 0.05 vs. 0.1% DMSO control. Effects of PD0325901 on the expression of protein levelh f ects o 03 590 o t e e p ess o o p ote e e The immunocytochemistry results are shown in Fig. 3. Treatment with PD0325901 downregulated PCNA and cyclin D1 expression but upregulated p27 expression. Figure 4 shows the western blotting results. PCNA and cyclin D1 expression were significantly lower, and p27 expression significantly higher, in both the 1 and 10 µM PD0325901 treatment groups than in the 0.1% DMSO control (all P < 0.001). Effects of PD0325901 on cell cycleh Effects of PD0325901 on a glaucoma filtration model in rabbitsti Representative photographs after glaucoma filtration surgery are shown in Fig. 7. The PD0325901 treatment group showed less hyperemia and more diffuse blebs than the control group. It also had a greater average bleb height on postoperative day 7 (Fig. 8). Although IOP tended to be lower in that group, it did not significantly differ from the control on any of the measurement dates (Fig. 9). No differences were observed between the treatment and control groups in slit-lamp and fundus examinations. g p p Hematoxylin and eosin (HE) staining revealed densely packed cells in the bleb area of controls, suggestive of fibrosis (Fig. 10a, b). Conversely, in the treatment group, the cells were relatively sparse and scar formation was suppressed (Fig. 10c, d). Elastica van Gieson (EVG) staining revealed densely stained collagen deposits and scar formation in both the conjunctiva and sclera of controls (Fig. 10e, f). By contrast, in the PD0325901 treatment p27 Cyclin D1 I P A D A N C P DAPI DAPI DMSO p27 Cyclin D1 I P A D A N C P DAPI DAPI PD0325901 DMSO 100 µm Figure 3. Effects of PD0325901 on immunocytochemistry. Expression of PCNA and cyclin D1 was lower, and p27 expression was higher, in PD0325901-treated HCFs compared to DMSO-treated HCFs. PD0325901 100 µm Figure 3. Effects of PD0325901 on immunocytochemistry. Expression of PCNA and cyclin D1 was lower, and p27 expression was higher, in PD0325901-treated HCFs compared to DMSO-treated HCFs. igure 3. Effects of PD0325901 on immunocytochemistry. Expression of PCNA and cyclin D1 was lower, and 27 expression was higher, in PD0325901-treated HCFs compared to DMSO-treated HCFs. https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| Scientific Reports \| (2024) 14:1871 \| www.nature.com/scientificreports/ 0.0 0.5 1.0 1.5 0µM 1µM 10µM 0.0 0.5 1.0 1.5 0µM 1µM 10µM β-tubulin 10 0 PD0325901 (µM) p27 1 Cyclin D1 PCNA Relative expression to β-tubulin Relative expression to β-tubulin Relative expression to β-tubulin * * * * * PCNA Cyclin D1 p27 0.0 1.0 2.0 3.0 0µM 1µM 10µM (n = 6) (n = 6) (n = 5) b a Figure 4. Effects of PD0325901 on Western blotting. (a) Representative bands from the western blotting assay. Results are expressed relative to the loading control (β-tubulin). (b) Western blotting revealed decreased PCNA and cyclin D1 expression, and increased p27 expression, after exposure to 1 and 10 µM PD0325901 exposure compared to controls (P < 0.001). Effects of PD0325901 on a glaucoma filtration model in rabbitsti P < 0.05 vs. 0.1% DMSO control. The original blots are presented in Supplementary Figs. 1 and 2. Samples from the same experiment were used, and gels and blots were processed simultaneously. 0.0 0.5 1.0 1.5 0µM 1µM 10µM 0.0 0.5 1.0 1.5 0µM 1µM 10µM Relative expression to β-tubulin Relative expression to β-tubulin Relative expression to β-tubulin * * * * * PCNA Cyclin D1 p27 0.0 1.0 2.0 3.0 0µM 1µM 10µM (n = 6) (n = 6) (n = 5) b β-tubulin 10 0 PD0325901 (µM) p27 1 Cyclin D1 PCNA a Figure 4. Effects of PD0325901 on Western blotting. (a) Representative bands from the western blotting assay. Results are expressed relative to the loading control (β-tubulin). (b) Western blotting revealed decreased PCNA and cyclin D1 expression, and increased p27 expression, after exposure to 1 and 10 µM PD0325901 exposure compared to controls (P < 0.001). P < 0.05 vs. 0.1% DMSO control. The original blots are presented in Supplementary Figs. 1 and 2. Samples from the same experiment were used, and gels and blots were processed simultaneously. Table 1. Effect of PD0325901 on cell cycle of HCFs. n = 4, mean ± SD. DMSO 1 µM of PD0325901 10 µM of PD0325901 G0/G1 phase (%) 56.3 ± 1.7 84.9 ± 2.0 85.9 ± 2.1 S phase (%) 28.9 ± 1.3 9.0 ± 0.9 8.4 ± 0.7 G2/M phase (%) 10.2 ± 1.5 3.3 ± 0.8 4.9 ± 1.3 DMSO 1 µM of PD0325901 10 µM of PD0325901 G0/G1 phase (%) 56.3 ± 1.7 84.9 ± 2.0 85.9 ± 2.1 S phase (%) 28.9 ± 1.3 9.0 ± 0.9 8.4 ± 0.7 G2/M phase (%) 10.2 ± 1.5 3.3 ± 0.8 4.9 ± 1.3 Table 1. Effect of PD0325901 on cell cycle of HCFs. n = 4, mean ± SD. 0 50 100 150 200 0 50 100 150 200 0 50 100 150 200 800 600 400 200 0 800 600 400 200 0 800 600 400 200 0 Cell number Cell number Cell number DNA content tn etn o c A N D tn etn o c A N D PD0325901 1µM PD0325901 10µM DMSO 0.1% G0/1 phase S phase G2/M phase Figure 5. Effects of PD0325901 on cell cycle. Effects of PD0325901 on a glaucoma filtration model in rabbitsti Treatment with both 1 and 10 μM PD035901 increased the percentage of HCFs in the G0/1 phase and decreased the percentages of HCFs in the S and G2/M phases compared to the 0.1% DMSO control (all P < 0.008). 0 50 100 150 200 800 600 400 200 0 Cell number tn etn o c A N D PD0325901 1µM 0 50 100 150 200 800 600 400 200 0 Cell number tn etn o c A N D PD0325901 10µM 0 50 100 150 200 800 600 400 200 0 Cell number DNA content DMSO 0.1% G0/1 phase S phase G2/M phase Figure 5. Effects of PD0325901 on cell cycle. Treatment with both 1 and 10 μM PD035901 increased the percentage of HCFs in the G0/1 phase and decreased the percentages of HCFs in the S and G2/M phases compared to the 0.1% DMSO control (all P < 0.008). Scientific Reports \| (2024) 14:1871 \| https://doi.org/10.1038/s41598-024-52359-y www.nature.com/scientificreports/ 0 100 200 300 400 500 600 DMSO 0.1 % PD03259011 µM PD0325901 10 µM DMSO 0.1% PD0325901 1µM 48 h 24 h 0 h PD0325901 10µM b a h 8 4 h 0 24 h Scratch width (µm) (n = 6) Figure 6. Effects of PD0325901 on migration assay. The cell-migration assay was performed on HCFs for 0, 24, and 48 h (n = 6). The baseline (0 h) represented the migration distance of cells with no stimulation. (a) The scratch width did not differ among the groups at 24 h. (b) At 48 h, all DMSO-group replicates showed reduced widths whereas all treatment replicates retained their widths; migration was significantly suppressed by 1 and 10 μM PD0325901 exposure at 48 h post-incubation compared to the DMSO control (P < 0.001). 0 100 200 300 400 500 600 DMSO 0.1 % PD03259011 µM PD0325901 10 µM b h 8 4 h 0 24 h Scratch width (µm) (n = 6) DMSO 0.1% PD0325901 1µM 48 h 24 h 0 h PD0325901 10µM a b Figure 6. Effects of PD0325901 on migration assay. The cell-migration assay was performed on HCFs for 0, 24, and 48 h (n = 6). The baseline (0 h) represented the migration distance of cells with no stimulation. (a) The scratch width did not differ among the groups at 24 h. Effects of PD0325901 on a glaucoma filtration model in rabbitsti (b) At 48 h, all DMSO-group replicates showed reduced widths whereas all treatment replicates retained their widths; migration was significantly suppressed by 1 and 10 μM PD0325901 exposure at 48 h post-incubation compared to the DMSO control (P < 0.001). control group PD0325901 treatment group 7 D O P 5 D O P 2 D O P Figure 7. Representative photographs after glaucoma filtration surgery. Representative photographs of the postoperative conjunctival bleb at postoperative days 2, 5, and 7. Less hyperemia and diffuse blebs were observed in the PD0325901 group compared to the control group. The circle lines indicate the extent of the bleb. control group Figure 7. Representative photographs after glaucoma filtration surgery. Representative photographs of the postoperative conjunctival bleb at postoperative days 2, 5, and 7. Less hyperemia and diffuse blebs were observed in the PD0325901 group compared to the control group. The circle lines indicate the extent of the bleb. group, collagen deposits were faintly stained, conjunctival tissue had a loose appearance, and the subconjunctival space was preserved (Fig. 10g, h). Discussion POD 1 POD 5 POD 7 control group PD0325901 treatment group 0 100 200 300 400 500 600 700 control group PD0325901 treatment group The height of the bleb (µm) a b (n = 6) Figure 8. Anterior segment optical coherence tomography (AS-OCT) after glaucoma filtration surgery. (a) The space in the bleb in the PD0325901 treatment group was better maintained compared to that in the DMSO control group. (b) Bleb heights on postoperative day 7 were significantly greater in the treatment group than in controls (P = 0.003). 0 100 200 300 400 500 600 700 control group PD0325901 treatment group The height of the bleb (µm) * b (n = 6) b Figure 8. Anterior segment optical coherence tomography (AS-OCT) after glaucoma filtration surgery. (a) The space in the bleb in the PD0325901 treatment group was better maintained compared to that in the DMSO control group. (b) Bleb heights on postoperative day 7 were significantly greater in the treatment group than in controls (P = 0.003). Intraocular pressure (mmHg) (n = 6) 0 2 4 6 8 10 12 14 pre POD 1 POD 2 POD 5 POD 7 treatment group control group Figure 9. Intraocular pressure after glaucoma filtration surgery. No significant differences in IOP were observed between the treatment and control groups at any time point. Figure 9. Intraocular pressure after glaucoma filtration surgery. No significant differences in IOP were observed between the treatment and control groups at any time point. xperimental results of our study reveal the effects of a MEK inhibitor in terms of arresting the HCF cell cycle nd inhibiting cell proliferation. 22 MEK and p38 are involved in the mitogen-activated protein kinase (MAPK) cascade22. p38 is activated by oxidative stress and inflammatory cytokines, and is involved in the regulation of cellular responses such as inflam- mation, apoptosis, and differentiation22. It has been shown that p38 inhibitors suppress myofibroblast transdif- ferentiation of HCFs and are effective for treating a rabbit model of glaucoma filtration surgery23,24. The MEK cascade differs from p38 in that it is activated by growth factors and hormones and is involved in regulating cell proliferation22. Wen et al. demonstrated that another MEK inhibitor, U0126, suppressed myofibroblast transdif- ferentiation of HCFs, as well as TGF-β1-induced phosphorylation of Smad2/3, p38/MAPK, and ERK1/217. Discussion We investigated whether PD0325901, a MEK inhibitor, could arrest the HCF cell cycle and prevent scar formation after glaucoma filtration surgery. The CCK-8 assay revealed concentration-dependent growth inhibition of HCFs by PD0325901. Immunocytochemistry, qPCR, and western blotting showed that PD0325901 downregulated cyc- lin D1 and PCNA expression, and increased p27 expression, in HCFs. Flow cytometry indicated that it arrested the HCF cell cycle in the G0/1 phase, while the migration assay revealed that it inhibited migration of HCFs in vitro, implying an inhibitory effect on wound healing. In a rabbit model of glaucoma filtration, PD0325901 effectively maintained bleb volume and prevented fibrotic scarring.h f y pi g ERK1/2 is ubiquitously expressed and is part of the Raf/MEK/ERK signal transduction cascade21. This cas- cade plays a crucial role in cell-cycle progression, cell migration, proliferation, and transcription21. The in vitro https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| www.nature.com/scientificreports/ POD 1 POD 5 POD 7 control group PD0325901 treatment group 0 100 200 300 400 500 600 700 control group PD0325901 treatment group The height of the bleb (µm) * a b (n = 6) Figure 8. Anterior segment optical coherence tomography (AS-OCT) after glaucoma filtration surgery. (a) The space in the bleb in the PD0325901 treatment group was better maintained compared to that in the DMSO control group. (b) Bleb heights on postoperative day 7 were significantly greater in the treatment group than in controls (P = 0.003). POD 1 POD 5 POD 7 control group PD0325901 treatment group 0 100 200 300 400 500 600 700 control group PD0325901 treatment group The height of the bleb (µm) * a b (n = 6) Figure 8. Anterior segment optical coherence tomography (AS-OCT) after glaucoma filtration surgery. (a) The space in the bleb in the PD0325901 treatment group was better maintained compared to that in the DMSO control group. (b) Bleb heights on postoperative day 7 were significantly greater in the treatment group than in controls (P = 0.003). Intraocular pressure (mmHg) (n = 6) 0 2 4 6 8 10 12 14 pre POD 1 POD 2 POD 5 POD 7 treatment group control group Figure 9. Intraocular pressure after glaucoma filtration surgery. No significant differences in IOP were observed between the treatment and control groups at any time point. Discussion EVG staining revealed dense collagen deposition and scar formation in the conjunctiva and sclera of the control group (e,f), whereas collagen deposition was faintly stained, conjunctival tissue was loose, and the subconjunctival space was maintained in the treatment group (g,h). Magnification: ×4 for (a,c,e,g); ×10 for (b,d,f,h). Arrows: the conjunctiva, arrow heads: the sclera. effectively suppress postoperative fibrosis in most, but not all, cases. During the time course of postoperative fibrosis formation, the inhibition of multiple complex cellular interactions is essential and should be regulated. The administration of multiple inhibitors has the potential for synergic effects26. For example, we previously reported the effects of an mTOR inhibitor on HCF proliferation and suppression of fibrosis after filtration surgery in rabbits27. Meanwhile, a synergistic antiproliferative effect has been reported for the combined use of an mTOR inhibitor and MEK inhibitor in tumor cells26. Thus, the combination of several inhibitors may be useful for the safe and effective suppression of HCF proliferation. f pp p 5-FU inhibits the production of deoxythymidine monophosphate, which is essential for DNA synthesis, contributing to its cytotoxic effects28. MMC is an alkylating agent that crosslinks DNA, thereby directly caus- ing DNA damage and interfering with DNA synthesis and cell division1. In contrast, MEK inhibitors target the MEK/ERK pathway, which is a key signaling pathway involved in cell proliferation and survival11. The use of DNA-damaging reagents such as MMC induces activation of ERK29, which has also been suggested to contribute to cell cycle re-entry after DNA damage-induced cell cycle arrest30,31. Hence, MEK inhibitors are expected to be useful as adjuncts to MMC. Further research is needed to explore this topic.if j p p Although in our study IOP tended to be lower in the PD0325901 group, there were no significant differences between the treatment and control groups at any time points. In the rabbit strain used in this study, the baseline IOP was so low that any IOP-lowering effect via sustainable bleb formation with PD0325901 could not be veri- fied. In a future study, we plan to use rabbit or other animal models of ocular hypertension to assess this effect.h i y p ypf This study had several limitations. First, we did not compare PD0325901 with currently used perioperative agents such as MMC or 5-FU. Second, the postoperative observation period was short. Discussion Thus, MEK inhibitors, as with p38 inhibitors, may effectively regulate postoperative fibrosis after glaucoma filtration surgery through the modulation of scar formation via the MAPK cascade. In addition to the above-described mechanisms via signaling-pathway suppression, we found that PD0325901 attenuated fibrosis by arresting the HCF cell cycle. PD0325901 significantly blocked the cell cycle and prevented cell proliferation but did not cause HCF cell death. It did not show high cytotoxicity compared to other antifi- brotic drugs, such as MMC, implying that it can be used safely in combination with other antifibrotic drugs in glaucoma surgery. We also showed the effective suppression of fibrosis in a rabbit model of glaucoma filtration surgery via subconjunctival injection of PD0325901. Regarding the relationship between cell-cycle regulation and the suppression of fibrosis after filtration surgery, a recent study also showed that subconjunctival injection of adrenaline during trabeculectomy or other filtration surgeries effectively suppressed the contractile properties of HCFs and inhibited fibroblast proliferation by blocking key cell-cycle genes25. MEK inhibitors have been widely used for the treatment of neoplastic disorders, with the expectation that they suppress tumor cell growth14–16, particularly in combination with other therapeutic agents. As currently used drug therapies, 5-FU and MMC Scientific Reports \| (2024) 14:1871 \| https://doi.org/10.1038/s41598-024-52359-y www.nature.com/scientificreports/ 1,000 µm a 4x 10x treatment group control group treatment group control group EVG staining HE staining c e g b d f h 200 µm Figure 10. Histopathologic evaluation of blebs. HE staining of eye tissue revealed dense cells and fibrotic scarring in the control group (a,b) and relatively sparse cells and milder scar formation in the treatment group (c,d). EVG staining revealed dense collagen deposition and scar formation in the conjunctiva and sclera of the control group (e,f), whereas collagen deposition was faintly stained, conjunctival tissue was loose, and the subconjunctival space was maintained in the treatment group (g,h). Magnification: ×4 for (a,c,e,g); ×10 for (b,d,f,h). Arrows: the conjunctiva, arrow heads: the sclera. EVG staining treatment group staining control group EVG sta e a x control group 1,000 µm treatment group g treatment group treatment group control group g e g c a d f 0x b h 200 µm h f d b Figure 10. Histopathologic evaluation of blebs. HE staining of eye tissue revealed dense cells and fibrotic scarring in the control group (a,b) and relatively sparse cells and milder scar formation in the treatment group (c,d). Quantitative PCR Following serum starvation, 5 × 104 cells were replated in a 24-well plate in DMEM/F-12 medium with 10% FBS with or without PD0325901 (1 or 10 µM). After incubation for 48 h, the cells were lysed using ISOGEN (NIP- PON GENE Ltd., Tokyo, Japan), and isopropyl alcohol and chloroform were used to extract mRNA. qPCR was performed as described previously27. Primers were purchased from Hokkaido System Science (Hokkaido, Japan). The primer sequences in this study were referred to those used in previous reports39–42. The sequences of the PCR primers were: GAPDH: forward, 5′-GTCTCCTCTGACTTCAACAGCG-3′ and reverse, 5′-ACCACCCTGTTG CTGTAGCCA-3′; PCNA: forward, 5′-CCTGCTGGGATATTAGCTCCA-3′ and reverse, 5′-CAGCGGTAGGTG TCGAAGC-3′; p27: forward, 5′-TAATTGGGGCTCCGGCTAACT-3′ and reverse, 5′-TTGCAGGTCGCTTCC TTATTC-3′; and cyclin D1: forward, 5′-AGCTGTGCATCTACACCGA-3′ and reverse, 5′-GAAATCGTGCGG GGTCATTG-3′. The expression levels of GAPDH were used to normalize those of PCNA, p27, and cyclin D1. Cell culture and passage b d p g Primary HCFs were obtained from human donor eyes and characterized as described previously38. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F-12 containing 10% FBS and antibiotic–antimycotic solution (100×) (Sigma-Aldrich, St. Louis, MO, USA) in a CO2 incubator at 37 °C. In all experiments, only cells from passages 3–6 were used. To arrest the cell cycle, HCFs were grown to confluence and starved for 48 h in DMEM/F-12 without FBS. Western blottingt g After cell-cycle arrest, 1 × 106 cells were replated in DMEM with 10% FBS with or without PD0325901 (1 or 10 µM) or 0.1% DMSO in a 60-mm dish, and incubated for 48 h. The cells were obtained using radioimmuno- precipitation assay buffer (RIPA buffer; Thermo Fisher Scientific K.K., Kanagawa, Japan) and protease inhibitors (Roche Diagnostics, Basel, Switzerland). A BCA Protein Assay Kit (Thermo Fisher Scientific K.K.) was used to ascertain the protein quantities in the supernatant. Western blotting was performed as described previously27. The following primary antibodies were incubated at 4 °C overnight: anti-PCNA (1:1000; Sigma-Aldrich), anti- p27 (1:1000; Cell Signaling Technology), anti-cyclin D1 (1:1000; Cell Signaling Technology), and anti-β-tubulin (1:1000; FUJIFILM Wako Pure Chemical). Horseradish peroxidase-conjugated secondary antibody (H goat anti-rabbit or anti-mouse IgG; 1:5000; Thermo Fisher Scientific) was incubated for 1 h at room temperature. An ImageQuant LAS 4000 mini-instrument (GE Healthcare, Chicago, IL, USA) was used to detect protein bands. WB Stripping Solution (Nacalai Tesque) was used to strip membranes of antibodies. The density of protein bands was measured using ImageJ ver. 1.53 (National Institutes of Health, Bethesda, MD, USA). Protein expression was quantified against that of β-tubulin. Cell cycle analysis Aft After serum starvation, 3 × 106 cells were replated in DMEM with 10% FBS with or without PD0325901 (1 or 10 µM) or 0.1% DMSO in a 100-mm dish, and then incubated for 48 h. The cells were trypsinized, washed in PBS, and fixed with 70% ethanol at –20 ℃. Fixed cells were stained with 50 mg/mL propidium iodide. Flow cytometric analysis was performed using a BD FACSAria™ III cell sorter (BD Biosciences, San Jose, CA, USA). The percentage of HCF in each cell cycle was obtained by the Watson Pragmatic algorithm in FlowJo v10. Immunocytochemistryt y y After serum starvation, 5 × 104 cells were replated in 24-well plates with glass coverslips in DMEM/F-12 medium containing 10% FBS with 1 μM PD0325901 or 0.1% DMSO. Following 48 h incubation, the cells were fixed in ice-cold 4% paraformaldehyde for 15 min. Permeabilization was performed with 0.3% Triton X-100 for 5 min. For blocking, Blocking One Histo (Nacalai Tesque, Kyoto, Japan) was used for 30 min at room temperature. Immunocytochemistry was performed as described previously27. The cells were incubated at 4 °C overnight with primary antibodies. The primary antibodies were anti-PCNA (1:200; Sigma-Aldrich), anti-p27 (1:1000; Cell Signaling Technology, Danvers, MA, USA), and anti-cyclin D1 (1:1000; Cell Signaling Technology). The cells were washed with PBS and incubated with the secondary antibody (Alexa Fluor 488; 1:1000; Thermo Fisher Scientific, Waltham, MA, USA) for 60 min at room temperature. Nucleus staining was performed using 4 μg/mL of 4′,6-diamidine-2-phenylindole dihydrochloride solution (DAPI; FUJIFILM Wako Pure Chemical). Cell proliferation assayh The cytotoxicity and growth-inhibition effects of PD0325901 were assessed using a CCK-8 assay. After cell- cycle arrest, 1 × 104 cells were replated in a 96-well plate in DMEM/F-12 medium with 10% FBS with or without PD0325901 (100 nM–100 µM dissolved in 0.1% DMSO or 0.1% DMSO in phosphate-buffered saline (PBS) in a CO2 incubator at 37 °C. After incubation for 24, 48, 72, or 96 h, 10 µL CCK‑8 solution was added to each well and incubated at 37 °C for 2 h. The absorbance at 450 nm (OD450) was measured using a multimode plate reader (Multi Microplate Reader, ARVOX3; Parkin Elmer, MA, USA). Discussion Because bleb failure in rabbits has been reported to occur within 10–14 days postoperatively without the use of antimetabolites, the observation period was set at 7 days32. Long-term observation is necessary for comparing the effect of PD0325901 and antimetabolites on scar formation. In addition, a sustained drug-delivery system may be useful and should be explored in the future. Third, the side effects of PD0325901 have not been adequately observed. MEK inhibi- tors can cause retinopathy and uveitis33. Vitreous injection of PD0325901 at concentrations much higher than those tested in this study causes retinopathy in rabbits34. Since there is a report of MEK inhibitor hindering wound healing of corneal epithelial cells35, corneal damage may need to be carefully monitored when using MEK inhibitor intraoperatively. Although we observed no adverse effects, additional studies are needed to determine the safe dose of PD0325901. Forth, two bands of different molecular weights are observed in the p27 membrane in the supplement figure. p27 isoform has been reported36, and bands at similar molecular weight positions was found in a previous report37. The results of this study may be due to the p27 isoform.i h In summary, PD0325901 arrests the cell cycle of HCFs and may be beneficial in suppressing scar formation after glaucoma filtration surgery. https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| www.nature.com/scientificreports/ Statistical analysish y The results are presented as means ± standard deviations. Statistical analyses were performed using R ver. 4.2.1 (The R Foundation for Statistical Computing, Vienna, Austria). The results of CCK-8 were analyzed using Dun- nett’s test, while the results of the other experiments were analyzed using t-tests. P < 0.05 was considered statisti- cally significant. Histologic evaluationh g The rabbits were euthanized by intravenous injection of secobarbital (150 mg/kg body weight) at 7 days after surgery. Frozen sections were made as described previously43. After enucleation, the eyes were immersed in 4% paraformaldehyde and fixed for 24 h. Then the eyes were dehydrated in 30% sucrose for 24 h. Subsequently, the eyes were embedded in optimal cutting temperature compound (Tissue-Tek; Sakura Finetek, Tokyo, Japan). The sections were stained with HE and EVG. The evaluation of fibrosis was conducted in a masked manner. Migration assay fl Confluent HCFs in a 24-well plate were serum starved for 48 h. Each well was scratched crosswise with a pipette tip. After washing with PBS, the medium was changed to DMEM/F-12 with PD0325901 (1 or 10 µM) or 0.1% DMSO. Scratches were photographed at 0, 24, and 48 h after stimulation using a microscope (BZ-9000; Keyence https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| www.nature.com/scientificreports/ Corporation, Osaka, Japan). The shortest distance between the edges of cells on either side of the scratch was measured using ImageJ ver. 1.53. Corporation, Osaka, Japan). The shortest distance between the edges of cells on either side of the scratch was measured using ImageJ ver. 1.53. Animalsh This study was carried out in accordance with the Association for Research in Vision and Ophthalmology State- ment for the Use of Animals in Ophthalmic and Vision Research, and was approved by the Institutional Animal Research Committee of the University of Tokyo. The research has been reported in accordance with ARRIVE guidelines. The right eyes of 12 male Japanese white rabbits (14–19 weeks old, 1.50–1.99 kg; Kitayama Labes Co., Ltd., Nagano, Japan) were used in the current study. Before the experiments started, every rabbit was allowed 6 days to acclimate. The rabbits were housed in a conventional animal room with a 12-h light/dark cycle, and temperature and humidity controls of 22 °C ± 3 °C and 55% ± 10%, respectively. The rabbits were given tap water and food ad libitum. Evaluation of the effect of PD0325901 in rabbitsh The rabbits were randomly divided into a treatment group (n = 6) and a control group (n = 6). Photographs of the filtering bleb were captured using a digital camera. Slit-lamp and fundus examination were conducted. The height of the filtering bleb was measured using anterior segment optical coherence tomography (AS-OCT; RS-3000; NIDEK, Inc., Fremont, CA, USA). IOP was measured using TONOVET (M.E. Technica, Tokyo, Japan) under topical anesthesia. Three measurements were taken, from which the mean IOP was recorded. Immediately after surgery and after examination on postoperative days 1, 2, and 5, 0.2 mL 10 µM PD0325901 or 0.1% DMSO was injected under the bleb. Data availability y All data generated or analyzed in this study are available from the corresponding author upon reasonable request. Received: 9 November 2023; Accepted: 17 January 2024 Glaucoma filtration surgery in rabbits All surgical procedures were conducted by JL. Rabbits were anesthetized with an intramuscular injection of ketamine hydrochloride (5 mg/kg body weight) and xylazine hydrochloride (5 mg/kg body weight). A fornix- based flap of the conjunctiva was made in the superior region. After sclerostomy, a fistula leading to the anterior chamber was made. The conjunctival flap was sutured using 10-0 nylon. References 1. Al Habash, A., Aljasim, L. A., Owaidhah, O. & Edward, D. P. A review of the efficacy of mitomycin C in glaucoma filtration surgery Clin. Ophthalmol. 9, 1945–1951 (2015). p 2. Holló, G. Wound healing and glaucoma surgery: Modulating the scarring process with conventional antimetabolites and new molecules. Dev. Ophthalmol. 59, 80–89 (2017). 2. Holló, G. Wound healing and glaucoma surgery: Modulating the scarring process with conventional antimetabolites and new molecules. Dev. Ophthalmol. 59, 80–89 (2017). 3. Al-Haddad, C. E., Abdulaal, M., Al-Moujahed, A., Ervin, A. M. & Ismail, K. Fornix-based versus limbal-based conjunctival trab- eculectomy flaps for glaucoma: Findings from a Cochrane systematic review Am J Ophthalmol 174 33–41 (2017) 3. Al-Haddad, C. E., Abdulaal, M., Al-Moujahed, A., Ervin, A. M. & Ismail, K. Fornix-based versus limbal-based conjunctival trab- eculectomy flaps for glaucoma: Findings from a Cochrane systematic review. Am. J. Ophthalmol. 174, 33–41 (2017). 4 Schlunck G Meyer ter Vehn T Klink T & Grehn F Conjunctival fibrosis following filtering glaucoma surgery Exp Eye Res 3. Al-Haddad, C. E., Abdulaal, M., Al-Moujahed, A., Ervin, A. M. & Ismail, K. Fornix-based versus limbal-based conjunctival trab- eculectomy flaps for glaucoma: Findings from a Cochrane systematic review. Am. J. Ophthalmol. 174, 33–41 (2017). 4. Schlunck, G., Meyer-ter-Vehn, T., Klink, T. & Grehn, F. Conjunctival fibrosis following filtering glaucoma surgery. Exp. Eye Res 142, 76–82 (2016).f ecu ecto yl aps o g auco a: d gs o a Coc a e syste at c ev ew. m. J. Ophthalmol. 7 , 33 ( 0 7). 4. Schlunck, G., Meyer-ter-Vehn, T., Klink, T. & Grehn, F. Conjunctival fibrosis following filtering glaucoma surgery. Exp. Eye Res 142, 76–82 (2016).f 5. Cordeiro, M. F., Bhattacharya, S. S., Schultz, G. S. & Khaw, P. T. TGF-beta1, -beta2, and -beta3 in vitro: Biphasic effects on Tenon’s fibroblast contraction, proliferation, and migration. Invest. Ophthalmol. Vis. Sci. 41, 756–763 (2000). i 6. CAT-152 0102 Trabeculectomy Study Group. A phase III study of subconjunctival human anti-transforming growth factor beta(2) monoclonal antibody (CAT-152) to prevent scarring after first-time trabeculectomy. Ophthalmology 114, 1822–1830 (2007).i y p gti y p gy 7. Plikus, M. V. et al. Fibroblasts: Origins, definitions, and functions in health and disease. Cell. 184, 3852–3872 (2021). i 8. Gauglitz, G. G., Korting, H. C., Pavicic, T., Ruzicka, T. & Jeschke, M. G. Hypertrophic scarring and keloids: Pathomechanisms and current and emerging treatment strategies. Mol. Med. Acknowledgements g Supported by the Japan Society for the Promotion of Science (JSPS) grant no. 22K09807 (M.H.) and JSPS KAK- ENHI grant no. 23H03058 (M.A.). The English in this document has been checked by at least two professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/ hNrspV. www.nature.com/scientificreports/ Commun 8, 2149 (2017).iti g p y pi 8, 2149 (2017). 38 Aoyama-Araki Y et al Sphingosine-1-phosphate (S1P)-related response of human conjunctival fibroblasts after filtration surgery 8, 2149 (2017). 38. Aoyama-Araki, Y. et al. Sphingosine-1-phosphate (S1P)-related response of human conjunctival fibroblasts after filtration surgery f l I t Ophth l l Vi S i 58 2258 2265 (2017) 38. Aoyama-Araki, Y. et al. Sphingosine-1-phosphate (S1P)-related response of human conjunctival fibroblasts after filtration sur for glaucoma. Invest. Ophthalmol. Vis. Sci. 58, 2258–2265 (2017).i g p 9. Bery, N., Miller, A. & Rabbitts, T. A potent KRAS macromolecule degrader specifically targeting tumours with mutant KRAS. Nat Commun. 11, 3233 (2020). ( ) 0. Chen, K. et al. Human bone marrow-derived stem cell proliferation is inhibited by hepatocyte growth factor via increasing the cell cycle inhibitors p53, p21 and p27. Bone 49, 1194–1204 (2011).h y p p p 1. Liu, Y. et al. The natural polyphenol curcumin induces apoptosis by suppressing STAT3 signaling in esophageal squamous cel carcinoma. J. Exp. Clin. Cancer Res. 37, 303 (2018). J p ( ) 2. Zhao, B. et al. Periodontal ligament stem cell-derived small extracellular vesicles embedded in matrigel enhance bone repair through the adenosine receptor signaling pathway. Int. J. Nanomed. 17, 519–536 (2022). g g g y 3. Nemoto, H., Honjo, M., Okamoto, M., Sugimoto, K. & Aihara, M. Potential mechanisms of intraocular pressure reduction by micropulse transscleral cyclophotocoagulation in rabbit eyes. Invest. Ophthalmol. Vis. Sci. 63, 3 (2022). 43. Nemoto, H., Honjo, M., Okamoto, M., Sugimoto, K. & Aihara, M. Potential mechanisms of intraocular pressure reduction by micropulse transscleral cyclophotocoagulation in rabbit eyes. Invest. Ophthalmol. Vis. Sci. 63, 3 (2022). www.nature.com/scientificreports/ www.nature.com/scientificreports/ 3. Madala, S. K. et al. MEK-ERK pathway modulation ameliorates pulmonary fibrosis associated with epidermal growth factor receptor activation. Am. J. Respir. Cell Mol. Biol. 46, 380–388 (2012). p p 14. Kobeissi, I. & Tarhini, A. A. Systemic adjuvant therapy for high-risk cutaneous melanoma. Ther. Adv. Med. Oncol 17588359221134088 (2022). ( ) 15. Guaitoli, G. et al. Non-small-cell lung cancer: How to manage BRAF-mutated disease. Drugs Context 12, 2022-11-3 (2013). 6. Hamidi, S. et al. Review article: New treatments for advanced differentiated thyroid cancers and potential mechanisms of dru resistance. Front. Endocrinol. (Lausanne) 14, 1176731 (2023).ifi 17. Wen, J. et al. MEK inhibition prevents TGF-β1-induced myofibroblast transdifferentiation in human tenon fibroblasts. Mol. Med. Rep. 19, 468–476 (2019). p 8. Strzalka, W. & Ziemienowicz, A. Proliferating cell nuclear antigen (PCNA): A key factor in DNA replication and cell cycle regula- tion. Ann. Bot. 107, 1127–1140 (2011). 9. Yang, K., Hitomi, M. & Stacey, D. W. Variations in cyclin D1 levels through the cell cycle determine the proliferative fate of a cell Cell Div. 18(1), 32 (2006). 0. Satoh, T. & Kaida, D. Upregulation of p27 cyclin-dependent kinase inhibitor and a C-terminus truncated form of p27 contributes to G1 phase arrest. Sci. Rep. 6, 27829 (2016). G p S p , 9 ( ) 21. Roskoski, R. Jr. ERK1/2 MAP kinases: Structure, function, and regulation. Pharmacol. Res. 66, 105–143 (2012). p p 21. Roskoski, R. Jr. ERK1/2 MAP kinases: Structure, function, an s: Structure, function, and regulation. Pharmacol. Res. 66, 105–143 21. Roskoski, R. Jr. ERK1/2 MAP kinases: Structure, function, and regulation. Pharmacol. Res. 66, 105 143 (2012). 22. Kim, E. K. & Choi, E. J. Compromised MAPK signaling in human diseases: An update. Arch. Toxicol. 89, 867–882 (2015).ifi E. J. Compromised MAPK signaling in human diseases: An update. Arch. Toxicol. 89, 867–882 (2015). l h b b d d fib bl d ff h fib 22. Kim, E. K. & Choi, E. J. Compromised MAPK signaling in human diseases: An update. Arch. Toxicol. 89, 867–if p g g p 23. Meyer-Ter-Vehn, T. et al. p38 inhibitors prevent TGF-beta-induced myofibroblast transdifferentiation in human tenon fibroblasts Invest. Ophthalmol. Vis. Sci. 47, 1500–1509 (2006).ti p 24. Nassar, K. et al. A p38 MAPK inhibitor improves outcome after glaucoma filtration surgery. J. References 17, 113–125 (2011).f g g g 9. Lin, X., Yu, M., Wu, K., Yuan, H. & Zhong, H. Effects of pirfenidone on proliferation, migration, and collagen contraction of human Tenon’s fibroblasts in vitro. Invest. Ophthalmol. Vis. Sci. 50, 3763–3770 (2009).hfi i p 10. Fischer, C. V. et al. The antiproliferative effect of bevacizumab on human tenon fibroblasts is not mediated by vascular endothelial growth factor inhibition. Invest. Ophthalmol. Vis. Sci. 57, 4970–4977 (2016). i 10. Fischer, C. V. et al. The antiproliferative effect of bevacizumab on human tenon fibrobla growth factor inhibition. Invest. Ophthalmol. Vis. Sci. 57, 4970–4977 (2016). 11. Akinleye, A., Furqan, M., Mukhi, N., Ravella, P. & Liu, D. MEK and the inhibitors: From bench to bedside. J. Hematol. Oncol. 6, 27 (2013).hi ( ) 2. Andrikopoulos, P. et al. The MEK inhibitor trametinib ameliorates kidney fibrosis by suppressing ERK1/2 and mTORC1 signaling J. Am. Soc. Nephrol. 30, 33–49 (2019). https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \| Author contributions Conception, design, data analysis, and writing and editing of the manuscript were carried out by J.L., M.H., and M.A.; J.L. conducted in vitro and in vivo experiments; M.H. reviewed and edited the manuscript; and all authors participated in its review and approval. www.nature.com/scientificreports/ Glaucoma 24, 165–178 (2015 Th K X l Ad li bl k k ll l d hibi ifib i d i ff i l , p pt gi g y J , ( ) 25. Thong, K. X. et al. Adrenaline blocks key cell cycle genes and exhibits antifibrotic and vasoconstrictor effects in glaucoma surgery. Exp. Eye Res. 233, 109561 (2023).fi p y 26. Varin, J. et al. Dual mTORC1/2 inhibition induces anti-proliferative effect in NF1-associated plexiform neurofibroma and malignant peripheral nerve sheath tumor cells. Oncotarget 7, 35753–35767 (2016).fit p p g 7. Igarashi, N., Honjo, M. & Aihara, M. Effects of mammalian target of rapamycin inhibitors on fibrosis after trabeculectomy. Exp Eye Res. 203, 108421 (2021).i y ( ) 28. Fan Gaskin, J. C., Nguyen, D. Q., Soon Ang, G., O’Connor, J. & Crowston, J. G. Wound healing modulation in glaucoma filtration surgery-conventional practices and new perspectives: The role of antifibrotic agents (Part I). J. Curr. Glaucoma Pract. 8, 37–45 (2014). 29. Lee, S. W. et al. Sustained activation of Ras/Raf/mitogen-activated protein kinase cascade by the tumor suppressor p53. Proc. Natl. Acad. Sci. USA 97, 8302–8305 (2000). 30. Wei, F., Yan, J. & Tang, D. Extracellular signal-regulated kinases modulate DNA damage response—A contributing factor to using MEK inhibitors in cancer therapy. Curr. Med. Chem. 18, 5476–5482 (2011). py 31. Razidlo, G. L. et al. KSR1 is required for cell cycle reinitiation following DNA damage. J. Biol. Chem. 284, 6705–6715 (2009) , q y g g J , ( ) 32. van Mechelen, R. J. S. et al. Animal models and drug candidates for use in glaucoma filtration surgery: A systematic review. Exp. Eye Res. 217, 108972 (2022). y 33. Arora, S. et al. Retinal toxicities of systemic anticancer drugs. Surv. Ophthalmol. 67, 97–148 (2022). y g p 34. Huang, W. et al. PD0325901, a mitogen-activated protein kinase kinase inhibitor, produces ocular toxicity in a rabbit animal model of retinal vein occlusion. J. Ocul. Pharmacol. Ther. 25, 519–530 (2009). h 5. Byun, Y. S. et al. Diquafosol promotes corneal epithelial healing via intracellular calcium-mediated ERK activation. Exp. Eye Res 143, 89–97 (2016). 6. Hirano, K. et al. Cloning and functional expression of a degradation-resistant novel isoform of p27Kip1. Biochem. J. 353, 51–57 (2001).i 7. Fang, J. S. et al. Shear-induced Notch-Cx37-p27 axis arrests endothelial cell cycle to enable arterial specification. Nat. Competing interests h p g The authors declare no competing interests. Additional information upplementary Information The online version contains supplementary material available at https://doi.org/ 0.1038/s41598-024-52359-y. Additional information Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-024-52359-y. Correspondence and requests for materials should be addressed to M.H. Scientific Reports \| (2024) 14:1871 \| https://doi.org/10.1038/s41598-024-52359-y www.nature.com/scientificreports/ Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2024 https://doi.org/10.1038/s41598-024-52359-y Scientific Reports \| (2024) 14:1871 \|
https://openalex.org/W3197399325	https://thrombosisjournal.biomedcentral.com/track/pdf/10.1186/s12959-021-00315-5	English	null	Frequency of deep vein thrombosis at admission for acute stroke and associated factors: a cross-sectional study	Thrombosis journal	2,021	cc-by	4,445	Frequency of deep vein thrombosis at admission for acute stroke and associated factors: a cross-sectional study Takahisa Mori1* , Kazuhiro Yoshioka1 and Yuhei Tanno1,2 p @ 1Department of Stroke Treatment, Shonan Kamakura General Hospital, Okamoto 1370-1, 247-8533 Kamakura City, Kanagawa, Japan Full list of author information is available at the end of the article * Correspondence: morit-koc@umin.net p @ 1Department of Stroke Treatment, Shonan Kamakura General Hospital, Okamoto 1370-1, 247-8533 Kamakura City, Kanagawa, Japan Full list of author information is available at the end of the article Abstract Background: Intermittent pneumatic compression (IPC) is commonly used to prevent deep vein thrombosis (DVT) during hospitalization in patients with acute stroke. However, if DVT exists at admission, IPC of the legs with DVT may cause migration of the thrombi, resulting in pulmonary emboli. Whole-leg ultrasonography (wl-US) is a practical tool to detect DVT; however, wl-US is not always performed at admission in all stroke patients. This retrospective cross-sectional study aimed to investigate DVT frequency and identify significant factors indicating the presence of DVT at admission for acute stroke. Methods: We included patients admitted within 24 h of stroke onset between 2017 and 2019. Patients who did not undergo blood tests for D-dimer or wl-US within 72 h of arrival were excluded. We collected patient data on age; sex; anthropometric variables; presence of DVT on wl-US; and biomarkers such as D-dimer, high-sensitivity C- reactive protein (hs-CRP), and lipids. Results: Of 1129 acute stroke patients, 917 met our inclusion criteria. DVT was detected in 161 patients (17.6 %). Patients with DVT were older; were more likely to be female; had lower body weight; had higher D-dimer and hs- CRP levels; had lower albumin, hemoglobin, and triglyceride levels; and had higher National Institutes of Health Stroke Scale and pre-stroke modified Rankin scale scores than patients without DVT (n = 756). In addition, multiple logistic regression analysis showed that sex (female) and D-dimer levels (≥1.52 µg/mL) were independent significant factors for the presence of DVT. Among 161 patients with DVT, 78 (48.4 %) had both these significant factors. Among 756 patients without DVT, 602 (79.6 %) had no or one significant factor. The odds ratio of the presence of DVT in patients with both significant factors was 6.29, using patients without any significant factors as the group for comparison. Conclusions: The frequency of DVT is high in acute stroke patients at admission. Female sex and a high D-dimer level were independent significant factors for the presence of DVT. Therefore, in patients with these two significant factors at admission, IPC should be avoided or wl-US should be performed before IPC. Keywords: Deep vein thrombosis, D-dimer, Frequency, Intermittent pneumatic compression, Stroke, Ultrasonography Mori et al. Thrombosis Journal (2021) 19:62 https://doi.org/10.1186/s12959-021-00315-5 Mori et al. Thrombosis Journal (2021) 19:62 https://doi.org/10.1186/s12959-021-00315-5 Methods To investigate DVT frequency at admission and identify related factors, we included patients admitted within 24 h of stroke onset between March 2017 and March 2019. We excluded patients whose plasma D-dimer level was not examined within 24 h of arrival or in whom whole-leg US was not performed within 72 h of arrival. We collected patient data on age, sex, anthropometric variables, and US findings of DVT. We evaluated bio- markers such as hemoglobin (Hb), serum albumin (Alb), high-sensitivity C-reactive protein (hs-CRP), glucose, HbA1c, total cholesterol, high-density lipoprotein chol- esterol, triglycerides (TG), aspartate aminotransferase (AST), alanine aminotransferase (ALT), AST/ALT ratio, and plasma D-dimer. In addition, we evaluated the Na- tional Institutes of Health Stroke Scale (NIHSS) score [8] and pre-stroke modified Rankin scale (mRS) score at admission [9]. The low-density lipoprotein cholesterol concentration was calculated using the Friedewald for- mula: low-density lipoprotein cholesterol = total choles- terol – high-density lipoprotein cholesterol – TG/5. D- dimer levels were measured using latex turbidimetric immunoassay (LIAS AUTO D-Dimer NEO, Sysmex Co., Statistical analysis Non-normally distributed continuous variables are expressed as medians and interquartile ranges. We com- pared all possible pairs of variables with significant dif- ferences between patients with and without DVT. A dummy variable was used to represent categorical data, such as data on sex, and Spearman rank correlation co- efficient (rs) was calculated to measure the strength of the relationships. We defined 0 ≤\|rs\| < 0.1 as no correl- ation, 0.1 ≤\|rs\| < 0.4 as a weak correlation, 0.4 ≤\|rs\| < 0.6 as a moderate correlation, and 0.6 ≤\|rs\| as a strong correlation. Multicollinearity was defined as the presence of a moderate or strong correlation between variables. When variables were moderately or strongly correlated with one another, we adopted the variable with a larger chi-squared value. After excluding variables with multi- collinearity, we conducted a multiple logistic regression analysis to identify independent variables indicating the presence of DVT. We estimated the threshold values of independent variables indicating the presence of DVT using area under the curve values derived from the re- ceiver operating characteristic curves of the logistic re- gression model. Statistical significance was set at a P- value < 0.05. We used JMP software (version 16.0; SAS Institute, Cary, NC, USA) for all statistical analyses. © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Mori et al. Thrombosis Journal (2021) 19:62 Mori et al. Thrombosis Journal (2021) 19:62 Page 2 of 6 Page 2 of 6 Background Hyogo, Japan) [10]. DVT was diagnosed according to the findings of wl-US (Xario, Canon Medical Systems Co., Tochigi, Japan), performed by trained radiologists [11]. DVT was diagnosed based on the following US findings: presence of a non-compressible segment or flow impairment on color Doppler imaging [12]. Com- pression was performed at 2 cm intervals. g Venous thromboembolism (VT) is a common cause of death and morbidity in patients with acute stroke during hospitalization [1]. Anticoagulants reduce the frequency of pulmonary emboli due to VT; however, this benefit is offset by an increase in the frequency of extracranial hemorrhage [2]. In addition, anticoagulants cannot be administered to patients with hemorrhagic stroke. Therefore, intermittent pneumatic compression (IPC) is commonly used to reduce the risk of deep vein throm- bosis (DVT) during hospitalization [3]. A DVT fre- quency of 8.0 or 8.7 % at admission was reported from Polish institutions [4, 5]; however, the DVT frequency at admission has been unknown in Japan. If the DVT fre- quency is high in Japan and IPC of the legs with DVT is started in patients with DVT at the time of admission, IPC may cause migration or fragmentation of thrombi and lead to pulmonary embolism. During hospitalization, D-dimer levels are often elevated in patients with DVT [6]. Whole-leg ultrasonography (wl-US) is a practical tool to detect DVT in outpatients or inpatients [7]. wl- US or D-dimer measurement should always be per- formed for detecting DVT at admission for stroke; how- ever, wl-US is not routinely performed at admission in many facilities. Therefore, a practical index to estimate the presence of DVT at stroke admission is necessary. Our retrospective cross-sectional study aimed to investi- gate DVT frequency at admission and identify significant factors specific to the presence of DVT at admission for acute stroke. Results q g Abbreviations: ALT alanine aminotransferase, AST aspartate aminotransferase, BW body weight, HDL high-density lipoprotein cholesterol, hs-CRP high- sensitivity C-reactive protein, DVT deep vein thrombosis, LDL low-density lipoprotein cholesterol, mRS modified Rankin scale score, NGSP National Glycohemoglobin Standardization Program, N number, NIHSS National Institutes of Health Stroke Scale, TC total cholesterol, TG triglyceride p y The incidence of DVT has been reported to be ap- proximately 50 % within 2 weeks in the absence of hep- arin prophylaxis in patients with acute hemiplegic stroke [13]. Patients with proximal subclinical DVT had a 15 % risk of fatal pulmonary embolism [13, 14]. Untreated below-knee DVT is associated with a 20 % risk of prox- imal extension [13], and the pulmonary embolism rate is reportedly 6.1 % in trauma patients with below-knee DVT [15]. On admission to the stroke rehabilitation unit, the prevalence of DVT in patients with stroke ranges from 12 to 40 % [16]. Therefore, the onset of DVT during hospitalization in primary stroke centers must be prevented. When DVT is not detected at admis- sion, IPC can be used safely. If DVT is detected at ad- mission in patients with ischemic stroke, anticoagulants may be started soon. The disuse of IPC may not induce thrombi fragmentation in patients with DVT at admis- sion. Early anticoagulant therapy may protect against thromboembolism caused by DVT present at admission, and early IPC or anticoagulant therapy may prevent DVT development after admission. Overall, the fre- quency of symptomatic or critical DVT may decrease had both these significant factors. Among 756 patients without DVT, 602 (79.6 %) had no or one significant fac- tor. The odds of the presence of DVT was 0.506 in pa- tients with both significant factors, 0.182 in patients with one significant factor, and 0.081 in patients without any significant factors. On using patients without any signifi- cant factors as the group for comparison, the odds ratio of the presence of DVT was 6.29 in patients with both significant factors (Table 5). had both these significant factors. Among 756 patients without DVT, 602 (79.6 %) had no or one significant fac- tor. The odds of the presence of DVT was 0.506 in pa- tients with both significant factors, 0.182 in patients with one significant factor, and 0.081 in patients without any significant factors. Results On using patients without any signifi- cant factors as the group for comparison, the odds ratio of the presence of DVT was 6.29 in patients with both significant factors (Table 5). Results During the study period, 1129 acute stroke patients were admitted, and 917 met our inclusion criteria. Patient characteristics are summarized in Table 1. DVT was de- tected in 161 (17.6 %) of 917 patients at admission. Pa- tients with DVT (n = 161) were older; were more likely to be female; had lower body weight; had higher plasma D-dimer and hs-CRP levels; had lower Hb, serum Alb, TG, and ALT levels; had a higher AST/ALT ratio; and had higher NIHSS and pre-stroke mRS scores than pa- tients without DVT (n = 756) (Table 2). After excluding variables with significant differences between the two groups and variables with multicollinearity (Additional file 1), multiple logistic regression analysis showed that sex and D-dimer levels were independent variables for the presence of DVT (Table 3). Receiver operating char- acteristic curves demonstrated that female sex and a D- dimer level of ≥1.52 µg/mL were independent factors for the presence of DVT at the time of stroke admission (Table 4). Among 161 patients with DVT, 78 (48.4 %) Page 3 of 6 Mori et al. Thrombosis Journal (2021) 19:62 Page 3 of 6 Table 1 Patient characteristics Variables Values N 917 DVT, n (%) 161 (17.6 %) Ischemic stroke, n (%) 734 (80.0 %) Female sex, n (%) 449 (49.0 %) Age, years 80 (71, 86) BMI, kg/m2 22.0 (19.5, 24.4) BW, kg 55 (47, 65) Hb, g/dL 13.3 (12, 14.6) Plt, /µL 20.8 (17.1, 25.5) Alb, mg/dL 3.9 (3.6, 4.2) AST, U/L 23 (19, 29) ALT, U/L 16 (12, 23) AST/ALT ratio 1.4 (1.1, 1.8) Glucose, mg/dL 122 (105, 151) HbA1c, % (NGSP) 5.8 (5.5, 6.4) TC, mg/dL 196 (168, 225) LDL, mg/dL 110 (85, 134) HDL, mg/dL 56.2 (46.1, 68.4) TG, mg/dL 93 (65, 136) hs-CRP, mg/dL 0.14 (0.05, 0.53) D-dimer, µg/mL 1.4 (0.7, 3.1) NIHSS at admission 5 (2, 16) Pre-stroke mRS 0 (0, 3) All values except for categorical data are represented as median (interquartile range) Abbreviations: ALT alanine aminotransferase, AST aspartate aminotransferase, BW body weight, HDL high-density lipoprotein cholesterol, hs-CRP high- sensitivity C-reactive protein, DVT deep vein thrombosis, LDL low-density lipoprotein cholesterol, mRS modified Rankin scale score, NGSP National Glycohemoglobin Standardization Program, N number, NIHSS National Institutes of Health Stroke Scale, TC total cholesterol, TG triglyceride both these significant factors at admission, IPC should be avoided or wl-US should be performed before IPC. Results wl-US has been shown to have high sensitivity (94.0 %) and specificity (97.3 %) for detecting DVT [7]. However, wl-US is not always performed for stroke patients in most facilities. In contrast, D-dimer levels can be mea- sured easily in any institution. A cut-off value of 0.5 µg/ mL for D-dimer levels showed a sensitivity of 82.9 % and specificity of 32.7 % for detecting DVT in patients during hospitalization [6]. Furthermore, our results demon- strated that D-dimer was a significant factor for the presence of DVT at stroke admission. Therefore, D- dimer levels should be examined routinely at stroke admission. both these significant factors at admission, IPC should be avoided or wl-US should be performed before IPC. wl-US has been shown to have high sensitivity (94.0 %) and specificity (97.3 %) for detecting DVT [7]. However, wl-US is not always performed for stroke patients in most facilities. In contrast, D-dimer levels can be mea- sured easily in any institution. A cut-off value of 0.5 µg/ mL for D-dimer levels showed a sensitivity of 82.9 % and specificity of 32.7 % for detecting DVT in patients during hospitalization [6]. Furthermore, our results demon- strated that D-dimer was a significant factor for the presence of DVT at stroke admission. Therefore, D- dimer levels should be examined routinely at stroke admission. According to previous studies, DVT was found on day 3 after stroke onset in 8.0 % of acute stroke patients and within 7 days of stroke onset in 8.7 % of acute stroke pa- tients in Polish institutions [4, 5]. In comparison, the DVT frequency was high in our study at 17.5 %; this may be because our patients were older than those in previous studies, and US was performed within 72 h of arrival in our patients, compared to the performance of US within 7 days of stroke onset in a previous study [4, 5]. Female sex, elevated CRP levels, and pre-stroke dis- ability were risk factors for DVT within 7 days of stroke onset, and elevated CRP and pre-stroke disability were independent risk factors for the presence of DVT [5]. However, D-dimer was not examined in the previous study [5]. Elevated CRP levels and pre-stroke disability were also found to be significant factors for the presence of DVT in our patients (Table 2). However, they were not independent because of multicollinearity. Discussion Thrombosis Journal (2021) 19:62 Page 4 of 6 Table 2 Comparison of variables between the two groups Patients with DVT Patients without DVT Chi-square value P-value N 161 756 Ischemic stroke, n (%) 132 (82.0 %) 602 (79.6 %) 0.5 0.4924 Female sex, n (%) 106 (65.8 %) 343 (45.4 %) 22.4 < 0.0001 Age, years 82 (77, 89) 79 (71, 86) 21.0 < 0.0001 BMI, kg/m2 21.5 (19.1, 23.7) 22.1 (19.6, 24.5) 3.6 0.0569 BW, kg 52 (44, 60) 56 (47, 65) 11.5 0.0007 Hb, g/dL 12.6 (11.5, 13.8) 13.4 (12.3, 13.9) 21.3 < 0.0001 Plt, /µL 20.5 (17.1, 25.5) 20.9(17.2, 25.7) 0.2 0.6772 Alb, mg/dL 3.7 (3.4, 4.1) 4.0 (3.7, 4.2) 25.7 < 0.0001 AST, U/L 22 (19, 31) 23 (19, 28) 0.0 0.9744 ALT, U/L 15 (12, 21) 17 (12, 23) 6.3 0.0123 AST/ALT ratio 1.53 (1.22, 1.87) 1.38 (1.13, 1.73) 10.7 0.0011 Glucose, mg/dL 122 (106, 149) 123 (105, 153) 0.4 0.5314 HbA1c, % (NGSP) 5.8 (5.5, 6.3) 5.8 (5.5, 6.4) 0.2 0.6810 TC, mg/dL 194 (159, 226) 197 (170, 224) 0.3 0.5873 LDL, mg/dL 110 (82, 131) 111 (87, 135) 1.2 0.2794 HDL, mgl/dL 56.7 (44.3, 71.7) 55.8 (46.2, 68.2) 0.1 0.7125 TG, mg/dL 82 (61, 123) 95 (66, 141) 4.8 0.0279 hs-CRP, mg/dL 0.17 (0.07, 0.77) 0.13 (0.05, 0.47) 6.6 0.0103 D-dimer, µg/mL 2.7 (1.3, 6.0) 1.2 (0.6, 2.6) 60.2 < 0.0001 NIHSS at admission 8 (3,18) 5 (2,15) 12.3 0.0004 Pre-stroke mRS 2 (0, 3.5) 0 (0, 3) 16.1 < 0.0001 All values except for categorical data are represented as median (interquartile range) Alb albumin, ALT alanine aminotransferase, AST aspartate aminotransferase, BMI body mass index, BW body weight, Hb hemoglobin, HDL high-density lipoprotein cholesterol, hs-CRP high-sensitivity C-reactive protein, DVT deep vein thrombosis, LDL low-density lipoprotein cholesterol, mRS modified Rankin scale score, N number, NGSP National Glycohemoglobin Standardization Program, NIHSS National Institutes of Health Stroke Scale, P probability, Plt platelet, TC total cholesterol, TG triglyceride Table 2 Comparison of variables between the two groups All values except for categorical data are represented as median (interquartile range) Alb albumin, ALT alanine aminotransferase, AST aspartate aminotransferase, BMI body mass index, BW body weight, Hb hemoglobin, HDL high-density lipoprotein cholesterol, hs-CRP high-sensitivity C-reactive protein, DVT deep vein thrombosis, LDL low-density lipoprotein cholesterol, mRS modified Rankin scale score, N number, NGSP National Glycohemoglobin Standardization Program, NIHSS National Institutes of Health Stroke Scale, P probability, Plt platelet, TC total cholesterol, TG triglyceride during hospitalization in primary stroke centers. Discussion Plasma D-dimer levels must be measured at the time of stroke, and patients with both significant factors, i.e., female sex and a D-dimer level ≥1.52 µg/mL, should immediately undergo wl-US, if possible. Discussion Our results demonstrated that DVT was present in 17.6 % of acute stroke patients at admission, and female sex and a high D-dimer level were significant factors in- dicating the presence of DVT at admission. The odds ra- tio of the presence of DVT in patients with both significant factors was 6.29. Therefore, in patients with Mori et al. References 1. House of Commons Health Committee. The prevention of venous thromboembolism in hospitalised patients. Second Report of Session 2004– 05, HC 99. London: Stationery Office; 2005. 2. Sandercock PA, Counsell C, Kamal AK. Anticoagulants for acute ischaemic stroke. Cochrane Database Syst Rev. 2008:CD000024. https://doi.org/10.1 002/14651858.CD000024.pub3. Limitations Study participant enrollment was based on an opt-out model, which the ethical committee permitted. presence of DVT. A prospective study including US and plasma D-dimer examination is required to determine the frequency of DVT at stroke admission and signifi- cant associated factors. Consent for publication Not applicable. Availability of data and materials 10. Suzuki H, Masaki Y, Okubo M, Yotsui S, Ogura M, Imanishi K, et al. A comparative study of Sysmex Latex Test BL-2 P-FDP and LIAS AUTO D- Dimer NEO with similar assay reagents of two other companies on the fully automated blood coagulation analyzer CS-5100. Sysmex J Int. 2014;24. https://www.sysmex.co.jp/en/products_solutions/library/journal/vol24_no1/ vol24_1_10.pdf . The datasets analyzed during the current study are available from the corresponding author on reasonable request. The datasets analyzed during the current study are available from the corresponding author on reasonable request. Acknowledgements W ld l k h 5. Bembenek JP, Karlinski M, Kobayashi A, Czlonkowska A. Deep venous thrombosis in acute stroke patients. Clin Appl Thromb Hemost. 2012;18: 258–64. We would like to thank Nozomi Chiba, BA., for her secretarial assistance and the ultrasonographers at our stroke center for specialized support. 6. Canan A, Halicioglu SS, Gurel S. Mean platelet volume and D-dimer in patients with suspected deep venous thrombosis. J Thromb Thrombolysis. 2012;34:283–7. Abbreviations Alb: Albumin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; DVT: Deep vein thrombosis; hs-CRP: High-sensitivity C- reactive protein; IPC: Intermittent pneumatic compression; mRS: Modified Rankin scale; NIHSS: National Institutes of Health Stroke Scale; US: Ultrasonography; VT: Venous thromboembolism; wl-US: Whole-leg ultrasonography Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 9. van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke. 1988;19:604–7. Additional file 1. 4. Bembenek J, Karlinski M, Kobayashi A, Czlonkowska A. Early stroke-related deep venous thrombosis: risk factors and influence on outcome. J Thromb Thrombolysis. 2011;32:96–102. Author details 1 1Department of Stroke Treatment, Shonan Kamakura General Hospital, Okamoto 1370-1, 247-8533 Kamakura City, Kanagawa, Japan. 2Department of Neurology, Nakatsugawa Municipal General Hospital, Komaba 1522-1, Gifu 508-8502 Nakatsugawa City, Japan. Received: 8 July 2021 Accepted: 22 August 2021 Received: 8 July 2021 Accepted: 22 August 2021 Authors’ contributions Conceptualization: T.M. Methodology: T.M. Validation: T.M. Formal analysis: T.M. Investigation: T.M., K.Y., Y.T. Resources: T.M., K.Y., Y.T. Data curation: T.M., K.Y., Y.T. Writing–original draft preparation: T.M. Writing–review and editing: T.M. Visualization: T.M. All authors have read and agreed to the published version of the manuscript. 7. Bhatt M, Braun C, Patel P, Patel P, Begum H, Wiercioch W, et al. Diagnosis of deep vein thrombosis of the lower extremity: a systematic review and meta-analysis of test accuracy. Blood Adv. 2020;4:1250–64. 8. Lyden P, Brott T, Tilley B, Welch KM, Mascha EJ, Levine S, et al. Improved reliability of the NIH Stroke Scale using video training. NINDS TPA Stroke Study Group. Stroke. 1994;25:2220–6. Competing interests The frequency of DVT at admission in acute stroke pa- tients was high at 17.6 % in our institution. Female sex and high D-dimer levels were significant factors for the presence of DVT. Therefore, in patients with these two significant factors at admission, IPC should be avoided or wl-US should be performed before IPC. p g The authors declare that they have no competing interests. Declarations Ethics approval and consent to participate Supplementary Information 3. CLOTS (Clots in Legs Or sTockings after Stroke) Trials Collaboration. Effectiveness of intermittent pneumatic compression in reduction of risk of deep vein thrombosis in patients who have had a stroke (CLOTS 3): a multicentre randomised controlled trial. Lancet. 2013;382:516–24. pp y The online version contains supplementary material available at https://doi. org/10.1186/s12959-021-00315-5. Limitations Our study had several limitations. First, the sample size was small, and the study had a retrospective, cross- sectional design. Second, although US has a sensitivity of 94 % for detecting DVT, it cannot always detect DVT. Third, because most of the patients were Japanese, generalization of the study outcomes to non-Japanese populations may not be possible. There might be racial differences in the association between DVT-related fac- tors and threshold values of factors associated with the Table 3 Multiple logistic regression for deep vein thrombosis presence at the admission of stroke using receiver operating characteristics curves Odds ratio P-value AUC BIC < 0.0001 0.687 774 Sex 2.04 (1.40–3.00) 0.0002 D-dimer 1.05 (1.02–1.08) 0.0003 TG 1.00 (0.99–1.00) 0.0560 hs-CRP 1.08 (0.99–1.17) 0.0680 ALT 0.99 (0.98–1.00) 0.2479 NIHSS 1.01 (0.99–1.03) 0.2698 ALT alanine aminotransferase, AUC area under the curve, BIC Bayesian information criterion, hs-CRP high-sensitivity C-reactive protein, NIHSS National Institutes of Health Stroke Scale at admission, P probability, TG triglyceride Table 3 Multiple logistic regression for deep vein thrombosis presence at the admission of stroke using receiver operating characteristics curves Table 4 Threshold values for DVT presence using receiver operating characteristics curves from logistic regression analysis Sens (%) Spec (%) Odds ratio P-value AUC BIC Sex (1 vs. < 1) 65.8 54.5 2.31 (1.62–3.31) < 0.0001 0.602 843 D-dimer (≥1.52 vs. < 1.52) μg/mL 70.0 59.6 1.06 (1.03–1.09) < 0.0001 0.695 836 AUC area under the curve, BIC Bayesian information criterion, DVT deep vein thrombosis, P probability, Sens sensitivity, Spec specificity, TG triglyceride Table 4 Threshold values for DVT presence using receiver operating characteristics curves from logistic regression analysi Mori et al. Thrombosis Journal (2021) 19:62 Page 5 of 6 Table 5 Cross tabulation table between patients with significant factors, DVT, Odds, and Odds ratio DVT presence DVT absence Odds Odds ratio N 161 756 Patients with the two significant factors 78 154 0.5065 6.29 Patients with one significant factor 62 341 0.1818 2.26 Patients without any significant factors 21 261 0.0805 The two significant factors are female sex and D-dimer ≥1.52 µg/mL DVT deep vein thrombosis, N number Ethics Committee approved this retrospective study with a waiver for informed consent (TGE01723-024). Study participant enrollment was based on an opt-out model, which the ethical committee permitted. Ethics Committee approved this retrospective study with a waiver for informed consent (TGE01723-024). Ethics approval and consent to participate 11. Konstantinides SV, Torbicki A, Agnelli G, Danchin N, Fitzmaurice D, Galie N, et al. 2014 ESC guidelines on the diagnosis and management of acute pulmonary embolism. Eur Heart J. 2014;35:3033–69, 3069a–3069k. All procedures were performed in accordance with the ethical standards of the institution and the 1964 Helsinki Declaration. The Tokushukai Group Page 6 of 6 Mori et al. Thrombosis Journal (2021) 19:62 Mori et al. Thrombosis Journal (2021) 19:62 12. Needleman L, Cronan JJ, Lilly MP, Merli GJ, Adhikari S, Hertzberg BS, et al. Ultrasound for lower extremity deep venous thrombosis: multidisciplinary recommendations from the Society of Radiologists in Ultrasound Consensus Conference. Circulation. 2018;137:1505-15. 12. Needleman L, Cronan JJ, Lilly MP, Merli GJ, Adhikari S, Hertzberg BS, et al. Ultrasound for lower extremity deep venous thrombosis: multidisciplinary recommendations from the Society of Radiologists in Ultrasound Consensus Conference. Circulation. 2018;137:1505-15. 13. Kelly J, Rudd A, Lewis R, Hunt BJ. Venous thromboembolism after acute stroke. Stroke. 2001;32:262–7. 13. Kelly J, Rudd A, Lewis R, Hunt BJ. Venous thromboembolism after acute stroke. Stroke. 2001;32:262–7. 14. Khan MT, Ikram A, Saeed O, Afridi T, Sila CA, Smith MS, et al. Deep vein thrombosis in acute stroke - a systemic review of the literature. Cureus. 2017;9:e1982. 14. Khan MT, Ikram A, Saeed O, Afridi T, Sila CA, Smith MS, et al. Deep vein thrombosis in acute stroke - a systemic review of the literature. Cureus. 2017;9:e1982. 15. Olson EJ, Zander AL, Van Gent JM, Shackford SR, Badiee J, Sise CB, et al. Below-knee deep vein thrombosis: an opportunity to prevent pulmonary embolism? J Trauma Acute Care Surg. 2014;77:459–63. 15. Olson EJ, Zander AL, Van Gent JM, Shackford SR, Badiee J, Sise CB, et al. Below-knee deep vein thrombosis: an opportunity to prevent pulmonary embolism? J Trauma Acute Care Surg. 2014;77:459–63. 16. Wilson RD, Murray PK. Cost-effectiveness of screening for deep vein thrombosis by ultrasound at admission to stroke rehabilitation. Arch Phys Med Rehabil. 2005;86:1941–8. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W4313271772	https://www.frontiersin.org/articles/10.3389/fimmu.2022.1045329/pdf	English	null	Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases	Frontiers in immunology	2,022	cc-by	10,061	Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases doi: 10.3389/fimmu.2022.1045329 OPEN ACCESS EDITED BY Lu Yang, Guangdong Provincial People’s Hospital, China REVIEWED BY Wangjun Liao, Southern Medical University, China Jian Li, Beijing Cancer Hospital, China Qi Wang, The Afﬁliated Hospital of Qingdao University, China CORRESPONDENCE Yuhong Li liyh@sysucc.org.cn Minshan Chen chenmsh@sysucc.org.cn Ziming Du duzm1@sysucc.org.cn †These authors have contributed equally to this work and share ﬁrst authorship SPECIALTY SECTION This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology RECEIVED 15 September 2022 ACCEPTED 14 November 2022 PUBLISHED 16 December 2022 CITATION Hu M, Chen Z, Hu D, Xi S, Wang D, Zhang X, Fong WP, Wen L, Cai Y, Yuan Y, Li B, Wu X, Lu Z, Chen G, Li L, Ding P, Pan Z, Wan D, Du Z, Chen M and Li Y (2022) Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases. Front. Immunol. 13:1045329. doi: 10.3389/fimmu.2022.1045329 Mingtao Hu 1,2†, Zhigang Chen 1,2†, Dandan Hu 1,3†, Shaoyan Xi 1,4†, Deshen Wang 1,2, Xiaolong Zhang 1, William Pat Fong 1,2, Lei Wen 1,2, Yanyu Cai 1,2, Yunfei Yuan 1,3, Binkui Li 1,3, Xiaojun Wu 1,5, Zhenhai Lu 1,5, Gong Chen 1,5, Liren Li 1,5, Peirong Ding 1,5, Zhizhong Pan 1,5, Desen Wan 1,5, Ziming Du 1,6, Minshan Chen 1,3* and Yuhong Li 1,2* 1State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-Sen University Cancer Center, Guangzhou, China, 2Department of Medical Oncology, Sun Yat-Sen University Cancer Center, Guangzhou, China, 3Department of Hepatobiliary Surgery, Sun Yat-Sen University Cancer Center, Guangzhou, China, 4Department of Pathology, Sun Yat-Sen University Cancer Center, Guangzhou, China, 5Department of Colorectal Surgery, Sun Yat- Sen University Cancer Center, Guangzhou, China, 6Department of Molecular Diagnostics, Sun Yat- Sen University Cancer Center, Guangzhou, China Background: Histopathological growth patterns (HGPs) have shown important prognostic values for patients with colorectal cancer liver metastases, but the potential molecular mechanisms remain largely unknown. Methods: We performed an exploratory analysis by conducting the RNA sequencing of primary colorectal lesions, colorectal liver metastatic lesions and normal liver tissues. Findings: We found that desmoplastic HGPs of the metastatic lesions were signiﬁcantly enriched in EMT, angiogenesis, stroma, and immune signaling pathways, while replacement HGPs were enriched in metabolism, cell cycle, and DNA damage repair pathways. TYPE Original Research PUBLISHED 16 December 2022 DOI 10.3389/fimmu.2022.1045329 TYPE Original Research PUBLISHED 16 December 2022 DOI 10.3389/fimmu.2022.1045329 Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases OPEN ACCESS EDITED BY Lu Yang, Guangdong Provincial People’s Hospital, China REVIEWED BY Wangjun Liao, Southern Medical University, China Jian Li, Beijing Cancer Hospital, China Qi Wang, The Afﬁliated Hospital of Qingdao University, China CORRESPONDENCE Yuhong Li liyh@sysucc.org.cn Minshan Chen chenmsh@sysucc.org.cn Ziming Du duzm1@sysucc.org.cn †These authors have contributed equally to this work and share ﬁrst authorship SPECIALTY SECTION This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology RECEIVED 15 September 2022 ACCEPTED 14 November 2022 PUBLISHED 16 December 2022 CITATION Hu M, Chen Z, Hu D, Xi S, Wang D, Zhang X, Fong WP, Wen L, Cai Y, Yuan Y, Li B, Wu X, Lu Z, Chen G, Li L, Ding P, Pan Z, Wan D, Du Z, Chen M and Li Y (2022) Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases. Front. Immunol. 13:1045329. doi: 10.3389/fimmu.2022.1045329 OPEN ACCESS EDITED BY Lu Yang, Guangdong Provincial People’s Hospital, China REVIEWED BY Wangjun Liao, Southern Medical University, China Jian Li, Beijing Cancer Hospital, China Qi Wang, The Afﬁliated Hospital of Qingdao University, China CORRESPONDENCE Yuhong Li liyh@sysucc.org.cn Minshan Chen chenmsh@sysucc.org.cn Ziming Du duzm1@sysucc.org.cn †These authors have contributed equally to this work and share ﬁrst authorship SPECIALTY SECTION This article was submitted to Cancer Immunity and Immunotherapy, a section of the journal Frontiers in Immunology RECEIVED 15 September 2022 ACCEPTED 14 November 2022 PUBLISHED 16 December 2022 CITATION Hu M, Chen Z, Hu D, Xi S, Wang D, Zhang X, Fong WP, Wen L, Cai Y, Yuan Y, Li B, Wu X, Lu Z, Chen G, Li L, Ding P, Pan Z, Wan D, Du Z, Chen M and Li Y (2022) Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases. Front. Immunol. 13:1045329. CITATION Hu M, Chen Z, Hu D, Xi S, Wang D, Zhang X, Fong WP, Wen L, Cai Y, Yuan Y, Li B, Wu X, Lu Z, Chen G, Li L, Ding P, Pan Z, Wan D, Du Z, Chen M and Li Y (2022) Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases. Front. Immunol. 13:1045329. doi: 10.3389/fimmu.2022.1045329 KEYWORDS colorectal cancer liver metastases, histopathological growth pattern, RNA sequencing, tumor microenvironment, soil and seeds Sample collection This study was approved by Sun Yat-sen University cancer center’s medical ethics committee and was performed according to the Helsinki Declaration of the World Medical Association. We collected fresh-frozen samples from a clinical cohort of patients who underwent simultaneous R0 resection for simultaneous colorectal liver metastases at the Sun Yat-sen University cancer center from July 2019 to September 2020. Histologically conﬁrmed primary colorectal lesions (C) and metastatic liver lesions (L), as well as normal liver tissues as far away from tumors as possible (Ln), were collected (Figure 1A). Clinical characteristics, such as age, gender, tumour location and TNM stage, were collected retrospectively from medical records. Patients were followed up until July 2021. Relapse-free survival (RFS) was deﬁned as the time from curative operation to the ﬁrst relapse or last follow-up. Accumulating studies have explored the pathological morphology heterogeneity between the two subtypes. Replacement and desmoplastic lesions have been shown to possess different forms of tumor vascularization (12). Replacement HGP can utilize vessel co-option rather than angiogenesis to create their blood supply, leading to resistance to anti-angiogenic therapy (13). The two subtypes also possess different tumor immune microenvironments. Our previous study showed that dHGP was correlated to a high immunoscore in pathological tissues and predicted a favorable prognosis independent of the immunoscore (14). However, the Introduction potential molecular mechanisms underlying HGPs remain largely unknown. Thus, this study aimed to compare the different molecular signatures between the dHGP and rHGP subtypes at the RNA level. Furthermore, we explored how the distinctions between the “seed” (the primary tumor) and “soil” (the normal liver tissue) of those two HGP subtypes might contribute to the different HGP formations. Approximately 50% of patients with colorectal cancer develop liver metastases during the course of their disease, making it the leading cause of death in the majority of patients (1). Hepatectomy is the main curative treatment option for colorectal cancer liver metastasis (CRLM). However, over 50% of patients suffer from recurrence within 2 years (2). In recent years, histopathological growth patterns (HGPs) have attracted extensive attention as prognostic factors in lesions with positive resection margins, recurrent disease and overall survival following complete resection (3–5). The HGPs are deﬁned due to the interface between the metastatic cancer cells and the surrounding normal liver cells examinating on standard hematoxylin & eosin-stained (H&E) tissue sections under the light microscopy, which are easily obtainable, inexpensive and reproducible (6). Actually, they are comprehensive parameters of tumor and surrounding tumor microenvironment (TME), including tumor invasion ability, angiogenesis, paracrine and autocrine effects of growth factors, ﬁbrosis, and tumor immune microenvironment. These particular patterns have been previously described in liver metastases originating from colorectal cancer, gastric cancer, pancreatic cancer, uveal melanoma and cutaneous melanoma (7–11). In the CRLM, the desmoplastic HGP (dHGP, a rim of desmoplastic stroma separating the tumor cells and the normal liver parenchyma) and replacement HGP (rHGP, the tumor cells invade along the liver cell plates and replace normal hepatocytes) are the two most common HGP subtypes. Delineating the molecular landscape of different histopathological growth patterns in colorectal cancer liver metastases With the exception of immune-related genes, the differentially expressed genes of the two HGPs from colorectal liver metastases were mostly inherited from the primary tumor. Moreover, normal liver tissue in the desmoplastic HGP subgroup was markedly enriched in the ﬁbrinous inﬂammation pathway. Conclusions: We surmised that HGPs are observable morphological changes resulting from the regulation of molecular expressions, which is the Frontiers in Immunology frontiersin.org 01 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 combined effect of the heterogeneity and remodeling of primary tumors seeds and liver soils. KEYWORDS Pathological assessment of HGPs Each available H&E-stained and formalin-ﬁxed parafﬁn- embedded (FFPE) liver lesion specimens were assessed in accordance with the international consensus guidelines by two independent pathologists using a light microscope (6). Examples for displaying the dHGP, rHGP, or mixed lesions are shown in Figure 1B. In this study, the liver lesion was categorized as dHGP or rHGP lesion if only dHGP or rHGP was present at the Frontiers in Immunology frontiersin.org 02 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 10.3389/ﬁmmu.2022.1045329 interface. The liver lesion with more than one HGP at the interface was categorized as an impure liver lesion. Patients presenting all dHGP or rHGP liver metastases were classiﬁed as the dHGP or rHGP patients, respectively, while the patients showing different HGP liver lesions or impure liver lesions were deﬁned as impure HGP patients (Figure 1C). Primary tumor and normal liver tissues from pure HGP patients, as well as all available liver lesions, were collected for RNA sequencing. In order to better compare the differences between dHGP liver metastases and rHGP liver metastases, only pure lesions were included. Similarly, only pure patients were included in the comparision for the primary tumor and normal liver tissue. A B D E C FIGURE 1 Study design. (A) Tissues were obtained from patients undergoing colectomy and hepatic resection in this study. (B) Evaluations of HGP were done for liver metastasis and each patient. (C) Patient enrolment and samples were included in the analyses. Of the 41 enrolled patients, a total of 90 liver metastasis lesions, 35 normal liver tissues, and 35 primary lesions were included in the transcriptome sequencing and analyses. (D) A heatmap depicting the HGP for each liver metastasis and each patient. (E) Pathological evaluation and RNA sequencing of all liver lesions. : the sample was sent to RNA sequencing. A B C C C B A D E gn. (A) Tissues were obtained from patients undergoing colectomy and hepatic resection in this study. (B) Evaluations of HGP were ver metastasis and each patient. (C) Patient enrolment and samples were included in the analyses. Of the 41 enrolled patients, a tota metastasis lesions, 35 normal liver tissues, and 35 primary lesions were included in the transcriptome sequencing and analyses. Results In total, 166 intraoperative liver lesions from 41 patients were found. 125 lesions were histologically conﬁrmed metastatic colorectal adenocarcinoma and assessed as pure dHGP or rHGP according to the pathological examination. Six patients had different HGP liver lesions, and 35 patients were pure HGP (20 dHGP and 15 rHGP) (Figures 1D, E). Pathological assessment of HGPs (D) A d i ti th HGP f h li t t i d h ti t (E) P th l i l l ti d RNA i f ll li l i D E D FIGURE 1 Study design. (A) Tissues were obtained from patients undergoing colectomy and hepatic resection in this study. (B) Evaluations of HGP were done for liver metastasis and each patient. (C) Patient enrolment and samples were included in the analyses. Of the 41 enrolled patients, a total of 90 liver metastasis lesions, 35 normal liver tissues, and 35 primary lesions were included in the transcriptome sequencing and analyses. (D) A heatmap depicting the HGP for each liver metastasis and each patient. (E) Pathological evaluation and RNA sequencing of all liver lesions. : the sample was sent to RNA sequencing. FIGURE 1 Study design. (A) Tissues were obtained from patients undergoing colectomy and hepatic resection in this study. (B) Evaluations of HGP were done for liver metastasis and each patient. (C) Patient enrolment and samples were included in the analyses. Of the 41 enrolled patients, a total of 90 liver metastasis lesions, 35 normal liver tissues, and 35 primary lesions were included in the transcriptome sequencing and analyses. (D) A heatmap depicting the HGP for each liver metastasis and each patient. (E) Pathological evaluation and RNA sequencing of all liver lesions. : the sample was sent to RNA sequencing. Frontiers in Immunology 03 frontiersin.org Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 Comparison of the transcriptome landscapes between the 47 dHGP and 43 rHGP liver metastatic lesions RNA purity was veriﬁed using the NanoPhotometer® spectrophotometer (IMPLEN, CA, USA), while RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA). The library preparations were sequenced on an Illumina Hiseq platform, and 125 bp/150 bp paired-end reads were generated. In total, 160 samples, including 90 liver metastases, 35 normal liver tissues, and 35 primary lesions, were successfully sequenced using this approach. Differential expression analysis between 47 dHGP and 43 rHGP liver metastatic lesions was performed (\|LogFC\| > 1, false discover rate (FDR) < 0·05; Table S2; Figure 2A). The GO-BP enrichment analysis were conducted to identify the biological function of DEGs (FDR < 0·05, Table S3; Figure 2B). 1460 DEGs were up-regulated in the dHGP subgroup and were found to be markedly enriched in immune-related biological processes. Among them, T-cell activation was the most signiﬁcantly enriched pathway. Of the 72 important immune genes, 55 were up-regulated in the dHGP liver metastases (FDR < 0·05; Figure 2C). Meanwhile, 427 DEGs were up-regulated in the rHGP subgroup and were mainly associated with cell proliferation and cell cycle, including the ERBB2 signaling pathway. Furthermore, comparisons of the HALLMARK pathways activity indicated that dHGP liver metastatic lesions were signiﬁcantly enriched in speciﬁc cancer-related pathways, including epithelial-mesenchymal transition (EMT), angiogenesis, KRAS signaling pathway, inﬂammatory response, and interferon-gamma response, while the rHGP liver metastatic lesions were associated with DNA repair, fatty acid metabolism, E2F targets, oxidative phosphorylation, G2M checkpoints, glycolysis and MYC target (Wilcoxon test, p < 0·05; Tables S4, S5; Figure 2D). With a higher immune and stromal score, dHGP liver metastases were characterized by a higher abundance of monocytes, ﬁbroblasts, M2 macrophages, CD8+ T cells, and naive B cells and obtained higher scores from 7 immune-related signatures, indicating that the TIME had a better anti-tumor ability (Wilcoxon test, p < 0·05; Tables S6, S7; Figure 2E). Notably, among all the 90 liver metastatic lesions, activity levels of EMT and angiogenesis were positively correlated with the immune score, stromal score, CD8+ T cells, ﬁbroblasts, and anti-tumor immune signatures (Spearman correlation analysis, p < 0·01; Table S8; Figure 2F). Similar differences between dHGP and rHGP liver metastatic lesions were further conﬁrmed in two impure patients (Figure S1). Statistical analysis All statistical analyses were performed using R Statistical Software (version 3.6.3). The Chi-square tests were employed to examine the associations of categorical variables, while the Wilcoxon test was utilized to compare two groups. The pROC package for R was applied to calculate AUC to examine the predictive accuracy of each differentially expressed gene (DEG), C-score, and Ln-score for HGP type of liver metastases. The interactions were estimated by Spearman correlation analyses. The associations of HGP or risk score with RFS were examined by Kaplan-Meier and Cox proportional hazard analyses. And the log-rank test was employed. The stats package for R was applied to conduct principal component analysis (PCA). All the above calculations were done with TPM data. All statistical p-values were two-sided. Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients Clinical characteristics dHGP patients (N Gender male 14 (70.0) Female 6 (30.0) Age mean ± standard deviation 56.9 ± 8.1 Location of primary tumor Left hemicolon and rectum 14 (70.0) Right hemicolon 6 (30.0) pT stage T1-2 1 (5.0) T3 17 (85.0) T4 2 (10.0) pN stage N0 12 (60.0) N1 7 (35.0) N2 1 (5.0) Extrahepatic metastasis Yes 5 (25.0) No 15 (75.0) Preoperative chemotherapy Yes 18 (90.0) No 2 (10.0) Combined with bevacizumab Yes 3 (15.0) No 17 (85.0) Combined with cetuximab Yes 7 (35.0) No 13 (65.0) History of hepatitis B Yes 3 (15.0) No 17 (85.0) Drinking History Yes 1 (5.0) No 19 (95.0) Smoking history Yes 1 (5.0) No 19 (95.0) History of Hypertension Yes 4 (20.0) No 16 (80.0) History of diabetes Yes 4 (20.0) No 16 (80.0) The bold values are Clinical charateristics dHGP patients (N = 20) Clinical characteristics dHGP patients (N = 20) rHGP patients (N = 15) P value Gender 0.829 male 14 (70.0) 11 (73.3) Female 6 (30.0) 4 (26.7) Age 0.398 mean ± standard deviation 56.9 ± 8.1 54.2 ± 10.5 Location of primary tumor 0.833 Left hemicolon and rectum 14 (70.0) 10 (66.7) Right hemicolon 6 (30.0) 5 (33.3) pT stage 0.660 T1-2 1 (5.0) 2 (13.3) T3 17 (85.0) 12 (80.0) T4 2 (10.0) 1 (6.7) pN stage 0.127 N0 12 (60.0) 5 (33.3) N1 7 (35.0) 6 (40.0) N2 1 (5.0) 4 (26.7) Extrahepatic metastasis 0.393 Yes 5 (25.0) 2 (13.3) No 15 (75.0) 13 (86.7) Preoperative chemotherapy 0.759 Yes 18 (90.0) 13 (86.7) No 2 (10.0) 2 (13.3) Combined with bevacizumab 0.698 Yes 3 (15.0) 3 (20.0) No 17 (85.0) 12 (80.0) Combined with cetuximab 0.599 Yes 7 (35.0) 4 (26.7) No 13 (65.0) 11 (73.3) History of hepatitis B 0.117 Yes 3 (15.0) 0 (0) No 17 (85.0) 15 (100) Drinking History 0.833 Yes 1 (5.0) 1 (6.7) No 19 (95.0) 14 (93.3) Smoking history 0.167 Yes 1 (5.0) 3 (20.0) No 19 (95.0) 12 (80.0) History of Hypertension 0.265 Yes 4 (20.0) 1 (6.7) No 16 (80.0) 14 (93.3) History of diabetes 0.265 Yes 4 (20.0) 1 (6.7) No 16 (80.0) 14 (93.3) The bold values are Clinical charateristics. The bold values are Clinical charateristics. pathways, including oxidative phosphorylation and fatty acid metabolism. Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients There were no signiﬁcant differences in gender, age, pT stage and pN stage between the two groups. The primary tumor of colorectal cancer in both groups were mainly located in the left colon and rectum [14 patients (70.0%) vs. 10 patients (66.7%), P =0.833]. Most patients had received preoperative chemotherapy [18 (90.0%) vs. 13 (86.7%), P =0.759]. All three patients with viral hepatitis were dHGP, but there was no statistically signiﬁcant difference [3 (15.0%) vs. 0 (0%), P =0.117] (Table 1). Analysis of the primary lesions showed that 1091 DEGs were up-regulated in dHGP subgroup and 452 DEGs were up- regulated in the rHGP subgroup (\|LogFC > 1\|, FDR < 0·05; Table S10; Figure 3A). GO-BP enrichment analysis of 1091 DEGs (FDR < 0·05; Table S11; Figure 3B) and HALLMARK Frontiers in Immunology frontiersin.org 04 Hu et al. 10.3389/ﬁmmu.2022.1045329 TABLE 1 Clinical characteristics of dHGP and rHGP patients. TABLE 1 Clinical characteristics of dHGP and rHGP patients. Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients pathways analysis (Wilcoxon test, p < 0·05; Tables S4, S5; Figure 3C) revealed that the primary lesions of dHGP patients were associated with EMT, angiogenesis, and transforming growth factor-b (TGF-b). Interestingly, primary lesions of rHGP patients were also enriched in metabolism-related Unlike metastatic liver lesions, dHGP patients showed a higher stromal score but not immune score in primary lesions. A higher abundance of ﬁbroblasts, M2 macrophages, M0 Frontiers in Immunology frontiersin.org 05 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 macrophages and neutrophils were present in primary lesions of dHGP patients (Wilcoxon test, p < 0·05; Tables S6, S7; Figure 3D). Primary lesions of rHGP patients exhibited a higher abundance of regulatory T cells. Consensus molecular subtypes (CMS)were different in two types of patients. Primary lesions of dHGP patients were characterized by CMS4 while rHGP patients were marked with CMS3 (Table S12; Chi-square test, p < 0·01). The common DEGs between primary and metastatic liver lesions included 243 up-regulated genes in the dHGP subgroup and 44 up-regulated genes in the rHGP subgroup (Figure 3F). A B D E F C FIGURE 2 Comparison of the transcriptome landscapes between dHGP and rHGP liver metastases. (A) Volcano plot displaying 1460 up-regulated DEGs in the dHGP subgroup and 427 up-regulated DEGs in the rHGP subgroup (\|LogFC\| > 1, FDR < 0·05). (B) GO-BP enrichment analysis of the differentially expressed genes (FDR < 0·05). (C) Circos plot shows 55/72 up-regulated immune genes in the dHGP samples (LogFC > 0, FDR < 0·05). (D) Heatmap depicting the differences of HALLMARK pathways activity between dHGP and rHGP subgroups (Wilcoxon test. ns, p > 0·05; p < 0·05; p < 0·01; p < 0·001; p < 0.0001). Rows represent GSVA scores of HALLMARK pathways, and columns represent samples. (E) Heatmap depicting the differences between two subgroups in TIME-relevant molecular signatures were linked to cellular estimates, immune cells and immune signatures (Wilcoxon test. ns, p > 0·05; p < 0·01; *p < 0·001; *p < 0.0001). Rows represent ssgsea scores of TIME- relevant molecular signatures and columns represent samples. (F) Correlations between selected HALLMARK pathways activity and TIME- relevant molecular signatures (Spearman correlation analysis. p < 0·05; p < 0·01; p < 0·001). A B B A C C F F E F E E FIGURE 2 Comparison of the transcriptome landscapes between dHGP and rHGP liver metastases. Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients (A) Volcano plot displaying 1460 up-regulated DEGs in the dHGP subgroup and 427 up-regulated DEGs in the rHGP subgroup (\|LogFC\| > 1, FDR < 0·05). (B) GO-BP enrichment analysis of the differentially expressed genes (FDR < 0·05). (C) Circos plot shows 55/72 up-regulated immune genes in the dHGP samples (LogFC > 0, FDR < 0·05). (D) Heatmap depicting the differences of HALLMARK pathways activity between dHGP and rHGP subgroups (Wilcoxon test. ns, p > 0·05; p < 0·05; p < 0·01; p < 0·001; *p < 0.0001). Rows represent GSVA scores of HALLMARK pathways, and columns represent samples. (E) Heatmap depicting the differences between two subgroups in TIME-relevant molecular signatures were linked to cellular estimates, immune cells and immune signatures (Wilcoxon test. ns, p > 0·05; p < 0·01; *p < 0·001; *p < 0.0001). Rows represent ssgsea scores of TIME- relevant molecular signatures and columns represent samples. (F) Correlations between selected HALLMARK pathways activity and TIME- relevant molecular signatures (Spearman correlation analysis. p < 0·05; p < 0·01; p < 0·001). frontiersin.org Frontiers in Immunology 06 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 A B D E F G H C FIGURE 3 Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients. (A) Volcano plot displaying 1091 up- regulated DEGs in the dHGP subgroup and up-regulated 452 DEGs in the rHGP subgroup (\|LogFC\| > 1, FDR < 0·05). (B) GO-BP enrichment analysis of the differentially expressed genes. (C) Heatmap depicting the differences of HALLMARK pathways activity between dHGP and rHGP subgroups (Wilcoxon test. ns, p > 0·05; p < 0·01; p < 0·001). Rows represent the mean GSVA scores of HALLMARK pathways, and columns represent HGP subgroups. (D) Heatmap depicting the differences between two subgroups in TIME-relevant molecular signatures were linked to cellular estimates, immune cells and immune signatures (Wilcoxon test. ns, p > 0·05; p < 0·01). Rows represent mean ssgsea scores of TIME- relevant molecular signatures, and columns represent HGP subgroups. (E) Proportion of different CMS subtypes of primary lesions in dHGP or rHGP subgroups (Chi-square test, p < 0·01). (F) Venn diagram illustrating the intersection of DEGs in dHGP and rHGP subgroups respectively between the primary lesions and liver metastases. (G) Volcano plot showing the common up-regulated DEGs in the EMT pathways in the dHGP subgroup between primary lesions and liver metastases. Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients (H) Volcano plot showing the intersection of DEGs up-regulated in angiogenesis pathway in the dHGP subgroup between primary lesions and liver metastases. p < 0.05. A B B D D C D E E F H G F H G GU 3 Comparison of the transcriptome landscapes of primary lesions between 20 dHGP and 15 rHGP patients. (A) Volcano plot displaying 1091 up- regulated DEGs in the dHGP subgroup and up-regulated 452 DEGs in the rHGP subgroup (\|LogFC\| > 1, FDR < 0·05). (B) GO-BP enrichment analysis of the differentially expressed genes. (C) Heatmap depicting the differences of HALLMARK pathways activity between dHGP and rHGP subgroups (Wilcoxon test. ns, p > 0·05; p < 0·01; p < 0·001). Rows represent the mean GSVA scores of HALLMARK pathways, and columns represent HGP subgroups. (D) Heatmap depicting the differences between two subgroups in TIME-relevant molecular signatures were linked to cellular estimates, immune cells and immune signatures (Wilcoxon test. ns, p > 0·05; p < 0·01). Rows represent mean ssgsea scores of TIME- relevant molecular signatures, and columns represent HGP subgroups. (E) Proportion of different CMS subtypes of primary lesions in dHGP or rHGP subgroups (Chi-square test, p < 0·01). (F) Venn diagram illustrating the intersection of DEGs in dHGP and rHGP subgroups respectively between the primary lesions and liver metastases. (G) Volcano plot showing the common up-regulated DEGs in the EMT pathways in the dHGP subgroup between primary lesions and liver metastases. (H) Volcano plot showing the intersection of DEGs up-regulated in angiogenesis pathway in the dHGP subgroup between primary lesions and liver metastases. *p < 0.05. Comparison of the transcriptome landscapes of normal liver tissues between 20 dHGP and 15 rHGP patients Both primary and metastatic lesions in the dHGP subgroup showed enrichment in the EMT and angiogenesis pathways (Table S13). As shown in Figures 3G, H, we summarized the intersecting up-regulated DEGs between the primary and metastatic liver lesions of the dHGP subgroup that were also enriched in the EMT and angiogenesis pathways. In normal liver tissues, 9 DEGs were up-regulated in dHGP patients but there was no DEGs in rHGP patients (\|LogFC\| > 0·5, Frontiers in Immunology frontiersin.org 07 Hu et al. 10.3389/ﬁmmu.2022.1045329 Hu et al. FDR < 0·05; Table S14; Figure 4A). GO-BP enrichment analysis revealed that LUM was associated with the regulation of transforming growth factor beta1 production, which also participates in the process of collagen ﬁbril organization along with LOXL4, EPHA3 and ITGBL1 were associated with cell- matrix adhesion. (FDR < 0·05; Table S15; Figure S2A). In addition, GSEA analysis of the KEGG pathways showed that normal liver tissues in the dHGP subgroup were enriched in the structural extracellular matrix (ECM)-receptor interaction (FDR < 0·05; Table S16; Figure S2B & Figure 4B). The above evidence prompted us to believe that the ﬁbrosis progression of normal liver tissues might be implicated in the formation of different HGPs. FDR < 0·05; Table S14; Figure 4A). GO-BP enrichment analysis revealed that LUM was associated with the regulation of transforming growth factor beta1 production, which also participates in the process of collagen ﬁbril organization along with LOXL4, EPHA3 and ITGBL1 were associated with cell- matrix adhesion. (FDR < 0·05; Table S15; Figure S2A). In addition, GSEA analysis of the KEGG pathways showed that normal liver tissues in the dHGP subgroup were enriched in the structural extracellular matrix (ECM)-receptor interaction (FDR < 0·05; Table S16; Figure S2B & Figure 4B). The above evidence prompted us to believe that the ﬁbrosis progression of normal liver tissues might be implicated in the formation of different HGPs. diseases was examined. Compared to the normal group, expression of all the 9 genes increased in the ﬁbrotic or cirrhotic group of at least 4 human datasets. The expressions of 7 genes (NALCN, MOXD1, LUM, ITGBL1, EPHA3, EFEMP1, DTNA) were increased in 6 or more human datasets. The mRNA expression of 3 genes (LUM, ITGBL1, EPHA3) was markedly elevated in the ﬁbrotic subgroup across humans and mice, regardless of the underlying etiology (LogFC > 0, FDR < 0·05; Figure 4C). The hypothesis of the formation of different HGPs formation in metastatic liver lesions from primary tumor cells. Combined with the previous ﬁndings, the Ln-score indirectly reﬂected the degree of ﬁbrosis in normal liver tissues. Compared to the rHGP patients, dHGP patients showed a much higher C-score and Ln-score (p < 0·01; Table S18; Figures S3B, S3C). In other words, the higher the C-score or Ln-score, the more likely that the liver metastatic lesions possessed a desmoplastic growth pattern. This was further conﬁrmed by the hierarchical clustering of individual C-score and Ln-score, which classiﬁed 20 dHGP and 15 dHGP patients (Figure 5A). 11 dHGP patients (100%, 11/11) displayed both a high C-score and a high Ln-score while 11 rHGP patients (91.7%, 11/12) were found to be associated with a low C-score and a low Ln-score (Chi-square test, p < 0·01; Figure 5B). In addition, 8 dHGP patients (75%, 8/12) were shown to be either associated with a high C-score or a high Ln-score. Patients with different HGPs were divided into separate groups based on the two scores, further conﬁrmed by PCA analysis (Figure 5C). Moreover, the AUC of the combination of C-score and Ln-score measured by ROC curves among the patients was 0.963 (95%CI: 0·800-0·973), which was higher than the AUC of either one independently (Figure 5D). To provide a biological interpretation of the different HGP types, we constructed two scoring systems to quantify the differences between the primary lesions (C) and normal liver tissues (Ln) of dHGP and rHGP patients. As shown in Figure S3A, the common DEGs between primary lesions and liver metastases included 243 up-regulated genes in the dHGP subgroup and 44 up-regulated genes in the rHGP subgroup. The scores of these up-regulated genes were quantiﬁed by the “ssGSEA” method (15) and were deﬁned as dHGP score and rHGP score, respectively. An approach similar to “Gene expression grade index” (16) was utilized to calculate the C-score of each pure patient: C-score = dHGP score - rHGP score, which represented each patient’s transcriptomic characterization of the intrinsic inheritance of the primary lesion respectively. Comparison of the transcriptome landscapes of normal liver tissues between 20 dHGP and 15 rHGP patients However, the extent of mRNA expression fold change of 9 genes varied among different cohorts, which would require further validation. In summary, the expression of 9 DEGs was up-regulated with advancing ﬁbrosis, while there were only modest changes in normal situations or mild ﬁbrosis. Therefore, we surmised that normal liver tissue ﬁbrosis might contribute to different HGP-type formations in liver metastases. Furthermore, the differential expression of 9 DEGs in 14 transcriptome datasets of ﬁbrotic livers caused by non-cancerous A B C FIGURE 4 Comparison of the transcriptome landscapes of normal liver tissues between 20 dHGP and 15 rHGP patients. (A) Volcano plot showing 9 upregulated DEGs in the dHGP subgroup (\|LogFC\| > 0·5, FDR < 0·05). (B) Enrichment plots showing ECM receptor interaction pathways in the dHGP group. (C) Heatmap depicting the differential expression of 9 DEGs in the ﬁbrosis or cirrhosis group versus the normal group in 14 public datasets of ﬁbrotic livers caused by non-cancerous diseases. Numbers in each square box represent the fold change. NAFLD indicates nonalcoholic fatty liver disease. HCV indicates the hepatitis C virus. HBV indicates the hepatitis B virus. BDL indicates bile duct ligation. CCL4 indicates carbon tetrachloride. Pdgf indicates platelet-derived growth factor (C) BDL, CCL4, and Pdgf-c are the modeling methods of liver ﬁbrosis in mice. B A B C FIGURE 4 Comparison of the transcriptome landscapes of normal liver tissues between 20 dHGP and 15 rHGP patients. (A) Volcano plot showing 9 upregulated DEGs in the dHGP subgroup (\|LogFC\| > 0·5, FDR < 0·05). (B) Enrichment plots showing ECM receptor interaction pathways in the dHGP group. (C) Heatmap depicting the differential expression of 9 DEGs in the ﬁbrosis or cirrhosis group versus the normal group in 14 public datasets of ﬁbrotic livers caused by non-cancerous diseases. Numbers in each square box represent the fold change. NAFLD indicates nonalcoholic fatty liver disease. HCV indicates the hepatitis C virus. HBV indicates the hepatitis B virus. BDL indicates bile duct ligation. CCL4 indicates carbon tetrachloride. Pdgf indicates platelet-derived growth factor (C) BDL, CCL4, and Pdgf-c are the modeling methods of liver ﬁbrosis in mice. Frontiers in Immunology frontiersin.org 08 Hu et al. 10.3389/ﬁmmu.2022.1045329 Hu et al. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients To our knowledge, this is the ﬁrst study that performed comprehensive analyses and comparisons at the RNA level on primary tumors, liver metastases and normal liver tissues, using two HGP subtypes in the CRLM. The mainly comparison of the transcriptome landscapes between the dHGP and rHGP were summarized in Table 2. We found that desmoplastic and replacement liver metastases showed different dysregulated mechanisms at the RNA level, mostly matching histopathological morphologies. The different expression genes in liver metastases might be partly inherited from the primary tumors, and the normal liver parenchymal seemed to possess different inﬂammatory microenvironments between the two HGP subtypes. Specimens from 41 patients were collected for this study. In addition to the 35 pure HGP patients, there were 6 impure HGP patients. Those impure patients and pure rHGP patients were deﬁned as non-dHGP patients. One pure dHGP patient was lost to follow-up, and 40 patients (19 dHGP vs 21 non-dHGP) were included in this part. In contrast to non-dHGP patients, dHGP patients were more likely to have a better 6 months RFS (RFS6M) (Log-rank test, p = 0·174, HR: 0.4 (95%CI: 0·2 to 1·4), Figure 6A; Chi-square test, p = 0·36, Figure 6B). Previous studies suggested that integrative molecular subtyping of the liver metastases could determine the prognosis of CRLM patients (17, 18). Based on the HALLMARK pathways and TIME-relevant molecular signatures, we performed unsupervised clustering to classify 90 liver metastases into three biologically distinct transcriptome subtypes, termed as High-IS (immune score and stromal score), Medium-IS and Low-IS (Figure S4; Figure 6C). Signiﬁcant differences were observed in the distribution of three transcriptome subtypes in dHGP or rHGP metastatic lesions (Chi-square test, p < 0·001; Figure 6D). Notably, rHGP metastases displayed no High-IS metastases subtype and more Medium-IS and Low-IS metastases subtypes than dHGP metastases. Higher inﬂammatory and immune responses were exhibited in the dHGP at the RNA level, consistent with previous studies showing that dHGP correlated with increased cytotoxic immune inﬁltrate in the immunohistochemistry and ﬂow cytometry (19). This might explain the superior survival of patients with dHGP. In contrast, the rHGP seemed to be good at proliferation and DNA damage repair at the RNA level, which was not directly shown in morphology. Correspondingly, rHGP lesions showed more aggressive invasiveness and metastasis potential. The hypothesis of the formation of different HGPs formation in metastatic liver lesions Similarly, 9 up-regulated DEGs in the normal liver tissues of the dHGP subgroup were quantiﬁed and considered as the Ln-score, which represented individual transcriptomic characterization of the normal liver microenvironment during the formation of liver metastases However, there were some exceptions; P38 was a dHGP patient who obtained a low C-score and Ln-score, which A B D C FIGURE 5 Exploration in the hypothesis of the formation of different HGPs in liver metastases. (A) Hierarchical clustering of individual C-score and Ln- score to classify 20 dHGP and 15 dHGP patients. “Euclidean” method was applied to calculate the distance. Column dendrograms were hide. Heatmap showing CMS subtypes of each patient. (B) Distribution of dHGP and rHGP patients stratiﬁed by both C-score and Ln-score (Chi- square test, p < 0·01). The two scores were divided into high and low groups based on the median. (C) PCA plot of patients in different HGP subgroups based on the individual C-score and Ln-score. (D) The predictive values for HGP types of the C-score, Ln-score, and both were measured by ROC curves among 35 patients. A A C B D C D B D B B FIGURE 5 Exploration in the hypothesis of the formation of different HGPs in liver metastases. (A) Hierarchical clustering of individual C-score and Ln- score to classify 20 dHGP and 15 dHGP patients. “Euclidean” method was applied to calculate the distance. Column dendrograms were hide. Heatmap showing CMS subtypes of each patient. (B) Distribution of dHGP and rHGP patients stratiﬁed by both C-score and Ln-score (Chi- square test, p < 0·01). The two scores were divided into high and low groups based on the median. (C) PCA plot of patients in different HGP subgroups based on the individual C-score and Ln-score. (D) The predictive values for HGP types of the C-score, Ln-score, and both were measured by ROC curves among 35 patients. frontiersin.org Frontiers in Immunology 09 Hu et al. 10.3389/ﬁmmu.2022.1045329 months after hepatectomy (Chi-square test, p = 0·18; Figure 6H). months after hepatectomy (Chi-square test, p = 0·18; Figure 6H). prompted us to believe that the factors inﬂuencing the formation of HGPs are more complex than the hypothesis we proposed in this study. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients Thus, histopathological variations are observable morphological changes resulting from the regulation of molecular expressions, and RNA sequencing could provide more details about the HGP features. We found that the primary tumors with dHGP liver metastases also showed upregulation of EMT, angiogenesis, and the TGF-b signal pathway, which indicated that many liver metastases signatures were inherited from their primary tumor. The tumor border conﬁguration of primary CRC also displayed different growth patterns. The pattern circumscribed with clearly desmoplastic stromal was consistently associated with a favorable prognosis and might indicate speciﬁc gene changes (20). According to a previous study, liver metastases originating from breast cancers mostly showed a predominance of rHGP (21). Therefore, it is reasonable to presume that the primary tumor may be responsible for the HGP in metastases. However, there was no signiﬁcant difference in immune inﬁltration between the two HGP types of primary tumors. Our previous study also found that there was no signiﬁcant correlation between the immune status of primary tumors and liver metastases, and the immune score of the primary tumor could not predict the prognosis of CRLM (22). However, in some metastatic lesions, the same HGP contains different transcriptomic isoforms, suggesting heterogeneity at the molecular level. Considering that the HGP type and transcriptome subtype contribute to patients’ prognosis, we deﬁned a novel risk score based on them and classiﬁed patients into two main groups (Figure 6E). The non-dHGP and transcriptome subtype with Medium-IS subtype were assigned 1 point, while the dHGP and transcriptome subtype without Medium-IS subtype were assigned 0 points. Patients with 0-1 points in total were deﬁned as low-risk, while those with 2 points were considered high risk. Kaplan-Meier curves for RFS6M patients with different risk scores showed that risk score positively correlated with early recurrence (Log-rank test, p = 0·16; Figure 6F). Furthermore, the patients in high-risk subgroup displayed shorter RFS6M than patients in low-risk subgroup (Log-rank test, p = 0·065, HR: 0.32 (0·09-1·21); Figure 6G). Sixteen percent of low-risk and 40% of high-risk patients developed recurrence within 6 Frontiers in Immunology frontiersin.org 10 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 A B D E F G H C on in the predictive value of the HGP type and transcriptome subtype of liver metastases for the survival of CRLM patients. (A) Kaplan- rves for 6 months Relapse-Free Survival (RFS6M) of 19 dHGP patients and 21 non-dHGP patients (Log-rank test, p = 0.174). Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients (B) Rate of pse events within 6 months (RFS_6M, Yes and No) in dHGP or non-dHGP subgroups (Chi-square test, p = 0.36). (C) Unsupervised g of important HALLMARK pathways and TIME-relevant molecular signatures among 90 liver metastases. Rows represent scores of gnatures, and columns represent samples. (D) Rate of three transcriptome subtypes in dHGP or non-dHGP subgroups (Chi-square test, 1). (E) Heatmap showing the transcriptome subtypes, HGP, risk score, risk subgroup, and early relapse event within 6 months of each F) Kaplan-Meier curves for RFS6M of patients with different risk scores (Log-rank test, p = 0.16). (G) Kaplan-Meier curves for RFS6M of with different risk subgroups (Log-rank test, p = 0.065). (H) Rate of early recurrence events within 6 months (RFS_6M, Yes and No) in or low-risk subgroups (Chi-square test, p = 0.18). A B B A D C D E F G H FIGURE 6 Exploration in the predictive value of the HGP type and transcriptome subtype of liver metastases for the survival of CRLM patients. (A) Kaplan- Meier curves for 6 months Relapse-Free Survival (RFS6M) of 19 dHGP patients and 21 non-dHGP patients (Log-rank test, p = 0.174). (B) Rate of early relapse events within 6 months (RFS_6M, Yes and No) in dHGP or non-dHGP subgroups (Chi-square test, p = 0.36). (C) Unsupervised clustering of important HALLMARK pathways and TIME-relevant molecular signatures among 90 liver metastases. Rows represent scores of various signatures, and columns represent samples. (D) Rate of three transcriptome subtypes in dHGP or non-dHGP subgroups (Chi-square test, E E H G G H F H G F FIGURE 6 Exploration in the predictive value of the HGP type and transcriptome subtype of liver metastases for the survival of CRLM patients. (A) Kaplan- Meier curves for 6 months Relapse-Free Survival (RFS6M) of 19 dHGP patients and 21 non-dHGP patients (Log-rank test, p = 0.174). (B) Rate of early relapse events within 6 months (RFS_6M, Yes and No) in dHGP or non-dHGP subgroups (Chi-square test, p = 0.36). (C) Unsupervised clustering of important HALLMARK pathways and TIME-relevant molecular signatures among 90 liver metastases. Rows represent scores of various signatures, and columns represent samples. (D) Rate of three transcriptome subtypes in dHGP or non-dHGP subgroups (Chi-square test, p < 0·001). (E) Heatmap showing the transcriptome subtypes, HGP, risk score, risk subgroup, and early relapse event within 6 months of each patient. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients (F) Kaplan-Meier curves for RFS6M of patients with different risk scores (Log-rank test, p = 0.16). (G) Kaplan-Meier curves for RFS6M of patients with different risk subgroups (Log-rank test, p = 0.065). (H) Rate of early recurrence events within 6 months (RFS_6M, Yes and No) in high-risk or low-risk subgroups (Chi-square test, p = 0.18). neoplastic lesions of the human liver. However, these genes were not signiﬁcantly up-regulated in mice models. This might be an important reason why few liver metastases exist the dHGP in diverse animal models of CRLM unless speciﬁc genetic The soil of the liver is a fertile and popular site for excessive and complicated immunological activities (23). Several speciﬁc molecules were up-regulated in the normal liver tissue of patients with dHGP, related to inﬂammations in non- Frontiers in Immunology frontiersin.org 11 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 of liver soil (one with increased inﬂammation and the other without) and two types of primary seeds (one with enhanced EMT and angiogenesis, and the other with enhanced metabolisms). If primary seeds with enhanced EMT and angiogenesis land on the inﬂammatory soil, they would probably result in desmoplastic lesions. In contrast, replacement lesions would arise when seeds with enhanced metabolisms fall into the non-inﬂammatory soil. And liver metastases would be very heterogeneous if seeds with enhanced EMT and angiogenesis go modiﬁcation (24). These results strongly suggest that the microenvironment of the liver plays an essential role in determining HGP. The mechanism of HGPs formations is still unknown. The widely accepted “seed and soil” hypothesis described the metastases, which are the tumor cells as the “seeds” while the “suitable soil” is the metastatic microenvironment (25, 26). Thus, we speculated that the heterogeneity or reprogramming of seeds and soils might contribute to different HGPs (Figure 7). There may be two types TABLE 2 The mainly comparison of the transcriptome landscapes between the dHGP and rHGP. ing: ↑: up-regulated in the dHGP; ↓: up-regulated in the rHGP;/, no analysis; -, no signiﬁcance. HGP, histopathological growth pattern; dHGP, desmoplastic histop tern; rHGP, replacement histopathological growth pattern; DEGs, differential expression genes; CMS, consensus molecular subtypes. Noting: ↑: up-regulated in the dHGP; ↓: up-regulated in the rHGP;/, no analysis; -, no signiﬁcance. HGP, histopathological growth pattern; dHGP, desmoplastic histopathological growth pattern; rHGP, replacement histopathological growth pattern; DEGs, differential expression genes; CMS, consensus molecular subtypes. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients In this study, we provided a comprehensive molecular pathological subtype for patients with CRLM, in which the patients possessing the dHGP subtype and overexpressed immune genes showed a low risk of early recurrence. Although pathological morphology and molecular phenotype were consistent, the combination still showed better predictive power, which might guide treatment strategy, including targeted therapy and immunotherapy. However, more sample sizes and external validation are needed to verify the accuracy of our model. The CMS was developed based on abundant and independent gene expression data, but most patients were TNM I-III (27). In our cohort, most primary tumors with the dHGP liver metastases were classiﬁed as CMS4, while most rHGP showing metabolic dysregulation were classiﬁed as CMS3. In this study, we perform unsupervised clustering to classify liver metastases into three biologically distinct transcriptome subtypes. The contribution of different molecular features was similar to a previous study on curable oligometastatic CRLM (17). Subtype Low-IS metastases displayed enrichment for expression patterns associated with low immune and stromal inﬁltration. However, they were markedly enriched in E2F/MYC pathways and abnormalities in cell cycle checkpoints and DNA repair signaling, similar to the canonical subtype described in the previous study (17). In addition, High-IS and Medium-IS metastases subtypes were enriched for EMT, angiogenesis, and KRAS signaling. However, the High-IS metastases subtype showed a higher immune score than the Medium-IS metastases subtype. Therefore, the High-IS metastases subtype was more similar to the immune subtype as mentioned in the previous study, while Medium-IS metastases were more similar to the stromal subtype (17). Nevertheless, our study has some limitations. Firstly, most patients in our cohort received preoperative systemic chemotherapy leading to potential confounding effects on the HGP of CRLM and remodeling of the immune microenvironment (28, 29). Some metastatic lesions were so sensitive to chemotherapy that large areas of tumor cells were necrotic and tumors even achieved pathological complete response, leading to HGP evaluation failure. Secondly, our sample size was not large enough, and no public data was available to verify the results so that the ﬁndings were mainly speculative and descriptive. The follow-up time of survival analysis was insufﬁcient, and overall survival data were not available. Thirdly, we did not perform in vivo or vitro experiments to uncover the critical molecules involved in the mechanism of HGP formation. This is because there is no accepted animal model of HGP to date. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients dHGP vs rHGPin metastatic liver lesions (n = 47 vs n = 43) dHGP vs rHGPin primary colorectal lesions (N = 20 vs N = 15) dHGP vs rHGPin normal liver tissues (N = 20 vs N = 15) Number of DEGs up-regulated in the dHGP 1460 (77.3%) 1091 (70.7%) 9 (100%) up-regulated in the rHGP 427 (22.6%) 452 (29.3%) 0 (0%) Gene sets analyses Inﬂammatory response ↑ ↑ – EMT ↑ ↑ – Angiogenesis ↑ ↑ – DNA repair ↓ – – Fatty acid metabolism ↓ ↓ – Oxidative phosphorylation ↓ ↓ – G2M_checkpoint ↓ – – KRAS signaling ↑ – – TGF-b signaling – ↑ – ECM-receptor interaction / / ↑ Cell types Tumor purity ↓ – – Immunes core ↑ – – Stromal score ↑ ↑ – Macrophage ↑ ↑ – Fibroblast ↑ ↑ – CD8+ T cell ↑ – – B cell ↑ – – CMS CMS 1 / 2 (10.5%) vs 1 (6.7%) / CMS 2 / 5 (26.3%) vs 4 (26.6%) / CMS 3 / 2 (10.5%) vs 9 (60.0%) / CMS 4 / 10 (52.7%) vs 1 (6.7%) / Noting: ↑: up-regulated in the dHGP; ↓: up-regulated in the rHGP;/, no analysis; -, no signiﬁcance. HGP, histopathological growth pattern; dHGP, desmoplastic histopathological growth pattern; rHGP, replacement histopathological growth pattern; DEGs, differential expression genes; CMS, consensus molecular subtypes. TABLE 2 The mainly comparison of the transcriptome landscapes between the dHGP and rHGP. TABLE 2 The mainly comparison of the transcriptome landscapes between the dHGP and rHGP. dHGP vs rHGPin metastatic liver lesions (n = 47 vs n = 43) dHGP vs rHGPin primary colorectal lesions (N = 20 vs N = 15) dHGP vs rHGP tissues (N = Frontiers in Immunology frontiersin.org 12 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 10.3389/ﬁmmu.2022.1045329 to the non-inﬂammatory soil or the inﬂammatory soil acquires seeds with enhanced metabolisms. However, we must admit that the hypothesis of the formation different HGPs is a bit far-fetched. For the C-score and Ln-score, only genes that were also signiﬁcantly different in the lesions are included, which meant we seemed to know that they were different between the HGP’s in liver metastases. All data should be seen as hypothesis-generating and we need validation data to prove the hypothesis. Publisher’s note The studies involving human participants were reviewed and approved by Sun Yat-sen University cancer center’s medical ethics committee. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Exploration of the predictive value of hepatic metastasis HGP type and transcriptome subtype for the survival of CRLM patients In the future, single-cell transcriptome spatial atlas and multiplexed tissue imaging are warranted to verify our ﬁndings further. Fourth, Some details are not well grasped. How normal were the areas selected for RNA-seq actually, and how far were these areas from the liver metastases? An array of technologies that offer improved resolution, often at the single cell level, including multiplexed FIGURE 7 Hypothesis for regulation of molecular expressions for HGPs. HGPs are observable morphological changes resulting from the regulation of molecular expressions, which is the combined effect of the heterogeneity and remodeling of primary tumors seeds and liver soils. FIGURE 7 Hypothesis for regulation of molecular expressions for HGPs. HGPs are observable morphological changes resulting from the regulation of molecular expressions, which is the combined effect of the heterogeneity and remodeling of primary tumors seeds and liver soils. frontiersin.org 13 Frontiers in Immunology Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 immunoﬂuorescence, single-cell RNA sequencing, and spatial transcriptional proﬁling, are now available and would be more appropriate analysis approaches than bulk RNA sequencing, which is confounded by differing abundances of different cell populations. In the future, we could use these techniques to further investigate. ZZP and DSW contributed to sample collection and data collection. All authors contributed to the article and approved the submitted version. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ ﬁmmu.2022.1045329/full#supplementary-material Funding In conclusion, We uncovered that the histopathological growth patterns of liver metastases were observable morphological changes arising from the regulation of molecular expressions through the combined effect of the heterogeneity of primary tumors (the seeds) and liver remodeling (the soil) (Figure 7). This research was supported by the National Natural Science Foundation of China (No. 81872010, 81874070). Author contributions YHL, MSC and ZMD were involved in designing study, supervising the work, and data interpretation. MTH, ZGC and DDH were involved in protocol development, data analysis, and wrote the original draft. SYX, and MTH evaluated the HGPs. DSW, XLZ, WPF, LW and YYC were involved in data collection and study design. DDH, YFY, BKL, XJW, ZHL, CG, LRL, PRD, Data availability statement We thank all the patients participating in this study. We thank Dr. Lei Fan for his help in drawing the diagram. All the raw data was deposited into Research Data Deposit (https://www.researchdata.org.cn/) with RDD number RDDB2022810006, and will be made available on request by the authors without undue reservation. Further inquiries can be directed to the corresponding authors. Conﬂict of interest The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 6. Van Dam PJ, Van Der Stok EP, Teuwen LA, Van den Eynden GG, Illemann M, Frentzas S, et al. International consensus guidelines for scoring the histopathological growth patterns of liver metastasis. Br J Cancer (2017) 117 (10):1427–41. doi: 10.1038/bjc.2017.334 4. Nielsen K, Rolff HC, Eefsen RL, Vainer B. The morphological growth patterns of colorectal liver metastases are prognostic for overall survival. Mod Pathol (2014) 27(12):1641–8. doi: 10.1038/modpathol.2014.4 5. Falcao D, Alexandrino H, Caetano Oliveira R, Martins J, Ferreira L, Martins R, et al. Histopathologic patterns as markers of prognosis in patients undergoing hepatectomy for colorectal cancer liver metastases - pushing growth as an independent risk factor for decreased survival. Eur J Surg Oncol (2018) 44 (8):1212–9. doi: 10.1016/j.ejso.2018.03.023 4. Nielsen K, Rolff HC, Eefsen RL, Vainer B. The morphological growth patterns of colorectal liver metastases are prognostic for overall survival. Mod Pathol (2014) 27(12):1641–8. doi: 10.1038/modpathol.2014.4 5. Falcao D, Alexandrino H, Caetano Oliveira R, Martins J, Ferreira L, Martins R, et al. Histopathologic patterns as markers of prognosis in patients undergoing hepatectomy for colorectal cancer liver metastases - pushing growth as an independent risk factor for decreased survival. Eur J Surg Oncol (2018) 44 (8):1212–9. doi: 10.1016/j.ejso.2018.03.023 6. Van Dam PJ, Van Der Stok EP, Teuwen LA, Van den Eynden GG, Illemann M, Frentzas S, et al. International consensus guidelines for scoring the histopathological growth patterns of liver metastasis. Br J Cancer (2017) 117 (10):1427–41. doi: 10.1038/bjc.2017.334 1. Leporrier J, Maurel J, Chiche L, Bara S, Segol P, Launoy G, et al. A population-based study of the incidence, management and prognosis of hepatic metastases from colorectal cancer. Br J Surg (2006) 93(4):465–74. doi: 10.1002/ bjs.5278 References 1. Leporrier J, Maurel J, Chiche L, Bara S, Segol P, Launoy G, et al. A population-based study of the incidence, management and prognosis of hepatic metastases from colorectal cancer. Br J Surg (2006) 93(4):465–74. doi: 10.1002/ bjs.5278 2. De Jong MC, Pulitano C, Ribero D, Strub J, Mentha G, Schulick RD, et al. Rates and patterns of recurrence following curative intent surgery for colorectal liver metastasis: an international multi-institutional analysis of 1669 patients. Ann Surg (2009) 250(3):440–8. doi: 10.1097/SLA.0b013e3181b4539b 3. Nierop PMH, Hoppener DJ, Van Der Stok EP, Galjart B, Buisman FE, Balachandran VP, et al. Histopathological growth patterns and positive margins after resection of colorectal liver metastases. HPB (Oxford) (2020) 22(6):911–9. doi: 10.1016/j.hpb.2019.10.015 frontiersin.org Frontiers in Immunology 14 Hu et al. Hu et al. 10.3389/ﬁmmu.2022.1045329 7. Vermeulen PB, Colpaert C, Salgado R, Royers R, Hellemans H, Van Den Heuvel E, et al. Liver metastases from colorectal adenocarcinomas grow in three patterns with different angiogenesis and desmoplasia. J Pathol (2001) 195(3):336– 42. doi: 10.1002/path.966 desmoplastic colorectal liver metastasis. Br J Cancer (2020) 123(2):196–206. doi: 10.1038/s41416-020-0881-z 20. Koelzer VH, Lugli A. The tumor border conﬁguration of colorectal cancer as a histomorphological prognostic indicator. Front Oncol (2014) 4:29. doi: 10.3389/ fonc.2014.00029 8. Temido MJ, Caetano Oliveira R, Martins R, Serodio M, Costa B, Carvalho C, et al. Prognostic factors after hepatectomy for gastric adenocarcinoma liver metastases: Desmoplastic growth pattern as the key to improved overall survival. Cancer Manag Res (2020) 12:11689–99. doi: 10.2147/CMAR.S264586 21. Stessels F, Van Den Eynden G, Van Der Auwera I, Salgado R, Van den Heuvel E, Harris AL, et al. Breast adenocarcinoma liver metastases, in contrast to colorectal cancer liver metastases, display a non-angiogenic growth pattern that preserves the stroma and lacks hypoxia. Br J Cancer (2004) 90(7):1429–36. doi: 10.1038/sj.bjc.6601727 9. Bohlok A, Vermeulen P, Leduc S, Latacz E, Botzenhart L, Richard F, et al. Association between the histopathological growth patterns of liver metastases and survival after hepatic surgery in breast cancer patients. NPJ Breast Cancer (2020) 6 (1):64. doi: 10.1038/s41523-020-00209-1 22. Lin HC, Shao Q, Liang JY, Wang Y, Zhang HZ, Yuan YF, et al. Primary tumor immune score fails to predict the prognosis of colorectal cancer liver metastases after hepatectomy in Chinese populations. Ann Trans Med (2021) 9 (4):310. doi: 10.21037/atm-20-4932 10. Barnhill R, Vermeulen P, Daelemans S, van Dam PJ, Roman-Roman S, Servois V, et al. COPYRIGHT © 2022 Hu, Chen, Hu, Xi, Wang, Zhang, Fong, Wen, Cai, Yuan, Li, Wu, Lu, Chen, Li, Ding, Pan, Wan, Du, Chen and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. References Replacement and desmoplastic histopathological growth patterns: A pilot study of prediction of outcome in patients with uveal melanoma liver metastases. J Pathol Clin Res (2018) 4(4):227–40. doi: 10.1002/cjp2.105 23. Robinson MW, Harmon C, O'farrelly C. Liver immunology and its role in inﬂammation and homeostasis. Cell Mol Immunol (2016) 13(3):267–76. doi: 10.1038/cmi.2016.3 11. Barnhill R, Van Dam PJ, Vermeulen P, Champenois G, Nicolas A, Rawson RV, et al. Replacement and desmoplastic histopathological growth patterns in cutaneous melanoma liver metastases: frequency, characteristics, and robust prognostic value. J Pathol Clin Res (2020) 6(3):195–206. doi: 10.1002/cjp2.161 24. Oliveira RC, Alexandrino H, Cipriano MA, Alves FC, Tralhao JG. Predicting liver metastases growth patterns: Current status and future possibilities. Semin Cancer Biol (2021) 71:42–51. doi: 10.1016/j.semcancer.2020.07.007 12. Haas G, Fan S, Ghadimi M, De Oliveira T, Conradi LC. Different forms of tumor vascularization and their clinical implications focusing on vessel Co-option in colorectal cancer liver metastases. Front Cell Dev Biol (2021) 9:612774. doi: 10.3389/fcell.2021.612774 25. Steeg PS. Targeting metastasis. Nat Rev Cancer (2016) 16(4):201–18. doi: 10.1038/nrc.2016.25 26. Paget S. The distribution of secondary growths in cancer of the breast. 1889. Cancer metastasis Rev (1989) 8(2):98–101. 13. Frentzas S, Simoneau E, Bridgeman VL, Vermeulen PB, Foo S, Kostaras E, et al. Vessel co-option mediates resistance to anti-angiogenic therapy in liver metastases. Nat Med (2016) 22(11):1294–302. doi: 10.1038/nm.4197 27. Guinney J, Dienstmann R, Wang X, de Reynies A, Schlicker A, Soneson C, et al. The consensus molecular subtypes of colorectal cancer. Nat Med (2015) 21 (11):1350–6. doi: 10.1038/nm.3967 14. Liang JY, Xi SY, Shao Q, Yuan YF, Li BK, Zheng Y, et al. Histopathological growth patterns correlate with the immunoscore in colorectal cancer liver metastasis patients after hepatectomy. Cancer Immunol Immunother (2020) 69 (12):2623–34. doi: 10.1007/s00262-020-02632-6 28. Nierop PM, Hoppener DJ, Buisman FE, van der Stok EP, Galjart B, Balachandran VP, et al. Preoperative systemic chemotherapy alters the histopathological growth patterns of colorectal liver metastases. J Pathol Clin Res (2021) 8(1):48–64. doi: 10.1002/cjp2.235 15. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinf (2013) 14:7. doi: 10.1186/1471-2105-14-7 29. Wu Y, Yang S, Ma J, Chen Z, Song G, Rao D, et al. Spatiotemporal immune landscape of colorectal cancer liver metastasis at single-cell level. Cancer Discovery (2021) 12(1):134–53. doi: 10.1158/2159-8290.Cd-21-0316 29. Wu Y, Yang S, Ma J, Chen Z, Song G, Rao D, et al. References Spatiotemporal immune landscape of colorectal cancer liver metastasis at single-cell level. Cancer Discovery (2021) 12(1):134–53. doi: 10.1158/2159-8290.Cd-21-0316 16. Sotiriou C, Wirapati P, Loi S, Harris A, Fox S, Smeds J, et al. Gene expression proﬁling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst (2006) 98(4):262–72. doi: 10.1093/jnci/djj052 17. Pitroda SP, Khodarev NN, Huang L, Uppal A, Wightman SC, Ganai S, et al. Integrated molecular subtyping deﬁnes a curable oligometastatic state in colorectal liver metastasis. Nat Commun (2018) 9(1):1793. doi: 10.1038/s41467-018-04278-6 18. Pitroda SP, Chmura SJ, Weichselbaum RR. Integration of radiotherapy and immunotherapy for treatment of oligometastases. Lancet Oncol (2019) 20(8):e434– e42. doi: 10.1016/S1470-2045(19)30157-3 19. Hoppener DJ, Nierop PMH, Hof J, Sideras K, Zhou G, Visser L, et al. Enrichment of the tumour immune microenvironment in patients with Frontiers in Immunology frontiersin.org 15
https://openalex.org/W2316825783	https://journals.vilniustech.lt/index.php/TEDE/article/download/4340/3685	English	null	TECHNOLOGY TRANSFER IN CLEAN DEVELOPMENT MECHANISM (CDM) PROJECTS: LESSONS FROM CHINA	Technological and economic development of economy	2,014	cc-by	11,979	Technological and economic development OF ECONOMY ISSN 2029-4913 print/ISSN 2029-4921 online 2013 Volume 19(Supplement 1): S471–S495 doi:10.3846/20294913.2013.879751 Technological and economic development OF ECONOMY ISSN 2029-4913 print/ISSN 2029-4921 online 2013 Volume 19(Supplement 1): S471–S495 doi:10.3846/20294913.2013.879751 Technological and economic development OF ECONOMY ISSN 2029-4913 print/ISSN 2029-4921 online 2013 Volume 19(Supplement 1): S471–S495 doi:10.3846/20294913.2013.879751 Received 11 November 2012; accepted 02 November 2013 Received 11 November 2012; accepted 02 November 2013 Abstract. China has become the largest host country of Clean Development Mechanism (CDM) in the world. This article provides an assessment of international technology transfer (TT) based on 500 registered Chinese CDM projects. It reveals that the projects hosted by large state-owned enterprises (SOEs), not Hydro and Wind projects, with foreign consultants or developers, commonly involve TT. Projects located in the comparatively developed regions such as Eastern China are more likely to involve TT. The findings indicate that the mitigation potential of non-SOEs, energy efficiency (EE) and other projects, has not been fully explored in China, which can be facilitated using advanced mitigation technologies. Keywords: Clean Development Mechanism, mitigation technology, technology transfer, China. Reference to this paper should be made as follows: Xie, L.; Zeng, S.; Zou, H.; Tam, V. W. Y.; Wu, Z. 2013. Technology transfer in Clean Development Mechanism (CDM) projects: lessons from China, Technological and Economic Development of Economy 19(Supplement 1): S471–S495. Reference to this paper should be made as follows: Xie, L.; Zeng, S.; Zou, H.; Tam, V. W. Y.; Wu, Z. 2013. Technology transfer in Clean Development Mechanism (CDM) projects: lessons from China, Technological and Economic Development of Economy 19(Supplement 1): S471–S495. JEL Classification: O33, P48, Q55, Q56. JEL Classification: O33, P48, Q55, Q56. Linna XIEa, Saixing ZENGa, Hailiang ZOUa, Vivian W. Y. TAMb, Zhenhua WUa a XIEa, Saixing ZENGa, Hailiang ZOUa, Vivian W. Y. TAMb, Zhenhua WUa a Antai School of Management, Shanghai Jiaotong University, 200052 Shanghai, China b School of Computing, Engineering and Mathematics, University of Western Sydney, Locked Bag 1797, NSW 2751 Penrith, Australia Introduction With the rapid economic growth, anthropogenic greenhouse gas (GHG) emissions from developing countries have become one of the main concerns around the world (La Rovere et al. 2011; Štreimikiene, Esekina 2008). Technology transfer (TT) is expected to play an important role in mitigating GHG emissions for developing countries (Halsnæs, Garg 2011; Kim et al. 2011; Marconi, Sanna-Randaccio 2011; Schneider et al. 2008). Clean Corresponding author Saixing Zeng E-mail: zengsaixing@sjtu.edu.cn Copyright © 2013 Vilnius Gediminas Technical University (VGTU) Press http://www.tandfonline.com/TTED S472 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... Development Mechanism (CDM) is one of the international instruments facilitating such transfers (Gangale, Mengolini 2011).hi The CDM fosters sustainable developments by channelling new financial resources to promote the use of technologies currently not available in the host countries (Reynolds 2012; Schneider et al. 2008). As one of the largest CO2 emission countries and one of the fastest developing countries, China and its government face the dilemma between economic growth and environmental conservation similar as most of the developing countries (Ward, Shively 2011). China has made use of CDM, not only to receive financial assistance, but also to obtain advanced technologies from developed countries (Wang 2010). Currently, China has become the largest CDM host country in the world (Richerzhagen, Scholz 2008). The number of registered and registering projects hosted in China is about 48% of all 7,520 re- gistered and registering projects around the world (UNFCCC 2013). The expected average annual certified emission reductions (CERs) of Chinese registered projects are about 64% of the total 905,682 ktCO2e per year (UNFCCC 2013). h The TT claims for CDM projects have been extensively studied (Dechezleprêtre et al. 2008; Seres et al. 2009; Wang 2010). However, this research is different from the previous studies in several important aspects. First, it attempts to differentiate the effects of CDM on TT, including: (1) form of ownership of Chinese participants; (2) foreign participants more than credit buyers; and (3) regional disparities of China. Second, it investigates the trends beyond individual projects and individual aspects. Third, it identifies several aspects that can be improved in the Chinese GHG mitigation activities. The aims are to gain an insight into the mitigation of TT and further understanding in the CDM at its current or modified form in the new negotiation stage. 1. Literature review The Intergovernmental Panel on Climate Change (IPCC) defined TT as “a broad set of processes covering flows of know-how, experience and equipment for mitigating and ad- apting climate change amongst different stakeholders such as governments, private sector entities, financial institutions, non-governmental organizations (NGOs) and research/ education institutions’’ (IPCC 2000). Technologies consist of not only the “hardware” such as machineries and equipment, but also the “software” including knowledge, skills, know-how, management arrangements and goods or services (Tébar Less, McMillan 2005). A project can involve both hardware and software (Dechezleprêtre et al. 2008). In CDM, “technology transfer” reported in the project design documents (PDDs) are not based on a specific or identical definition, but based on the interpretation of TT by project parti- cipants. According to former research, in general, it can be assumed that TT means “the use of equipment and/or knowledge not previously available in the host country” (Haites et al. 2006; Seres, Haites 2008; Seres et al. 2009). An extensive body of literature has reported economic, political, methodological and sustainable development aspects of the CDM project performance (Olsen 2007; Schneider et al. 2008). Currently, there are two main streams of literature studying TT in CDM projects. The first stream reports empirical analyses which study TT in CDM projects using the data S473 ogical and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 from the UNEP Risoe Center CDM Pipeline and the project design documents (PDDs), which explore TT on parameters such as project sizes, project types, host countries, tech- nology suppliers, local technology capabilities and partnerships. The second stream collects information from interviews, case studies, policy documents, government information and other sources, using qualitative approaches rather than quantitative, giving us a wide open mind and a big map of TT in CDM projects.iih In the studies of the first stream, some significant results are found. The data show that TT takes place in less than half of the CDM projects (Dechezleprêtre et al. 2008; Seres, Haites 2008; Seres et al. 2009, 2010). Large projects are more likely to involve TT than unilateral or small-scale projects (Dechezleprêtre et al. 2008; Haites et al. 2006; Seres et al. 2010). Slightly varying with the samples, TT possibility is high in agriculture, energy efficiency (EE) own generation, landfill gas, N2O, HFCs and wind projects, and low in biomass energy, cement, fugitive, hydro, and transportation projects (Das 2011; Dechezleprêtre et al. 1. Literature review 2008; Seres, Haites 2008; Seres et al. 2009). This reflects the variation of average sizes under different project types and technology characteristics. As more projects of a given type are located in the host country, subsequent projects rely more on local knowledge and equipment than imported technologies (Dechezleprêtre et al. 2008; Seres et al. 2009, 2010). Meanwhile, countries with more experience in the development and applications of mitigation technologies tend to rely on domestic technologies or developing technologies accompanied by foreign partners (Doranova et al. 2010). These indicate that the well-developed local technology facilitates the use of foreign technologies and implies availability of technology locally (Dechezleprêtre et al. 2008, 2009). A host country can influence TT in CDM through their criteria for approving CDM projects and other factors such as tariff, and protection of intellectual property rights (Seres, Haites 2008; Seres et al. 2009, 2010). Projects hosted by the subsidiary of a foreign company or with foreign consultants are more inclined to involve TT (Das 2011; Dechezle- prêtre et al. 2008, 2009; Doranova et al. 2010). In the second stream, insights about TT in CDM projects are displayed. Four TT bar- riers: (1) lack of commercial viability; (2) lack of access to capital; (3) lack of information; and (4) lack of institutional framework are identified and the CDM does contribute to TT by lowering these barriers except the last one (Schneider et al. 2008). The four barriers are used by other researchers to explain project characteristics’ effects on TT in CDM projects (Dechezleprêtre et al. 2009; Seres et al. 2009). Besides the four barriers, local technology capability is identified by Schneider et al. (2008) to explain the TT distribution across geo- graphies and project types. But they could not identify low local technology capability as the fifth barrier of TT in CDM projects. Local technology capability already can be measured by standard index, such as the ArCo technology index (Dechezleprêtre et al. 2008, 2009), or a set of indicators specially made for research purposes (Doranova et al. 2010). Although the studies focusing on CDM projects hosted in China are of paucity, some features of TT in CDM projects in China still could be found. 2.2. Project types The TT claim varies widely across project types (Das 2011). On average, there are 40% projects claiming TT, but the share of TT ranges from 13% to 100% across different project types (Seres et al. 2010).fi First, industrial regulations and policies directly affect TT in specific project types. For example, the Chinese government implements the large-scale wind farm plan, requiring about 70% local contents for the eligibility of concession bidding, which significantly spurs the localization of wind turbine manufacture (Wang 2010). Second, as the number of CDM projects in a given type grows in a host country, TT probability will gradually be reduced in later projects (Dechezleprêtre et al. 2008; Seres et al. 2009, 2010). It can be suggested that TT is more likely to happen in some project types than others. Hence, we propose that: y j y Hypothesis 2: The attribute of project types is positively associated with TT in CDM. 2.1. Project sizes According to previous empirical research, project sizes are one of the most important factors that influence the possibility of TT in CDM projects. It is suggested that the larger the project sizes, the more possible the project involving TT (Dechezleprêtre et al. 2008, 2009; Doranova et al. 2010; Haites et al. 2006; Seres et al. 2010). Therefore:h Hypothesis 1: The project size is positively associated with TT in CDM. 1. Literature review When the certified emission reduction (CER) income is low and most of the technologies are locally available, time effect, technology diffusion, governmental involvement, and investors’ and brokers’ participation play an important role in deciding TT (Wang 2010). The domestic regulations and policies have important influences on the CDM projects, e.g. CDM projects are heavily concentrated S474 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... in government priority areas such as renewable energy and energy efficiency in industrial applications (Marconi, Sanna-Randaccio 2011). It also points out that China’s technology localization strategies will eventually reduce TT in CDM projects and advance the level of technologies adopted (Wang 2010). Overall, this subject is not sufficiently discussed by previous researchers. This research follows the empirical analyses from the first stream and also inspired by the findings of the second stream researches. 2.3. Form of ownership of Chinese participants It is restricted by National Development and Reform Commission (NDRC) and three other ministries of Chinese government that only wholly-owned Chinese companies or Chinese holding companies are eligible for CDM projects in China (NDRC et al. 2005, 2011). There must be at least 51% of the company owned by Chinese entities (Wang 2010). In this condition, the CDM owners in China can be classified into three types: state-owned enterprises (SOEs), collective-owned enterprises (COEs) and private-owned enterprises (POEs) (Nee 1992). In China, although the SOEs undertake more complex socio-economic mission (Nolan 2001) and more commonly-faced multiple tasks than non-SOEs (Bai et al. 2000, 2006), they have the facilities in accessing capital resources (Brandt, Li 2003) and gaining political support from the state and local governments (Li, Zhou 2005). The COEs have fewer advantages than S4 nological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 SOEs, but are still ranked higher in terms of accessing political and financial support than POEs (Poncet et al. 2010). They have structural advantages over both SOEs and POEs (Xia et al. 2009; Xin, Pearce 1996), because they are affiliated with and are able to gain protection from the local government (Kung, Lin 2007; Nee 1992; Peng et al. 2004) and meanwhile they sell products in competitive markets that encourage efficiency (Kornai 1986; Kung, Lin 2007). COEs did benefit from their structural advantages in the early years of reformation (Xia et al. 2009). After undergoing a decline in the mid-1990s as the SOEs (Jefferson, Su 2006), COEs have experienced transformations under the central government policy since 1995, including both privatization and corporatization (Lin, Zhu 2001). In the transformation times, local governments may still play an important role in dealing with the agency problems as the controllers of COEs, and the close relationship with the government becomes an obstacle rather than advantages in improving firm performance (Xia et al. 2009; Zeng et al. 2012). The POEs gain less local policy support from the government and having weak influential power (Xin, Pearce 1996). They cannot enter into certain industries, and are commonly with less tax relief (ADB 2002; Ralston et al. 2006). They are harder to obtain loans from state-owned banks (Poncet et al. 2010), have less access to market information and have more problems in getting land which is owned by the state and other resources from the government (ADB 2002; Gregory et al. 1 According to Seres (2010) and a list of Annex I counties available on UNFCCC website (http://unfccc.int/ parties_and_observers/parties/annex_i/items/2774.php) which include the United States and Canada. This paper will include the United States and Canada for the analysis of the Annex I counties. 2.3. Form of ownership of Chinese participants 2000; Ralston et al. 2006). They can also be large and sophisticated, with much more discretion on hiring, firing people, and exercising market responses than SOEs (Pyke et al. 2000). The alternative resources, such as reputation and relationships, are used by POEs as alternative financing channels and governance mechanisms to overcome the imbalance among the three sectors (Allen et al. 2005). It is also suggested that the situation is changing over time, the institutional and market infrastructure inspiring both SOEs and non-SOEs are starting to be established in China (Carney et al. 2009). In general, the more support they could gain from the government and the stronger their financial power, the easier they can lower the four barriers of TT. It can be supposed that TT is likely to happen in SOEs, less likely in COEs, and least likely in POEs. It is hypothesized that:h Hypothesis 3: The form of ownership of Chinese participants is associated with TT in CDM. 3 See in the PDDs download from UNFCCC website using the titles of sampled projects. If there is no statement which confirms TT in “Section A.4.3.”, the keywords “technology”, “technologies”, “transfer”, “equipment”, “service”, “train”, “serve” “supplier”, “import”, “manufacturer” are used to make this process more efficient. If no information is found, the whole PDDs are covered and been read accordingly. If there is any equipment and/ or knowledge from another country involved, it is recorded as involving TT. 2.5. Disparity of host regions In this paper, China is the host country of the investigated projects. In China, the degree of reform and openness, geographical locations and infrastructure investments significantly affect economic growth performance across provinces (Démurger 2001). The heavy industry development strategies in China formed a rural-urban gap in the pre-reform period, while openness and decentralization induce and exacerbate the inland-coastal disparity in the reform period (Kanbur, Zhang 2005). The regional disparity of China may affect TT in CDM projects by lowering barriers on commercial viability, accessing to capital, information and institutional framework. First, the CDM projects located in the economically developed regions can take advantage of act- ive business activities and facility of access to capital. Second, the comparatively developed regions in China have better local technology capability (Qi et al. 2012), as they have better educated engineers and better trained skilled workers, and more financial support on research and development. High technology capabilities are necessary to adopt new technologies, but it also implies that the technologies needed may already be available in the local market (Dechezleprêtre et al. 2008, 2009). It is hypothesized that: Hypothesis 5: The regional disparity is positively associated with TT in CDM. 2.4. Foreign participants Cooperating with foreign companies can increase the possibility of TT in CDM projects. Beside technology suppliers, there are at least four types of foreign participants, such as project developers, financers, consultants and credit buyers. One foreign entity can play one or more roles at one time in a CDM project. The project developer clearly favours TT, if it is the subsidiary of a company from an Annex I country1 (Das 2011; Dechezleprêtre et al. 2008, 2009; Doranova et al. 2010). The subsidiaries have stronger effects on TT than credit buyers (Dechezleprêtre et al. 2008). And, the involvement of foreign consultants can also increase the possibility of TT in CDM projects (Das 2011). Thus: S476 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... Hypothesis 4: The Involvement of foreign companies is positively associated with TT in CDM. 2 CDM pipeline can be downloaded from the website of UNEP RISØ CENTRE (http://cdmpipeline.org/) 3.2.1. Technology transfer The projects in number and estimated annual emission reductions can be used as crude proxies to measure TT (Seres, Haites 2008). This paper analyses the TT claims made by CDM projects participants in their PDDs. In the following regression analysis, if any foreign equipment and/or knowledge are employed in the project activities, a value of ‘1’ will be assigned, ‘0’ otherwise. Overall, there are 195 projects out of the 500 CDM projects (about 39%) claimed TT. These 39% projects account for almost 61% of the estimated annual emission reductions which are 49,346 ktCO2e per year out of all 80,948 ktCO2e per year of the sampled 500 pro- jects. The average size of projects involving TT and not involving TT are about 253.06 and 103.61 ktCO2e per year, significantly different under independent-samples T test, p < 0.01. 3.1. Data collection In this paper, 500 recently-registered projects were chosen from the CDM pipeline by the end of 2010. There are 27 projects without credit buyers (unilateral project), or located in more than one province of China which were excluded and replaced by another 27 re- cently-registered projects. The information about host regions, project types, credit buyers, consultants, methods and planned annual reductions were listed in the CDM pipeline2. TT claims made by CDM projects participants can be seen in “Section A.4.3. Technology to be employed by the project activity” of their PDDs and other possible parts of PDDs are also covered in “Section A.2.”, “Section A.4.2.”, “Section B.5.”, “Section B.7.”, “Section E.”, “Annex”, et al.3 Although a recent survey suggests that the actual rate of TT may be higher than it is reported in PDDs (Kirkman et al. 2013), this research follows the former empirical analyses using the TT information from PDDs, and only the projects with information confirming S4 nological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 TT are coded as involving TT. The form of ownership of Chinese participants and foreign project participants are verified by searching the Internet with the information of project owners and foreign participants in “Section A.3. Project participants” of PDDs. 4 The EE projects include 25 EE own generation and 2 EE households projects. The rest 65 projects include 13 landfill gas, 12 biomass energy, 11 coal bed/mine methane, 8 fossil fuel switch, 7 N2O, 5 methane avoidance, 3 solar, 2 cement, 2 reforestation, 1 fugitive and 1 HFCs project. The four categories are classified with the limit of the sample. 3.2.2. Project sizes In this paper, the project size is measured by the estimated annual emission reductions in terms of ktCO2e per year. The projects are classified into small-scale and large-scale based on the methodologies used in calculating emission reductions. There are two projects used both small- and large-scale methodologies, which are classified as large-scale projects. Both the natural log of the estimated annual emission reductions and the classification of small-scale and large-scale projects based on the methodology of calculating emission reductions are used as the proxies of project size in the following regression analysis. The results of project size analysis are shown in Table 1. Table 1. Project size analysis Size Number of projects % Annual emission reductions % Number of TT projects % TT projects as percentage of Average size (ktCO2e per year) Number of projects% Annual emission reductions % Small-scale 25.40 5.35 11.28 17.32 19.89 34.093 Large-scale 74.60 94.65 88.72 46.38 63.28 205.412 Total 100.00 100.00 100.00 39.00 60.96 161.897 Table 1. Project size analysis Table 1 indicates that over 74% of projects are large-scale, accounting for more than 94% of the annual emission reductions. More than 88% TT happen in large-scale projects. About 46% large-scale projects and only 17% small-scale projects involve TT, which means large- scale projects are indeed more likely to involve TT than small-scale projects. In large-scale S478 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... projects, about 46% projects involve TT accounting for about 63% annual emission reduc- tions of all large-scale projects. In small-scale projects, about 17% projects involve TT which account for about 20%. The average sizes of small-scale and large-scale projects are 34.09 and 205.41 ktCO2e respectively per year, significantly different under independent-samples T test, p < 0.001. 3.2.3. Project types According to UNFCCC, the CDM projects can be classified into 26 different project types, of which 15 are involved in this study. Limited by this sample, they are classified into four categories: 252 hydro projects, 156 wind projects, 27 EE projects and 65 other projects.4 Three dummy variables are used to measure Hydro, EE and the Others with Wind as the base cat- egory. If the project is from one of the three categories except Wind, a value ‘1’ is assigned, ‘0’ otherwise. The variable Similar which is natural log of the projects in number using the same technology of the 15 project types in China, is used as a proxy of local mitigation technology availability of China in the regression analysis. Project type analysis is reported in Table 2. Table 2. Project type analysis Type Number of projects % Annual emission reductions % Number of TT projects % TT projects as percentage of Average size (ktCO2e per year) Number of projects % Annual emission reductions % Hydro 50.40 36.28 32.31 25.00 40.93 116.547 Wind 31.20 25.65 36.92 46.15 46.06 133.118 EE 5.40 4.52 8.72 62.96 83.69 135.383 Others 13.00 33.55 22.05 66.15 90.95 417.796 Total 100.00 100.00 100.00 39.00 60.96 161.897 From Table 2, Hydro and Wind are the two main CDM project types in China, account for more than 81% projects. Hydro projects have the lowest rate of projects involving TT, only 25%. Though the TT rate of Wind projects drops in the recent years (Wang 2010), it still apportions about 46%. The EE and the Other projects have the highest TT rate, more than 63% and 66% of the projects involving TT and respectively account for about 84% and 91% of annual emission reductions. Except the high rate of TT in EE projects, the average size of EE projects is not much different with Hydro and Wind projects. The average size of the Other projects is significantly larger than Hydro, Wind and EE projects, tested by one-way ANOVA, p < 0.05. The Other projects category contains some very large projects such as HFCs and fossil fuel switch which all involve TT. Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 In the end of 2012, 3,479 CDM projects in China are registered. There are 1,445 Wind and 1,244 Hydro projects, which are 42% and 36% of the 3,479 projects. 3.2.3. Project types The other projects are the 191 EE own generation projects, 132 solar projects and 121 biomass energy projects. The rest projects are less than 100 in each project type. The Wind projects are significantly growing in China. In the end of 2010, there were only 338 Wind projects registered. In 2011 and 2012, 477 and 473 Wind projects started their first CDM comment, and 274 and 833 Wind projects were registered.5 The dynamic changes of Wind project located in China from 2005 to2012 are shown in Figure 1. 3.2.4. Form of ownership of Chinese participants The CDM project involves the effect of the ownership of the project owners. According to the entities from the Chinese side, they are classified into three basic types of ownerships including SOEs, COEs and POEs. Two dummy variables are used to measure different own- erships in the following regression analysis with COEs as the base category. If the project is from one of the two categories except COEs, a value of ‘1’ is assigned, ‘0’ otherwise. The results of ownership analysis are reported in Table 3. Table 3. Ownership analysis Owner- ship Number of projects % Annual emission reductions % Number of TT projects % TT projects as percentage of Average size (ktCO2e per year) Number of projects % Annual emission reductions % SOEs 73.40 73.15 80.51 42.78 61.52 161.343 COEs 17.60 11.73 9.23 20.45 29.48 107.867 POEs 9.00 15.12 10.26 44.44 82.71 272.069 Total 100.00 100.00 100.00 39.00 60.96 161.897 Table 3. Ownership analysis From Table 3, over 73% CDM project hosts are SOEs, accounting for the same amount of annual emission reductions. Over 80% TT projects are hosted by SOEs. But the projects hosted by POEs have the highest portion of projects involving TT, more than 44% and accounting for 83% annual emission reductions. The average size of the projects hosted by POEs is larger than projects hosted by SOEs, and significantly larger than projects hosted by COEs, tested by one-way ANOVA, p < 0.1. 5 Data are collected from CDM pipeline. URL: http://cdmpipeline.org/ 3.2.5. Foreign participants In this research, all the 500 samples are the projects with at least one credit buyer. There are 184 projects which involve foreign consultants in this sample. In 105 projects, the foreign consultants act as both the consultants and the credit buyers. There are 17 projects in which L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... S480 the foreign participants act as project developers or financers.6 Thirteen out of the 17 projects hire foreign consultants, the rest 4 projects hire local consultants, and in two projects the foreign consultants also act as credit buyers. All the 17 projects are classified into the category of projects with foreign project developers. All samples are classified into the four categories. They are consultants (including 68 projects), consultants and buyers (including 103 projects), project developers (including 17 projects) and the Other projects (including 312 projects). Three dummy variables are used to measure different foreign participants in the following regression analysis with the Other 312 projects as the base category. If the project is from one of the three categories except the Other projects, a value of ‘1’ is assigned, ‘0’ otherwise. The results of foreign participant analysis are reported in Table 4. Table 4. Foreign participant analysis Foreign participant Number of projects % Annual emission reductions % Number of TT projects % TT projects as percentage of Average size (ktCO2e per year) Number of projects % Annual emission reductions % Consultant 13.60 17.70 16.41 47.06 81.46 210.709 Consultant and buyer 20.60 23.46 27.69 52.43 70.46 184.370 Project developer 3.40 6.61 6.67 76.47 93.42 314.936 Other projects 62.40 52.23 49.23 30.77 45.64 135.500 Total 100.00 100.00 100.00 39.00 60.96 161.897 Table 4. Foreign participant analysis Except the technology suppliers and credit buyers, there are about 38% of the projects involve other types of foreign participants (consultants, consultants and buyers, and project developers). These projects have higher TT shares than the Other projects, and the TT projects in these three categories account for larger portions of emission reductions than the Other projects. Especially the projects in consultants and project developers categor- ies, the TT projects account for 81% and 93% of the annual emissions reductions in each category. The average sizes of the project with foreign participants are much larger than the Other projects. p j p j g pi 7 Data is collected from China Statistical Yearbook published on the website of National Bureau of Statistics of China (http://www.stats.gov.cn/tjsj/ndsj/). 6 In four of these 17 projects, the projects’ foreign developers are also the financers. 8 The ArCo technology index includes eight sub-indexes (a1 patens, a2 scientific articles, b1 internet penetration, b2 telephone penetration, b3 electricity consumption, c1 tertiary science and engineering enrolment, c2 mean years of schooling, c3 literacy rate) (Archibugi, Coco 2004). We can only acquire data of sub-indexes from 2009 and 2010. From the limited available data, we use the average of two years standardized indicators to compute the local technology capability. 3.2.6. Disparity of host regions In the host country, the regions or provinces have different economic development level and technology capabilities, which can influence TT in the projects. The GDP per capita (in 10 thousand of RMB) and the GDP growth are used as the proxies of local economic development, which are the average data of each province from 2006 to 2010.7 The research Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 S481 and development (R&D) investment (the average percentage of R&D per GDP from 2006 to 20107) of each province and the ArCo technology indexes of each province8 are used as the proxies of local technology capabilities. The variable Similar in province which is the natural log of the projects in number adopting the same type of technology in each province is used as the proxy of local mitigation technology availability in provinces in the following regression analysis. In this paper, the CDM projects involve 30 regions/provinces in China. As National Bureau of Statistics of China stated, all the regions/provinces can be classified into four economic regions (NBSC 2011) according to their geographical locations, economic development and institutional environment.9 The East is a region with a comparatively high level of economic and institutional development. The West is a region with comparatively low level of economic and institutional development. The development level of Midland is between the East and West. And the Northeast is a region of traditional heavy industry and large-scale agricultural production. There are 56% of the projects that are located in the West. The West region can be further classified into two sub-regions, which are the Mid-west (including Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan and Shaanxi) and the Wild-west (includ- ing Gansu, Xinjiang, Ningxia, Qinghai and Xizang). The results of host region analysis are reported in Table 5 and Table 6. Table 5. 9 NBSC (2011) stated that the East includes Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong and Hainan; the Midland includes Shanxi, Anhui, Jiangxi, Henan, Hubei and Hunan; and the West includes Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Shaanxi, Gansu, Qinghai, Ningxia and Xinjiang; the Northeast includes Liaoning, Jilin and Heilongjiang. 3.2.6. Disparity of host regions Host region analysis Region Number of projects % Annual emission reductions % Number of TT projects % TT projects as percentage of Average size (ktCO2e per year) Number of projects % Annual emission reductions % Wild-west 12.80 10.46 11.79 35.94 54.40 132.334 Mid-west 42.80 36.02 35.90 32.71 42.00 136.262 East 20.40 30.64 28.72 54.90 87.23 243.136 Midland 16.40 15.13 14.36 34.15 60.32 149.398 Northeast 7.60 7.74 9.23 47.37 55.37 164.958 Total 100.00 100.00 100.00 39.00 60.96 161.897 From Table 5, more than 42% of the CDM projects are located in the Mid-west and account for about 36% of the annual emission reductions. Only 20% of the projects are located in the East but account for 31% of the annual emission reductions. More than one third of the TT L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... S482 projects are located in the Mid-west, but the highest TT rates are in the East region. Over 54% of the projects in the East involve TT which account for 87% of the annual emission reductions in this region. The average size of projects located in the East is much larger than other regions. Table 6. Regional disparity of China China Min. Max. East Mid- land North- east Mid- west Wild- west GDP per capita 2.23 0.97 6.45 3.45 1.84 2.63 1.95 1.48 GDP growth 13.43 10.58 17.60 12.77 13.23 14.23 14.29 11.39 R&D 1.00 0.26 5.35 1.39 1.04 1.17 0.82 0.82 ArCo index 0.24 0.12 0.84 0.40 0.20 0.27 0.19 0.19 Similar in province 2.78 0.00 5.02 1.85 2.00 2.06 3.69 2.62 Table 6. Regional disparity of China Table 6 shows the regional disparity of China in economic development, local technology capabilities and local mitigation technology availability described by five variables. The GDP per capital of the East is 3.45 which is significantly higher than all other four regions. The GDP growth of the Wild-west is 11.39 which is significantly smaller than all other four regions. The R&D investment of the Mid-west and the Wild-west are both 0.82 which is significantly lower than other three regions. And the ArCo index of the East is 0.40 which is significantly higher than all other four regions. In the opposite, the values of the Similar in the provinces of the Mid-west and Wild-west are significantly larger than other three regions. All are tested by one-way ANOVA, p < 0.001. 3.2.6. Disparity of host regions h The dynamic changes of the distribution of Wind projects and the distribution of all CDM projects in China are shown in Figure 1. Most of the 1,445 registered Wind projects are located in Inner Mongolia, Hebei, Shandong, Liaoning, and Ningxia. The number of Wind projects was growing very fast in these regions during 2011 and 2012. The Wind projects located in other provinces are fewer than 100 in each province. Most of the registered CDM projects are located in the Mid-west, the Wild-west, the Bohai Rim (Shandong, Hebei and Liaoning) and Hunan province. 4. Results and analysis Table 7 shows correlations among variables. All the bivariate correlations are lower than the recommended 0.7 threshold. The variance inflation factors (VIFs) are below the recommended ceiling of 10 (Cohen et al. 2003). The individual variables can be used in the regression analysis. The dependent variable has values of either ‘0’ or ‘1’. The logistic regression analysis is used for the following regression analysis. The results are reported in Table 8. S483 ogical and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 Fig. 1. The distribution of wind projects in cumulated numbers from 2005 to 2012 and the all registered CDM projects in China Note: The numbers in the maps are the cumulated numbers of registered projects located in each province; these numbers are divided equally into eight ranks coloured by blue; the data is collected from the CDM pipeline. Fig. 1. The distribution of wind projects in cumulated numbers from 2005 to 2012 and the all registered CDM projects in China Note: The numbers in the maps are the cumulated numbers of registered projects located in each province; these numbers are divided equally into eight ranks coloured by blue; the data is collected from the CDM pipeline. istribution of wind projects in cumulated numbers from 2005 to 2012 and the all registered CDM projects in China mbers in the maps are the cumulated numbers of registered projects located in each province; these numbers are divided equally into eight ranks blue; the data is collected from the CDM pipeline. L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... S484 Table 7. Means, standard deviations, VIFs, and correlationsa Mean Std. 4. Results and analysis De- viation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 TT 1.000 2 Project sizes 11.41 0.95 0.337 1.000 3 Small-scale projects 0.25 0.44 –0.259 –0.663 1.000 4 Hydro 0.50 0.50 –0.289 –0.271 0.358 1.000 5 EE 0.05 0.23 0.117 –0.043 –0.017 –0.241 1.000 6 Others 0.13 0.34 0.215 0.233 –0.062 –0.390 –0.092 1.000 7 SOEs 0.73 0.44 0.126 0.137 –0.107 0.043 –0.048 –0.024 1.000 8 POEs 0.09 0.29 0.043 –0.076 0.087 –0.088 0.179 0.102 –0.546 1.000 9 Consultants 0.14 0.34 0.066 –0.045 0.077 0.020 0.112 0.090 –0.051 0.132 1.000 10 Consultants and credit buyers 0.21 0.40 0.140 0.094 –0.025 –0.108 0.010 –0.006 0.009 0.037 –0.202 1.000 11 Project devel- opers 0.03 0.18 0.144 0.103 –0.084 –0.167 –0.045 0.321 –0.006 0.008 –0.074 –0.096 1.000 12 GDP per capita 2.23 1.00 0.243 0.155 –0.185 –0.557 0.105 0.104 0.024 0.067 –0.028 0.110 –0.026 1.00 13 GDP growth 13.43 1.90 0.006 0.069 –0.124 –0.275 –0.061 –0.114 0.071 –0.122 –0.125 0.168 –0.046 0.407 1.000 14 Similar 5.15 1.27 –0.301 –0.230 0.133 0.536 –0.234 –0.676 0.037 –0.184 –0.128 0.020 –0.263 –0.226 0.037 1.000 15 R&D 1.00 0.50 0.145 0.010 0.091 –0.037 0.113 0.137 –0.005 0.076 0.089 –0.019 0.028 0.383 –0.147 –0.144 1.000 16 Local technol- ogy capability 0.24 0.12 0.228 0.113 –0.095 –0.420 0.138 0.174 –0.019 0.140 0.034 0.065 –0.018 0.606 0.100 –0.268 0.579 1.000 17 Similar in prov- ince 2.78 1.42 –0.227 –0.025 0.033 0.400 –0.246 –0.558 0.015 –0.150 –0.095 0.064 –0.215 –0.229 0.235 0.695 –0.305 –0.390 1.000 Variance Infla- tion Factors VIFs b 2.072 2.096 2.818 1.477 3.550 1.481 1.584 1.120 1.120 1.175 6.797 2.453 4.551 1.886 5.157 2.811 Notes: a N = 500; b None of these correlations exceed 0.70 and all of the VIFs are much less than the recommended maximum threshold of 10. S485 ogical and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 Table 8. 4. Results and analysis Results of logistic regression analyses Variables Model 1 Model 2 Model 3 Model 4 Model 5 Constants –8.329* (2.15) –6.061* (2.34) –5.039 (2.46) –7.975* (2.27) –5.689 (2.55) Project sizes 0.685* (0.17) 0.695* (0.17) 0.666* (0.17) 0.738* (0.18) 0.706* (0.18) Small-scale projects –0.382 (0.38) –0.385 (0.39) –0.432 (0.39) –0.369 (0.39) –0.396 (0.39) Project types: Hydro –0.323 (0.30) –0.132 (0.32) –0.046 (0.35) –0.072 (0.35) 0.025 (0.35) EE 0.891* (0.49) 0.343 (0.54) 0.298 (0.54) 0.648 (0.51) 0.280 (0.54) Others 0.463 (0.39) –0.608 (0.58) –0.617 (0.59) 0.110 (0.45) –0.629 (0.59) Ownerships: SOEs 0.921* (0.32) 0.866* (0.32) 0.848* (0.32) 0.858* (0.32) 0.832*** (0.32) POEs 0.829* (0.45) 0.653 (0.46) 0.698 (0.46) 0.832* (0.46) 0.704 (0.46) Foreign participants: Con- sultants 0.689 (0.31) 0.661 (0.31) 0.667 (0.31) 0.725 (0.32) 0.694 (0.32) Consultants and credit buy- ers 0.849* (0.27) 0.913* (0.27) 0.954* (0.28) 0.935* (0.27) 0.974* (0.28) Project developers 1.522 (0.66) 1.487 (0.66) 1.445 (0.67) 1.384 (0.67) 1.402 (0.67) GDP per capita 0.469* (0.14) 0.472* (0.15) 1.143 (0.49) 1.269 (0.49) 1.262 (0.50) GDP growth –0.151 (0.06) –0.159 (0.06) –0.232 (0.09) –0.197 (0.09) –0.222 (0.09) Similar –0.419 (0.17) –0.454* (0.17) –0.369 (0.19) R&D 0.547* (0.31) 0.447 (0.30) 0.513* (0.31) Local technology capability –6.621* (3.92) –7.952** (4.03) –7.785* (4.09) Similar in province –0.248** (0.12) –0.136 (0.13) Pseudo R2 0.3479 0.3577 0.3642 0.3597 0.3658 Correctly classified 81.00% 82.40% 82.80% 81.60% 82.80% Notes: Standard errors are in parentheses; N = 500; * p < 0 .10; p < 0.05; * p < 0 .01. L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... S486 Table 9. 10 The average of five years’ data is used, which from 2006 to 2010. Most of the data is from China Statistical Yearbook published on the website of National Bureau of Statistics of China. URL: http://www.stats.gov.cn/tjsj/ ndsj/. The fossil fuel consumption data is from China Energy Yearbook 2011. 4. Results and analysis S487 Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 In Model 4, the variable Similar is substituted by Similar in province. The result of Model 1 and Model 2 are stable. The TT possibility is significantly negative related to Similar in province. In Model 5, both Similar and Similar in province are added. The significance of Similar in province becomes weak. Other results are stable. In Table 9, the regression models only include 156 Wind projects. The relation between TT and Similar in province becomes not significant, but the Similar is still significantly related to TT. p gi gi y Robust tests are conducted. When the GDP per capita is substituted by log of GDP, log of FDI, or log of export and import, the results of the regression models are stable. And it is shown in Model 1 that Hydro projects are significantly less inclined to involve TT than Wind projects. When the Local technology capability is substituted by the percentage of tertiary educated local population, the results of the regression models are stable. Other influence of regional disparity have also been considered, such as the percentage of fossil fuel in local energy consumption, the local resource consumption (the log of energy consumption and the log of water consumption), the local pollution control investment, and the local emission compliance (the pollution control percentage of industrial wastewater discharge, industrial and domestic emissions, industrial solid waste disposal, and domestic rubbish disposal).10 The variables are added in the regression model, the results are still stable. TT is significantly positively related to fossil fuel percentage and energy consumption, but negatively related to pollution control investment, the pollution control percentage of industrial and domestic emissions, and the pollution control percentage of industrial solid waste disposal. Four dummy variables East, Northeast, Mid-west and Wild-west are used as the proxies of regional disparity, with Mid- land as the base category. When other regional variables are substituted by these four dummy variables, the regression results are still stable. It is shown that Hydro projects are significantly less inclined and projects located in the East region are more inclined to involve TT. 4. Results and analysis Results of logistic regression analyses, with Wind projects only Variables Model 6 Model 7 Model 8 Constant 4.372 (5.70) –2.777 (4.82) 5.159 (6.27) Project sizes 0.262 (0.40) 0.311 (0.40) 0.240 (0.41) Small-scale projects –0.076 (1.23) –0.266 (1.24) –0.056 (1.22) Ownerships: SOEs 0.565 (0.46) 0.681 (0.45) 0.562 (0.46) POEs –0.780 (1.02) –0.799 (1.02) –0.768 (1.01) Foreign participants: Consultants –0.210 (0.72) –0.248 (0.71) –0.219 (0.72) Consultants and credit buyers 1.175* (0.44) 0.921 (0.42) 1.168*** (0.44) Project developers 0.406 (1.09) 0.375 (1.11) 0.419 (1.09) GDP per capita 1.328* (0.73) 1.668** (0.82) 1.198 (0.85) GDP growth –0.290* (0.15) –0.248 (0.15) –0.299* (0.16) Similar –1.238 (0.55) –1.358 (0.68) R&D –0.036 (0.56) –0.069 (0.54) 0.010 (0.59) Local technology capability –4.537 (6.09) –6.871 (6.45) –3.692 (6.70) Similar in province –0.308 (0.28) 0.109 (0.36) Pseudo R2 0.2711 0.2519 0.2715 Correctly classified 77.95% 74.10% 78.59% Notes: Standard errors are in parentheses; N = 500; * p < 0 .10; p < 0.05; * p < 0 .01. Table 9. Results of logistic regression analyses, with Wind projects only From Table 8, the results of Model 1 indicate that TT commonly increases with project sizes. EE projects are significantly more likely to involve TT than Wind projects. TT is significantly more likely to happen in the project hosted by SOEs and POEs than COEs. The projects with foreign participants which act as consultants, both the consultants and credit buyers, and the project developers are significantly more likely to involve TT. The projects hosted in the regions with high GDP per capita are significantly more likely to involve TT, such as projects in Chinese Eastern provinces. But the TT possibility is significantly negatively related to the host province’s GDP growth rate. Of the five hypotheses, four are well supported by the results of Model 1 except Hypothesis 5. In Model 2, the variable Similar is added. The results of Model 1 are stable, except the significance of project types which becomes weak. Similar is a variable which has stronger influence on TT in CDM projects than project type. TT is significantly negatively related to the number of projects using the same type of technology. h In Model 3, the R&D and Local technology capability are added. The results of Model 1 and Model 2 are stable. The TT possibility is significantly positively related to the R&D investment of the host province, but significantly negatively related to Local technology capability. 5. Discussion Project sizes are one of the main factors affecting TT in CDM projects. Large-scale projects can obtain more financial support and have more opportunities in gaining investment to use advanced mitigation technologies because large-scale projects can supply CERs more stead- ily than small-scale projects. Globally, about 40% of CDM projects are small-scale projects, about 25% of small-scale projects involve TT and overall 40% of projects involve TT (Seres et al. 2010; UNFCCC 2013). In China, only 25% are small-scale projects, 17% of small-scale projects involve TT, and the overall percentage of TT is about 39% which is close to the global level. The percentages of small-scale projects and its TT in China are comparatively low. This means that the large entities in China are more likely to implement CDM projects and involve TT than small entities. This also means that the mitigation potential of small-scale projects is not fully explored in China. In small-scale CDM projects, the transaction costs of CDM and TT has a higher impact on its commercial viability than in large-scale projects at current S488 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... CER price (Das 2011; Dechezleprêtre et al. 2008, 2009; Schneider et al. 2008). And CDM is criticized for its project-by-project crediting process which is inefficient to avoid dangerous climate change (Lewis 2010). To reduce the transaction costs of mitigation activities and in response to these criticisms, bundling and programming are allowed in CDM (Lewis 2010). But these forms of activities are not widely used in CDM. Till the end of June 2013, there were only 220 programme activities registered and only 33 of them were located in China (UNFCCC 2013). The increasing implementation of these reformed activities will help to explore the mitigation potential of China. But it still needs flexible and diversiform emission reductions. To maintain the integrity of the emission reduction credits in the more flexible and diversiform crediting forms, the future mechanism should devote to the reduction of the asymmetric information between the participants and regulator (MacKenzie, Ohndorf 2012).f y p p g ( , ) TT is different among CDM project types in China. EE projects are more likely to involve TT than wind projects. This category includes the project types (EE own generation and EE households projects) which are new in China, for more than 81% CDM projects in China are Hydro and Wind projects. 5. Discussion The technologies of these new types of projects are not available in the local market or inefficient as foreign technologies. China is the main hydro technology supplier for CDM projects in the world (Seres et al. 2010). Most of the hydro technologies are locally available. The Hydro projects hosted in China do not incline to adopt foreign technologies than other project types. There are still about 25% of Hydro projects involving TT. This is the result of the efficiency, quality or other virtues of foreign technologies. Wind projects are more likely to involve TT than Hydro projects, but still less likely involve TT than EE and Other projects. The large-scale wind farm plans and the local content requirements of the Chinese government significantly spur the localization of wind turbine manufacture, but the inferiority of components’ quality still troubles the local turbine providers in China (Wang 2010). Hence, there are still about 46% of the Wind projects involving TT. The Other projects category includes some very large projects, such as HFC and N2O projects. But it doesn’t show that the project types have any significant influence on TT in CDM projects. There are about 66% of projects in this category involving TT for their large project sizes, not for their project types. The influence of project types becomes weak when the variable Similar is added into the regression model. Similar is a proxy of local mitigation technologies availability of the 15 project types in our sample. It shows that TT is significantly negatively related with the number of projects in the same type. The more projects are implemented, the later projects in this type are more inclined to use local technology, especially for wind projects which grow very fast in China. The GHG mitigation potential of EE and Other project types have not been fully explored in China. Giving policy support of TT in these project types will help explore the mitigation potential. When the number of projects of certain types is small, the policies, such as tax relief for these projects or lowering tariff on foreign equipment, will help spur their development and increase the possibility of involving TT. When the number of projects in this type is growing, policies, such as the requirements of local content levels or the use of local equipment, will help spur the localization of foreign advanced technologies and promote local technology diffusion. 5. Discussion hil f The form of ownership of Chinese participants is a significant factor which influences TT in CDM projects in China. Projects hosted by SOEs and POEs are more likely to involve TT Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 S489 than COEs. Most SOEs are very large and in a monopoly position nationally or locally. They have advantages in financing mitigation activities (Xu et al. 2012; Zeng et al. 2012), gaining support and resources from the government, and obtaining investment and long term loans, which make them more capable to involve TT than POEs. POEs can gain the least support from the government and banks, but they are more likely to involve TT than COEs. There are only 9% of the 500 CDM projects hosted by POEs and only 20 of them involve TT. It is hard to exclude that these projects may actually be very special and they may be hosted by POEs which have some similar attributes as SOE, e.g. some EE and cement projects are hosted by large-scale POEs. The advantages of SOEs cannot be learned entirely by other enterprises. POEs and COEs should implement mitigation projects according to their actual needs and capabilities. If the government supports the mitigation activities according to the importance and actual need of projects other than enterprise’s ownership, the results will be better for both mitigation activities and public resources efficiency. Nearly three quarters of the CDM projects in China are hosted by SOEs. The participations of non-SOEs in mitigation activities are insufficient. The institutional and market infrastructure which inspire both SOEs and non-SOEs are very important to promote economic developments in China, and are also important to fully explore the mitigation potential of China. In addition to the credit buyers, some types of foreign participants are important to increase TT possibility in bilateral CDM projects, which are foreign consultants, both con- sultants and buyers, and project developers. The entities from the Annex I countries which participate in CDM projects more than being credit buyers in these ways, cannot only promote the transfer of advanced mitigation technologies to developing countries, but also reduce the asymmetric information between CER buyers and suppliers. 5. Discussion It is good for enhancing the environmental integrity of the CER suppliers, but may increase the risk of losing control of the integrity of the CER buyers and widening the information gaps between the participants and regulators. Today, only Chinese or Chinese holding companies can be the host of CDM projects in China (NDRC et al. 2005, 2011). The Chinese government should encourage additional foreign participants to engage in mitigation activities in China, but should also enhance their abilities to supervise these activities. The projects located in the developed regions, such as Chinese Eastern provinces, can easily involve TT. The general explanations are that the developed regions have better fin- ancial ability to support the utilization of foreign technologies, have better ability to access the information of these foreign technologies, are much easier to access capital, gain more support from the government for their environmental conservation activities, and even have better local technology capabilities which facilitates the use of foreign technologies than other regions (Zeng et al. 2010a, b). However, additional findings have been obtained in this research. The high level of economic development indeed can increase the possibility of TT in CDM projects. But high economic growth may reduce this possibility. This might be for the reason that the high economic growth is earned by increasingly investing resources in economic growth and reducing the investment of resources in environmental conservation. The local technology capability in China is in a relatively high level which can reduce the possibility of TT, because the technologies are already available locally. But the R&D invest- ment is still one of the positive factors that influence TT in CDM projects. The influence S490 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... of Similar in province on TT is much weaker than Similar. This implies that the mitigation technologies are diffused at the country level rather than the province level, though it might be diffused at the province level first, such as the number of Wind projects first increases in Inner Mongolia then in the whole country as shown in Figure 1. In general, the advantages of the developed regions cannot be copied by all other local governments and enterprises of the undeveloped regions, which are limited by their geographical locations, technology cap- ability, as well as economic and institutional development (Xu et al. 2012). Conclusion This article focused on GHG mitigation technologies transferred by the CDM in China. China as one of the biggest developing countries with fast economic growth plays a significant role in GHG mitigation. Technology is an important and irreplaceable factor in striking a balance between economic development and environmental conservation. The Chinese government has tried its best to promote the mitigation of TT to China. CDM project characteristics (such as project sizes, project types, the form of ownership of Chinese participants, and the participation of foreign companies) and the characteristics of host regions (such as economic development levels, local technology capabilities, and local mitigation technologies availability) can affect the possibility of TT in CDM projects. The projects in large sizes, of comparatively new project types (like EE own generation and EE households projects), hosted by SOEs, or with foreign participants (such as foreign consultant and project developer), are more likely to involve TT. It was clear that the projects located in the comparatively-developed regions such as Eastern China are more likely to involve TT than in other regions, because of their advantages in economic development and R&D investment. g g p In addition to the above results, three lessons were identified. First, the non-SOEs rarely participate in mitigation activities in China. More than 73% of CDM projects were hosted by SOEs. The non-SOEs are significantly growing in China, which nearly 72% of the industrial outputs were produced by non-SOEs in 2008 (NBSC 2009). If mitigation activities of non-SOEs are supported by the government policies and banks equally as SOEs, these activities and techno- logy transfer of non-SOEs will immensely be spurred. Second, the mitigation potential of CDM projects in EE and Other projects categories was not fully explored in China. More than 81% of CDM projects in China were Hydro and Wind projects with only 18% of CDM projects being EE and Other projects. The mitigation potential of these CDM projects can extensively be ex- plored by the transfer, utilization and development of new mitigation technologies. Third, some projects are growing very fast and the technologies are widely diffused and localized in China, such as Wind projects. It is important to develop mitigation technologies locally and promote the mitigation activities, but the Wind projects in recent years are growing too fast in China (Li 2012). 5. Discussion Cooperating with companies or institutions from developed regions is a good way of CDM project hosts from other regions to take advantage of economic, technological and institutional development of the developed regions. The cooperation of different regions can help explore the mitigation potential of undeveloped regions and promote the mitigation technology diffusion in China. Conclusion The rate of abandoned wind energy is very high, which is about 11.12% on average and about 22.99% in Inner Mongolia at the end of 2011 in official documents (Li 2012), and is S491 S491 ogical and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 unofficially estimated as high as 40% to 50% in the middle of 2012 (Tong 2012). The current crisis of the photovoltaic industry (Liu 2012) and the recent deficit of wind turbine manufac- ture industry combining with the high rate of abandoned wind energy in China demonstrate that with the absence of general plans and industrial regulations on the mitigation activities, the fast growth rate can end in tragedy. When the government signals the policy preference on mitigation technologies and green energy, the capital will flow into these industries and trigger rapid growth in these industries, which may result in overcapacity and industry crises. If the government has not prepared for the fast growth rate and no general plan to regulate the growth, the faster the growth, the more serious the crises it will trigger. Other developing countries can learn from China. First, the mitigation activities should cooperate with the country’s resource endowment and industrial development. Before the implementation of CDM projects, the Chinese government has invested a lot of money in investigating and exploiting hydro resource. The CDM helps further explore the mitigation potential of hydro power in China. If the country has no such resource endowment or has no adequate local technology capability to use the mitigation technology, the CDM cannot help much to explore the mitigation potential of the country just using CER income or TT. Second, the implementation of mitigation activities should accommodate the country’s development. The expansion of large wind farms in China is much beyond the absorption capability of the Chinese market. The CDM does support the development of local mitigation activities, but it cannot support the development of the whole industry of a country, such as wind energy in China.fi CDM was criticized for its problems in excluding new projects, its inefficiency in GHG emission reductions, its insufficiency in supporting sustainable development, and its negative effects on the development of domestic low carbon policies in developing countries and the low-carbon transformation in developed countries (Lewis 2010; Vasa, Neuhoff 2011). Conclusion Some new mechanisms need to be effective, encourage mitigation actions, eliminate bottlenecks and bureaucracy, and maintain the integrity of emission reduction credits (Lewis 2010). When the new mechanisms are not established and matured, CDM will still play a significant role in GHG emission reduction in its present or reformed structures (Bakker et al. 2011). Further research is required in several areas. Some factors which may have important influence on TT are not sufficiently discussed in our study, such as the regulation of local government on environmental conservation activities, the regional disparity of institutional development, the technology capability of project owners with different ownerships. Local mitigation technology capability is not directly measured only using the number of similar projects as a crude proxy, and it has not been sufficiently discussed in this study. Besides these issues, improving the efficiency of TT in/by CDM and the reforming of CDM are interesting issues which need to be explored in further studies. References ADB. 2002. The development of private enterprise in the People’s Republic of China. Manila: Asian Devel- opment Bank.180 p. Allen, F.; Qian, J.; Qian, M. 2005. Law, finance, and economic growth in China, Journal of Financial Economics 77(1): 57–116. http://dx.doi.org/10.1016/j.jfineco.2004.06.010 Archibugi, D.; Coco, A. 2004. A new indicator of technological capabilities for developed and developing countries (ArCo), World Development 32(4): 629–654. http://dx.doi.org/10.1016/j.worlddev.2003.10.008 Bai, C. E.; Li, D. D.; Tao, Z.; Wang, Y. 2000. A multitask theory of state enterprise reform, Journal of Comparative Economics 28(4): 716–738. http://dx.doi.org/10.1006/jcec.2000.1681 Bai, C. E.; Lu, J.; Tao, Z. 2006. The multitask theory of state enterprise reform: empirical evidence from China, The American Economic Review 96(2): 353–357. http://dx.doi.org/10.1257/000282806777212125 Bakker, S.; Haug, C.; Van Asselt, H.; Gupta, J.; Saïdi, R. 2011. The future of the CDM: same same, but differentiated?, Climate Policy 11(1): 752–767. http://dx.doi.org/10.3763/cpol.2009.0035 Brandt, L.; Li, H. 2003. Bank discrimination in transition economies: ideology, information, or incentives?, Journal of Comparative Economics 31(3): 387–413. http://dx.doi.org/10.1016/S0147-5967(03)00080-5 Carney, M.; Shapiro, D.; Tang, Y. 2009. Business group performance in China: ownership and temporal considerations, Management and Organization Review 5(2): 167–193. http://dx.doi.org/10.1111/j.1740-8784.2009.00139.x Cohen, J.; Cohen, P.; West, S. G.; Aiken, L. S. 2003. Applied multiple regression/correlation analysis for the behavioral sciences. 3rd ed. Mahwah, New Jersey: Lawrence Erlbaum Associates. 703 p. Das, K. 2011. Technology transfer under the Clean Development Mechanism: an empirical study of 1000 CDM projects, The Governance of Clean Development Working Paper Series. UK: University of East Anglia. http://dx.doi.org/10.2139/ssrn.1887727 Dechezleprêtre, A.; Glachant, M.; Ménière, Y. 2008. The Clean Development Mechanism and the inter- national diffusion of technologies: an empirical study, Energy Policy 36(4): 1273–1283. http://dx.doi.org/10.1016/j.enpol.2007.12.009 Dechezleprêtre, A.; Glachant, M.; Ménière, Y. 2009. Technology transfer by CDM projects: a comparison of Brazil, China, India and Mexico, Energy Policy 37(2): 703–711. http://dx.doi.org/10.1016/j.enpol.2008.10.007 Démurger, S. 2001. Infrastructure development and economic growth: an explanation for regional dis- parities in China?, Journal of Comparative Economics 29(1): 95–117. http://dx.doi.org/10.1006/jcec.2000.1693 Doranova, A.; Costa, I.; Duysters, G. 2010. Knowledge base determinants of technology sourcing in Clean Development Mechanism projects, Energy Policy 38(10): 5550–5559. http://dx.doi.org/10.1016/j.enpol.2010.04.055 Gangale, F.; Mengolini, A. 2011. CDM contribution to RES penetration in the power generation sector of China and India, International Journal of Sustainable Development and World Ecology 18(4): 283–290. http://dx.doi.org/10.1080/13504509.2011.553354 Gregory, N. F.; Tenev, S.; Wagle, D. M. 2000. China’s emerging private enterprises: prospects for the New Century. Washington, D.C.: International Finance Corporation. 82 p. Acknowledgements We acknowledge the helpful comments and suggestions given by two anonymous reviewers which substantially improved this article. This research is supported by the National Natural Science Foundation of China (Grant No. 71025006, 71373161, 71390525), and the Ministry of Education of China (Grant No. 20120073110030). L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... S492 References Haites, E.; Duan, M.; Seres, S. 2006. Technology transfer by CDM projects, Climate Policy 6(3): 327–344. http://dx.doi.org/10.3763/cpol.2006.0621 Halsnæs, K.; Garg, A. 2011. Assessing the role of energy in development and climate policies – conceptual approach and key indicators, World Development 39(6): 987–1001. http://dx.doi.org/10.1016/j.worlddev.2010.01.002 IPCC. 2000. Methodological and technological issues in technology transfer. 1st ed. Cambridge, UK: Cam- bridge University Press. 450 p. S493 Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 S493 Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 S493 Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 Jefferson, G. H.; Su, J. 2006. Privatization and restructuring in China: evidence from shareholding own- ership, 1995–2001, Journal of Comparative Economics 34(1): 146–166. http://dx.doi.org/10.1016/j.jce.2005.11.008 Kanbur, R.; Zhang, X. 2005. Fifty years of regional inequality in China: a journey through central planning, reform, and openness, Review of Development Economics 9(1): 87–106. http://dx.doi.org/10.1111/j.1467-9361.2005.00265.x Kim, S. K.; Lee, B. G.; Park, B. S.; Oh, K. S. 2011. The effect of R&D, technology commercialization cap- abilities and innovation performance, Technological and Economic Development of Economy 17(4): 563–578. http://dx.doi.org/10.3846/20294913.2011.603481 Kirkman, G. A.; Seres, S.; Haites, E.; Spalding-Fecher, R. 2013. Benefits of the Clean Development Mech- anism 2012. [online], [cited 20 February 2013]. Available from Internet: http://cdm.unfccc.int/about/ dev_ben/ABC_2012.pdf Kornai, J. 1986. The Hungarian reform process: visions, hopes, and reality, Journal of Economic Literature 24(4): 1687–1737. Kung, J. K; Lin, Y. 2007. The decline of township-and-village enterprises in China’s economic transition, World Development 35(4): 569–584. http://dx.doi.org/10.1016/j.worlddev.2006.06.004 La Rovere, E. L.; Pereira, A. S.; Simões, A. F. 2011. Biofuels and sustainable energy development in Brazil, World Development 39(6): 1026–1036. http://dx.doi.org/10.1016/j.worlddev.2010.01.004 hi Lewis, J. I. 2010. The evolving role of carbon finance in promoting renewable energy development in China, Energy Policy 38(6): 2875–2886. http://dx.doi.org/10.1016/j.enpol.2010.01.020 Li, H.; Zhou, L. A. 2005. Political turnover and economic performance: the incentive role of personnel control in China, Journal of Public Economics 89(9): 1743–1762. http://dx.doi.org/10.1016/j.jpubeco.2004.06.009 Li, J. 2012. China Wind Power Outlook 2012. Beijing, China: China Environmental Science Publis Lin, Y. M.; Zhu, T. 2001. Ownership restructuring in Chinese state industry: an analysis of evidence on initial organizational changes, The China Quarterly 166: 305–341. http://dx.doi.org/10.1017/S000944390100016X Liu, Y. 2012, August 13, 2012. Photovoltaic industry powering ahead, China Daily [online], [cited 9 October 2012]. Available from Internet: http://www.chinadaily.com.cn/cndy/2012-08/13/content_15669054. htm MacKenzie, I. A.; Ohndorf, M. 2012. Optimal monitoring of credit-based emissions trading under asym- metric information, Journal of Regulatory Economics 42(2): 180–203. References http://dx.doi.org/10.1007/s11149-012-9196-1 Marconi, D.; Sanna-Randaccio, F. 2011. Global corporations, CDM and technology transfer to China, in Proc. of the International Workshop on “Foreign Direct Investment and Climate Change: New Research Directions” (ICCG-2011), 7–8 October, 2011, Island of San Giorgio Maggiore, Venice, Italy. NBSC. 2009. The Constituent Ratios of State-owned and Non-state-owned Industrial Enterprises in Recent Years [online], [cited 9 October 2012]. Beijing, China: National Bureau of Statistics of China. Available from Internet: http://www.stats.gov.cn/was40/reldetail.jsp?docid=402602506 NBSC. 2011. The East, West, Midland and Northeast of China [online], [cited 12 February 2012]. Beijing, China: National Bureau of Statistics of China. Available from Internet: http://www.stats.gov.cn/tjzs/ t20110613_402731597.htm NDRC; MOST; MOFA; MOF. 2005. Measures for Operation and Management of Clean Development Mechanism Projects in China [online], [cited 9 October 2012]. Beijing, China: National Development and Reform Commission, Ministry of Science and Technology, Ministry of Foreign Affairs, Ministry of Finance. Available from Internet: http://www.sdpc.gov.cn/zcfb/zcfbl/zcfbl2005/t20051206_52626.htm NDRC; MOST; MOFA; MOF. 2011. Measures for Operation and Management of Clean Development Mechanism Projects in China [online], [cited 9 October 2012]. Revised ed. Beijing, China: Na- tional Development and Reform Commission, Ministry of Science and Technology, Ministry of S494 L. Xie et al. Technology transfer in Clean Development Mechanism (cdm) projects... Foreign Affairs, Ministry of Finance. Available from Internet: http://www.gov.cn/flfg/2011-09/22/ content_1954044.htm Foreign Affairs, Ministry of Finance. Available from Internet: http://www.gov.cn/flfg/2011-09/22/ content_1954044.htm Nee, V. 1992. Organizational dynamics of market transition: hybrid forms, property rights, and mixed economy in China, Administrative Science Quarterly 37(1): 1–27. http://dx.doi.org/10.2307/2393531 Nee, V. 1992. Organizational dynamics of market transition: hybrid forms, property rights, and mixed economy in China, Administrative Science Quarterly 37(1): 1–27. http://dx.doi.org/10.2307/2393531 Nolan, P. 2001. China and the global economy: national champions, industrial policy and the big business revolution. Basingstoke, UK: Palgrave Macmillan. http://dx.doi.org/10.1057/9780230524101 Olsen, K. H. 2007. The Clean Development Mechanism’s contribution to sustainable development: a review of the literature, Climatic Change 84(1): 59–73. http://dx.doi.org/10.1007/s10584-007-9267-y Peng, M. W.; Tan, J.; Tong, T. W. 2004. Ownership types and strategic groups in an emerging economy, Journal of Management Studies 41(7): 1105–1129. http://dx.doi.org/10.1111/j.1467-6486.2004.00468.x Poncet, S.; Steingress, W.; Vandenbussche, H. 2010. Financial constraints in China: firm-level evidence, China Economic Review 21(3): 411–422. http://dx.doi.org/10.1016/j.chieco.2010.03.001 Pyke, D.; Robb, D.; Farley, J. 2000. Manufacturing and supply chain management in China: a survey of state-, collective-, and privately-owned enterprises, European Management Journal 18(6): 577–589. http://dx.doi.org/10.1016/S0263-2373(00)00051-7 Qi, G. Y.; Zeng, S. X.; Li, X. D.; Tam, C. M. 2012. References R.; Pearce, J. L. 1996. Guanxi: connections as substitutes for formal institutional support, Academy of Management Journal 39(6): 1641–1658. http://dx.doi.org/10.2307/257072 Xu, X. D.; Zeng, S. X.; Tam, C. M. 2012. Stock market’s reaction to disclosure of environmental violations: evidence from China, Journal of Business Ethics 107(2): 227–237. http://dx.doi.org/10.1007/s10551-011-1035-2 Zeng, S. X.; Xie, X. M.; Tam, C. M. 2010a. Evaluating innovation capabilities for science parks: a system model, Technological and Economic Development of Economy 16(3): 397–413. http://dx.doi.org/10.3846/tede.2010.25 Zeng, S. X.; Xie, X. M.; Tam, C. M. 2010b. Relationship between cooperation networks and innovation per- formance of SMEs, Technovation 30(3): 181–194. http://dx.doi.org/10.1016/j.technovation.2009.08.003 Zeng, S. X.; Xu, X. D.; Yin, H. T.; Tam, C. M. 2012. Factors that drive Chinese listed companies in voluntary disclosure of environmental information, Journal of Business Ethics 109(3): 309–321. http://dx.doi.org/10.1007/s10551-011-1129-x Linna XIE. A PhD student in Antai School of Management at Shanghai Jiaotong University, China. She has received her Master’s degree in 2010 and passed the PMP certification in 2008. She has published four chapters in An Experimental Course in Project Management by China Renmin University Press in 2010. Research interests include technology management, project management. Saixing ZENG. Doctor, Head and Professor in Antai School of Management at Shanghai Jiaotong Uni- versity, China. As a researcher in Technology Management and related fields, he has been in charge of a large number of research projects, and has published more than 100 journal and conference papers, books, and reports on technology management and project management. Research interests include technology management, project management. Hailiang ZOU. A PhD student in Antai School of Management at Shanghai Jiaotong University, China. He has received his Master’s degree in 2009. He has been involved in two research projects supported by the National Natural Science Foundation of China. Research interests include technology management, environmental management. Vivian W. Y. TAM. Doctor, A Senior Lecturer in School of Computing, Engineering and Mathematics at University of Western Sydney, Australia. She has received her PhD in City University of Hong Kong. She is a Principal Supervisor for doctoral projects, and has published more than 160 international refereed journal papers. Research interests include environmental management, sustainable construction. Zhenhua WU. A graduated student in Antai School of Management at Shanghai Jiaotong University, China. He has received his Master’s degree in 2011. He is currently working in the China Securities Depository and Clearing Corporation Limited. References Role of internalization process in defining the relation- ship between ISO 14001 certification and corporate environmental performance, Corporate Social Responsibility and Environmental Management 19(3): 129–140. http://dx.doi.org/10.1002/csr.258 Ralston, D. A.; Terpstra-Tong, J.; Terpstra, R. H.; Wang, X.; Egri, C. 2006. Today’s state-owned enterprises of China: are they dying dinosaurs or dynamic dynamos?, Strategic Management Journal 27(9): 825–843. http://dx.doi.org/10.1002/smj.545 Reynolds, T. W. 2012. Institutional determinants of success among forestry-based carbon sequestration projects in Sub-Saharan Africa, World Development 40(3): 542–554. http://dx.doi.org/10.1016/j.worlddev.2011.09.001 Richerzhagen, C.; Scholz, I. 2008. China’s capacities for mitigating climate change, World Development 36(2): 308–324. http://dx.doi.org/10.1016/j.worlddev.2007.06.010 Schneider, M.; Holzer, A.; Hoffmann, V. H. 2008. Understanding the CDM’s contribution to technology transfer, Energy Policy 36(8): 2930–2938. http://dx.doi.org/10.1016/j.enpol.2008.04.009 Seres, S.; Haites, E. 2008. Analysis of Technology Transfer in CDM Projects [online], [cited 15 June 2011]. Available from Internet: http://cdm.unfccc.int/Reference/Reports/TTreport/TTrep08.pdf Seres, S.; Haites, E.; Murphy, K. 2009. Analysis of technology transfer in CDM projects: an update, Energy Policy 37(11): 4919–4926. http://dx.doi.org/10.1016/j.enpol.2009.06.052 Seres, S.; Haites, E.; Murphy, K. 2010. The Contribution of the Clean Development Mechanism under the Kyoto Protocol to Technology Transfer [online], [cited 15 June 2011]. Available from Internet: http:// cdm.unfccc.int/Reference/Reports/TTreport/TTrep10.pdf Štreimikiene, D.; Esekina, B. 2008. EU pollution reduction strategies and their impact on atmospheric emissions in Lithuania, Technological and Economic Development of Economy 14(2): 162–170. Tébar Less, C.; McMillan, S. 2005. Achieving the successful transfer of environmentally sound technologies: trade-related aspects, OECD Trade and Environment Working Papers No. 2005–02. OECD, Paris. 38 p. Tong, K. 2012, June 6, 2012. The rate of abandoned wind energy is as high as 50% – it should be incor- porated into the power grid, Securities Times of China [online], [cited 20 July 2013]. Available from Internet: http://epaper.stcn.com/paper/zqsb/html/2012-06/06/content_376220.htm UNFCCC. 2013. CDM Insights [online], [cited 20 July 2013]. Available from Internet: http://cdm.unfccc. int/Statistics/Public/index.html Vasa, A.; Neuhoff, K. 2011. The Role of CDM Post 2012: carbon pricing for low-carbon investment project. Berlin: Climate Policy Initiative, Deutsches Institut für Wirtschaftsforschung. 36 p. S495 Technological and Economic Development of Economy, 2013, 19(Supplement 1): S471–S495 Wang, B. 2010. Can CDM bring technology transfer to China? An empirical study of technology transfer in China’s CDM projects, Energy Policy 38(5): 2572–2585. http://dx.doi.org/10.1016/j.enpol.2009.12.052 Ward, P.; Shively, G. 2011. Vulnerability, income growth and climate change, World Development 40(5): 916–927. http://dx.doi.org/10.1016/j.worlddev.2011.11.015 Xia, J.; Li, S.; Long, C. 2009. The transformation of collectively owned enterprises and its outcomes in China, 2001–05, World Development 37(10): 1651–1662. http://dx.doi.org/10.1016/j.worlddev.2009.03.007 Xin, K. References He is responsible for the technical development and maintenance of the security depository and clearing system. Research interests include project management.
https://openalex.org/W4239261239	https://www.researchsquare.com/article/rs-156682/latest.pdf	English	null	Impact Of Pre-event Testing And Quarantine On Reducing The Risk Of COVID-19 Epidemic Rebound: A Modelling Study	Research Square (Research Square)	2,021	cc-by	5,392	Abstract BACKGROUND With the evolving growth of the COVID-19 epidemic, travel restriction policies would need to be adjusted accordingly. Prohibition of mass event may be relaxed for social and economic benefits when virus transmission stops but could bear the risk of epidemic rebound. Against the background of the varied SARS-CoV-2 prevalence internationally, we modelled the potential impacts of pre-event interventions on epidemic risk of holding a mass event when COVID-19 is under control. METHODS We developed a mathematical model of SARS-CoV-2 transmission in Guangdong Province, China, where local virus transmission ceased to occur. A large-scale international trade fair was assumed to be held, with influx of people from overseas and rest of China over a short period of time, who participated for 2-week. Scenarios of pre-event intervention (none, quarantine arrangement and polymerase chain reaction (PCR) testing for participants) were compared. The influence of contact pattern, SARS-CoV-2 prevalence outside the province and China, and testing coverage were examined in sensitivity analyses. RESULTS In basecase scenario (no event), the epidemic has been under control since March 2020. The event would lead to the detection of 1% more confirmed cases by 31 July when community contact rate increases to pre-epidemic level. In event scenario without additional interventions, there would be 599 (93%) more new infections comparing with basecase scenario. To avert new infections, quarantining all participants before the event would be the most effective strategy, followed by quarantining all overseas participants and testing all other participants, and testing all participants before the event and on day 7. However, testing strategy is likely to be affected by the SARS-CoV-2 prevalence outside the event province. CONCLUSIONS Pre-event interventions are effective for reducing the risk of epidemic rebound caused by an international large-scale event. Universal testing for participants is likely to be an effective and feasible intervention. Research article Posted Date: February 2nd, 2021 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Version of Record: A version of this preprint was published at BMC Infectious Diseases on January 24th, 2022. See the published version at https://doi.org/10.1186/s12879-021-06963-2. Page 1/13 Overview We developed a basecase deterministic compartmental model to simulate COVID-19 epidemic under implementation of interventions and gathering restrictions in Guangdong Province, China from 28 December 2019 to 31 August 2020. We then developed hypothetical scenarios of holding a large-scale international event (event hereafter) involving participants from Guangdong, other provinces in Mainland China, and overseas, and other Guangdong residents not participating the event. Sensitivity analyses were performed to analyze the epidemic impacts of key model parameters. Background The epidemic of COVID-19 varied temporally and geographically. In December 2020, most places in the Western Pacific were classified as belonging to the level of ‘cluster of cases’ while other places such as North America and Europe were classified as ‘community transmission’ by the World Health Organization. [1] International travel restriction have been adopted in most places to suppress the number of imported cases. A wide range of non-pharmaceutical interventions have also been implemented in different degrees over time to control the epidemic. One of the interventions is restriction of mass gathering and mass events, the setting of which involved a large number of people who may not be known to each other and are in contact for extended duration in indoor environment.[2] Outbreaks have been reported in mass events in the earlier phase of the COVID-19 pandemic, as exemplified by clusters in a business conference in Singapore,[3] and a Muslim missionary movement in Malaysia.[4] In coping with the pandemic, major Page 2/13 Page 2/13 international activities, such as large-scale trade, politics, religious, cultural, academic and sports events, have been cancelled, postponed or become virtual worldwide.[5] When the COVID-19 epidemic was under control locally, intervention strategies were often adjusted to reduce the negative impacts in social and economic aspects.[6,7] For instance, efforts have been made to partially relax the restriction by implementing travel bubbles for people between selected places for enabling re-connections.[8] Local gathering restrictions have been lifted, enabling organization of local mass events. However, international large-scale events are generally disallowed in consideration of the anticipated risk for widespread transmission. To inform COVID-19 epidemic control strategies, mathematical modelling is a useful approach for risk assessment and simulation of possible outcomes under different intervention scenarios. Previous modelling studies have analyzed the impact of non-pharmaceutical interventions including contact tracing, testing, and quarantine. Interventions for special settings such as refugee camp have also been analyzed in models.[9] Exit strategies for COVID-19 were examined in a modelling study in Singapore.[10] So far, investigations specifically for mass event setting have not yet been fully explored in modelling studies. Against this background, we undertook to analyze the potential epidemiologic impacts of organizing a large-scale international event under different intervention strategies. The hypothetical scenarios of holding the China Import and Export Fair (Canton Fair) in Guangdong Province of China, a trade event attracting more than 180,000 participants from more than 40 countries per year, were developed. Study area Guangdong, a province with 113.46 million population in 2018,[11] has reported 1641 confirmed cases including 8 deaths by 30 June 2020.[12] Imported cases, defined as COVID-19 patients whose infection originated outside Guangdong, accounted for around 76% of confirmed cases.[12] Contact tracing, testing, confirmed case isolation, and quarantine were implemented throughout the epidemic period. COVID-19 epidemic in Guangdong has been under control with less than 2 daily new confirmed cases since 21 March 2020. Basecase model structure and assumptions This is an open model with individuals entering and leaving the province (Figure 1, Appendix p.2). Besides quarantining close contacts, health quarantine of inbound travelers from listed origins have been imposed upon travelers’ arrival, and polymerase chain reaction (PCR) testing for SARS-CoV-2 (95% sensitivity) performed. We assumed all symptomatic travelers testing positive were directed to the hospital compartment, while those testing negative or positive without symptoms were directed to the quarantine compartment. After 14-days’ quarantine, individuals diagnosed would be directed to the hospital compartment, while the rest flowed back to the susceptible compartment. The quarantine arrangement applied to travelers who were from a) Hubei Province, China, between 14 February 2020 and 23 March 2020; b) overseas since 1 March 2020; and c) Hong Kong and Macau, both Special Administrative Regions (SAR) of China, as from 27 March 2020. A net proportion of travelers, who were susceptible and not fulfilling quarantine criteria at the time of arrival, entered the susceptible compartment. Infected individuals not fulfilling quarantine criteria at the time of arrival (i.e. not in the list of designated countries) entered either the pre-infectious compartment or symptomatic compartment. Page 3/13 Page 3/13 Page 3/13 Data Source Model inputs included reproduction number (R0), biological parameters (latent period, time to recovery in asymptomatic infections, time from hospitalization and intensive care units (ICU) to recovery in symptomatic infections, mortality rate of COVID-19, asymptomatic proportion), demographic parameters (permanent resident population in 2018 in Guangdong, travelers staying overnight in 2018 in Guangdong), policy and healthcare system parameters (contact tracing, quarantine, and testing), and event parameters (eTable 1). Other interventions such as school closure, group gathering restrictions, and business and premises restrictions are absorbed in background and reflected by the proportion of contact rate reduction comparing with pre-epidemic level. Model parameter values or ranges were derived from provincial yearbooks, local government reports, scientific literature and assumptions.[13-18] We fitted model predictions over time to daily number of confirmed cases in the province by 29 March 2020, using mle function under negative log likelihood in R (stats4 package). In this calibration process, we simultaneously varied the proportion of asymptomatic infection and proportion of reduction of contact rate comparing with pre-epidemic level. Modelling results were validated with the daily number of confirmed cases between 30 March and 15 May 2020. IRB approval and a waiver of consent were obtained from Dermatology Hospital of Southern Medical University, China. Event scenarios The setting of the event was the biannual Canton Fair, which was organized virtually in 2020 because of the COVID-19 epidemic. Event scenario was developed on the assumption that this Fair was held on-site in a large conference venue in Guangzhou between 1 and 14 June 2020. In the event, we assumed the contact rate among participants doubled with reference to that in the community. Assuming only half of the overseas participants (100,000) in the past two years have joined the event, and the number of local Page 4/13 Page 4/13 participants remained the same as in the past for Guangdong (38,000) and other provinces of China (100,000), six scenarios were developed to examine the impacts of different pre-event interventions of quarantine and testing (assumptions in Appendix p.2). participants remained the same as in the past for Guangdong (38,000) and other provinces of China (100,000), six scenarios were developed to examine the impacts of different pre-event interventions of quarantine and testing (assumptions in Appendix p.2). 1. None of participants were quarantined before the event, unless they were contact traced; 2. All participants travelling from overseas were quarantined for 14 days be All participants (regardless of origins) were quarantined before the event; 4. Scenario 2, and all Mainland participants were tested before attending th 5. Scenario 1, and all participants were tested before attending the event; 6. Scenario 1, and all participants were tested before and on day 7 following opening of the event Model outcomes The main modelling outcomes were the cumulative number and proportion change of cumulative new infections generated locally (regardless of symptom presentation) above basecase scenario by 31 July 2020, to allow enough time for demonstrating the impact on incidence caused by the event. The secondary outcome was the cumulative number of confirmed cases. Sensitivity analyses To account for variability of R0, contact rate, and proportion of asymptomatic infections over time, we have performed 5000 simulations in the basecase model (Appendix p.18). We performed one-way sensitivity analysis around the key parameters in the following scenarios: contact rate reduction from baseline ranged between 0% and 90% since 1 May 2020, prevalence of SARS-CoV-2 in overseas participants (0.00005-0.001) and other provinces’ participants (0.00001-0.005), proportion of all infected individuals in Guangdong who attended the event (0.001-0.1), and testing coverage (0%-100%) for participants in scenarios 4-6. We also varied the contact rate at the event from doubled, tripled and ten- fold with reference to that in the community. To examine the impact of duration with very low contact rate (10%) through intense interventions at epidemic peak (9 February 2020) and the change of contact rate after relaxing the interventions, two-way sensitivity analysis was performed. Event scenario If the Canton Fair was organized without any pre-event interventions and the contact rate remained at 70% of pre-epidemic level (scenario 1), the estimated cumulative number of new infections and confirmed cases would be 93% higher (599 more infections) and 23% higher (388 more cases) than basecase scenario by 31 July 2020, respectively (Figure 4 a-b). The daily number of new infections and confirmed cases would increase linearly in the event period, reaching the peak after three days, following which local participants would have spent all time in the community (Figure 4 c-d). If all overseas participants were quarantined before the event (scenario 2), the estimated cumulative number of new infections would be 16% higher than basecase scenario. By expanding the quarantine criteria to all participants (scenario 3), no new infections attributable to the event was estimated. However, the total number of confirmed cases would be 2% higher, all of which confirmed once they arrived at the province / event venue (Figure 4 b, d). When all overseas participants were quarantined while all participants from Mainland China were tested before the event (scenario 4), the estimated total new infections would be 1% higher than basecase scenario. Comparing with basecase scenario, the estimated number would be 3% higher if all participants were tested before the event (scenario 5), and 2% higher if all participants were tested both before the event and on day 7 of the event. In scenarios involving testing strategies, a decrease of the testing coverage from 100% to 10%, would lead to 14%, 81%, and 79% of additional infections in scenarios 4, 5, and 6, respectively (eFigure 3). Sensitivity analysis was performed on various scenarios when contact rate was back to the pre-epidemic level (i.e. 100%) from 1 May 2020. Compared to basecase scenario, organizing the event would result in no change in the estimated cumulative number of new infections under scenario 3, minimal change (1% higher than basecase scenario) in scenario 4, and less than 10% more new infections under scenarios 5-6 (Figure 4). However, the cumulative number of new infections would be 39% higher than basecase scenario in scenario 2 by quarantining overseas participants only, and 2.5 folds higher without interventions (scenario 1). The model has assumed that participants would double their contacts in the event comparing with that in the community. Basecase scenario Our basecase model (without event) provided a reasonable fit to the daily number of newly confirmed (symptomatic only) cases in Guangdong from 15 January to 29 March 2020, and prediction from 30 March to 15 May 2020 (Figure 2). In basecase scenario, the daily number of confirmed cases reached its peak in early February and declined rapidly within two weeks. The model estimated that the number of cumulative local new infections reached the plateau of between 600 and 650 in mid-February. By 31 July, Page 5/13 Page 5/13 the estimated cumulative number of new infections and confirmed cases would be 644 and 1680, respectively. To account for the changes of contact rate from pre-epidemic level, one-way sensitivity analysis was performed. By changing the contact rate from 70% (status quo) to 100% from 1 May 2020, the cumulative number of new infections and confirmed cases was 1% and 0.4% higher respectively than basecase scenario (eFigure 1). On the contrary, if contact rate decreased to 10%, the estimated cumulative number of new infections and confirmed cases would be 2% and 1% lower than basecase scenario respectively. Assuming that intense intervention had not achieved very low contact rate (10% of pre-epidemic level) from 9 February 2020, the estimated cumulative number of new infections would be 8% higher than basecase scenario (Figure 3, eFigure 2), as shown in two-way sensitivity analysis. Event scenario When the contact rate at the event is triple or ten-fold higher, substantial increases in total new infections were estimated in all scenarios except scenario 3. In Page 6/13 scenario 1 without pre-event interventions, the estimated total number of new infections by 31 July 2020 would be 235% and 13,339% higher than basecase scenario if contact rate in the event was triple and ten- fold higher, respectively (eTable 2). The impact of the SARS-CoV-2 prevalence in other provinces and overseas on the percentage change of total new infections would be high, as shown in sensitivity analyses. The cumulative number of new infections in scenario 1 would rise from 80% higher to 350% higher than basecase scenario when the prevalence in other provinces increased from 0.00001 to 0.001 (eFigure 4). When prevalence in other provinces rose from 0.00001 to 0.001 in scenario 2, 264% more new infections were estimated. Similarly, there would be 253% more new infections in scenario 1 than basecase scenario if the prevalence overseas increased from 0.00005 to 0.001. However, there would be minimal impact (<2% change on total new infections) when proportion of local infected cases participating in the event changed from 0.1% to 10% in all scenarios. Discussion The effectiveness of testing intervention to avert new infections was shown in a modelling study,[23] suggesting that universal testing alone without lockdown could reduce the amplitude of the peak by 40%.[23] Universal testing has been suggested as a lockdown exit strategy in the United Kingdom, with feasibility study planned.[24] With improvement of sensitivity and specificity of SARS-CoV- 2 tests and availability of point-of-care tests,[24] accurate and timely test results would be available to facilitate efficient implementation of control. This would increase the feasibility of positioning universal testing as a key preventive intervention to minimize adverse economic impacts of COVID-19. the capability of detecting pre-infectious cases through testing. At higher prevalence, the estimated number of new infections would be larger even the rise would not be as high as of scenario 2 (quarantining overseas participants but no intervention for Mainland participants). Strategically adding one more time point for testing all participants (scenario 6) could avert a proportion of new infections estimated in scenario 5. The effectiveness of testing intervention to avert new infections was shown in a modelling study,[23] suggesting that universal testing alone without lockdown could reduce the amplitude of the peak by 40%.[23] Universal testing has been suggested as a lockdown exit strategy in the United Kingdom, with feasibility study planned.[24] With improvement of sensitivity and specificity of SARS-CoV- 2 tests and availability of point-of-care tests,[24] accurate and timely test results would be available to facilitate efficient implementation of control. This would increase the feasibility of positioning universal testing as a key preventive intervention to minimize adverse economic impacts of COVID-19. The infection risk of an international event rests not just with potential virus exposure at the event setting but also transmission in the community. With low contact rate in the community, limited ongoing transmissions could occur in association with exposure to imported cases. The difference of estimated new infections with and without different pre-event interventions would be small. However, when community contact rate before the event approaches the pre-epidemic level, the likelihood of ongoing transmission could be higher. To control possible ongoing transmissions from imported cases, pre-event intervention strategies would play a key role. Discussion Since the emergence of the COVID-19 pandemic, intense interventions have been implemented to minimize gathering of people and travel between places.[10,19,20] Large-scale international events were cancelled or postponed, resulting in epidemic control but anticipated economic loss. This modelling study simulated the potential epidemiologic impact of organizing an international trade event with and without pre-event interventions. At basecase without event, the epidemic is under control even if the contact rate increased to the pre-epidemic level in Guangdong, China. Without additional pre-event interventions, organizing the Canton Fair at the end of the local epidemic could double the cumulative number of new infections if SARS-CoV-2 prevalence in other provinces is 0.00006 and overseas is 0.0003. The increase of new infections would be proportional to the prevalence during event organization. Our study suggested that additional pre-event interventions including quarantine and testing for participants could avert a substantial proportion of new infections. Quarantining all participants regardless of origins would be most effective. It is least likely to be affected by other factors including community contact rate and SARS-CoV-2 prevalence in other places. Such positive outlook is achievable if all infected cases could be identified, quarantined or hospitalized before the event. There would be minimal chance for outbreaks in the community and at the event venue. The effectiveness of quarantine intervention on epidemic growth is consistent with the results of previous modelling studies.[15,21,22] However its feasibility is doubtful because of capacity need to quarantine more than 0.2 million people for 14 days at the same time. Testing all Mainland participants and quarantining all overseas participants (scenario 4) would be similarly effective as the strategy of quarantining all participants (scenario 3), if the prevalence in other provinces remains low. The strategy of testing everyone (scenario 5) would result in slightly higher proportion of total new infections (~2%) but could keep the quarantined number at the minimum. However, the impact of testing intervention would be affected by the current state of the epidemic, and Page 7/13 Page 7/13 the capability of detecting pre-infectious cases through testing. At higher prevalence, the estimated number of new infections would be larger even the rise would not be as high as of scenario 2 (quarantining overseas participants but no intervention for Mainland participants). Strategically adding one more time point for testing all participants (scenario 6) could avert a proportion of new infections estimated in scenario 5. Discussion The importance of contact rate on epidemic growth could be reflected from the implementation of physical distancing measures in the real-world with effectiveness shown in modelling studies.[25-27] To minimize the adverse impact of organizing an event, the contact rate in the community would have to be taken into consideration. There are a few limitations in this modelling study. First, the logistics for local quarantine arrangement, and the provision of SARS-CoV-2 tests (on days 1, 4, 7 and 14) have been simplified.[13] Nonetheless, with limited number of quarantined persons eventually diagnosed with COVID-19, the impact of simplified logistics could be minimal. We also acknowledged the possibility of introducing new infections in household members of home-quarantined individuals for residents in the organizer city, the occurrence of which has been absorbed in the background. Second, we used contact rate in comparison with pre- epidemic level to evaluate the overall impacts of physical distancing measures in the community. There were no further breakdowns of contact rate reduction as differentiated by the types of measures, such as school closure, workplace closure, and restriction of gathering in public area, while only quarantine and PCR testing for SARS-CoV-2 were specified as pre-event interventions in modelling scenarios. We acknowledged the difficulties and inherent limitations of evaluating the specific impacts of different contact rates and their association with respective social distancing measure. Third, the event scenarios had assumed that many overseas individuals would still participate in the event despite the 14-day pre- event quarantine. Although we assumed a 50% reduction of the number of overseas participants comparing with the previous years, the actual number may be even lower, and the model may have overestimated the number of new infections in the event scenarios. We have also oversimplified the flow of participants coming in and leaving the event by assuming that all participants would stay for the whole period. The downside, again, was the overestimation of the epidemic impacts of the model. Finally, Page 8/13 Page 8/13 Page 8/13 the modelling study has focused on the growth of the COVID-19 epidemic at an international event that took place after the outbreak has gone to a quiescent phase. Other public health impacts, for example, cost-effectiveness analysis may be needed to account for the cost and utility involved in each strategy. Conclusion When restriction of international large-scale event organization is lifted, different forms of pre-event interventions could be considered to effectively reduce the risk of widespread SARS-CoV-2 transmission. Mathematical modelling is a useful approach for risk assessment and planning of pre-event interventions. Availability of data and materials: The parameter values used in the model are show eTable 1. Competing interests: The authors declare that they have no competing interests Competing interests: The authors declare that they have no competing interests Funding: Guangdong Medical Research Foundation (A2019402) Funding: Guangdong Medical Research Foundation (A2019402) Authors' contributions: CW conceptualized the study, collected data and supervise the study. NSW performed the analysis and drafted the manuscript. KMM, SSL and EY provided technical support and interpreted the results. All authors read, reviewed and approved the final manuscript Acknowledgements: The authors would like to thank Wenjun Ma, Pengjian Xiao, Jianxiong Hu, and Tao Liu at Guangdong Provincial Institute of Public Health for assistance in data collection. Li Ka Shing Institute of Health Sciences and Stanley Ho Centre for Emerging Infectious Disease of The Chinese University of Hong Kong provided technical support in conducting the research. Abbreviations ICU – intensive care units PCR – polymerase chain reaction PCR – polymerase chain reaction R0 – reproduction number R0 – reproduction number Discussion the modelling study has focused on the growth of the COVID-19 epidemic at an international event that took place after the outbreak has gone to a quiescent phase. Other public health impacts, for example, cost-effectiveness analysis may be needed to account for the cost and utility involved in each strategy. Declarations Ethics approval and consent to participate: IRB approval and a waiver of consent were obtained from Dermatology Hospital of Southern Medical University, China. Consent for publication: Not applicable. Consent for publication: Not applicable. Availability of data and materials: The parameter values used in the model are shown in Appendix and eTable 1. Page 9/13 Page 9/13 1. World Health Organization. WHO Coronavirus Disease (COVID-19) Dashboard. https://covid19.who.int/ (accessed 16 December 2020). 2. Wong NS, Lee SS, Kwan TH, Yeoh EK. Settings of virus exposure and their implications in the propagation of transmission networks in a COVID-19 outbreak. The Lancet Regional Health Western Pacific. 2020;4:100052. 3. Pung R, Chiew CJ, Young BE, Chin S, Chen MI, Clapham HE, et al. Investigation of three clusters of COVID-19 in Singapore: implications for surveillance and response measures. Lancet. 2020;395:1039–46. 4. Che Mat NF, Edinur HA, Abdul Razab MKA, Safuan S. A single mass gathering resulted in massive transmission of COVID-19 infections in Malaysia with further international spread. J Travel Med. 2020;27:taaa059. 5. Barnes M, Sax PE. Challenges of "Return to Work" in an Ongoing Pandemic. N Engl J Med. 2020;383(8):779–86. 6. Department of Health of Australian Government. 3-Step Framework for a COVIDSafe Australia. 2020. https://www.health.gov.au/resources/publications/3-step-framework-for-a-covidsafe-australia (accessed 30 May 2020). 7. Petersen E, Wasserman S, Lee SS, Go U, Holmes AH, Al-Abri S, et al. COVID-19-We urgently need to start developing an exit strategy. Int J Infect Dis. 2020;96:233–9. 7. Petersen E, Wasserman S, Lee SS, Go U, Holmes AH, Al-Abri S, et al. COVID-19-We urgently need to start developing an exit strategy. Int J Infect Dis. 2020;96:233–9. 8. The Lancet Infectious Diseases. Air travel in the time of COVID-19. Lancet Infect Dis. 2020;20:993. 8. The Lancet Infectious Diseases. Air travel in the time of COVID-19. Lancet Infect Dis. 2020;20:993. 9. Gilman RT, Mahroof-Shaffi S, Harkensee C, Chamberlain AT. Modelling interventions to control COVID-19 outbreaks in a refugee camp. BMJ Glob Health. 2020;5:e003727. 9. Gilman RT, Mahroof-Shaffi S, Harkensee C, Chamberlain AT. Modelling interventions to control COVID-19 outbreaks in a refugee camp. BMJ Glob Health. 2020;5:e003727. 10. Dickens BL, Koo JR, Lim JT, Park M, Quaye S, Sun H, et al. Modelling lockdown and exit strategies for COVID-19 in Singapore. The Lancet Regional Health Western Pacific. 2020;1:100004. 10. Dickens BL, Koo JR, Lim JT, Park M, Quaye S, Sun H, et al. Modelling lockdown and exit strategies for COVID-19 in Singapore. The Lancet Regional Health Western Pacific. 2020;1:100004. 11. Guangdong Bureau of Statistics. [Economic situation report in the first quarter of 2020 in Guangdong Province, 28 April 2020] http://stats.gd.gov.cn/tjfx/index.html. (accessed 4 July 2020). 11. Guangdong Bureau of Statistics. [Economic situation report in the first quarter of 2020 in Guangdong Province, 28 April 2020] http://stats.gd.gov.cn/tjfx/index.html. (accessed 4 July 2020). 12. Page 9/13 Health Commission of Guangdong Province. [COVID-19 situation report in Guangdong Province, 30 June 2020] http://wsjkw.gd.gov.cn/zwyw_yqxx/content/post_3025638.html (accessed 4 July 2020). 12. Health Commission of Guangdong Province. [COVID-19 situation report in Guangdong Province, 30 June 2020] http://wsjkw.gd.gov.cn/zwyw_yqxx/content/post_3025638.html (accessed 4 July 2020). 13. Bi Q, Wu Y, Mei S, Ye C, Zou X, Zhang Z, et al. Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study. Lancet Infect Dis. 2020;S1473-3099:30287–5. 13. Bi Q, Wu Y, Mei S, Ye C, Zou X, Zhang Z, et al. Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study. Lancet Infect Dis. 2020;S1473-3099:30287–5. 14. Mizumoto K, Kagaya K, Zarebski A, Chowell G. Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board the Diamond Princess cruise ship, Yokohama, Japan, 2020. Euro Surveill. 2020;25:2000180. 14. Mizumoto K, Kagaya K, Zarebski A, Chowell G. Estimating the asymptomatic proportion of coronavirus disease 2019 (COVID-19) cases on board the Diamond Princess cruise ship, Yokohama, Japan, 2020. Euro Surveill. 2020;25:2000180. 15. Ferguson NM, Laydon D, Nedjati-Gilani G, Imai N, Ainslie K, Baguelin M, et al. Impact of non- pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. Imperial College London (16-03-2020). https://doi.org/10.25561/77482. 15. Ferguson NM, Laydon D, Nedjati-Gilani G, Imai N, Ainslie K, Baguelin M, et al. Impact of non- pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand. Imperial College London (16-03-2020). https://doi.org/10.25561/77482. Page 10/13 16. Li Q, Guan X, Wu P, Wang X, Zhou L, Tong Y, et al. Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected Pneumonia. N Engl J Med. 2020;382:1199–207. 17. Health Commission of Guangdong Province. [Guangdong Province decides to activate the first level emergency response]. 2020. http://wsjkw.gd.gov.cn/zwyw_yqxx/content/post_2878895.html (accessed 30 May 2020). 18. Health Commission of Guangdong Province. [The 31st press conference of the Guangdong’s fighting against COVID-19]. 2020. http://wsjkw.gd.gov.cn/zwyw_xwfbh/content/post_2908689.html (accessed 30 May 2020). 19. Lai S, Ruktanonchai NW, Zhou L, Prosper O, Luo W, Floyd JR, et al. Effect of non-pharmaceutical interventions to contain COVID-19 in China. Nature. 2020;585(7825):410–3. 19. Lai S, Ruktanonchai NW, Zhou L, Prosper O, Luo W, Floyd JR, et al. Effect of non-pharmaceutical interventions to contain COVID-19 in China. Nature. 2020;585(7825):410–3. 20. Davies NG, Kucharski AJ, Eggo RM, Gimma A, Edmunds WJ. Centre for the Mathematical Modelling of Infectious Diseases COVID-19 working group. Figure 2 Model estimates of (a) daily number of new COVID-19 confirmed cases and (b) cumulative number of new SARS-CoV-2 infections in Guangdong Province, 28 December 2019 – 31 August 2020 Figure 1 Basecase model diagram Page 9/13 Effects of non-pharmaceutical interventions on COVID-19 cases, deaths, and demand for hospital services in the UK: a modelling study. Lancet Public Health. 2020;5:e375–85. 21. Nussbaumer-Streit B, Mayr V, Dobrescu AI, Chapman A, Persad E, Klerings I, et al. Quarantine alone or in combination with other public health measures to control COVID-19: a rapid review. Cochrane Database Syst Rev. 2020;4:CD013574. 21. Nussbaumer-Streit B, Mayr V, Dobrescu AI, Chapman A, Persad E, Klerings I, et al. Quarantine alone or in combination with other public health measures to control COVID-19: a rapid review. Cochrane Database Syst Rev. 2020;4:CD013574. 22. Hou C, Chen J, Zhou Y, Hua L, Yuan J, He S, et al. The effectiveness of quarantine of Wuhan city against the Corona Virus Disease 2019 (COVID-19): A well-mixed SEIR model analysis. J Med Virol. 2020;92(7):841–8. 23. Goscé L, Phillips PA, Spinola P, Gupta DRK, Abubakar PI. Modelling SARS-COV2 Spread in London: Approaches to Lift the Lockdown. J Infect. 2020;S0163-4453:30315–7. 24. Peto J, Alwan NA, Godfrey KM, Burgess RA, Hunter DJ, Riboli E, et al. Universal weekly testing as the UK COVID-19 lockdown exit strategy. Lancet. 2020;395:1420–1. 25. Jarvis CI, Van Zandvoort K, Gimma A, Prem K, CMMID COVID-19 working group, Klepac P, et al. Quantifying the impact of physical distance measures on the transmission of COVID-19 in the UK. BMC Med. 2020;18:124. 26. Tuite AR, Fisman DN, Greer AL. Mathematical modelling of COVID-19 transmission and mitigation strategies in the population of Ontario, Canada [published online ahead of print, 2020 Apr 8]. CMAJ. 2020;cmaj.200476. 27. Badr HS, Du H, Marshall M, Dong E, Squire MM, Gardner LM. Association between mobility patterns and COVID-19 transmission in the USA: a mathematical modelling study. Lancet Infect Dis. 2020;S1473-3099:30553–3. Figures Page 11/13 Figure 3 Appendix20201217.docx Figure 4 Impact of Canton Fair on epidemic with different interventions under different contact rate reduction background. the percentage change of (a) cumulative number of new infections and (b) cumulative number of confirmed cases by 31 July 2020 under different contact rate from 1 May 2020 in different event scenario; and the daily number of (c) new infections and (d) newly confirmed cases in different event scenarios with contact rate remained to be 70% of pre-epidemic level; Figure 3 Impact of contact rate and time of changing the contact rate in two-way sensitivity analysis Percentage change of cumulative number of new infections above basecase scenario by 31 July 2020, along the change of contact rate (10%-100% comparing with pre-epidemic level in x-axis) from dates of 9 February, 24 February, 31 March and 30 April 2020 (shown in color scale). Parameter values used in basecase scenario is changing to 70% contact rate from 24 February 2020. Impact of contact rate and time of changing the contact rate in two-way sensitivity analysis Percentage change of cumulative number of new infections above basecase scenario by 31 July 2020, along the change of contact rate (10%-100% comparing with pre-epidemic level in x-axis) from dates of 9 February, 24 February, 31 March and 30 April 2020 (shown in color scale). Parameter values used in basecase scenario is changing to 70% contact rate from 24 February 2020. Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. Page 13/13
https://openalex.org/W4220916907	https://www.nature.com/articles/s41598-022-08174-4.pdf	English	null	Cold tolerance strategies of the fall armyworm, Spodoptera frugiperda (Smith) (Lepidoptera: Noctuidae)	Scientific reports	2,022	cc-by	12,417	Cold tolerance strategies of the fall armyworm, Spodoptera frugiperda (Smith) (Lepidoptera: Noctuidae) OPEN Mohammad Vatanparast & Youngjin Park* The fall armyworm (FAW), Spodoptera frugiperda, is native to the tropical and subtropical areas of the American continent and is one of the world’s most destructive insect pests and invaded Africa and spread to most of Asia in two years. Glycerol is generally used as a cryoprotectant for overwintering insects in cold areas. In many studies, the increase in glycerol as a main rapid cold hardening (RCH) factor and enhancing the supercooling point was revealed at low temperatures. There are two genes, including glycerol-3-phosphate dehydrogenase (GPDH) and glycerol kinase (GK), that were identified as being associated with the glycerol synthesis pathway. In this study, one GPDH and two GK sequences (GK1 and GK2) were extracted from FAW transcriptome analysis. RNA interference (RNAi) specific to GPDH or GK1 and GK2 exhibited a significant down-regulation at the mRNA level as well as a reduction in survival rate when the RNAi-treated of FAW larvae post a RCH treatment. Following a cold period, an increase in glycerol accumulation was detected utilizing high-pressure liquid chromatography and colorimetric analysis of glycerol quantity in RCH treated hemolymph of FAW larvae. This research suggests that GPDH and GK isozymes are linked to the production of a high quantity of glycerol as an RCH factor, and glycerol as main cryoprotectant plays an important role in survival throughout the cold period in this quarantine pest studied. The fall armyworm (FAW), Spodoptera frugiperda (Smith), is native to the American continent’s tropical and subtropical regions1. It is a polyphagous insect, and due to its wide host range, it is one of the most dangerous pests affecting tropical annual crops2,3. They are usually composed of two genetically distinct strains, such as rice (R-strain) and corn (C-strain)4–6. FAW was reported in a number of Southeast Asian countries in 2018 and 2019, including India, Thailand, Myanmar, China, Japan, the Philippines, Indonesia, and most recently, Aus- tralia. The first invaded populations of FAW in South Korea were genetically confirmed using a mitochondrial cytochrome oxidase subunit I (COI) gene in 20197. The presence of ideal climatic conditions for FAW in many parts of Africa and Asia, as well as an abundance of suitable host plants, indicates that the pest can produce many generations in a single season, and that the pest is likely to become endemic8. The chance of FAW spreading would be greatly increased by its long-distance migration. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Plant Quarantine Technology Center, Animal and Plant Quarantine Agency, Gimcheon 39660, Republic of Korea. email: parky1127@korea.kr Results Gl l Glycerol content in plasma in response to low temperatures and exposure time. When fifth instar larvae of FAW were incubated at low temperatures (5 and 10 °C) compared to higher temperatures, their glycerol content increased more than threefold (15 and 20 °C) (Fig. 1A). The exposure period was also linked to an increase in the amount of glycerol in plasma, with the most glycerol found after 24 h of incubation (Fig. 1B). The high glycerol level indicated that it is a major component of plasma after cold stress. Molecular architecture of glycerol biosynthesis genes. To explain the increase in glycerol content, we attempted to identify the enzymes involved in glycerol biosynthesis (Fig. 1C). Based on a previous study30 we chose dihydroxyacetone-3-phosphate (DHAP) as a precursor of glycerol biosynthesis from glycolysis intermedi- ates. The catalytic activity of GPDH and GK converts DHAP to glycerol (Fig. 1D). As a result, the GPDH and GK genes were predicted to be involved in the synthesis of glycerol, which is an important cryoprotectant in insects when temperatures are extremely low. The transcriptome of FAW (NCBI accession number: GSE175545) were used to determine full open reading frames (ORFs) of GPDH (Sf-GPDH) and GK (Sf-GK1 and Sf-GK2) of FAW. Sf-GPDH, Sf-GK1 and Sf-GK2 ORFs encode for 353, 343, and 332 amino acid residues, respectively. Sf- GPDH protein contains a bi-domain protein structure, as illustrated in Fig. 2A that it encoded NAD+-dependent GPDHs with an N-terminal NAD+-binding domain and a C-terminal NAD+-dependent GPDH domain. Both identified FAW glycerol kinases shared N-terminal (FGGY-N) and C- terminal (FGGY-C) domains, which are colored blue and red, respectively, as shown in Fig. 2B, confirming that the targeted proteins are members of the FGGY carbohydrate kinase family. Based on comparisons with other well-known insect proteins, the three- dimensional structures of Sf-GPDH, Sf-GK1, and Sf-GK2 proteins were predicted using the homology modeling method (Fig. 3). These findings indicated that the sequences of Sf-GPDH and two Sf-GKs closely matched the homologous templates on the server, indicating that these protein models were reliable. The GPDH domain structure of Sf-NAD+-binding revealed two key components: a spatially symmetric β-sheet core and multiple helices (α1–α17) wrapping on both sides of the β-sheet core. The bioinformatics analysis indicated that four functional amino acids including Arg99, Glu100, Phe155, and Asn266 in Sf-GK1 and Arg92, Glu93, Phe148, and Asp258 in Sf-GK2 which are as glycerol binding residues (Fig. 3C, E). www.nature.com/scientificreports/ rate at – 10 °C16. Temperature is a vital abiotic variable that influences organisms’ geographic distribution and seasonal activity patterns1–5 and it has a big impact on pest biology, and abundance17. With behavioral avoidance, migration, diapause or in an extremely altered physiological state, insects escape extreme temperatures18. Since insect development takes place within a certain temperature, a change in temperature can affect the development rate, lifespan and ultimately survival of the insects19. Because of their poikilothermic nature, low temperatures act as a physical barrier preventing insects from expanding their habitats20. The insect’s survival capability is characterized as cold hardiness after exposure to low temperature levels. This procedure leads to the develop- ment of particular compounds known as cryoprotectants, which are polyols and sugars21. Insects withstand cold temperatures by holding their body fluids liquid below their normal melting point (freeze-intolerant) or by avoiding ice formation in their tissues (freeze-tolerance)22–24. The primary strategy of freeze-intolerant insects is to avoid exposure to lethal temperatures. In contrast, freeze-tolerance insects are able to overcome freezing by employing a variety of mechanisms, such as reducing ice formation in cells or delaying ice formation21,25. Another strategy, termed ‘supercooling’ is focused on the ability of insects to be cooled until spontaneous ice nucleation happens within their body fluids. The supercooling point (SCP) is the temperature at which body water spontaneously freezes24,26. While body fluid cools below its freezing point during the supercooling state, no crystallization occurs. However, in many situations, death is likely to happen at temperatures far above the SCP27,28. Insects may quickly change their response to low temperatures, either by preventing chilling injury or by modifying their behavior, a process known as rapid cold hardening (RCH) that it is associated with chemical changes in hemolymph composition to increase polyols29. Exposure to 5 °C for 6 h in Spodoptera exigua (Hüb- ner) caused a major RCH in all developmental stages, from egg to adult, which was accompanied by a strong increase in glycerol titers in hemolymph30. RCH for 2 h at 5 °C of the newly-emerged adult of five coleopteran grain-related species substantially increased the survival at different temperatures below zero as compared to the non-acclimated period31. www.nature.com/scientificreports/ According to a recent study on the cold hardiness of invasive FAW species in China, pupae and older larvae have a much higher survival rate than eggs and younger larvae, and FAW can live in some southern areas of China’s subtropical zone during the winter based on the SCPs of developmental stages at low temperatures and China’s climatic regionalization. Supercooling capacity of S. exigua32 and its RCH33,34 allows it to live at low temperatures in temperate areas during the winter. Based on high pressure liquid chromatography (HPLC) analysis of glycerol titers in response to pre-exposure to a low temperature, it was demonstrated that glycerol is a key cryoprotectant in RCH in S. exigua30.fi y y p g Because of FAW’s dispersal ability and high spreading efficiency, as well as its large reproductive capacity and wide host plant range, the pest is likely to become one of the most important migratory insect pests in South Korea that already categorized as a quarantine pest. FAW may increase cryoprotectant contents in hemolymph, such as glycerol, and may be able to endure cold seasons, but its high spreading efficiency and dispersal ability may also assist it in migrating from harsh to moderate environments. In this study we hypothesized that FAW use RCH and glycerol as an associated factor to survive to low temperatures. To investigate the function of glycerol, we used RNA interference (RNAi) to knock down genes involved in glycerol biosynthesis and then examined the intensity of RCH and glycerol accumulation. Cold tolerance strategies of the fall armyworm, Spodoptera frugiperda (Smith) (Lepidoptera: Noctuidae) OPEN The first confirmation of the invasion of FAW in Yun- nan Province (western area) of China was documented on January 11, 2019. FAW had spread to most provinces in southern China by May 20199. In reality, due to low temperatures, FAW can only successfully breed in the summer and cannot survive the winter in most areas of mainland China, Japan, and the Korean, so these areas will need to be reinvaded on an annual basis10,11. The East Asian migration area includes the Japanese Islands, Korea, and eastern China. The geographical place, ecological climate, and climatic conditions of these areas are all intertwined. Many seasonal pests, such as rice plant hoppers (Nilaparvata lugens (Stål), Sogatella furcifera (Horváth), and Laodelphax striatellus (Fallén)) and the oriental armyworm, Mythimna separate (Walker), can fly from China to Japan and the Korea12–14. Now that FAW has made its way into Southeast Asia and southern China and southern Korea, the pest has a better possibility of invading Japan and Korea9. FAW has posed a significant threat to local corn and other crop production, as well as food security. Modeling the insect’s rate of expansion and future potential migratory range using a trajectory analytical method and meteorological data during five years (2014–2018) revealed a very high probability that FAW will annually invade Korea, potentially causing a substantial decrease in agricultural productivity. g p y Since FAW do not diapause, they migrate to areas with better environmental conditions4. Even though Sparks estimated the minimum temperature for survival to be 10 °C15, it was discovered in 1979 that temperatures below 13 °C at FAW overwintering sites do not enable larvae and pupae to survive1. After exposing all stages of FAW to low temperatures for three hours, it was discovered that the egg was the most resistant, with a 30% survival \| https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 www.nature.com/scientificreports/ Results Gl l Three-dimensional analysis indicates 66% homology of Sf-GPDH with Tribolium castaneum (Herbert) GPDH under 73% coverage. When the Sf-GK1 and Sf-GK2 were compared by Spodoptera litura (Fabricus) glycerol kinase, the homology was 45 and 48% under 54 https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ (A) (B) Incubation temperature (oC) Free Glycerol (nmol/ml) of L5 haemolymph 0 20 40 60 80 100 5 10 15 25 a b b a Incubation time (h) at 5 oC 0 20 40 60 80 0 1.5 3 6 12 24 a b c c d e Free Glycerol (nmol/mL) of L5 haemolymph (C) (D) (A) (B) Incubation temperature (oC) Free Glycerol (nmol/ml) of L5 haemolymph 0 20 40 60 80 100 5 10 15 25 a b b a Incubation time (h) at 5 oC 0 20 40 60 80 0 1.5 3 6 12 24 a b c c d e Free Glycerol (nmol/mL) of L5 haemolymph (B) (2022) 12:4129 \| https://doi.org/10.1038/s41598-022-08174-4 Incubation temperature (oC) Free Glycerol (nmol/ml) of L5 haemolymph 0 20 40 60 80 100 5 10 15 25 a b b a Incubation time (h) at 5 oC 0 20 40 60 80 0 1.5 3 6 12 24 a b c c d e Free Glycerol (nmol/mL) of L5 haemolymph (C) (D) Figure 1. Measurement of free glycerol in plasma of L5 of Spodoptera frugiperda (A) at diffe larvae incubated for 24 h. (B) Effect of exposure time on glycerol content of plasma when the Each treatment was replicated three times with 10 larvae per replication. Different letters ind among means at (Type I error = 0.05, LSD test). (C) Chromatograms of hemolymph extracted for 24 h. (D) A putative glycerol production pathway. Glycerol is formed by catabolizing gluc phosphate (DHAP), which is then reduced to glycerol-3-phosphate (G3P). TRE, PGM, PGI, phosphoglucomutase, phosphoglucoisomerase, and glycerol‐3‐phosphate phosphatase, respe (2022) 12:4129 \| https://doi.org/10.1038/s41598-022-08174-4 Incubation temperature (oC) 0 5 10 15 25 Incubation time (h) at 5 oC 0 0 1.5 3 6 12 24 F (C) (D) Figure 1. Measurement of free glycerol in plasma of L5 of Spodoptera frugiperda (A) at different temperature when the larvae incubated for 24 h. (B) Effect of exposure time on glycerol content of plasma when the larvae were incubated at 10 °C Each treatment was replicated three times with 10 larvae per replication. Results Gl l In these two glycerol kinases, several amino acids were conserved including ATP- binding motif and FGGY signature motives (Figs. 3, 4). A phylogenetic analysis indicated that the Sf-GPDH and Sf-GK1 were clustered with lepidopteran insects quite distinct from other insect orders. However, interestingly Sf-GK2 was clustered with Homopteran insect (Fig. 5). Expression profile of glycerol biosynthesis genes and inducible expression in response to low temperature in FAW. Three glycerol biosynthesis genes were expressed in FAW (Fig. 6). They were expressed from egg to adult in whole stages of development (Fig. 6A, C, E). In larval stage, they were expressed in different tissues such as hemocytes, fat body, midgut, and epidermis (Fig. 6B, D, F). However, their expression levels were varied among treatments during the developmental stages. All the three genes showed high expres- sion level at adult stages of female insects. The highest expression level of all three genes was detected at midgut tissue. The expression levels of all three glycerol biosynthesis genes were inducible in response to low tempera- ture (5 °C), and they showed a positive correlation with increasing incubation time. (Fig. 6G). Glycerol content is reduced by RNAi targeting glycerol biosynthesis genes following RCH. RNAi was done on each glycerol biosynthesis gene (Sf-GPDH, Sf-GK1, and Sf-GK2) by injecting gene specific double-stranded RNAs (dsRNAs) into L5 larvae (Fig. 7). All three genes showed significant decreases (P < 0.05) with incubation time when one µg of dsRNA for each gene was injected into each larva. In all three genes, the strongest RNAi effect was observed at 48 h post injection, with a ⁓ 40–80 percent drop in mRNA expression levels (Fig. 7A).i p ( g ) RNAi downregulation of glycerol biosynthesis gene expression significantly suppressed glycerol amount (P < 0.05) in plasma at 48 h post-dsRNA injection after RCH treatment (Fig. 7B). The larvae treated with dsRNA for three genes had a basal amount of glycerol (29–35 mmol/mL), but control larvae (injected with dsRNA to enhanced green fluorescent protein (EGFP) gene) had approximately 73 mmol/mL glycerol (Fig. 7C). After RCH treatment and RNAi, the cryoprotectant(s) was monitored in hemolymph of fifth instar larvae using HPLC (Figs. 1C, 7B). Glycerol content significantly increased from 17.1 to 44.0 mM (Table 1) when the larvae were incubated at 5 °C. RNAi treatment larvae also showed a reduction in glycerol level when compare with control treatment (EGFP). Results Gl l Different letters indicate significant differences among means at (Type I error = 0.05, LSD test). (C) Chromatograms of hemolymph extracted from larvae exposed to 5 °C for 24 h. (D) A putative glycerol production pathway. Glycerol is formed by catabolizing glucose to dihydroxyacetone-3- phosphate (DHAP), which is then reduced to glycerol-3-phosphate (G3P). TRE, PGM, PGI, and GPP represent for trehalas phosphoglucomutase, phosphoglucoisomerase, and glycerol‐3‐phosphate phosphatase, respectively. ( ) (C) (C) (D) (D) Figure 1. Measurement of free glycerol in plasma of L5 of Spodoptera frugiperda (A) at different temperature when the larvae incubated for 24 h. (B) Effect of exposure time on glycerol content of plasma when the larvae were incubated at 10 °C. Each treatment was replicated three times with 10 larvae per replication. Different letters indicate significant differences among means at (Type I error = 0.05, LSD test). (C) Chromatograms of hemolymph extracted from larvae exposed to 5 °C for 24 h. (D) A putative glycerol production pathway. Glycerol is formed by catabolizing glucose to dihydroxyacetone-3- phosphate (DHAP), which is then reduced to glycerol-3-phosphate (G3P). TRE, PGM, PGI, and GPP represent for trehalase, phosphoglucomutase, phosphoglucoisomerase, and glycerol‐3‐phosphate phosphatase, respectively. https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ Figure 2. Protein domain analysis of glycerol biosynthesis genes of Spodoptera frugiperda. (A) Prediction of signature motifs of Sf-GPDH. (B) Prediction of signature motifs of Sf-GK1 and Sf-GK2. The functional domains were predicted using the NCBI conserved domain database and the EMBL-EBI HMMER database. igure 2. Protein domain analysis of glycerol biosynthesis genes of Spodoptera frugiperda. (A) Prediction of ignature motifs of Sf-GPDH. (B) Prediction of signature motifs of Sf-GK1 and Sf-GK2. The functional domains were predicted using the NCBI conserved domain database and the EMBL-EBI HMMER database. Figure 2. Protein domain analysis of glycerol biosynthesis genes of Spodoptera frugiperda. (A) Prediction of signature motifs of Sf-GPDH. (B) Prediction of signature motifs of Sf-GK1 and Sf-GK2. The functional domains were predicted using the NCBI conserved domain database and the EMBL-EBI HMMER database. and 59% coverage (Fig. 3D, F). In these two glycerol kinases, several amino acids were conserved including ATP- binding motif and FGGY signature motives (Figs. 3, 4). A phylogenetic analysis indicated that the Sf-GPDH and Sf-GK1 were clustered with lepidopteran insects quite distinct from other insect orders. However, interestingly Sf-GK2 was clustered with Homopteran insect (Fig. 5). and 59% coverage (Fig. 3D, F). Results Gl l Injection of dsGK2 resulted in a significant reduction in glycerol levels of more than seven times (6.08 mM) (Table 1, Fig. 7B). RNAi of glycerol biosynthesis genes increases the mortality of treated larvae of FAW. Larvae at 48 h post-dsRNA injection did increase their mortality after RCH treatment (Fig. 7C, D). There was no sig- nificant difference in mortality between RCH and control (no RCH) treatment after RNAi of either Sf-GPDH or https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ p Figure 3. Three-dimensional analysis of glycerol biosynthesis genes of Spodoptera frugiperda. (A, C, and E) The functional domains of Sf-GPDH, Sf-GK1, and Sf-GK2 were demonstrated, respectively. (B, D, and F) the Sf-GPDH, Sf-GK1, and Sf-GK2 proteins respectively, were compared with same protein from another well- known insect, including Tribolium castaneum and Spodoptera litura. Blue and pink region in (A) indicate beta sheet and alpha helices, respectively. In (C and E), the glycerol binding residues were indicated with blue atoms as well as yellow part that showing ATP-binding domains. N and C are an abbreviation for N-terminus and C-terminus of amino acid sequences. These models were made using SWISS-model web database. Three dimensional constructs were made using Chimera, version 1.13.1. Figure 3. Three-dimensional analysis of glycerol biosynthesis genes of Spodoptera frugiperda. (A, C, and E) The functional domains of Sf-GPDH, Sf-GK1, and Sf-GK2 were demonstrated, respectively. (B, D, and F) the Sf-GPDH, Sf-GK1, and Sf-GK2 proteins respectively, were compared with same protein from another well- known insect, including Tribolium castaneum and Spodoptera litura. Blue and pink region in (A) indicate beta sheet and alpha helices, respectively. In (C and E), the glycerol binding residues were indicated with blue atoms as well as yellow part that showing ATP-binding domains. N and C are an abbreviation for N-terminus and C-terminus of amino acid sequences. These models were made using SWISS-model web database. Three dimensional constructs were made using Chimera, version 1.13.1. Sf-GK2, However the mortality decreased significantly when the larvae injected with dsRNA specific to Sf-GK1 after RCH treatment than to control (Fig. 7C). Following RCH treatment, the SCP increased. The effect of RCH on SCP was evaluated in all devel- opmental stages including both sexes in pupal and adult stages (Table 1). Egg, first instar and pupal stages exhib- ited SCP at lowest temperature than other developmental stages. Results Gl l The data showed that supercooling capacity was unaffected by RCH treatment in egg, first and second instar, male pupae, and female adult whereas SCP temperature in the others was significantly reduced (Table 2). From these data, we found that RCH treatment is often accompanied by elevated SCPs. To investigate the involvement of glycerol biosynthesis genes in SCP, RNAi-treated larvae (L3 to L6) were incubated at RCH conditions and their SCP was assessed (Table 3). The SCPs of larvae injected with dsRNA specific to glycerol biosynthesis genes, specifically dsGPDH and dsGK2, were significantly lower than those of dsEGFP-injected larvae, suggesting that glycerol biosynthesis genes elevate https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ Figure 4. Alignment of amino acid sequences of glycerol biosynthesis genes including (A) glyce phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera f other well-known insect. Sequence alignment used Clustal W program of MegAlign (DNASTAR The abbreviation explanation and GenBank accession numbers of sPLA2 amino acids are listed i Figure 4. Alignment of amino acid sequences of glycerol biosynthesis genes including (A) glycerol-3- phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with other well-known insect. Sequence alignment used Clustal W program of MegAlign (DNASTAR, Version 7.0). The abbreviation explanation and GenBank accession numbers of sPLA2 amino acids are listed in Table S2. Figure 4. Alignment of amino acid sequences of glycerol biosynthesis genes including (A) glycerol-3- phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with other well-known insect. Sequence alignment used Clustal W program of MegAlign (DNASTAR, Version 7.0). The abbreviation explanation and GenBank accession numbers of sPLA2 amino acids are listed in Table S2. https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 5. Phylogenetic analysis of glycerol biosynthesis genes including (A) glycerol-3-phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with species from different orders. The analysis was performed using MEGA6.06. Each node contains bo value after 1000 replications to support branching and clustering. Accession numbers and the abbre explanation are shown in Table S2. Figure 5. Phylogenetic analysis of glycerol biosynthesis genes including (A) glycerol-3-phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with other insect species from different orders. The analysis was performed using MEGA6.06. Results Gl l Each node contains bootstrap value after 1000 replications to support branching and clustering. Accession numbers and the abbreviation explanation are shown in Table S2. Figure 5. Phylogenetic analysis of glycerol biosynthesis genes including (A) glycerol-3-phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with other insec species from different orders. The analysis was performed using MEGA6.06. Each node contains bootstrap value after 1000 replications to support branching and clustering. Accession numbers and the abbreviation explanation are shown in Table S2. Figure 5. Phylogenetic analysis of glycerol biosynthesis genes including (A) glycerol-3-phosphate dehydrogenase (GPDH) and (B) two glycerol kinases (GK1 and GK2) of Spodoptera frugiperda with other insect species from different orders. The analysis was performed using MEGA6.06. Each node contains bootstrap value after 1000 replications to support branching and clustering. Accession numbers and the abbreviation explanation are shown in Table S2. https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ (A) (C) (B) (A) (C) (B) (D) (E) (F) Time (h) at 5°C 0 2 4 6 8 10 12 14 16 18 20 22 24 Relative mRNA expression level 0 2 4 6 8 10 12 14 16 GPDH GK1 GK2 a a a b c c c a a b c d d d A B B BC C C C (G) (C) (D) Figure 6. Expression analysis of glycerol biosynthesis genes from Spodoptera frugiperda. RT-PCR analysis of Sf-GPD Sf-GK2 in different developmental stages (A, C, and E, respectively) and tissues (B, D, and F, respectively). The gels in E) were cropped from various gels and were cleared with vertical white space. The full-length gels are included in Supp Information (Fig. S2). EF1 was used to validate cDNA integrity. Different developmental stages included Egg (‘Eg’), la (ʻL1–L6ʼ), pre-pupa (‘PP’), pupa of female (ʻPFʼ), pupa of male (‘PM’), adult of female (ʻAFʼ) and adult of male (‘AM’ tissues included hemocyte, fat body, midgut, and epidermis. (G) RT-PCR analysis of Sf-GPDH, Sf-GK1 and Sf-GK2 ex temperature (5 °C) at different exposure time (h). Agarose gel (1%) was used for electrophoresis. Each treatment was r times. Different letters above standard deviation bars indicate significant difference among means at Type I error = 0.0 (D) (C) (F) Time (h) at 5°C Figure 6. Expression analysis of glycerol biosynthesis genes from Spodoptera frugiperda. Results Gl l RT-PCR analysis of Sf-GPDH, Sf-GK1, and Sf-GK2 in different developmental stages (A, C, and E, respectively) and tissues (B, D, and F, respectively). The gels in (A, C, and E) were cropped from various gels and were cleared with vertical white space. The full-length gels are included in Supplementary Information (Fig. S2). EF1 was used to validate cDNA integrity. Different developmental stages included Egg (‘Eg’), larval instars (ʻL1–L6ʼ), pre-pupa (‘PP’), pupa of female (ʻPFʼ), pupa of male (‘PM’), adult of female (ʻAFʼ) and adult of male (‘AM’). Different tissues included hemocyte, fat body, midgut, and epidermis. (G) RT-PCR analysis of Sf-GPDH, Sf-GK1 and Sf-GK2 expression at low temperature (5 °C) at different exposure time (h). Agarose gel (1%) was used for electrophoresis. Each treatment was replicated three times. Different letters above standard deviation bars indicate significant difference among means at Type I error = 0.05 (LSD test). https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ Figure 7. The effect of RNA interference (RNAi) specific to genes linked with glycerol biosynthesis, GPDH and GK of Spodoptera frugiperda. All dsRNAs specific to target glycerol biosynthesis genes were constructed at ~ 300–400 bp and injected to each L5 larva at 3 μg. Control RNAi (‘dsEGFP’) was injected with dsRNA specific to EGFP gene. (A) Quantitative real-time PCR to monitor changes in mRNA levels of Sf-GPDH, Sf-GK1, and Sf-GK2 after RNAi. EF1 was used to validate cDNA integrity. (B) Chromatogram of HPLC that shows the effect of RNAi specific to glycerol biosynthesis genes, (48 h post dsRNAs injection) and RCH treatment (5 °C for 6 h), on glycerol content in hemolymph of fifth instar larvae. (C) Suppression of cold tolerance after RNAi treatment of either GPDH or GK1 or GK2. The glycerol content in hemolymph of fifth instar larvae was measured after 48 h of dsRNAs injection. (D, E) After RNAi injection (48 h post injection) and RCH treatment (5 °C for 6 h), the larvae were incubated at − 10 °C for 1 h and the mortality was recorded. Each treatment was replicated three times with 10 individuals per replication. Asterisks indicate significant difference between RCH and no RCH treatments (Type I error = 0.05, LSD test). n.s. means no significant difference. Figure 7. The effect of RNA interference (RNAi) specific to genes linked with glycerol biosynthesis, GPDH and GK of Spodoptera frugiperda. Results Gl l All dsRNAs specific to target glycerol biosynthesis genes were constructed at ~ 300–400 bp and injected to each L5 larva at 3 μg. Control RNAi (‘dsEGFP’) was injected with dsRNA specific to EGFP gene. (A) Quantitative real-time PCR to monitor changes in mRNA levels of Sf-GPDH, Sf-GK1, and Sf-GK2 after RNAi. EF1 was used to validate cDNA integrity. (B) Chromatogram of HPLC that shows the effect of RNAi specific to glycerol biosynthesis genes, (48 h post dsRNAs injection) and RCH treatment (5 °C for 6 h), on glycerol content in hemolymph of fifth instar larvae. (C) Suppression of cold tolerance after RNAi treatment of either GPDH or GK1 or GK2. The glycerol content in hemolymph of fifth instar larvae was measured after 48 h of dsRNAs injection. (D, E) After RNAi injection (48 h post injection) and RCH treatment (5 °C for 6 h), the larvae were incubated at − 10 °C for 1 h and the mortality was recorded. Each treatment was replicated three times with 10 individuals per replication. Asterisks indicate significant difference between RCH and no RCH treatments (Type I error = 0.05, LSD test). n.s. means no significant difference. https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ (C) Relative glycerol content in hemolymph of L5 0 20 40 60 80 100 dsEGFP dsGPDH dsEGFP dsEGFP dsSf-GK1 dsSf-GK2 * * (D) (E) dsEGFP dsGPDH dsGK1 dsGK2 Relative mortalityof treated L5 with dsRNAs (%) 0 20 40 60 80 100 RCH No RCH * n.s. n.s. * Figure 7. (continued) (C) Relative glycerol content in hemolymph of L5 0 20 40 60 80 100 dsEGFP dsGPDH dsEGFP dsEGFP dsSf-GK1 dsSf-GK2 * * * 10 s \| (2022) 12:4129 \| https://doi.org/10.1038/s41598-022-08174-4 Relative glyc 0 20 dsEGFP dsGPDH dsEGFP dsEGFP dsSf-GK1 dsSf-GK2 (D) (E) dsEGFP dsGPDH dsGK1 dsGK2 Relative mortalityof treated L5 with dsRNAs (%) 0 20 40 60 80 100 RCH No RCH * n.s. n.s. * Figure 7. (continued) Table 1. Change polyol content in Spodoptera frugiperda fifth instar hemolymph in response to exposure to 5 °C. Each treatment was replicated three times with 10 individuals per replication. Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). Results Gl l Stage RCH treatment N SCP (°C) Egg No 90 − 17.50 ± 0.56a Yes 90 − 17.80 ± 0.35a L1 No 90 − 17.20 ± 1.40a Yes 90 − 17.50 ± 0.64a L2 No 60 − 14.60 ± 1.10a Yes 60 − 15.80 ± 0.85a L3 No 10 − 10.62 ± 0.59b Yes 10 − 14.12 ± 0.73a L4 No 5 − 12.15 ± 0.62b Yes 5 − 14.15 ± 0.45a L5 No 5 − 8.58 ± 0.93b Yes 5 − 13.88 ± 0.48a L6 No 5 − 8.35 ± 0.49b Yes 5 − 13.39 ± 0.45a Pre-pupae No 5 − 14.62 ± 0.67b Yes 5 − 16.19 ± 0.57a Pupae (male) No 5 − 18.47 ± 1.19a Yes 5 − 18.87 ± 0.53a Pupae (female) No 5 − 16.90 ± 0.35b Yes 5 − 18.20 ± 1.12a Adult (male) No 5 − 12.33 ± 1.20b Yes 5 − 14.15 ± 0.58a Adult (female) No 5 − 13.23 ± 1.52a Yes 5 − 14.53 ± 0.82a Stage RCH treatment N SCP (°C) Egg No 90 − 17.50 ± 0.56a* Yes 90 − 17.80 ± 0.35a L1 No 90 − 17.20 ± 1.40a Yes 90 − 17.50 ± 0.64a L2 No 60 − 14.60 ± 1.10a Yes 60 − 15.80 ± 0.85a L3 No 10 − 10.62 ± 0.59b Yes 10 − 14.12 ± 0.73a L4 No 5 − 12.15 ± 0.62b Yes 5 − 14.15 ± 0.45a L5 No 5 − 8.58 ± 0.93b Yes 5 − 13.88 ± 0.48a L6 No 5 − 8.35 ± 0.49b Yes 5 − 13.39 ± 0.45a Pre-pupae No 5 − 14.62 ± 0.67b Yes 5 − 16.19 ± 0.57a Pupae (male) No 5 − 18.47 ± 1.19a Yes 5 − 18.87 ± 0.53a Pupae (female) No 5 − 16.90 ± 0.35b Yes 5 − 18.20 ± 1.12a Adult (male) No 5 − 12.33 ± 1.20b Yes 5 − 14.15 ± 0.58a Adult (female) No 5 − 13.23 ± 1.52a Yes 5 − 14.53 ± 0.82a Table 2. Changes in supercooling points of Spodoptera frugiperda after rapid cold hardening treatment (5 °C for 6 h). Each treatment was replicated three times with different individuals per replication. Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Table 3. Results Gl l Stage RCH treatment N SCP (°C) dsEGFP dsGPDH dsGK1 dsGK2 L3 No 18 − 17.20 ± 0.53a − 12.30 ± 0.87c − 15.76 ± 0.45b − 8.65 ± 0.44d Yes 18 − 17.30 ± 0.51a − 13.70 ± 0.73b − 16.95 ± 0.34a − 9.75 ± 0.61c L4 No 18 − 16.80 ± 2.40a − 10.24 ± 0.70b − 16.21 ± 0.42a − 7.88 ± 0.68c Yes 18 − 17.70 ± 0.62a − 10.32 ± 1.12b − 16.80 ± 1.10a − 8.54 ± 0.48c L5 No 18 − 14.10 ± 0.80a − 8.76 ± 0.92c − 12.76 ± 0.62b − 7.36 ± 0.74c Yes 18 − 15.70 ± 1.85a − 10.23 ± 0.65b − 13.48 ± 0.74a − 9.87 ± 0.56c L6 No 18 − 11.82 ± 0.32a − 7.94 ± 0.62c − 9.54 ± 0.76b − 7.24 ± 0.23c Yes 18 − 13.02 ± 0.52a − 9.39 ± 0.45b − 10.88 ± 0.33b − 7.56 ± 0.36c Stage RCH treatment N SCP (°C) dsEGFP dsGPDH dsGK1 dsGK2 L3 No 18 − 17.20 ± 0.53a* − 12.30 ± 0.87c − 15.76 ± 0.45b − 8.65 ± 0.44d Yes 18 − 17.30 ± 0.51a − 13.70 ± 0.73b − 16.95 ± 0.34a − 9.75 ± 0.61c L4 No 18 − 16.80 ± 2.40a − 10.24 ± 0.70b − 16.21 ± 0.42a − 7.88 ± 0.68c Yes 18 − 17.70 ± 0.62a − 10.32 ± 1.12b − 16.80 ± 1.10a − 8.54 ± 0.48c L5 No 18 − 14.10 ± 0.80a − 8.76 ± 0.92c − 12.76 ± 0.62b − 7.36 ± 0.74c Yes 18 − 15.70 ± 1.85a − 10.23 ± 0.65b − 13.48 ± 0.74a − 9.87 ± 0.56c L6 No 18 − 11.82 ± 0.32a − 7.94 ± 0.62c − 9.54 ± 0.76b − 7.24 ± 0.23c Yes 18 − 13.02 ± 0.52a − 9.39 ± 0.45b − 10.88 ± 0.33b − 7.56 ± 0.36c Table 3. Changes in supercooling points of Spodoptera frugiperda (third to sixth instar) after RNAi and rapid cold hardening treatment. All dsRNAs specific to target glycerol biosynthesis genes were constructed at ~ 300–400 bp and injected to each L5 larva at 3 μg. Control RNAi (‘dsEGFP’) was injected with dsRNA specific to EGFP gene. The supercooling point was measured after RNAi injection (48 h post injection) and RCH treatment (5 °C for 6 h). Results Gl l RNAi treatment groups Glycerol (mM)/hemolymph of fifth instar Trehalose (mM)/hemolymph of fifth instar dsEGFP-noRCH 17.10 ± 0.98d* 4.42 ± 0.10d dsEGFP-RCH 44.07 ± 0.95a 5.99 ± 0.13c dsGPDH-RCH 22.58 ± 0.82c 6.89 ± 0.08b dsGK1-RCH 40.36 ± 0.30b 5.70 ± 0.17c dsGK2-RCH 6.08 ± 0.03e 21.53 ± 0.23a (D) (E) dsEGFP dsGPDH dsGK1 dsGK2 Relative mortalityof treated L5 with dsRNAs (%) 0 20 40 60 80 100 RCH No RCH * n.s. n.s. * Figure 7. (continued) (D) (E) Figure 7. (continued) 10 34567890) Scientific Reports \| (2022) 12:4129 \| https://doi.org/10.1038/s41598-022-08174-4 Table 1. Change polyol content in Spodoptera frugiperda fifth instar hemolymph in response to exposure to 5 °C. Each treatment was replicated three times with 10 individuals per replication. Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). RNAi treatment groups Glycerol (mM)/hemolymph of fifth instar Trehalose (mM)/hemolymph of fifth instar dsEGFP-noRCH 17.10 ± 0.98d 4.42 ± 0.10d dsEGFP-RCH 44.07 ± 0.95a 5.99 ± 0.13c dsGPDH-RCH 22.58 ± 0.82c 6.89 ± 0.08b dsGK1-RCH 40.36 ± 0.30b 5.70 ± 0.17c dsGK2-RCH 6.08 ± 0.03e 21.53 ± 0.23a Table 1. Change polyol content in Spodoptera frugiperda fifth instar hemolymph in response to exposure to 5 °C. Each treatment was replicated three times with 10 individuals per replication. Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each polyol (Type I error = 0.05, LSD test). https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ Table 2. Changes in supercooling points of Spodoptera frugiperda after rapid cold hardening treatment (5 °C for 6 h). Each treatment was replicated three times with different individuals per replication. Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Results Gl l Changes in supercooling points of Spodoptera frugiperda (third to sixth instar) after RNAi and rapid cold hardening treatment. All dsRNAs specific to target glycerol biosynthesis genes were constructed at ~ 300–400 bp and injected to each L5 larva at 3 μg. Control RNAi (‘dsEGFP’) was injected with dsRNA specific to EGFP gene. The supercooling point was measured after RNAi injection (48 h post injection) and RCH treatment (5 °C for 6 h). Each treatment was replicated three times with 18 individuals per replication. Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Results Gl l Stage RCH treatment N SCP (°C) Egg No 90 − 17.50 ± 0.56a Yes 90 − 17.80 ± 0.35a L1 No 90 − 17.20 ± 1.40a Yes 90 − 17.50 ± 0.64a L2 No 60 − 14.60 ± 1.10a Yes 60 − 15.80 ± 0.85a L3 No 10 − 10.62 ± 0.59b Yes 10 − 14.12 ± 0.73a L4 No 5 − 12.15 ± 0.62b Yes 5 − 14.15 ± 0.45a L5 No 5 − 8.58 ± 0.93b Yes 5 − 13.88 ± 0.48a L6 No 5 − 8.35 ± 0.49b Yes 5 − 13.39 ± 0.45a Pre-pupae No 5 − 14.62 ± 0.67b Yes 5 − 16.19 ± 0.57a Pupae (male) No 5 − 18.47 ± 1.19a Yes 5 − 18.87 ± 0.53a Pupae (female) No 5 − 16.90 ± 0.35b Yes 5 − 18.20 ± 1.12a Adult (male) No 5 − 12.33 ± 1.20b Yes 5 − 14.15 ± 0.58a Adult (female) No 5 − 13.23 ± 1.52a Yes 5 − 14.53 ± 0.82a Table 2. Changes in supercooling points of Spodoptera frugiperda after rapid cold hardening treatment (5 °C for 6 h). Each treatment was replicated three times with different individuals per replication. Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Discussion Many insect species can develop cold-hardiness well below freezing temperatures, and various features of insect cold-hardiness have been studied23,35. The most significant part of acclimatization for cold resistance is low temperature exposure22,36. Low-weight molecular molecules, often known as cryoprotectant, such as polyols and sugars, are produced during this procedure21. The most prevalent cryoprotectants include polyols (glycerol, sorbitol, and manitol), sugars (glucose, trehalose, and fructose), and amino acids37–40. High polyol concentrations not only lower the temperature at which an insect’s body fluids crystallize but also stabilize the state of proteins, even when collected in relatively low concentrations41. Polyols regulate the amount of water accessible for freez- ing, which reduces the amount of cell dehydration caused by extracellular freezing. They protect biological membranes and proteins from freezing-induced dehydration by preserving their structures41,42. In the present work, the tolerance of FAW was analyzed by rapid cold hardening (RCH). In insects without diapause, RCH is especially important for overcoming lethal cold shock by rapidly increasing cold tolerance20. RCH has been induced in a variety of insects at temperatures ranging from 0 to 5 °C30,43–47. Glycerol production is divided into two distinct pathways, depending on the insect. In Epiblema scudderiana (Clemens), a moth belongs to Tortrici- dae family, polyol dehydrogenase catalyzes the reaction of glyceraldehyde with NADPH + H+ in one route48. The other pathway converts dihydroxyacetone-3-phosphate to glycerol via GPDH/GK (S. exigua)30. Identification of key genes associated with overwintering in Anoplophora glabripennis (Motschulsky) larva, a coleopteran species, using gene co-expression network analysis, was demonstrated that, fatty acid desaturase, glycerol phosphate dehydrogenase, glycerol kinase, and trehalose phosphate synthase were among the 15 genes implicated in the control of antifreeze protectants49. We studied on GPDH and GK genes expression to investigate the glycerol production pathway. In the FAW transcriptome, we discovered two GK isoforms and one GPDH isoform. It was discovered that both genes expressed and associated with glycerol biosynthesis pathway. The whole Plutella xylostella (Linnaeus) genome was used to predict four GKs and one GPDH50. The genome of FAW contains only one type of GPDH, indicating that it is a unigene with a conserved biological function in metabolism. Because we obtained these sequences from transcriptome data and there are likely no other endogenous genes of GPDH and GK, we believe our expression and functional analysis are associated with these isozymes. GPDH and both GK isoforms were discovered to be widely expressed in different studied tissues. Discussion As we know at low temperatures, most gene expression decreases51. However, in 5 °C, real-time PCR of cold-exposed larvae revealed that GPDH, GK1, and GK2 were expressed at relatively high levels (Fig. 6G). This suggests that these proteins are important for cold tolerance to the low temperature by RCH. As found in other insects52,53, cold tolerance rose as acclima- tization time increased, which could be in line with our findings, that mRNA expression levels of analyzed genes increased as incubation time increased (Fig. 6G).i g RNAi is a non-invasive way of delivering dsRNA into insects to knockdown specific gene expression54–57. We have shown that injecting RNAi is feasible and can suppress the transcription level of target genes in FAW larvae. Our system confirmed the effective knockdown of three genes at the mRNA expression level.hi if Their expression was knocked-down by specific dsRNAs associated with glycerol biosynthesis genes. In response to pre-exposure to a low temperature, this RNAi treatment reduced RCH and prevented glycerol accumulation. According to the RNAi experiments, sorbitol dehydrogenase, trehalose-6-phosphate synthase, and glycerol kinase are all involved in the overwintering stage of Chinese white pine larvae (Dendroctonus armandi (Tsai and Li))58. Glycerol phosphorylation, which is essential for glycerol consumption, is catalyzed by GK59,60. GK has a function in overwintering termination in Hyalophora cecropia (Linnaeus) eggs that accumulate glycerol by converting glycerol to glycerol-3-phosphate for other intermediary metabolism61. In A. glabripennis larvae, the gene expression level of glycerol kinase increased sharply at the midpoint of the overwintering stage, and then declined at the latter, which corresponded to the change in glycerol content. The findings suggest that glycerol kinase is involved in the synthesis of glycerol, which could help this insect adapt to low temperatures49.i Because RNAi targeting GK1 and GK2 significantly reduced glycerol accumulation in a 5 °C pretreatment, it was shown that both GKs catalyze the dephosphorylation of glycerol-3-phosphate to generate glycerol, as reported earlier in S. exigua, a near Noctuidae species to FAW30,62. However, it was discovered that GK2 has a greater effect on glycerol production based on RNAi data for mortality, glycerol accumulation, and HPLC results. The significant increase in Sf-GK2 expression vs Sf-GK1 shows that it has physiological significance in RCH, as evidenced by the RNAi functional study. P. xylostella GK1 showed a significant increase in expression in response to 5 °C exposure vs other three isozyme of GKs50. Results Gl l Each treatment was replicated three times with 18 individuals per replication. Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). Different superscript letters indicate significant difference between means for each SCP (Type I error = 0.05, LSD test). their SCP by accumulating extracellular cryoprotectant including glycerol in their bodies. In compared to the dsEGFP treatment group, the SCPs of larvae injected with dsGK1 were significantly lower than to dsGPDH and dsGK2 injected larvae (Table 3). their SCP by accumulating extracellular cryoprotectant including glycerol in their bodies. In compared to the dsEGFP treatment group, the SCPs of larvae injected with dsGK1 were significantly lower than to dsGPDH and dsGK2 injected larvae (Table 3). https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ Materials and methods Insect rearing, exposure temperatures and sample preparation. The larvae of FAW were obtained from Frontier Agriculture Sciences (Newark, DE, USA) and used F4 and F5 generations, which were main- tained in laboratory. They were raised until pupation under laboratory-controlled conditions (26 °C, 70% RH, and a photoperiod of 14 h:10 h [L:D]). The larvae were fed an artificial diet (Newark, DE, USA) during their development67 and larval instar (L1-L5) were determined by head capsule sizes and molting times. The diet was changed every day for larvae. These larvae were grown in plastic containers with aerated lids measuring 40 × 20 × 15 cm. From the third instar onwards, larvae were reared separately to prevent cannibalism. This was carried out in Petri dishes (8.5 cm diameter). Bioinformatics to predict Glycerol synthase genes. The mRNA sequences for all genes were obtained from a previous study of transcriptome analysis of whole body of FAW using TruSeq RNA Sample Prep Kit v2 (Macrogen, Seoul, South Korea). The predicted amino acid sequences were aligned using Clustal W program of MegAlign (DNASTAR Version 7.0). The phylogenetic trees were constructed with Neighbor-joining method and Poisson correction model (1,000 bootstrap repetitions to support branching clusters) using MEGA 7.0 software (www.megasoftware.net). Conserved domains of glycerol synthase genes were predicted using NCBI Conserved Domain Database (www.ncbi.nlm.nih.gov/cdd). UCSF Chimera (https://www.cgl.ucsf.edu/chimera/) was used for protein motif analysis and making 3D structure. RNA extraction and RT‑qPCR. RNAs samples were extracted from FAW larvae using Trizol reagent (Inv- itrogen, Carlsbad, CA, USA)68. After RNA extraction, it was resuspended in nuclease-free water and quantified using a spectrophotometer (NanoDrop, Thermo Scientific, Wilmington, DE, USA). cDNA was then synthesized from RNA (1 μg) using RT PreMix (Intron Biotechnology, Seoul, Korea) containing oligo dT primer according to the manufacturer’s instruction. All quantitative PCRs (qPCRs) in this study were determined using a Real time PCR machine (CFX Connect Real-Time PCR Detection System, Bio-Rad, Hercules, CA, USA) and iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) according to the guideline of manufacture. The reaction mixture (20 μL) contain 10 μL of iQ SYBR Green Supermix, 1 μL of cDNA template (100 ng), and 1 μL each of forward and reverse primers (Table S1) and 7 μL nuclease free water. Materials and methods RT-qPCR cycling began with 95 °C heat treatment for 10 min followed by 40 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C temperature for 30 s, and extension at 72 °C for 20 s. Expression level of EF1 as reference gene was used to normalize target gene expression levels69 under different treatments. PCR products were assessed by melting curve analysis. Quantita- tive analysis was performed using comparative CT (2−ΔΔCT) method70. RCH bioassay. RCH was measured according to a previous method30. The each developmental stage from eggs to adults was exposed to 5 °C for 6 h prior to − 10 °C for 1 h. For each treatment group, test individuals were placed in a Petri dish (10 × 15 mm). After 2 h of recovery at 25 °C after cold treatment, the survival rates of all developmental stages were determined. After gentle probing on the abdomen with a stick, autonomous move- ment of individuals was the criterion for being categorized as alive. Hatching in the 25 °C recovery state was used to assess egg survival. Adult emergence in the 25 °C recovery state was used to assess pupal survival. SCP measurement. SCPs were measured using a thermocouple (BTM-4208SD, LT Lutron, Taipei, Taiwan) to detect the release of the latent heat of fusion as body water froze, as described previously32,71. In SCP meas- urement, all developmental stages of FAW were examined after RCH treatment (exposed to 5 °C for 6 h prior to − 10 °C for 1 h). The thermocouples were kept in contact with the cuticle by putting the insect in a 1.5 mL tube and filling it with cotton wool to keep the insect and thermocouple together (Figure S1). They were then put in a styrofoam box (30 × 30 × 20 cm), and the box was placed into a freezer at − 80 °C. The cooling rate was measured as 1 °C min−1. Glycerol analysis in FAW hemolymph. The glycerol content of the samples was determined using the Glycerol Assay Kit (BioVision, Milpitas, CA, USA). We followed the manufacturer’s instruction for fluorometric measurements. In summary, a hemolymph from 10 fifth instar larvae (Day 1) was collected by cutting prologs of the treated larvae and mixed with a 100 µL volume of anticoagulant buffer (ACB). ACB was prepared with 186 mM NaCl, 17 mM Na2EDTA, and 41 mM citric acid72. www.nature.com/scientificreports/ on glycerol accumulation and insect mortality in response to low temperature pre-exposure, revealed that glyc- erol is a substantial cryoprotectant in RCH in FAW. Increased glycerol concentrations may contribute to whole animal freeze tolerance by enhancing cell survival by freeze-tolerant. However, each cryoprotectant may have a distinct non-overlapping function and contribute to freeze tolerance through memchanisms distinct from those of others with different potency. In addition, the permeability of different tissues to cryoprotectants can be vary, affecting their ability to protect cells and this constituents. Supercooling data clearly demonstrated that FAW can endure very low temperatures, and as a key agricultural pest, it may be able to become one of most important migratory insect pests in Korea. To limit the impact of this pest, it is critical to create pest manage- ment strategies and detecting systems. In addition, more research on migration behavior is needed to predict source areas and migration times. Discussion In Bombyx mori (Linnaeus), at least three GK isozymes have been discovered, but only one, GK3, appears to be connected to glycerol utilization63. The knockdown of the target genes Sf-GPDH, Sf-GK1, and Sf-GK2 not only reduces their transcription levels but also affects larval cold-tolerance capacity, leading to an increase in low-temperature mortality. The most obvious explanation for these findings is that these genes are necessary for overwintering larvae’s cold tolerance. As there is no existing evidence of systemic spread in Lepidoptera64, we were unable to totally silence these three genes, however, the partial knockdown had a clear effect on low-temperature mortality.h f The hemolymph polyol analysis revealed that trehalose was the primary blood sugar, with a concentration of 4.42 mmol−1 in hemolymph and a slight increase with low temperature exposure (5.99 mmol−1 after 6 h at 4 °C). Trehalose titers in insect hemolymph are relatively high in general, but very considerably between insects (rang- ing from 0.1 to 133 mmol−1)30. We detected 5.7 mM of trehalose following dsGK1-RCH treatment, whereas the titer increased considerably to 21.53 mM following dsGK2-RCH treatment. This may be a compensating effect of the glycerol depletion. However, we believe that in order to obtain more precise results, we need incorporate trehalose(s) (which catalyzes the conversion of trehalose to two glucose monomers) into our future studies. In conclusion, due to a lack of a diapause mechanism, FAW cannot overwinter in area with a cold winter, despite the fact that they can disperse thousands of kilometers north during the growth season65,66. However, in this study, RNAi investigation of two types of important genes linked to glycerol production and their effects Scientific Reports \| (2022) 12:4129 \| https://doi.org/10.1038/s41598-022-08174-4 www.nature.com/scientificreports/ Materials and methods The ACB was adjusted to pH 8.0 by the addition of NaOH. The resultant hemolymph was centrifuged at 13,500 rpm for 10 min at 4 °C. The supernatant (100 µL) were mixed with 100 µL of glycerol assay buffer (GAB) (provided by the kit) for 10 min on ice. The amount of 10 µL resulted supernatant (12,000 rpm, 5 min) was mixed with 86 µL GAB followed by 2 µL Probe (provided by https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ the kit) and 2 µL glycerol enzyme mix (GEM) (provided by the kit) in a 96 well plate. The background control mixture was prepared as described above without GEM. In this assay, glycerol in the presence of glycerol enzyme mix is converted to an intermediate after incubation at 37 °C for 60 min, which reduces a colorless probe to a colored product with strong absorbance at 450 nm. Down‑regulation of associated glycerol biosynthesis genes by RNA interference (RNAi). RNAi was performed using dsRNA prepared with Megascript RNAi Kit (Ambion, Austin, TX, USA) according to the manufacturer’s instruction and a previous method73. Partial segments were amplified with gene-specific primers containing T7 promoter sequence at 5′ end (Table S1). dsRNAs (dsGPDH, dsGK1 and dsGK2) were synthesized at 37 °C for 4 h and then left at 70 °C for 5 min to inactivate T7 RNA polymerase. As control dsRNA (‘dsCON’), 300 bp fragment of enhanced green fluorescent protein (EGFP) was synthesized74. Three µg of dsRNA (1 µg/ µL) was injected into each fifth instar larva with a Hamilton micro syringe. RNAi efficiency was determined by qPCR described above at 24, 48, 72, and 96 h post injection. For each treatment, at least 10 larvae were used. Each treatment was replicated three times. Sample preparation and HPLC condition. Hemolymph of 10 fifth instar (Day 1) was collected by cut- ting prologs of the treated larvae and mixed with a 100 µL volume of ACB. The resultant hemolymph was centri- fuged at 13,500 rpm for 10 min at 4 °C. The supernatant (500 µL) was transferred to a new 1.5 mL tube, and then the same volume of acetonitrile (ACN) was added into the tubes and were shaken for 15 s. The tubes were incu- bated at room temperature for 10 min and then centrifuged as described above. The upper phase was collected in a new 1.5 mL tube. Materials and methods The previous step was repeated with the addition of 250 µL ACN to increase the purifica- tion. The final supernatant was filtered out by 0.22 µM syringe filters. The purified samples were directly used for HPLC in Metabolomics Research Center for Functional Materials, Kyungsung Univeristy (Busan, Korea). A reversed-phase HPLC connected to an evaporative light scattering detector (ELSD) (ELSD-LT II, Shimadzu, Japan) was optimized for simultaneous determination of cryoprotectant. HPLC separation was achieved using a Unison UK-Amino column (250 × 4.6 mm). Water and acetonitrile were used as the mobile phase. The ideal flow rate of 0.7 mL/min and a proportion of acetonitrile of 90% over 30 min were used to optimize the separation of cryoprotectant using isocratic elution conditions. The temperature of the ELSD detector was set at 30 °C with a temperature of column oven at 64 °C. Calibration curves were generated in GraphPad Prism by plotting the area against cryoprotectant concentration. Averages of the areas for each standard were calculated and plotted against the known concentrations. Data analysis. All studies were performed in three independent biological replicates. Results were plotted using Sigma plot 10.0. Means were compared by least squared difference (LSD) test of one-way analysis of vari- ance (ANOVA) using PROC GLM of SAS program75,76 and discriminated at Type I error = 0.05. Received: 14 November 2021; Accepted: 3 March 2022 Received: 14 November 2021; Accepted: 3 March 2022 References 1. Sparks, A. N. A review of the biology of the fall armyworm. Florida Entomol. 62, 82–87 (1979).h 1. Sparks, A. N. A review of the biology of the fall armyworm. Florida Entomol. 62, 82–87 (1979).h 2. Andrews, K. L. The whorlworm, Spodoptera frugiperda, in Central America and neighboring areas. Florida Entomol. 456–467 (1980).f 2. Andrews, K. L. The whorlworm, Spodoptera frugiperda, in Central America and neighborin (1980). 3. Cruz, I., Figueiredo, M. L. C., Oliveira, A. C. & Vasconcelos, C. A. Damage of Spodoptera frugiperda (Smith) in different maize genotypes cultivated in soil under three levels of aluminium saturation. Int. J. Pest Manag. 45, 293–296 (1999).hf 4. Du Plessis, H., Schlemmer, M. L. & Van den Berg, J. The effect of temperature on the development of Spodoptera frugiperda (Lepi- doptera: Noctuidae). Insects 11, 228 (2020). 5. Quisenberry, S. S. Fall armyworm (Lepidoptera: Noctuidae) host strain reproductive compatibility. Florida Entomol. 74, 194–199 (1991). 6. Nagoshi, R. N. & Meagher, R. L. Seasonal distribution of fall armyworm (Lepidoptera: Noctuidae) host strains in agricultural and turf grass habitats. Environ. Entomol. 33, 881–889 (2004). 7. Lee, G.-S. et al. First report of the fall armyworm, Spodoptera frugiperda (Smith, 1797) (Lepidoptera, Noctuidae), a new migratory pest in Korea. Korean J. Appl. Entomol. 59, 73–78 (2020). pest in Korea. Korean J. Appl. Entomol. 59, 73–78 (2020). p pp 8. Sidana, J., Singh, B. & Sharma, O. Occurrence of the new invasive pest, fall armyworm, Spodoptera frugiperda (JE Smith)(Lepi- doptera: Noctuidae), in the maize fields of Karnataka, India. Curr. Sci. 115, 621–623 (2018). pi 9. Ma, J. et al. High risk of the Fall armyworm invading into Japan and the Korean Peninsula via overseas migration. bioRxiv https:// doi.org/10.1101/662387 (2019). g 10. Early, R., González-Moreno, P., Murphy, S. T. & Day, R. Forecasting the global extent of invasion of the cereal pest Spodoptera frugiperda, the fall armyworm (2018). 1. Li, X. et al. Prediction of migratory routes of the invasive fall armyworm in eastern China using a trajectory analytical approach Pest Manag. Sci. 76, 454–463 (2020). g 2. Otuka, A. Migration of rice planthoppers and their vectored re-emerging and novel rice viruses in East Asia. Front. Microbiol. 4 309 (2013). 13. Kisimoto, R. Synoptic weather conditions inducing long-distance immigration of planthoppers, Sogatella furcifera Horvath and Nilaparvata lugens Stal. Ecol. Entomol. 1, 95–109 (1976). p g 4. Jiang, X., Luo, L., Zhang, L., Sappington, T. W. www.nature.com/scientificreports/ Regulation of supercooling and nucleation in a freeze intolerant beetle (Tenebrio molitor). Cryobiol- ( ) y, y, J , g , p p J , ( ) 26. Johnston, S. L. & Lee, R. E. Jr. Regulation of supercooling and nucleation in a freeze intolerant beetle (Tenebrio molitor). Cryobio ogy 27, 562–568 (1990). 6. Johnston, S. L. & Lee, R. E. Jr. Regulation of supercooling and nucleation in a freeze intolerant beetle (Tenebrio molitor). Cryobiol ogy 27, 562–568 (1990).f gy 27. Nedve, D. O., Lavy, D. & Verhoef, H. A. Modelling the time–temperature relationship in cold injury and effect of high-temper interruptions on survival in a chill-sensitive collembolan. Funct. Ecol. 12, 816–824 (1998). y g p p j yf g p interruptions on survival in a chill-sensitive collembolan. Funct. Ecol. 12, 816–824 (1998). . Turnock, W. J. & Fields, P. G. Winter climates and coldhardiness W. J. & Fields, P. G. Winter climates and coldhardiness in terrestria 29. Li, A. & Denlinger, D. L. Rapid cold hardening elicits changes in brain protein profiles of the flesh fly, Sarcophaga crass Insect Mol. Biol. 17, 565–572 (2008). 0. Park, Y. & Kim, Y. RNA interference of glycerol biosynthesis suppresses rapid cold hardening of the beet armyworm, Spodoptera exigua. J. Exp. Biol. 216, 4196–4203 (2013).i 31. Burks, C. S. & Hagstrum, D. W. Rapid cold hardening capacity in five species of coleopteran pests of stored grain. J. Stored Prod. Res. 35, 65–75 (1999). 32. Kim, Y. & Kim, N. Cold hardiness in Spodoptera exigua (Lepidoptera: Noctuidae). Environ. Entomol. 26, 1117–1123 (1997)f 33. Kim, Y. & Song, W. Effect of thermoperiod and photoperiod on cold tolerance of Spodoptera exigua (Lepidoptera: Noctuidae). Environ. Entomol. 29, 868–873 (2000).f 34. Song, W. R. Physiological factors affecting rapid cold hardening of the beet armyworm, Spodoptera exigua (Hubner). Kore Appl. Entomol. 36, 249–255 (1997). pp 35. Danks, H. V., Kukal, O. & Ring, R. A. Insect cold-hardiness: Insights from the Arctic. Arctic 47, 391–404 (1994). 36. Lee, R. E. Principles of insect low temperature tolerance BT—Insects at low temperature. in (eds. Lee, R. E. & Denlinger, D. L.) 17–46 (Springer US, 1991). https://doi.org/10.1007/978-1-4757-0190-6_2. 37. Sømme, L. Supercooling and winter survival in terrestrial arthropods. Comp. Biochem. Physiol. Part A Physiol. 73, 519–543 (1 37. Sømme, L. Supercooling and winter survival in terrestrial arthropods. Comp. Biochem. Physiol. Part A Physiol. 73, 519 543 (1982). 38. Renault, D., Bouchereau, A., Delettre, Y. www.nature.com/scientificreports/ 48 L R E J I C ld h di T f f H i i l Bi i 39 308 313 (1989) 8. Lee, R. E. Jr. Insect Cold-hardiness: To freeze or not to freeze: How insects survive low temperatures. Bioscience 39, 308–313 (1989) 9. Xu, Y. et al. Identification of key genes associated with overwintering in Anoplophora glabripennis larva using gene co-expression network analysis. Pest Manag. Sci. 77, 805–816 (2021).i y g 0. Park, Y. & Kim, Y. A specific glycerol kinase induces rapid cold hardening of the diamondback moth, Plutella xylostella. J. Insec Physiol. 67, 56–63 (2014).it y 1. Choi, B., Park, Y. & Kim, Y. Suppression of gene expression in the fifth instar larvae of Spodoptera exigua at low developmenta threshold temperature. Korean J. Appl. Entomol. 52, 295–304 (2013).f 2. Smith, L. B. Effects of cold-acclimation on supercooling and survival of the rusty grain beetle, Cryptolestes ferrugineus (Stephens (Coleoptera: Cucujidae), at subzero temperatures. Can. J. Zool. 48, 853–858 (1970). Smith, L. B. Effects of cold-acclimation on supercooling and surviv j 53. Waagner, D., Holmstrup, M., Bayley, M. & Sørensen, J. G. Induced cold-tolerance mechanisms depend on duration of acclimation in the chill-sensitive Folsomia candida (Collembola). J. Exp. Biol. 216, 1991–2000 (2013). 53. Waagner, D., Holmstrup, M., Bayley, M. & Sørensen, J. G. Induced cold-tolerance mechanisms depend on duration of acclim in the chill sensiti e Folsomia candida (Collembola) J Exp Biol 216 1991 2000 (2013) 4. Chen, J. et al. Feeding-based RNA interference of a trehalose phosphate synthase gene in the brown planthopper, Nilaparvata lugens. Insect Mol. Biol. 19, 777–786 (2010). g 5. Vatanparast, M. & Kim, Y. Optimization of recombinant bacteria expressing dsRNA to enhance insecticidal activity against a lepidopteran insect, Spodoptera exigua. PLoS ONE 12, e0183054 (2017).f 56. Vatanparast, M., Kazzazi, M., Sajjadian, S. M. & Park, Y. Knockdown of Helicoverpa armigera protease genes affects its growth and mortality via RNA interference. Arch. Insect Biochem. Physiol. n/a, e21840 (2021). 7. Gasmi, L. et al. Horizontally transmitted parasitoid killing factor shapes insect defense to parasitoids. Science (80-) 373, 535–541 (2021). 8. Wang, J., Zhang, R. R., Gao, G. Q., Ma, M. Y. & Chen, H. Cold tolerance and silencing of three cold-tolerance genes of overwinter ing Chinese white pine larvae. Sci. Rep. 6, 34698 (2016). g 59. Agosto, J. A. M. & McCabe, E. R. B. Conserved family of glycerol kinase loci in Drosophila melanogaster. References & Hu, Y. Regulation of migration in Mythimna separata (Walker) in China: A review integrating environmental physiological hormonal genetic and molecular factors Environ Entomol 40 516–533 (2011) integrating environmental, physiological, hormonal, genetic, and molecular factors. Environ. Entomol. 40, 516–533 (2011). 5. Perkins, W. D. Laboratory rearing of the fall armyworm. Florida Entomol. 62, 87–91 (1979). g g p y g g 15. Perkins, W. D. Laboratory rearing of the fall armyworm. Florida Entomol. 62, 87–91 (1979). 15. Perkins, W. D. Laboratory rearing of the fall armyworm. Florida Entomol. 62, 87–91 (1979). 16 Foster R E & Cherry R H Survival of fall Armyworm Spodoptera frugiperda (Lepidoptera: Noctuidae) exposed to cold tem 15. Perkins, W. D. Laboratory rearing of the fall armyworm. Florida Entomol. 62, 87–91 (1979). 16. Foster, R. E. & Cherry, R. H. Survival of fall Armyworm, Spodoptera frugiperda, (Lepidoptera: Noctuidae) exposed to cold tem- 15. Perkins, W. D. Laboratory rearing of the fall armyworm. Florida Entomol. 62, 87 91 (1979). 16. Foster, R. E. & Cherry, R. H. Survival of fall Armyworm, Spodoptera frugiperda, (Lepidoptera: Noctuidae) exposed to cold tem- peratures. Florida Entomol. 70, 419–422 (1987). y g y 6. Foster, R. E. & Cherry, R. H. Survival of fall Armyworm, Spodoptera frugiperda, (Lepidoptera: Noctuidae) exposed to cold tem- peratures. Florida Entomol. 70, 419–422 (1987). https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ 7. Tobin, P. C., Nagarkatti, S. & Saunders, M. C. Phenology of grape berry moth (Lepidoptera: Tortricidae) in cultivated grape a selected geographic locations. Environ. Entomol. 32, 340–346 (2003).h g g p ( ) Gullan, P. J. & Cranston, P. S. The Insects: An Outline of Entomology (Wiley, 2014).f g g p 18. Gullan, P. J. & Cranston, P. S. The Insects: An Outline of Entomology (Wiley, 2014).f h f gy y W. Temperature effects on embryonic development in insects. Ann pf y p ( ) 20. Lee, R. E., Chen, C. & Denlinger, D. L. A rapid cold-hardening process in insects. Science 80(238), 1415–1417 (198 20. Lee, R. E., Chen, C. & Denlinger, D. L. A rapid cold-hardening process in insects. Science 80(238), 1415–1417 (1987). 21 Andreadis S S & Athanassiou C G A review of insect cold hardiness and its potential in stored product insect control Cr 20. Lee, R. E., Chen, C. & Denlinger, D. L. A rapid cold hardening process in insects. Science 80(238), 1415 1417 (1987). 21. Andreadis, S. S. & Athanassiou, C. G. A review of insect cold hardiness and its potential in stored product insect control. Cro 20. Lee, R. E., Chen, C. & Denlinger, D. L. A rapid cold hardening process in insects. Science 80(238), 1415 1417 (1987). 21. Andreadis, S. S. & Athanassiou, C. G. A review of insect cold hardiness and its potential in stored product insect control. Cr Prot 91 93 99 (2017) g p g p 21. Andreadis, S. S. & Athanassiou, C. G. A review of insect cold hardiness and its potential in stored product insect co Prot. 91, 93–99 (2017). 22. Salt, R. W. Principles of insect cold-hardiness. Annu. Rev. Entomol. 6, 55–74 (1961). p 23. Lee, R. E. Principles of insect low temperature tolerance. in Insects at Low Temperature 17–46 (Springer, 1991). 23. Lee, R. E. Principles of insect low temperature tolerance. in Insects at Low Temperature en, K. E. Physiology of cold tolerance in insects. Physiol. Rev. 65, 7 y gy y 25. Storey, K. B. & Storey, J. M. Insect cold hardiness: Metabolic, gene, and protein adaptation. Can. J. Zool. 90, 456–475 (2012) 25. Storey, K. B. & Storey, J. M. Insect cold hardiness: Metabolic, gene, and protein adaptation. Can. J. Zool. 90, 456–475 (2012). 26. Johnston, S. L. & Lee, R. E. Jr. www.nature.com/scientificreports/ R., Hervant, F. & Vernon, P. Changes in free amino acids in Alphitobius diaperinus (Coleoptera: Tenebrionidae) during thermal and food stress. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 143, 279–285 (2006).hf 8. Re au t, ., ouc e eau, ., e ett e, . R., e va t, . & Ve o , . C a ges ee a o ac ds lphitobius diaperinus (Coleoptera: Tenebrionidae) during thermal and food stress. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 143, 279–285 (2006). ( ) 39. Fields, P. G. et al. The effect of cold acclimation and deacclimation on cold tolerance, trehalose and free amino acid levels in Sit- ophilus granarius and Cryptolestes ferrugineus (Coleoptera). J. Insect Physiol. 44, 955–965 (1998). l k l d h ld l l h h l 0. Clark, M. S. & Worland, M. R. How insects survive the cold: molecular mechanisms-a review. J. Comp. Physiol. B 178, 917–933 (2008). (2008). 1. Koštál, V., Zahradníčková, H., Šimek, P. & Zelený, J. Multiple component system of sugars and polyols in the overwintering spruce 1. Koštál, V., Zahradníčková, H., Šimek, P. & Zelený, J. Multiple component system of sugars and polyols in the overwintering spruce bark beetle, Ips typogrphus. J. Insect Physiol. 53, 580–586 (2007). p yp g p y ( ) 42. Storey, K. B., Storey, J. M., Lee, R. E. & Denlinger, D. L. Insects at low temperature. (1991). 43. Cui, F., Wang, H., Zhang, H. & Kang, L. Anoxic stress and rapid cold hardening enhance cold tolerance of the migratory lo Cryobiology 69, 243–248 (2014).i 4. Broufas, G. D. & Koveos, D. S. Rapid cold hardening in the predatory mite Euseius (Amblyseius) finlandicus (Acari: Phytoseiidae) J. Insect Physiol. 47, 699–708 (2001).i y 5. Cha, W. H. & Lee, D. Identification of rapid cold hardening-related genes in the tobacco budworm, Helicoverpa assulta. J. Asia Pac. Entomol. 19, 1061–1066 (2016). 6. Ju, R.-T., Xiao, Y.-Y. & Li, B. Rapid cold hardening increases cold and chilling tolerances more than acclimation in the adults o the sycamore lace bug, Corythucha ciliata (Say) (Hemiptera: Tingidae). J. Insect Physiol. 57, 1577–1582 (2011).li 7. Koveos, D. S. Rapid cold hardening in the olive fruit fly Bactrocera oleae under laboratory and field conditions. Entomol. Exp. Appl 101, 257–263 (2001). p gl y yi p pp 101, 257–263 (2001). www.nature.com/scientificreports/ M., Vatanparast, M. & Kim, Y. Toll/IMD signal pathways mediate cellular immune responses via induction of intracel lular PLA 2 expression. Arch. Insect Biochem. Physiol. 101, e21559 (2019). 73. Sajjadian, S. M., Vatanparast, M. & Kim, Y. Toll/IMD signal pathways mediate cellular immune responses via induction of intracel- lular PLA 2 expression. Arch. Insect Biochem. Physiol. 101, e21559 (2019). y 4. Vatanparast, M., Lee, D. H. & Kim, Y. Biosynthesis and immunity of epoxyeicosatrienoic acids in a lepidopteran insect, Spodoptera exigua. Dev. Comp. Immunol. 107, 103643 (2020). g p 75. [SAS], S. A. S. SAS/STAT User guide, version 9.1. 2. (2004).l g p 75. [SAS], S. A. S. SAS/STAT User guide, version 9.1. 2. (2004). 6. Al Baki, M., Vatanparast, M. & Kim, Y. Male-biased adult production of the striped fruit fly, Zeugodacus scutellata, by feeding dsRNA specific to Transformer-2. Insects 11, 211 (2020). Acknowledgementsh g This research was supported by funds (PQ20180A006 and B-1543086-2021-23) by Research of Animal and Plant Q i A K g This research was supported by funds (PQ20180A006 and B-1543086-2021-23) by Research of Animal and Plant Q ti A K g This research was supported by funds (PQ20180A006 and B-1543086-2021-23) by Research of Animal and Plant Quarantine Agency, Korea. h Quarantine Agency, Korea. Author contributions Y.P. and M.V. conceived the idea, designed the experiments; M.V. and Y.P. performed the experiments; M.V. and Y.P. analyzed the data; M.V. and Y.P. co-wrote the manuscript; M.V. and Y.P. discussed the results and com- mented on the manuscript. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 64. Christiaens, O., Whyard, S., Vélez, A. M. & Smagghe, G. Double-stranded RNA technology to control insect pests: Current status and challenges. Front. Plant Sci. 11, 451 (2020). b ll Th ll f l g 65. Luginbill, P. The Fall Army Worm. (US Department of Agriculture, 1928). g 65. Luginbill, P. The Fall Army Worm. (US Department of Agriculture, 1928). h 6. Johnson, S. J. Migration and the life history strategy of the fall armyworm, Spodoptera frugiperda in the western hemisphere. Int J. Trop. Insect Sci. 8, 543–549 (1987). p 7. Vatanparast, M., Ahmed, S., Sajjadian, S. M. & Kim, Y. A prophylactic role of a secretory PLA 2 of Spodoptera exigua against entomopathogens. Dev. Comp. Immunol. 95, 108–117 (2019). p g p ( ) 8. Vatanparast, M. et al. EpOMEs act as immune suppressors in a lepidopteran insect, Spodoptera exigua. Sci. Rep. 10, 1–19 (2020) h l d fi f d h d f d l b d h p g p 68. Vatanparast, M. et al. EpOMEs act as immune suppressors in a lepidopteran insect, Spodoptera exigua. Sci. Rep. 10, 1–19 (2020). 69. Shu, B. et al. Identification of azadirachtin responsive genes in Spodoptera frugiperda larvae based on RNA-seq. Pestic. Biochem. Physiol. 172, 104745 (2021). 68. Vatanparast, M. et al. EpOMEs act as immune suppressors in a lepidopteran insect, Spodoptera exigua. Sci. Rep. 10, 1–19 (2020). 69. Shu, B. et al. Identification of azadirachtin responsive genes in Spodoptera frugiperda larvae based on RNA-seq. Pestic. Biochem. y ( ) 0. Livak, K. J. & Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative pcr and the 2−ΔΔCT Method Methods 25, 402–408 (2001). ( ) 1. Sinclair, B. J., Alvarado, L. E. C. & Ferguson, L. V. An invitation to measure insect cold tolerance: methods, approaches, and work- flow. J. Therm. Biol. 53, 180–197 (2015). l Jh ( ) 72. Hrithik, M. T. H., Vatanparast, M., Ahmed, S. & Kim, Y. Repat33 acts as a downstream component of eicosanoid signaling path di ti i f S d t i L id t i t I t 12 449 (2021) lh 72. Hrithik, M. T. H., Vatanparast, M., Ahmed, S. & Kim, Y. Repat33 acts as a downstream component of eicosanoid signaling pathway mediating immune responses of Spodoptera exigua, a Lepidopteran insect. Insects 12, 449 (2021). mediating immune responses of Spodoptera exigua, a Lepidopte 3. Sajjadian, S. www.nature.com/scientificreports/ Mol. Genet. Metab. 88, 334–345 (2006). 60. Stanczak, C. M., Chen, Z., Zhang, Y., Nelson, S. F. & McCabe, E. R. B. Deletion mapping in Xp21 for patients with complex glycerol kinase deficiency using SNP mapping arrays. Hum. Mutat. 28, 235–242 (2007). i y g pp g y 61. Wyatt, G. R. Regulation of protein and carbohydrate metabolism in insect fat body. Verh Ungender Dtsch Zool Ges (1975). y g p y y g 62. Shetty, P., Gitau, M. M. & Maróti, G. Salinity stress responses and adaptation mechanisms in eukaryotic green microalgae. Cells 8, 1657 (2019). y g y y g 62. Shetty, P., Gitau, M. M. & Maróti, G. Salinity stress responses and adaptation mechanisms in eukaryotic green microalgae. 8, 1657 (2019). ( ) 3. Kihara, F. et al. Glycerol kinase activity and glycerol kinase-3 gene are up-regulated by acclimation to 5°C in diapause eggs of the silkworm. Bombyx mori. Insect Biochem. Mol. Biol. 39, 763–769 (2009). https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \| © The Author(s) 2022 Competing interests h The authors declare no competing interests. Additional informationh Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-022-08174-4. Correspondence and requests for materials should be addressed to Y.P. Correspondence and requests for materials should be addressed to Y.P. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. https://doi.org/10.1038/s41598-022-08174-4 Scientific Reports \| (2022) 12:4129 \|
https://openalex.org/W2607137135	http://dspace.stir.ac.uk/bitstream/1893/25281/1/Jin%20PLoS%20ONE%202017%20%28black%20seabream%29.pdf	English	null	Dietary DHA/EPA ratio affected tissue fatty acid profiles, antioxidant capacity, hematological characteristics and expression of lipid-related genes but not growth in juvenile black seabream (Acanthopagrus schlegelii)	PloS one	2,017	cc-by	12,269	RESEARCH ARTICLE Abstract An 8-week feeding trial was conducted to investigate the effects of dietary docosahexaenoic to eicosapentaenoic acid ratio (DHA/EPA) on growth performance, fatty acid profiles, anti- oxidant capacity, hematological characteristics and expression of some lipid metabolism related genes of juvenile black seabream (Acanthopagrus schlegelii) of initial weight 9.47 ± 0.03 g. Five isonitrogenous and isolipidic diets (45% crude protein and 14% crude lipid) were formulated to contain graded DHA/EPA ratios of 0.65, 1.16, 1.60, 2.03 and 2.67. There were no differences in growth performance and feed utilization among treatments. Fish fed higher DHA/EPA ratios had higher malondialdehyde (MDA) contents in serum than lower ratios. Serum triacylglycerol (TAG) content was significantly higher in fish fed the lowest DHA/EPA ratio. Tissue fatty acid profiles reflected the diets despite down-regulation of LC- PUFA biosynthesis genes, fatty acyl desaturase 2 (fads2) and elongase of very long-chain fatty acids 5 (elovl5), by high DHA/EPA ratios. Expression of acetyl-CoA carboxylase alpha (accα) and carnitine palmitoyl transferase 1A (cpt1a) were up-regulated by high DHA/EPA ratio, whereas sterol regulatory element-binding protein-1 (srebp-1) and hormone-sensitive lipase (hsl) were down-regulated. Fatty acid synthase (fas), 6-phosphogluconate dehydro- genase (6pgd) and peroxisome proliferator-activated receptor alpha (pparα) showed highest expression in fish fed intermediate (1.16) DHA/EPA ratio. Overall, this study indicated that dietary DHA/EPA ratio affected fatty acid profiles and significantly influenced lipid metabo- lism including LC-PUFA biosynthesis and other anabolic and catabolic pathways, and also had impacts on antioxidant capacity and hematological characteristics. Editor: Silvia Martı´nez-Llorens, Universitat Politècnica de València, SPAIN Editor: Silvia Martı´nez-Llorens, Universitat Politècnica de València, SPAIN Received: March 1, 2017 Accepted: March 21, 2017 Published: April 21, 2017 Copyright: © 2017 Jin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. OPEN ACCESS Citation: Jin M, Monroig O´, Lu Y, Yuan Y, Li Y, Ding L, et al. (2017) Dietary DHA/EPA ratio affected tissue fatty acid profiles, antioxidant capacity, hematological characteristics and expression of lipid-related genes but not growth in juvenile black seabream (Acanthopagrus schlegelii). PLoS ONE 12(4): e0176216. https://doi.org/10.1371/journal. pone.0176216 Dietary DHA/EPA ratio affected tissue fatty acid profiles, antioxidant capacity, hematological characteristics and expression of lipid-related genes but not growth in juvenile black seabream (Acanthopagrus schlegelii) a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Min Jin1,2, O´ scar Monroig3, You Lu1, Ye Yuan1, Yi Li1, Liyun Ding1, Douglas R. Tocher3, Qicun Zhou1,2* 1 Laboratory of Fish Nutrition, School of Marine Sciences, Ningbo University, Ningbo, China, 2 Collaborative Innovation Center for Zhejiang Marine High-efficiency and Healthy Aquaculture, Ningbo University, Ningbo, China, 3 Institute of Aquaculture, School of Natural Sciences, University of Stirling, Stirling, Scotland, United Kingdom * zhouqicun@nbu.edu.cn * zhouqicun@nbu.edu.cn Introduction C. Wong Magna Fund and K. C. Wong Education Foundation at Ningbo University. The C18 polyunsaturated fatty acids (PUFA) linoleic acid (LA, 18:2n-6) and α-linolenic acid (ALA, 18:3n-3) cannot be synthesized de novo in vertebrates and consequently they are regarded as dietary essential fatty acids (EFA) [1,2]. However, marine carnivorous fish have limited ability to convert LA and ALA into the physiologically important long-chain (C20-24) PUFA (LC-PUFA) such as eicosapentaenoic acid (EPA, 20:5n-3), arachidonic acid (ARA, 20:4n-6) and docosahexaenoic acid (DHA, 22:6n-3), and therefore these compounds must be supplied in their diet to ensure normal growth and development [3]. As marine ecosystems are naturally rich in LC-PUFA [4], adaptation to high dietary input of LC-PUFA in marine fish has been postulated as the evolutionary driver accounting for the loss of LC-PUFA biosyn- thetic capability in these species [3]. At a molecular level, specific deficiencies in one or more enzymes, namely fatty acyl desaturases (Fads) with and elongases of very long-chain fatty acids (Elovl), involved in LC-PUFA biosynthesis underpin the abovementioned limitation in the biosynthetic capability of marine fish [1]. Studies have shown that the black seabream Acanthopagrus schlegelii, a commercially important species for intensive culture in China, Japan, Korea and other countries in Southeast Asia [5–11], possesses a Fads2 with Δ6 desatur- ase activity [12], as well as an Elovl5 with high efficiency for elongation of C18 and C20 PUFA [13]. Competing interests: The authors have declared that no competing interests exist. As DHA has important structural roles in biomembranes, especially in neural tissues such as brain and eye, where it is a major component of polar lipids [14,15], it is expected that DHA requirements are high in fast growing stages of development in order to satisfy the demands of rapidly forming tissues that accumulate DHA. While EPA has a major role as a precursor of highly bioactive compounds such as eicosanoids, it can also partly satisfy DHA requirements in species with adequate elongase and desaturase activities to convert EPA to DHA [1]. In addition, previous studies reported that the absolute requirement for n-3 LC-PUFA decreased with increased dietary DHA/EPA ratio [16,17]. Thus, in addition to the absolute dietary levels of DHA and EPA, their relative proportion is also an important aspect for consideration in feed formulations for fast-growing stages of fish [18]. Dietary DHA/EPA ratio impact on black seabream Data Availability Statement: All relevant data are within the paper. Funding: This research was supported by the National Natural Science Foundation of China (Grant Nos. 31272670 and 41476125), Major Spark Plan Project of the National Ministry of Science and Technology (2014GA701001), the Open Fund of Zhejiang Provincial Top Key Discipline of Aquaculture in Ningbo University, K. 1 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Introduction The requirements of dietary DHA/EPA ratios for marine fish has been reported to range from 0.5 to 2.0 according to NRC [3]. p g g Lipid metabolism involves anabolic (biosynthetic) and catabolic processes involving bio- chemical reactions catalyzed by key enzymes and regulated by, among others, transcriptional factors [19]. Acetyl-CoA carboxylase alpha (Accα) is a cytosolic enzyme that controls the pro- duction of malonyl-CoA and thus plays an important role in the biosynthesis of long-chain fatty acids [20–22]. Both 6-phosphogluconate dehydrogenase (6pgd) and glucose 6-phosphate dehydrogenase (G6pd) are key regulatory enzymes involved in NADPH production, essential for fatty acid biosynthesis [19,23]. Fatty acid synthase (Fas) catalyzes de novo fatty acid synthe- sis [24], whereas sterol regulatory element-binding protein-1 (Srebp-1) is a major regulator of fatty acid and lipid biosynthesis [25]. Among catabolic enzymes, lipoprotein lipase (Lpl) hydrolyzes triacylglycerols (TAG) in plasma lipoproteins and provides free fatty acids for either further storage in adipose tissue or oxidation in other tissues. Studies also confirm that Lpl plays a crucial role in regulating the content of body lipids [19,26]. Carnitine palmitoyl- transferase (Cpt1) is regarded as the main regulatory enzyme in fatty acid oxidation catalyzing the conversion of cytosolic fatty acyl-CoA to fatty acyl-carnitine for entry into mitochondria [27,28]. Adipose triglyceride lipase (Atgl) and hormone-sensitive lipase (Hsl) are important enzymes involved in lipogenesis and lypolysis, respectively [29]. Peroxisome proliferator-acti- vated receptor alpha (Pparα) can modulate expression of genes encoding several mitochon- drial fatty acid-catabolizing enzymes in addition to mediating inducible mitochondrial and peroxisomal fatty acid β-oxidation [30]. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 2 / 20 Dietary DHA/EPA ratio impact on black seabream Oxidative stress implies an increase in cellular production of free radicals, often resulting in cell and tissue damage [31,32]. There is concern that dietary PUFA, especially omega-3 LC-PUFA such as EPA and DHA, might increase oxidative stress [33,34] due to PUFA being susceptible to oxidation because oxygen easily attacks double bonds, producing lipoperoxides [35]. To protect cells and tissues from oxidative damage, animals including fish have endoge- nous antioxidant defense systems to help counteract the activity of free radicals. Protective enzymes include superoxide dismutase (SOD), which accelerates the dismutation rate of O2 - to H2O2 as a first line of enzymatic anti-oxidant defense, and glutathione peroxidase (GSH-PX) that reduces all organic lipid peroxides in a reaction that also requires glutathione (GSH) as a hydrogen donor [36–39]. Ethics statement Animal experimentation in the present study was conducted in accordance with the Standard Operation Procedures (SOPs) of the Guide for Use of Experimental Animals of Ningbo Uni- versity, and approved by the Institutional Animal Care and Use Committee of Ningbo Univer- sity, China. Before handling and sacrificing, experimental fish were first anesthetized with tricaine methane sulfonate (MS-222). PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Introduction Lipid oxidation products and their metabolites can be assayed in blood, urine and tissues as markers of endogenous lipid peroxidation/oxidative stress [40]. For instance, malondialdehyde (MDA), derived from the oxidation of fatty acids bearing more than two methylene interrupted double bonds [41], is an important metabolite derived from lipid peroxidation [42]. Overall, the dietary DHA/EPA ratio is of vital importance in fast growing stages of farmed fish and may affect key physiological and biochemical pathways that can ultimately compro- mise growth and normal development. Moreover, as far as we are concerned, no studies have measured lipid anabolism and catabolism genes to explore mechanisms related to the physio- logical effects of dietary DHA/EPA ratio in black seabream. Hence, the present study aimed to determine the effects of dietary DHA/EPA ratio (0.65–2.67) on growth performance, antioxi- dant capacity, fatty acid profiles and expression of some lipid-related genes of juvenile black seabream, A. schlegelii. Dietary DHA/EPA ratio impact on black seabream Table 1. Formulation and composition of the experimental diets (% dry matter). Ingredient (%) Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 Casein 16.23 16.23 16.23 16.23 16.23 Defatted Fishmeala 18.00 18.00 18.00 18.00 18.00 Soybean meal 20.00 20.00 20.00 20.00 20.00 Wheat ﬂour 30.00 30.00 30.00 30.00 30.00 Palmitinb 8.73 8.51 8.37 8.29 8.17 ARA-enriched oilc 1.00 1.00 1.00 1.00 1.00 DHA-enriched oild 0.00 0.58 0.93 1.16 1.45 EPA-enriched oile 1.34 0.98 0.77 0.62 0.45 Soybean lecithin 1.00 1.00 1.00 1.00 1.00 Vitamin premixf 0.50 0.50 0.50 0.50 0.50 Mineral premixg 1.50 1.50 1.50 1.50 1.50 Choline chloride 0.20 0.20 0.20 0.20 0.20 Ca (H2PO4)2 1.50 1.50 1.50 1.50 1.50 Proximate composition (%) Dry matter 92.20 92.60 92.40 92.50 91.00 Crude protein 45.46 45.10 45.50 45.67 45.09 Crude lipid 13.89 14.55 14.46 14.30 14.34 Ash 7.00 6.90 7.00 7.00 7.00 Table 1. Formulation and composition of the experimental diets (% dry matter). a Defatted Fishmeal: 84% crude protein and 2.0% crude lipid d DHA enriched oil: DHA content, 437.7 mg g−1 oil; EPA content, 10.85 mg g−1 oil, as methyl esters; Jiangsu Tiankai Biotechnology Co., Ltd., China. e EPA enriched oil: EPA content, 501.3 mg g−1 oil; DHA content, 254.0 mg g−1 oil, as triglycerides; Hebei Kaiyuankangjian Biological Science and Technology Co., Ltd., China. Vitamin premix based on Zhou et al. [9] g Mineral mixture (g kg−1 premix): FeC6H5O7, 11.43; ZnSO4!7H2O, 11.79; MnSO4!H2O (99%), 2.49; CuSO4!5H2O (99%), 1.06; MgSO4!7H2O (99%), 27.31; KH2PO4, 233.2; NaH2PO4, 228.39; C6H10CaO6!5H2O (98%), 34.09; CoCl2!6H2O (99%), 0.54. KIO3 (99%), 0.06; zeolite, 449.66. g Mineral mixture (g kg−1 premix): FeC6H5O7, 11.43; ZnSO4!7H2O, 11.79; MnSO4!H2O (99%), 2.49; CuSO4!5H2O (99%), 1.06; MgSO4!7H2O (99%), 27.31; KH2PO4, 233.2; NaH2PO4, 228.39; C6H10CaO6!5H2O (98%), 34.09; CoCl2!6H2O (99%), 0.54. KIO3 (99%), 0.06; zeolite, 449.66. https://doi.org/10.1371/journal.pone.0176216.t001 b Palmitin: 97% of total fatty acids as palmitic acid methyl ester; Shanghai Yiji Chemical Co., Ltd., China. 1 Diet preparation Five isonitrogenous (~ 45% crude protein) and isolipidic (~ 14% crude lipid) diets were for- mulated to contain different ratios of DHA/EPA, with the diets named according to their respective DHA/EPA ratios, as “0.65”, “1.16”, “1.60”, “2.03” and “2.70” (Table 1). Casein, defat- ted fishmeal and soybean meal were used as protein sources, whereas palmitin (TAG contain- ing palmitic acid as the sole fatty acid), soybean lecithin, and purified ARA, EPA and DHA were used as the main lipid sources. All ingredients except palmitin, ARA, EPA and DHA (details provided in Table 1) were purchased from Ningbo Tech-Bank Feed Co. Ltd., Ningbo, China. The fatty acid compositions of the diets are shown in Table 2. All dry ingredients were ground into fine powder with particle size < 177 μm, micro components such as minerals and vitamins premix were added followed by lipid and distilled water (35%, w/w). The ground ingredients were mixed in a Hobart type mixer and cold-extruded pellets produced (F-26, Machine factory of South China University of Technology) with pellet strands cut into uni- form sizes (2 mm and 4 mm diameter pellets were prepared) (G-250, Machine factory of South China University of Technology). Pellets were steamed for 30 min at 90˚C, and then air-dried to approximately 10% moisture, sealed in vacuum-packed bags and stored at −20˚C until use in the feeding trial. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 3 / 20 a Defatted Fishmeal: 84% crude protein and 2.0% crude lipid a Defatted Fishmeal: 84% crude protein and 2.0% crude lipid itic acid methyl ester; Shanghai Yiji Chemical Co., Ltd., China. c ARA enriched oil: ARA content 40 mg g−1 oil, as triglycerides; Changsha Kenan Biotechnology Co., Ltd., China. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 97% of total fatty acids as palmitic acid methyl ester; Shanghai Yiji Chemical Co., Ltd., China. Dietary DHA/EPA ratio impact on black seabream Table 2. Fatty acid compositions (% total fatty acids) of the experimental diets with different dietary DHA/EPA ratios. Name Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 14:0 1.17 1.33 1.49 1.56 1.71 16:0 66.45 65.09 64.01 63.86 63.08 18:0 2.45 2.42 2.42 2.49 2.55 ΣSFA 70.07 68.84 67.92 67.91 67.34 16:1n-7 0.46 0.41 0.41 0.41 0.41 18:1n-9 5.83 5.56 5.45 5.31 5.26 20:1n-9 0.69 0.66 0.69 0.66 0.65 22:1n-9 1.37 1.43 1.49 1.49 1.48 24:1n-9 0.11 0.14 0.2 0.15 0.13 ΣMUFA 8.46 8.20 8.24 8.02 7.93 18:2n-6 7.85 8.01 8.08 8.06 8.14 20:4n-6 3.58 3.88 3.88 3.92 3.94 Σn-6PUFA 11.43 11.89 11.96 11.98 12.08 18:3n-3 0.73 0.76 0.78 0.78 0.84 20:5n-3 (EPA) 4.02 3.26 2.82 2.38 2.00 22:6n-3 (DHA) 2.62 3.77 4.51 4.82 5.33 Σn-3PUFA 7.37 7.79 8.11 7.98 8.17 n-3/n-6PUFA 0.64 0.66 0.68 0.67 0.68 n-3LC-PUFA 6.64 7.03 7.33 7.20 7.33 DHA/EPA 0.65 1.16 1.60 2.03 2.67 Some fatty acids, of which the contents are minor, trace amount or not detected, such as 12:0, 20:0, 20:2n−6, 20:3n−6, 22:5n−3, were not listed. DHA/EPA, 22:6n−3/20:5n−3; LC-PUFA, long-chain PUFA (C20-24); MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. atty acid compositions (% total fatty acids) of the experimental diets with different dietary DHA/EPA ratios. Some fatty acids, of which the contents are minor, trace amount or not detected, such as 12:0, 20:0, 20:2n−6, 20:3n−6, 22:5n−3, were not listed. DHA/EPA, 22:6n−3/20:5n−3; LC-PUFA, long-chain PUFA (C20-24); MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. Some fatty acids, of which the contents are minor, trace amount or not detected, such as 12:0, 20:0, 20:2n−6, 20:3 22:6n−3/20:5n−3; LC-PUFA, long-chain PUFA (C20-24); MUFA, monounsaturated fatty acids; PUFA, polyunsaturat −80˚C until further analysis of fatty acid compositions (pools of 3 fish per cage, n = 3), antioxi- dant enzyme activity (pools of 3 fish per cage, n = 3) and liver gene expression (pools of 5 fish per cage, n = 3). Blood samples were taken from the caudal vasculature of 10 fish per cage and using 2 ml syringes. Among the 10 fish samples, eight were collected with non-heparinized syringes for essays on serum, whereas blood samples from 2 fish were taken using heparinized syringes for whole blood biochemical analyses. Proximate composition analysis The crude protein, crude lipid, moisture and ash contents of diets and whole fish were deter- mined according to the methods of the Association of Official Analytical Chemists (AOAC, 2006) [44]. Briefly, moisture content was determined by drying the samples to a constant weight at 105˚C. Crude protein (N × 6.25) was determined via the Dumas combustion method with a protein analyzer (Leco FP528, St. Joseph, USA). Crude lipid was determined by the ether extraction method using the Soxhlet Method (Soxtec System HT6, Tecator, Sweden), and ash content was determined using a muffle furnace at 550˚C for 8 h. Feeding trial Juvenile black seabream (initial weight 9.47 ± 0.03 g) were obtained from a local commercial hatchery at Xiangshan Bay, Ningbo, China. Prior to the experiment, the black seabream juve- niles were acclimated for two weeks (26.0 ± 0.5˚C) and fed on a commercial diet (45% dietary protein, 12% crude lipid, Ningbo Tech-Bank Corp.). A completely randomized trial design was implemented. A total of 600 black seabream juveniles were randomly allocated to 15 float- ing net cages (1.5 m × 1.5 m × 2.0 m) corresponding to triplicate cages of the five dietary treat- ments. Fish were hand-fed to apparent satiation twice daily at 5:00 am and 17:00 pm during eight weeks. During the experimental period, physico-chemical conditions including tempera- ture 26.8–32.6˚C, salinity (21–24 ‰), ammonia nitrogen (< 0.05 mg l−1) and dissolved oxygen (6.4–7.0 mg l−1) were monitored daily (YSI Proplus, YSI, Yellow Springs, Ohio, USA). At the end of the feeding trial, fish were anesthetized with tricaine methane sulfonate (MS- 222) at 100 mg l−1, the dose referred to Topic Popovic et al. [43]. Five fish from each cage (15 per treatment) were pooled (n = 3) and used for analyzing the proximal composition of whole body, where three fish (nine per treatment) were used to determine morphological parameters including condition factor (CF), viscerosomatic index (VSI), hepatosomatic index (HSI) and intraperitoneal fat (IPF) ratio. Muscle and liver samples were also collected and stored at PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 4 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Fatty acid composition The fatty acid composition profiles of diets and fish tissues (liver and muscle) were determined as described by Zuo et al. [45] with minor modifications after tests to ensure that all fatty acids were esterified using the following procedures. Briefly, freeze-dried samples (liver samples ~80 mg and muscle samples ~120 mg) were added to a 12 ml volumetric glass tube with a screw 5 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Dietary DHA/EPA ratio impact on black seabream top containing a teflon gasket. Three ml 1 M potassium hydroxide in methanol was added and the mixture incubated at 72˚C for 20 min in a water bath. After cooling, 3 ml 2 M HCL in methanol was added and the mixture incubated at 72˚C for a further 20 min. Finally, 1 ml hex- ane was added to the mixture, shaken vigorously for 1 min, and then allowed to separate into two layers. Fatty acid methyl esters were separated and measured by GC-MS (Agilent Technol- ogies GC-MS 7890B-5977A, USA) with results presented as percentages of total fatty acids. Oxidation and antioxidant parameter assays The blood was assayed within 24 h from collection after storage at 4˚C, with serum collected by centrifugation at 956 g for 10 min at 4˚C (Eppendorf, Centrifugal 5810 R, Hamburg, Ger- many). All serum samples were used undiluted, with the exception of samples for GSH-PX analysis, which were subject to two-fold dilutions prior use. Liver samples were homogenized in nine volumes (w/v) of ice-cold physiological saline 0.89% (w/v), and then centrifuged as above. The contents of MDA, GSH and protein, as well as enzymatic activities of GSH-PX and T-SOD, were determined in serum and liver homogenates using assay kits (Nanjing Jiancheng Bioengineering Institute, China). The liver homogenate concentrations in each index assay were as following: MDA (10%), GSH (10%), protein (1%, was diluted 10-folds), GSH-PX (10%) and T-SOD (0.125%, was diluted 80-folds). Prior calculations, the content of protein was adjusted according to different concentration of liver homogenate. All the content of MDA, GSH and protein, as well as enzymatic activities of GSH-PX and T-SOD, were calcu- lated according to the manufacturer’s instructions. Haematological determinations Total protein (TP), glucose (GLU), triacylglycerol (TAG) and cholesterol (CHOL) contents were measured in serum samples. Moreover, whole blood samples were used for red blood cell (RBC) and white blood cell (WBC) counts, hemoglobin (HGB) and hematocrit (HCT), with analyses carried out with an automatic blood analyzer (Hitachi 7600–110 Ltd, Japan) at Ningbo University Hospital. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 RNA extraction and reverse-transcriptase quantitative PCR Gene expression was determined by reverse-transcriptase quantitative PCR (qPCR) as follows. Total RNA was extracted from the liver of juvenile black seabream using TRIzol reagent (Takara, Japan) according to the manufacturer’s instructions. Quantity and quality of isolated RNA were determined spectrophotometrically (Nanodrop 2000, Thermo Fisher Scientific, USA) and on a 1.2% denaturing agarose gel, respectively. The cDNA was prepared from 1000 ng of DNAase-treated RNA and synthesized using PrimeScript™RT Reagent Kit with gDNA Eraser (Perfect Real Time) (Takara). The stability of potential references genes including β- actin, gapd and 18S rRNA was tested using Bestkeeper [46]. The results confirmed that β-actin was very stable (stability value was 0.255) and was subsequently use as reference gene to nor- malize the expression levels of the candidate genes. Specific primers for the candidate genes including accα, 6pgd, g6pd, fas, srebp-1, lpl, cpt1a, atgl, hsl, pparα, elovl5 and fads2 used for qPCR were designed using Primer Premier 5.0 (Table 3). The primer specificity assay of the candidate genes was performed as described by Bustin et al. [47]. The primer specificity was checked by systematically running melting curve assays after the qPCR program, running the qPCR products on a 1% (w/v) agarose gel, and DNA sequencing technology (BGI, China). Amplification was performed using a quantitative thermal cycler (Lightcycler 96, Roche, Swit- zerland). The qPCR assays were performed in a total volume of 20 μL, containing 1.0 μL of each primer, 10 μL of 2× conc SYBR Green I Master (Roche), 2 μL of 1/5 diluted cDNA and PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 6 / 20 https://doi.org/10.1371/journal.pone.0176216.t003 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Biometric indices The viscerosomatic (VSI) and hepatosomatic (HSI) indices were significantly higher in fish fed intermediate dietary DHA/EPA ratios (1.60 and 2.03), but fish fed the highest DHA/EPA ratio (2.67) showed no significant differences to fish fed the lowest DHA/EPA ratio (Table 4). In contrast, dietary DHA/EPA ratio had no effect on condition factor (CF) or intraperitoneal fat ratio (IPF) (P > 0.05). Calculations Specific growth ratioÖSGR; % day#1Ü à 100 ⇥ÖÖLn final body weight ÖgÜ # Ln initial body weightÜ ÖgÜÜ=daysÜ: Feed efficiencyÖFEÜ à weight gain Ög; wet weightÜ=feed consumed Ög; dry weightÜ: Survival Ö%Ü à 100 ⇥Öfinal fish number=initial fish numberÜ: Condition factor ÖCF; g=cm3Ü à Body weight ÖgÜ ⇥100=body length ÖcmÜ 3: Viscerosomatic index ÖVSI; %Ü à 100 ⇥Övisceral weight=wet body weightÜ: Condition factor ÖCF; g=cm3Ü à Body weight ÖgÜ ⇥100=body length ÖcmÜ 3: Viscerosomatic index ÖVSI; %Ü à 100 ⇥Övisceral weight=wet body weightÜ: Hepatosomatic indexÖHSI; %Ü à 100 ⇥Öliver weight=wet body weightÜ: Intraperitoneal fat ratio ÖIPF; %Ü à 100 ⇥Öintraperitoneal fat weight=wet body weightÜ: Growth performance and feed utilization The impacts of dietary DHA/EPA ratio on growth performance and feed utilization are pre- sented in Table 4. Specific growth rate (SGR), feed efficiency (FE) and survival did not show any statistical differences among dietary treatments (P > 0.05). Nevertheless, the highest value of SGR was found in fish fed a diet with an intermediate DHA/EPA ratio (1.60), whereas the diet with a DHA/EPA ratio of 2.03 showed the highest FE and survival (Table 4). Statistical analysis Results are presented as means and SEM (number of replicates as indicated). The relative gene expression results (qPCR analyses) were expressed as mean normalized ratios (± SEM) corre- sponding to the ratio between the copy numbers of the target genes and the copy numbers of the reference gene, β-actin. The homogenity of variances (Levene’s test) were checked prior ANOVA tests. Effects of dietary DHA/EPA ratios were analyzed by one-way analysis of vari- ance (ANOVA) followed by Tukey’s HSD test at a significance level of P 0.05 (IBM SPSS Statistics 20). Dietary DHA/EPA ratio impact on black seabream Table 3. Primers for real-time quantitative PCR for lipid related genes and β-actin of black seabream (Acanthopagrus schlegelii). Gene Nucleotide sequence (5’ –3’) Size (bp) GenBank reference or Publication accα 1 F: AGTAGCCTGATTCGTTGGT 154 KX066238 R: GATTGAGGAGTCTGTTCGC 6pgd 2 F: GAAGGGCTTGCTGTTTGTT 101 KX066237 R: GTGGCCAGGCCTCTATATG g6pd 3 F: TCTCGCCAAGAAGAAAATC 297 KX078573 R: GCCAGGTAGAACAAACGGT fas 4 F: AAGAGCAGGGAGTGTTCGC 213 KX066240 R: TGACGTGGTATTCAGCCGA srebp-15 F: TGGGGGTAGGAGTGAGTAG 247 KX066235 R: GTGAAGGGTCAGTGTTGGA lpl6 F: CTGCTACTCCTCTGCCCA 204 KX078571 R: ACATCCCTGTTACCGTCC cpala 7 F: TGCTCCTACACACTATTCCCA 203 KX078572 R: CATCTGCTGCTCTATCTCCCG atgl8 F: GCATCCAGTTCACCCTCAC 241 KX078570 R: TTTGCCTCATCTTCATCGC hsl9 F: AGCAACTAAGCCCTCCCCATC 179 KX066236 R: TCTTCACCCAGTCCGACACAC pparα 10 F: ACGACGCTTTCCTCTTCCC 183 KX066234 R: GCCTCCCCCTGGTTTATTC elovl5 11 F: TCGTACTTCGGTGCCTCCCT 176 KU372149 R: GGCCATATGACTGCAAATATTGTC fads2 12 F: AGCCAGGACCGAAATAAAA 113 KX058437 R: AGGTGGAGGCAGAAGAACA β-actin F: ACCCAGATCATGTTCGAGACC 212 Jiao et al. [50] R: ATGAGGTAGTCTGTGAGGTCG AY491380 1 accα, acetyl-CoA carboxylase alpha; 2 6pgd, 6-phosphogluconate dehydrogenase; 3 g6pd, glucose 6-phosphate dehydrogenase; 4 fas, fatty acid synthase; 5 srebp-1, sterol regulator element-binding protein-1; 6 lpl, lipoprotein lipase; 7 cpt1a, carnitine palmitoyltransferase 1A; 8 atgl, adipose triglyceride lipase; 9 hsl, hormone-sensitive lipase; 10 pparα, peroxisome proliferators-activated receptor alpha; 11 elovl5, elongase of very long-chain fatty acids 5; 12 fads2, fatty acyl desaturase 2. https://doi.org/10.1371/journal.pone.0176216.t003 5 srebp-1, sterol regulator element-binding protein-1; 6 μL DEPC-water. The thermal-cycling conditions used for qPCR were as follows: 95˚C for 2 min, followed by 45 cycles of 95˚C for 10 s, 58˚C for 10 s and 72˚C for 20 s. Standard curves were generated using six different dilutions (in triplicate) of the cDNA samples, and the ampli- fication efficiency was analyzed using the equation E = 10(–1/Slope)-1 [48]. The amplification efficiencies of all genes were approximately equal and ranged from 93 to 102%. In the present study, all the gene expression data were presented as relative gene expression with regards to the expression of the 0.65 DHA/EPA group (reference group). The relative expression levels were calculated using the 2-ΔΔCt method as described by Livak and Schmittgen [49]. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 7 / 20 Dietary DHA/EPA ratio impact on black seabream Dietary DHA/EPA ratio impact on black seabream Table 4. Growth response, feed utilization and biometric indices of juvenile black seabream (Acanthopagrus schlegelii) fed diets containing differ- ent DHA/EPA ratios. Parameter Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM IBW (g) 9.45 0.00 9.47 0.00 9.49 0.04 9.46 0.00 9.46 0.01 0.415 2.57 FBW (g) 36.87 0.26 36.78 0.01 37.36 0.16 36.77 0.24 36.90 0.11 0.194 1.86 SGR (%/d) 2.43 0.01 2.42 0.00 2.45 0.01 2.43 0.01 2.43 0.01 0.168 2.01 FE 0.72 0.01 0.72 0.01 0.71 0.01 0.73 0.01 0.70 0.02 0.294 1.43 Survival (%) 95.83 2.20 95.83 3.00 96.67 0.83 99.17 0.83 94.17 1.67 0.294 1.43 CF (g/cm3) 2.90 0.03 2.91 0.08 2.93 0.01 2.84 0.04 2.94 0.10 0.842 0.35 VSI (%) 5.44a 0.05 5.73ab 0.13 6.11b 0.09 6.07b 0.08 5.42a 0.02 0.000 16.93 HSI (%) 1.34a 0.01 1.34a 0.01 1.41a 0.02 1.56b 0.04 1.33a 0.06 0.002 9.10 IPF (%) 1.18 0.07 1.22 0.10 1.21 0.09 1.31 0.08 1.21 0.06 0.824 0.37 ion and biometric indices of juvenile black seabream (Acanthopagrus schlegelii) fed diets containing differ- Table 4. Growth response, feed utilization and biometric indices of juvenile black seabream (Aca ent DHA/EPA ratios. Data are reported as the mean and SEM (n = 3 for IBW, FBW, SGR, FE and Survival; n = 9 for CF, VSI, HSI and IPF). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). FBW ﬁnal body weight; FE feed efﬁciency; IBG initial body weight; SGR speciﬁc growth rate https://doi.org/10.1371/journal.pone.0176216.t004 dietary DHA/EPA ratio increased from 0.65 to 1.16, and then decreased as the ratio increased further from 1.60 to 2.67. dietary DHA/EPA ratio increased from 0.65 to 1.16, and then decreased as the ratio increased further from 1.60 to 2.67. Liver and muscle fatty acid compositions reflected the dietary DHA/EPA ratios and thus as the dietary ratio increased, the DHA/EPA ratio in the tissues increased (Tables 6 and 7). The increasing dietary DHA/EPA ratio was achieved by a combination of increasing DHA and decreasing EPA (Table 2) and this was also reflected in the tissue compositions (Tables 6 and 7). Whole body composition and tissue fatty acid compositions Dry matter, crude protein and ash contents in the whole body did not show any statistical dif- ferences among the dietary treatments (Table 5). However, crude lipid level increased as the PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 8 / 20 Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly di ndex; IPF, intraperitoneal fat ratio; VSI, viscerosomatic index. ) cy; IBG, initial body weight; SGR, speciﬁc growth rate. Data are reported as the mean and SEM (n = 3 for IBW, FBW, SGR, FE and Survival; n = 9 for CF, VSI, HSI and IPF). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). FBW ﬁ l b d i ht FE f d fﬁi IBG i iti l b d i ht SGR iﬁ th t as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). https://doi.org/10.1371/journal.pone.0176216.t004 https://doi.org/10.1371/journal.pone.0176216.t005 Data are reported as the mean and SEM (n = 3 for IBW, FBW, SGR, FE and Survival; n = 9 for CF, VSI, HSI and IPF). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). FBW, ﬁnal body weight; FE, feed efﬁciency; IBG, initial body weight; SGR, speciﬁc growth rate. Hematological parameters Serum TAG (mmol/L) level was reduced by higher dietary DHA/EPA ratios with the level in fish fed the low ratio (0.65) having significantly higher serum TAG than fish fed the higher ratios (Table 9). In contrast, dietary DHA/EPA ratio had no effect on serum protein, choles- terol or glucose levels. Furthermore, dietary DHA/EPA ratio did not influence WBC (⇤109/L) or RBC (⇤1012/L) counts, or HGB (g/L) level and, although there was a trend for increasing HCT, this was not significant. Fatty acid compositions (% total fatty acids) of liver of juvenile black seabream (Acanthopagrus ratios. Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). Some fatty acids in trace amount such as 12:0, 20:0, 22:0, 24:0, 20:1n−9, 22:1n−11, 16:2n−6, 18:3n−6, 18:4n−3, 20:3n−3 and 20:4n−3 were not listed. DHA/EPA, 22:6n−3/ 20:5n−3; long-chain PUFA (C20-24); MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. Oxidation and antioxidant parameters The activities of the antioxidant enzymes SOD (units/ml and units/mg protein) and GSH-PX (units/ml and units/mg protein) as well as the levels of GSH (mg GSH/L and mg GSH/g pro- tein) in both serum and liver were not impacted by dietary DHA/EPA ratio (Table 8). In con- trast, the level of MDA (nmol/ml and nmol/mg protein) in serum increased as dietary DHA/ EPA increased from 0.65 up to 2.03, with no further increase observed beyond that point (Table 8). The level of MDA in liver showed a similar trend although this was not statistically significant (Table 9). Table 5. Whole body composition of the juvenile black seabream (Acanthopagrus schlegelii) (% wet weight) fed different dietary DHA/EPA ratios. Parameter Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM Dry matter 30.14 0.18 30.54 0.10 30.13 0.21 29.99 0.37 30.37 0.68 0.84 0.35 Protein 18.51 0.07 18.47 0.13 18.32 0.28 18.27 0.12 18.19 0.46 0.89 0.27 Lipid 7.22ab 0.28 7.81b 0.05 7.73b 0.04 7.63ab 0.22 6.82a 0.26 0.03 4.27 Ash 4.89 0.04 4.66 0.04 4.68 0.07 4.70 0.09 4.95 0.13 0.08 2.87 Data are reported as means and SEM (n = 3) Values in the same line with different superscripts are signiﬁcantly different (P < 0 05) Whole body composition of the juvenile black seabream (Acanthopagrus schlegelii) (% wet weight) fed differe venile black seabream (Acanthopagrus schlegelii) (% wet weight) fed different dietary DHA/EPA ratios. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 9 / 20 https://doi.org/10.1371/journal.pone.0176216.t006 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Dietary DHA/EPA ratio impact on black seabream Table 6. Fatty acid compositions (% total fatty acids) of liver of juvenile black seabream (Acanthopagrus schlegelii) fed different dietary DHA/EPA ratios. Name Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM 14:0 2.69 0.11 2.62 0.10 2.81 0.06 2.82 0.11 2.76 0.01 0.52 0.87 16:0 15.56 0.36 15.06 0.23 15.40 0.39 15.65 0.34 15.10 0.20 0.60 0.72 18:0 14.32 0.54 13.95 0.37 14.12 0.08 14.35 0.10 13.34 0.56 0.41 1.10 ∑SFA 32.57 0.97 31.64 0.63 32.33 0.36 32.81 0.17 31.20 0.41 0.32 1.36 16:1n-7 5.30 0.19 4.79 0.21 5.00 0.21 4.98 0.03 4.76 0.12 0.23 1.68 18:1n-9 25.89 1.32 25.56 0.50 24.65 0.66 24.65 0.31 23.83 0.98 0.47 0.96 ∑MUFA 31.19 1.51 30.35 0.48 29.65 0.81 29.63 0.29 28.59 1.10 0.43 1.04 18:2n-6 6.95 0.10 6.91 0.27 6.49 0.01 6.20 0.32 6.62 0.19 0.15 2.18 20:2n-6 4.12 0.12 4.16 0.13 4.03 0.18 3.90 0.28 4.05 0.24 0.90 0.25 20:3n-6 1.48a 0.08 1.60ab 0.06 1.68ab 0.01 1.62ab 0.09 1.77b 0.03 0.07 3.13 20:4n-6 3.83 0.22 4.10 0.16 4.23 0.18 4.17 0.25 4.49 0.25 0.36 1.22 22:5n-6 0.27a 0.11 1.13b 0.11 1.60c 0.05 1.87c 0.07 2.33d 0.11 0.00 71.05 ∑n-6PUFA 16.65a 0.34 17.90ab 0.51 18.04ab 0.21 17.76ab 0.43 19.27b 0.31 0.01 6.24 18:3n-3 0.29 0.03 0.43 0.06 0.31 0.03 0.29 0.02 0.41 0.04 0.06 3.19 20:5n-3(EPA) 2.48b 0.18 2.22b 0.10 1.79a 0.06 1.47a 0.04 1.43a 0.10 0.00 17.83 22:5n-3 2.42 0.68 1.96 0.52 1.88 0.46 1.77 0.26 1.64 0.23 0.80 0.41 22:6n-3(DHA) 4.81a 0.46 5.81ab 0.45 6.19ab 0.05 6.75b 0.49 7.23b 0.46 0.02 4.97 ∑n-3PUFA 9.99 1.31 10.42 0.62 10.17 0.49 10.29 0.63 10.71 0.83 0.98 0.11 n-3/n-6 PUFA 0.60 0.07 0.58 0.05 0.56 0.04 0.58 0.05 0.55 0.03 0.97 0.13 n-3 LC-PUFA 9.70 1.29 9.99 0.68 9.87 0.47 9.87 0.47 10.30 0.79 0.99 0.07 DHA/EPA 1.94a 0.07 2.61b 0.15 3.46c 0.10 4.57d 0.23 5.05d 0.09 0.00 86.28 Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). Some fatty acids in trace amount such as 12:0, 20:0, 22:0, 24:0, 20:1n−9, 22:1n−11, 16:2n−6, 18:3n−6, 18:4n−3, 20:3n−3 and 20:4n−3 were not listed. DHA/EPA, 22:6n−3/ 20:5n−3; long-chain PUFA (C20-24); MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. Table 6. Expression of lipid metabolism genes The expression of genes related to several lipid metabolism pathways including anabolism (panel A), catabolism (panel B) and LC-PUFA biosynthesis (panel C) in the liver of black seab- ream juveniles is shown in Fig 1. Among genes related to lipid biosynthesis pathways, the die- tary DHA/EPA ratio significantly affected the expression levels of accα, 6pgd, fas and srebp-1 (Fig 1A). The liver expression level of accα increased significantly with increase dietary DHA/ EPA ratios of 1.60 or beyond, with an opposite trend, i.e. down-regulation with increasing DHA/EPA ratios, observed for srebp-1. Transcript levels of 6gpd in the liver of juvenile black seabream fed diets with intermediate DHA/EPA ratios (1.16, 1.60 and 2.03) were significantly PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 10 / 20 Dietary DHA/EPA ratio impact on black seabream Table 7. Fatty acid compositions (% total fatty acids) of muscle of juvenile black seabream (Acanthopagrus schlegelii) fed different dietary ratios of DHA/EPA. Name Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM 14:0 1.83a 0.19 2.16ab 0.08 2.31b 0.09 2.33b 0.08 2.31b 0.03 0.04 3.97 16:0 18.52 0.13 19.15 0.91 19.44 0.69 19.04 0.30 19.30 0.26 0.79 0.43 18:0 9.17 0.04 8.95 0.06 8.95 0.25 8.90 0.07 8.83 0.19 0.57 0.77 ∑SFA 29.52 0.26 30.26 0.80 30.71 0.53 30.27 0.30 30.44 0.24 0.52 0.87 16:1n-7 3.50 0.13 3.57 0.16 3.55 0.10 3.72 0.11 3.30 0.06 0.22 1.73 18:1n-9 19.51 0.35 20.12 0.39 20.05 0.45 19.54 0.51 18.49 0.31 0.11 2.53 ∑MUFA 23.00 0.43 23.69 0.23 23.61 0.56 23.26 0.41 21.79 0.26 0.04 3.74 18:2n-6 9.38 0.24 9.63 0.40 9.67 0.14 9.58 0.14 9.51 0.04 0.91 0.24 20:2n-6 1.89 0.13 2.01 0.07 1.84 0.05 1.94 0.08 1.76 0.08 0.35 1.25 20:3n-6 1.45 0.10 1.38 0.01 1.39 0.07 1.52 0.02 1.34 0.08 0.40 1.11 20:4n-6 5.83 0.28 5.43 0.20 5.38 0.33 5.50 0.23 5.86 0.14 0.52 0.87 22:5n-6 0.95a 0.05 2.26b 0.13 2.89c 0.09 3.52d 0.16 4.09e 0.05 0.00 128.42 ∑n-6PUFA 19.50a 0.24 20.70ab 0.48 21.17ab 0.57 22.05b 0.51 22.56b 0.25 0.00 7.68 18:3n-3 0.78 0.06 0.91 0.12 0.82 0.03 0.81 0.02 0.86 0.01 0.62 0.69 20:5n-3 (EPA) 5.21c 0.13 3.70b 0.34 3.27b 0.17 3.52b 0.16 2.50a 0.06 0.00 25.73 22:5n-3 2.56b 0.13 1.87a 0.03 1.73a 0.17 1.81a 0.16 1.39a 0.09 0.00 11.42 22:6n-3 (DHA) 9.76a 0.21 9.46a 0.38 10.46a 0.41 10.81a 0.37 12.46b 0.19 0.00 13.22 ∑n-3PUFA 18.31b 0.24 15.94a 0.35 16.28a 0.72 16.95ab 0.39 17.21ab 0.26 0.02 4.56 n-3/n-6 PUFA 0.94b 0.02 0.77a 0.02 0.77a 0.02 0.77a 0.01 0.76a 0.01 0.00 22.39 n-3 LC-PUFA 17.52b 0.18 15.03a 0.26 15.47a 0.70 16.14ab 0.37 16.35ab 0.26 0.01 5.75 DHA/EPA 1.88a 0.09 2.62b 0.36 3.21b 0.12 3.07b 0.04 4.99c 0.17 0.00 36.65 Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). Some fatty acids in trace amount such as12:0, 20:0, 22:0, 24:0, 20:1n−9, 22:1n−11, 16:2n−6, 18:3n−6, 18:4n−3, 20:3n−3 and 20:4n−3 were not listed. DHA/EPA, 22:6n−3/ 20:5n−3; long-chain PUFA (C20-24); MUFA, mono-unsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. thione; GSH-PX, glutathione peroxidase; MDA, malondialdehyde; SOD, superoxide dismutase. Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). https://doi.org/10.1371/journal.pone.0176216.t008 eroxidase; MDA, malondialdehyde; SOD, superoxide dismutase. Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly di GSH, glutathione; GSH-PX, glutathione peroxidase; MDA, malondialdehyde; SOD, superoxide dismutase. GSH, glutathione; GSH-PX, glutathione peroxidase; MDA, malondialdehyde; SOD, superoxide dismutase. h //d i /10 13 1/j l 01 6216 008 Dietary DHA/EPA ratio impact on black seabream (include the serum and the whole blood) of juvenile black seabream (Acanthopagrus schlegelii) fed differen Table 9. Hematological parameters (include the serum and the whole blood) of juvenile black seabream (Acanthopagrus schlegelii) fed different dietary DHA/EPA ratios. Parameter Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM Serum Total protein (g/L) 39.43 0.85 38.53 2.13 37.03 1.71 37.80 2.00 36.33 1.43 0.72 0.52 Cholesterol (mmol/L) 5.26 0.22 4.80 0.18 5.21 0.08 5.18 0.03 5.25 0.11 0.21 1.80 TAG (mmol/L) 3.67b 0.13 3.09a 0.01 2.78a 0.18 2.90a 0.05 2.91a 0.11 0.00 10.14 Glucose (mmol/L) 7.00 0.48 6.82 0.18 6.57 1.01 7.59 0.55 6.50 0.30 0.69 0.57 Whole blood WBC (109/L) 105.77 4.76 107.33 6.93 105.57 3.48 104.75 1.41 104.73 6.38 1.00 0.05 RBC (1012/L) 3.67 0.16 3.88 0.26 3.72 0.21 3.67 0.03 3.46 0.11 0.56 0.78 HGB (g/L) 103.67 4.48 101.50 1.44 109.50 4.63 103.67 3.84 102.50 2.02 0.94 0.19 HCT 0.47 0.03 0.46 0.02 0.53 0.03 0.50 0.03 0.56 0.03 0.09 2.73 Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P <0.05). HCT, hematocrit; HGB, hemoglobin; RBC, red blood cell; TAG, triacylglycerol; WBC, white blood cell. Table 9. Hematological parameters (include the serum and the whole blood) of juvenile black seabream (Ac dietary DHA/EPA ratios. Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P <0.05). HCT, hematocrit; HGB, hemoglobin; RBC, red blood cell; TAG, triacylglycerol; WBC, white blood cell. Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcant HCT, hematocrit; HGB, hemoglobin; RBC, red blood cell; TAG, triacylglycerol; WBC, white blood cell. higher than those of fish fed diets with lower (0.65) and higher (2.67) DHA/EPA ratios. The highest expression level of fas was observed in fish fed the diet with a dietary DHA/EPA ratio of 1.16. With regards to genes related to lipid catabolism, the expression levels of cpt1a, atgl, hsl and pparα were all significantly affected by dietary DHA/EPA ratio (Fig 1B). The expression of cpt1a significantly increased with increased dietary DHA/EPA ratios of 1.16 and over. The lowest expression levels of atgl and hsl were found in fish fed the diets containing DHA/EPA ratios of 1.16 and 1.60, respectively, with high expression levels obtained in fish fed the lowest dietary DHA/EPA ratio (0.65). In contrast, pparα expression increased as dietary DHA/EPA ratio increased from 0.65 to 1.16 and then decreased with further increases in the ratio. Dietary DHA/EPA ratio had a clear impact on the genes encoding the enzymes of LC-PUFA biosynthesis, with both elovl5 and fads2 showing graded decreased expression with increasing dietary DHA/EPA ratio (Fig 1C). Data are reported as means and SEM (n = 3). Values in the same line with different superscripts are signiﬁcantly different (P < 0.05). Some fatty acids in trace amount such as12:0, 20:0, 22:0, 24:0, 20:1n−9, 22:1n−11, 16:2n−6, 18:3n−6, 18:4n−3, 20:3n−3 and 20:4n−3 were not listed. DHA/EPA, 22:6n−3/ 20:5n−3; long-chain PUFA (C20-24); MUFA, mono-unsaturated fatty acids; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. https://doi.org/10.1371/journal.pone.0176216.t007 Table 8. Serum and liver oxidation and antioxidant parameters of juvenile black seabream (Acanthopagrus schlegelii) fed different dietary DHA/ EPA ratios. Parameter Dietary DHA/EPA ratio 0.65 1.16 1.60 2.03 2.67 ANOVA P F-value Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM Serum GSH-PX (units/ml) 23.61 1.21 25.91 0.80 23.75 0.20 23.31 1.69 22.93 1.27 0.63 0.68 SOD (units/ml) 77.48 4.14 81.51 4.15 91.96 8.00 85.37 6.48 83.69 3.89 0.49 0.92 GSH (mg GSH/L) 39.57 2.89 38.57 1.62 41.90 1.59 36.33 4.47 37.65 6.01 0.86 0.32 MDA (nmol/ml) 10.38a 1.36 10.38a 0.89 10.69a 0.98 15.50b 0.49 15.07b 0.28 0.00 8.90 Liver GSH-PX (units/mg protein) 181.43 13.81 203.68 16.25 232.30 12.94 218.65 17.09 207.02 7.96 0.20 1.83 SOD (units/mg protein) 77.92 3.84 70.37 4.55 82.25 7.57 76.41 4.42 75.38 3.53 0.59 0.74 GSH (mg GSH/g protein) 9.54 0.71 9.37 0.67 9.02 0.31 9.65 0.66 10.03 0.09 0.76 0.46 MDA (nmol/mg protein) 1.85 0.16 1.95 0.11 1.98 0.10 1.98 0.07 2.51 0.25 0.08 2.88 Table 8. Serum and liver oxidation and antioxidant parameters of juvenile black seabream (Acanthopag EPA ratios. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 11 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 https://doi.org/10.1371/journal.pone.0176216.t009 Discussion Several studies have demonstrated a higher biological value (e.g. enhancing growth perfor- mance and immunity) for DHA than for EPA during first feeding of marine fish species such as red seabream (Pagrus major), gilthead seabream (Sparus aurata L.) and turbot (Scophthal- mus maximus) [51–55]. These results suggested that n-3 LC-PUFA requirements might not only be a function of the total amount of these fatty acids in the diet, but also of the relative proportions of essential LC-PUFA like DHA and EPA [56]. However, the results of the present study indicated that dietary DHA/EPA ratios within the tested dietary range (0.65–2.67) had no significant impact on growth performance or feed utilization of juvenile black seabream. These results are somewhat contradictory with those of a recent study on juveniles of large yel- low croaker (Larmichthys crocea), where it was reported that high dietary ratios of DHA/EPA (2.17–3.04) significantly improved growth performance [45]. While the reasons for such an apparent discrepancy cannot be established, it cannot be ruled out that dietary DHA/EPA ratio above 2.6 could have also resulted in growth enhancing effect in black seabream juveniles. More clearly though, the present study revealed that dietary DHA/EPA ratio has impacts on PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 12 / 20 Dietary DHA/EPA ratio impact on black seabream Fig 1. Effects of dietary DHA/EPA ratio on relative mRNA expression of genes involved in lipid metabolism pathways including anabolism (A), catabolism (B) and LC-PUFA biosynthesis (C) in the liver of juvenile black seabream (Acanthopagrus schlegelii). The control group (0.65 DHA/EPA) was used as the reference group, and the mRNA expression levels of target genes were normalized relative to the expression of β-actin. Values are means (n = 3), with standard errors represented by vertical bars. Mean values for the same gene with different letters were significantly different (P < 0.05). accα, acetyl-CoA carboxylase alfa; 6pgd, 6-phosphogluconate dehydrogenase; g6pd, glucose 6-phosphate dehydrogenase; fas, fatty acid synthase, srebp-1, sterol regulatory element-binding protein-1; lpl, lipoprotein lipase; cpt1a, carnitine palmitoyltransferase 1A; atgl, adipose triglyceride lipase; hsl, hormone-sensitive lipase; pparα, peroxisome proliferator-activated receptor alpha; fads2, fatty acyl desaturase 2 and elovl5, elongase of very long-chain fatty acids 5. NS, not significant. Fig 1. Effects of dietary DHA/EPA ratio on relative mRNA expression of genes involved in lipid metabolism pathways including anabolism (A), catabolism (B) and LC-PUFA biosynthesis (C) in the liver of juvenile black seabream (Acanthopagrus schlegelii). Discussion The accα and fas genes play impor- tant roles in fatty acid biosynthesis [21] and, in the present study, the expression of accα gene, which encodes the enzyme responsible for the production of malonyl-CoA, a key early step in the biosynthesis of fatty acids [19,20], increased with increased dietary DHA/EPA ratio. How- ever, subsequent steps in fatty acid synthesis are catalyzed by the fas gene product [19,24] and, conversely, fas gene expression was reduced in fish fed the higher DHA/EPA ratios. Therefore, although there are conflicting data, the gene expression results generally confirmed the above results for HSI, VSI and whole body lipid content, and showed that lipogenesis and lipolysis might be directly related to the expression levels of atgl, hsl, fas and 6pgd in black seabream. The fatty acid compositions of liver and muscle of black seabream showed similar results, largely reflecting the fatty acid compositions of the diets. For example, the level of DHA and the DHA/EPA ratio were significantly increased with increased dietary DHA/EPA ratio in both liver and muscle. On the contrary, EPA levels of liver and muscle decreased significantly with increased dietary DHA/EPA ratio, in agreement with a previous study in large yellow croaker using similar dietary formulations [57]. Many studies have reported that Fads2 and Elovl5 are two key enzymes in the LC-PUFA biosynthesis pathways [1,68–69]. In the present study, fish fed the lowest dietary DHA/EPA ratio (0.65) showed significantly higher expression of both fads2 and elovl5. Consistently, the fads2 expression of large yellow croaker also increased with decreasing dietary DHA/EPA [69] and moreover, these results are in agreement with the effect that DHA had on down-regulating fads2 and elovl5 in rainbow trout (Oncor- hynchus mykiss) [70]. However, the increased expression of fads2 and elovl5 observed in the present study for low DHA/EPA dietary treatments did not resulted in increased enzymatic activity to compensate for lower dietary input. This is in agreement with a study on gilthead seabream in which tissue fatty acid profiles did not reflect up-regulation of fads2 in fish fed vegetable oil [71]. Overall, this clearly indicates that diet is a major factor determining tissue fatty acid profiles in comparison to LC-PUFA endogenous production (biosynthesis). Lipid peroxidation is caused by free radicals leading to oxidative destruction of PUFA con- stitutive of cellular membranes [72]. Discussion The control group (0.65 DHA/EPA) was used as the reference group, and the mRNA expression levels of target genes were normalized relative to the expression of β-actin. Values are means (n = 3), with standard errors represented by vertical bars. Mean values for the same gene with different letters were significantly different (P < 0.05). accα, acetyl-CoA carboxylase alfa; 6pgd, 6-phosphogluconate dehydrogenase; g6pd, glucose 6-phosphate dehydrogenase; fas, fatty acid synthase, srebp-1, sterol regulatory element-binding protein-1; lpl, lipoprotein lipase; cpt1a, carnitine palmitoyltransferase 1A; atgl, adipose triglyceride lipase; hsl, hormone-sensitive lipase; pparα, peroxisome proliferator-activated receptor alpha; fads2, fatty acyl desaturase 2 and elovl5, elongase of very long-chain fatty acids 5. NS, not significant. https://doi.org/10.1371/journal.pone.0176216.g001 body composition (also including fatty acid composition), oxidation and antioxidant parame- ters, hematological characteristics and lipid metabolism gene expression of genes involved in lipid and fatty acid metabolism was also studied. Significant differences were found in VSI and HSI, with highest values in fish fed interme- diate DHA/EPA ratios (1.60–2.03) and lower values found in fish fed both the lowest and 13 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Dietary DHA/EPA ratio impact on black seabream highest DHA/EPA ratios. Since high HSI is associated with high energy reserves and metabolic activity [57], these results indicated that fish fed diet with intermediate DHA/EPA ratios were in a good nutritional status. A similar trend was found in the lipid content of whole body with fish fed intermediate DHA/EPA ratios having the highest levels of lipid. Body fat deposition may be influenced by the levels of endogenous lipid synthesis or catabolism or a combination of both mechanisms. The atgl, hsl, fas and 6pgd genes encode important enzymes involved in mechanisms of lypolysis and lipogenesis [29]. The gene expression or activity of lipogenic enzymes, including Fas (fas), Accα (accα) and 6pgd (6pgd) have been described previously [58–62] and, likewise, studies have indicated that gene expression of lipolytic enzymes such as Atgl (atgl) and Hsl (hsl) were regulated by dietary modifications [29,62–67]. In the present study, the relative expression of atgl and hsl showed similar trends, with higher expression lev- els in fish fed the lowest DHA/EPA ratio (0.65) and lower expression in fish fed intermediate ratio (1.16). In contrast, 6pgd showed the opposite pattern. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Discussion Evidence of lipid peroxidation in the form of increased MDA production, a surrogate marker of oxidative stress, has been noted [73]. In the present study, the serum MDA level was increased with increased dietary DHA/EPA ratio, and liver MDA level showed a similar trend. These results indicated that high dietary DHA/EPA ratio might induce higher oxidative stress in juvenile black seabream. Previous studies demon- strated that excess PUFA in liver may lead to increased lipid peroxidation [22,74–75]. The defenses against free radical-mediated injury include enzymatic deactivation and direct reac- tion with free radicals, such as SOD and GPX [73,76]. SOD is the first line of defense against oxygen derived free radicals and GPX catalyzes reductive destruction of hydrogen and lipid hydroperoxides, using GSH as an electron donor [36–39]. Furthermore, GSH is the most PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 14 / 20 Dietary DHA/EPA ratio impact on black seabream abundant non-enzymatic antioxidant present in the cell, plays an important role in the defense against oxidative-stress-induced cell injury [72]. However, in our study, the activities of SOD and GPX, as well as the content of GSH both in the liver and serum, showed no significantly different among groups. In this study, we demonstrated that higher DHA/EPA ratios could cause higher oxidative stress level, but could not enhance the antioxidant defense capability, at least through the activation of SOD and GSH-PX measured in this study. β-oxidation is regulated mainly by the transcription factor ppara, and it regulates the expression of genes that participate in fatty acid oxidation, for instance, the cpt1a [77–78]. Cpt1a is regarded as the main regulatory enzyme in fatty acid β-oxidation [19,27]. Previous studies demonstrated that mitochondrial β-oxidation can be inhibited by reducing the activity of CptI [75, 79–80], and lipid accumulation mainly occurred because the excess lipids that were consumed could not be oxidized [75]. In this study, although, the expression level of pparα increased significantly to the maximum levels with dietary ratio of DHA/EPA increased from 0.65 to 1.16 and then decreased significantly with further increased of dietary DHA/EPA ratio. However, the highest expression level of cpt1a and the lowest lipid content were all recorded in fish fed the highest DHA/EPA ratio. Therefore, our study indicated that fatty acid oxidation is increased by up-regulating cpt1a, therefore resulting in reduced lipid content as previously reported elsewhere [75,79–80]. Discussion We speculate that dietary DHA/EPA ratio could affect the gene relative expression of cpt1a, and then influenced the fatty acid oxidation in black seabream. Although there are relatively few data on the effects of dietary DHA/EPA ratio on hemato- logical characteristics of marine fish, it has been demonstrated that they can be affected by diet and thus be good indicators of nutrition, stress, and the overall health of fish [81–82]. Zhou et al. [10] stated that RBC are both mechanical and biochemical barriers against infections, bacteria, and blood parasites and immune reactions are regulated to ensure harmony between the RBC and WBC populations. In addition, HGB in aquatic animals operates over wide and independent variations in oxygen at the sites of loading and unloading and shows adaptations both to environmental conditions and metabolic requirements, which govern oxygen avail- ability and transport to tissues [10,83]. In the present study, however, the whole blood indices (WBC, RBC, HGB and HCT) were not affected by dietary DHA/EPA ratio suggesting that this dietary parameter did not affect fish health condition. We herein observed that srebp-1 expression decreased with increased dietary DHA/EPA ratio, whereas expression of lpl was not affected by the ratio. Srebp-1 is a transcription factor regulating fatty acid and lipid biosynthesis pathways [19,84] and lpl hydrolyzes TAG in plasma lipoproteins providing fatty acids for storage in adipose tissue [19,26]. Serum TAG was highest in fish fed the diet with the lowest DHA/EPA ratio (0.65) whereas serum cholesterol was unaf- fected by the ratio. Yan et al [79] suggested that higher gene expression of cpt1 might reduce the content of TAG, similar result was obtained in this current study. Few data are available on the effects of dietary DHA/EPA ratio on serum indices and, therefore, further studies are required to better understand the regulation of indices related to lipid metabolism in black seabream. There is evidence in both fish and humans that serum cholesterol might be influ- enced by dietary n3-LC-PUFA level [74,85–90], although, in the present study, the overall die- tary n-3-LC-PUFA level was essentially the same in all diets. From the above, we suggest that increased dietary DHA/EPA ratio (with constant dietary n-3 LC-PUFA) may reduce serum TAG, but with no impact on serum cholesterol and further investigations are required. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Visualization: MJ. Visualization: MJ. Writing – original draft: MJ. Writing – original draft: MJ. Writing – review & editing: OM DRT. Discussion In conclusion, the present study showed that although the dietary ratio of DHA/EPA did not affect growth performance or feed utilization, it did impact tissue fatty acid profiles, anti- oxidant capacity, hematological characteristics and expression of lipid related genes in juvenile black seabream. The study is the first to measure lipid anabolism and catabolism genes to PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 15 / 20 Dietary DHA/EPA ratio impact on black seabream explore mechanisms related to the physiological effects of dietary DHA/EPA ratio in black seabream. Project administration: MJ QCZ. Resources: MJ Y. Lu YY Y. Li LYD. Software: MJ. Supervision: QCZ. Validation: MJ. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 References Isolation and functional characterization of polyunsaturated fatty acid elongase (AsELOVL5) gene from black seabream (Acanthopagrus schlege- lii). Biotechnol Lett. 2012; 34: 261–268. https://doi.org/10.1007/s10529-011-0746-x PMID: 22198347 14. Seelig A, Seelig J. Dynamic structure of fatty acyl chains in a phospholipid bilayer measured by deute- rium magnetic resonance. Biochemistry. 1974; 13: 4839–4845. PMID: 4371820 15. Wassall SR, Stillwell W. Docosahexaenoic acid domains: the ultimate non-raft membrane domain. Chem Phys Lipids. 2008; 153: 57–63. https://doi.org/10.1016/j.chemphyslip.2008.02.010 PMID: 18343224 16. Rodrı´guez C, Pe´rez JA, Izquierdo MS, Mora J, Lorenzo A, Fernandez-Palacios H. Essential fatty acid requirements of larval gilthead seabream, Sparus aurata (L.). Aquacult Res. 1994; 25: 295–304. 17. Rodrı´guez C, Pe´rez JA, Badı´a P, Izquierdo MS, Ferna´ndez-Palacios H, Herna´ndez AL. The n-3 highly unsaturated fatty acids requirements of gilthead seabream (Sparus aurata L.) larvae when using an appropriate DHA/EPA ratio in the diet. Aquaculture, 1998; 169: 9–23. 18. Rodriguez C, Cejas JR, Martin MV, Badia P, Samper M, Lorenzo A. Influence of n-3 highly unsaturated fatty acid deficiency on the lipid composition of broodstock gilthead seabream (Sparus aurata L.) and on egg quality. Fish Physiol Biochem. 1998; 18: 177–187. 19. Zheng JL, Luo Z, Zhu QL, Tan XY, Chen QL, Sun LD, et al. Molecular cloning and expression pattern of 11 genes involved in lipid metabolism in yellow catfish Pelteobagrus fulvidraco. Gene. 2013; 531: 53– 63. https://doi.org/10.1016/j.gene.2013.08.028 PMID: 23988502 20. Takai T, Saito Y, Yamamoto K, Tanabe T. Developmental changes of the content of acetyl-CoA carbox- ylase mRNA in chicken liver. Arch Biochem Biophys. 1988; 266: 313–318. PMID: 2903720 21. Wakil SJ. Mechanism of fatty acid synthesis. J Lipid Res. 1961; 2: 1–24. 22. Qian Y, Li XF, Zhang DD, Cai DS, Tian HY, Liu WB. Effects of dietary pantothenic acid on growth, intes- tinal function, anti-oxidative status and fatty acids synthesis of juvenile blunt snout Bream Megalobrama amblycephala. PloS one. 2015; 10: e0119518. https://doi.org/10.1371/journal.pone.0119518 PMID: 25781913 23. Chen QL, Luo Z, Liu X, Song YF, Liu CX, Zheng JL, et al. Effects of waterborne chronic copper expo- sure on hepatic lipid metabolism and metal-element composition in Synechogobius hasta. Arch Environ Contam Toxicol. 2013; 64: 301–315. https://doi.org/10.1007/s00244-012-9835-7 PMID: 23229194 24. Cowey CB, Walton MJ. Intermediary metabolism. Fish Nutrition, 1989; 2: 259–329. 25. Minghetti M, Leaver MJ, Tocher DR. Transcriptional control mechanisms of genes of lipid and fatty acid metabolism in the Atlantic salmon (Salmo salar L.) established cell line, SHK-1. BBA-Mol Cell Bioll. References 1. Castro LFC, Tocher DR, Monroig O´ . Long-chain polyunsaturated fatty acid biosynthesis in chordates: insights into the evolution of Fads and Elovl gene repertoire. Prog Lipid Res. 2016; 62: 25–40. https:// doi.org/10.1016/j.plipres.2016.01.001 PMID: 26769304 2. Tocher DR. Metabolism and functions of lipids and fatty acids in teleost fish. Rev Fish Sci. 2003; 11, 107–184. 3. National Research Council (NRC) Nutrient Requirements of Fish and Shrimp. National Academies Press, Washington, DC; 2011. pp. 102–125. 4. Sargent JR. Origin and functions of eggs lipids: nutritional implications. In: Bromage N.R., Roberts R.J., Eds., Broodstock Management and Egg and Larval Quality. Blackwell, London, 1995. pp. 353–372. 5. Nip TH, Ho WY, Wong CK. Feeding ecology of larval and juvenile black seabream (Acanthopagrus schlegeli) and Japanese seaperch (Lateolabrax japonicus) in Tolo Harbour, Hong Kong. Environ Biol Fish. 2003; 66: 197–209. 6. Gonzalez EB, Umino T, Nagasawa K. Stock enhancement programme for black seabream, Acanthopa- grus schlegelii (Bleeker), in Hiroshima Bay, Japan: a review. Aquac Res. 2008; 39: 1307–1315. 7. Ma JJ, Xu ZR, Shao QJ, Xu J Z, Hung SS, Hu WL, et al. Effect of dietary supplemental l-carnitine on growth performance, body composition and antioxidant status in juvenile black seabream, Sparus macrocephalus. Aquaul Nutr. 2008; 14: 464–471. 8. Shao QJ, Ma JJ, Xu Z, Hu WL, Xu JZ, Xie SQ. Dietary phosphorus requirement of juvenile black seab- ream, Sparus macrocephalus. Aquaculture. 2008; 277: 92–100. 9. Zhou F, Xiao JX, Hua Y, Ngandzali BO, Shao QJ. Dietary l-methionine requirement of juvenile black seabream (Sparus macrocephalus) at a constant dietary cystine level. Aquaul Nutr. 2011; 17: 469– 481. 10. Zhou F, Shao J, Xu R, Ma J, Xu Z. Quantitative l-lysine requirement of juvenile black seabream (Sparus macrocephalus). Aquaul Nutr., 2010a; 16: 194–204. 11. Zhou F, Xiong W, Xiao JX, Shao QJ, Bergo ON, Hua Y, et al. Optimum arginine requirement of juvenile black seabream, Sparus macrocephalus. Aquacult Res. 2010b; 41: e418–e430. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 16 / 20 Dietary DHA/EPA ratio impact on black seabream 12. Kim SH, Kim JB, Kim SY, Roh KH, Kim HU, Li KR, et al. Functional characterization of a delta 6-desatur- ase gene from the black seabream (Acanthopagrus schlegeli). Biotechnol Lett. 2011; 33: 1185–1193. https://doi.org/10.1007/s10529-011-0555-2 PMID: 21318631 13. Kim S, Kim JB, Jang YS, Kim SY, Roh KH, Kim HU, et al. References Rev Fish Biol Fisher. 2005; 15: 75–88. 40. Gue´raud F, Tache´ S, Steghens JP, Milkovic L, Borovic-Sunjic S, Zarkovic N, et al. Dietary polyunsatu- rated fatty acids and heme iron induce oxidative stress biomarkers and a cancer promoting environment in the colon of rats. FREE Radical Bio Me. 2015; 83: 192–200. 41. Esterbauer H, Schaur RJR, Zollner H. Chemistry and biochemistry of 4-hydroxynonenal, malonalde- hyde and related aldehydes. Free Radical Bio Me. 1990; 11: 81–128. 42. Zuo RT, Ai QH, Mai KS, Xu W. Effects of conjugated linoleic acid on growth, non-specific immunity, anti- oxidant capacity, lipid deposition and related gene expression in juvenile large yellow croaker (Lar- michthys crocea) fed soyabean oil-based diets. Br J Nutr. 2013; 110: 1220–1232. https://doi.org/10. 1017/S0007114513000378 PMID: 23452520 43. Topic Popovic N, Strunjak-Perovic I, Coz-Rakovac R, Barisic J, Jadan M, Persin Berakovic A, et al. Tri- caine methane-sulfonate (MS-222) application in fish anaesthesia. J Appl Ichthyol. 2012; 28: 553–564. 44. AOAC, Official Methods of Analysis, 18th ed. Association of Official Analytical Chemists. Arlington, VA. 2006 45. Zuo RT, Ai QH, Mai KH, Xu W, Wang J, Xu HG, et al. Effects of dietary docosahexaenoic to eicosapen- taenoic acid ratio (DHA/EPA) on growth, nonspecific immunity, expression of some immune related genes and disease resistance of large yellow croaker (Larmichthys crocea) following natural infestation of parasites (Cryptocaryon irritans). Aquaculture. 2012; 334: 101–109. 46. Pfaffl MW, Tichopad A, Prgomet C, Neuvians TP. Determination of stable housekeeping genes, differ- entially regulated target genes and sample integrity: BestKeeper–Excel-based tool using pair-wise cor- relations. Biotechnol lett. 2004; 26: 509–515. PMID: 15127793 47. Bustin SA, Beaulieu JF, Hugget J, Jaggi R, Kibenge FS, Olsvik P A, Olsvik PA., Penning LC, Toegel S. MIQE precis: Practical implementation of minimum standard guidelines for fluorescence-based quanti- tative real-time PCR experiments. BMC Mol Biol. 2010; 11: 74. https://doi.org/10.1186/1471-2199-11- 74 PMID: 20858237 48. Jothikumar N, Cromeans TL, Robertson BH, Meng XJ, Hill VR. A broadly reactive one-step real-time RT-PCR assay for rapid and sensitive detection of hepatitis E virus. J Virol Methods. 2006; 131: 65–71. https://doi.org/10.1016/j.jviromet.2005.07.004 PMID: 16125257 49. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. methods, 2001; 25: 402–408. https://doi.org/10.1006/meth.2001.1262 PMID: 11846609 50. Jiao B, Huang X, Chan CB, Zhang L, Wang D, Cheng CH. References 2011; 1811: 194–202. 26. Nilsson-Ehle P, Garfinke AS, Schotz MC. Lipolytic enzymes and plasma lipoprotein metabolism. Annu Rev Biochem. 1980; 49: 667–693. https://doi.org/10.1146/annurev.bi.49.070180.003315 PMID: 6996570 27. Kerner J, Hoppel C. Fatty acid import into mitochondria. BBA-Mol Cell Bioll. 2000; 1486: 1–17. 28. Yang JH. Perfluorooctanoic acid induces peroxisomal fatty acid oxidation and cytokine expression in the liver of male Japanese medaka (Oryzias latipes). Chemosphere. 2010; 81: 548–552. https://doi. org/10.1016/j.chemosphere.2010.06.028 PMID: 20594573 29. Ma JJ, Shao QJ, Xu Z, Zhou F. Effect of Dietary n-3 highly unsaturated fatty acids on Growth, Body Composition and Fatty Acid Profiles of Juvenile Black Seabream, Acanthopagrus schlegeli (Bleeker). J World Aquacult Soc. 2013; 44: 311–325. 30. Aoyama T, Peters JM, Iritani N, Nakajima T, Furihata K, Hashimoto T, et al. Altered constitutive expres- sion of fatty acid-metabolizing enzymes in mice lacking the peroxisome proliferator-activated receptor α (PPARα). J Biol Chem. 1998; 273: 5678–5684. PMID: 9488698 31. Kohen R, Nyska A, Oxidation of biological systems: oxidative stress phenomena, antioxidants, redox reactions, and methods for their quantification. 2002; 30: 620–50. 32. Gao J, Koshio S, Ishikawa M, Yokoyama S, Mamauag REP. Interactive effects of vitamin C and E sup- plementation on growth performance, fatty acid composition and reduction of oxidative stress in juvenile Japanese flounder Paralichthys olivaceus fed dietary oxidized fish oil. Aquaculture. 2014; 422–423: 84–90. 33. McNamara RK, Strawn JR. Role of long-chain omega-3 fatty acids in psychiatric practice. Pharma Nutri- tion. 2013; 1: 41–49. 34. Visioli F. Pharma and nutrition: crossing the Rubicon. Pharma Nutrition. 2013; 1: 9. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 17 / 20 Dietary DHA/EPA ratio impact on black seabream 35. Visioli F, Galli C. Evaluating oxidation processes in relation to cardiovascular disease: a current review of oxidant/antioxidant methodology. Nutr Metab Cardiov. 1997; 7: 459–466. 36. Locatelli F, Canaud B, Eckardt KU, Stenvinkel P, Wanner C, Zoccali C. Oxidative stress in end-stage renal disease: an emerging threat to patient outcome. Nephrol Dial Trans. 2003; 18: 1272–1280. 37. Canaud B, Cristol JP, Morena M, Leray-Moragues H, Bosc JY, Vaussenat F. Imbalance of oxidants and antioxidants in haemodialysis patients. Blood Purif 1999; 17: 99–106.PMID: 10449867 38. Bhor VM, Raghuram N, Sivakami S. Oxidative damage and altered antioxidant enzyme activities in the small intestine of streptozotocin-induced diabetic rats. Int J Biochem Cell B. 2004; 36: 89–97. 39. Martinez-Alvarez RM, Morales AE, Sanz A. Antioxidant defenses in fish: biotic and abiotic factors. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 References The co-existence of two growth hormone receptors in teleost fish and their differential signal transduction, tissue distribution and hormonal regu- lation of expression in seabream. J Mol Endocrinol. 2006; 36: 23–40. https://doi.org/10.1677/jme.1. 01945 PMID: 16461924 51. Watanabe T, Izquierdo MS, Takeuchi T, Satoh S, Kitajima C. Comparison between eicosapentaenoic acid and docosahexaenoic acid in terms of essential fatty acid efficacy in larval red seabream. Nippon Suisan Gakk. 1989; 55: 1635–1640. 52. Watanabe T. portance of docosahexaenoic acid in marine larval fish. J World Aquacult Soc. 1993; 24: 152–161. 53. Koven WM, Tandler A, Sklan D, Kissil GW. The association of eicosapentaenoic and docosahexaenoic acids in the phospholipids of different age Sparus aurata larvae with growth. Aquaculture. 1993; 116: 71–82. 54. Reitan KI, Rainuzzo JR, Olsen Y. Influence of lipid composition of live feed on growth, survival and pig- mentation of turbot larvae. Aquacult Int. 1994; 2: 33–48. 55. Rodrı´guez C, Pe´rez JA, Dı´az M, Izquierdo MS, Ferna´ndez-Palacios H, Lorenzo A. Influence of the EPA/DHA ratio in rotifers on gilthead seabream Sparus aurata. larval development. Aquaculture. 1997; 150: 77–89. 18 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Dietary DHA/EPA ratio impact on black seabream 56. Rodrı´guez C, Pe´rez JA, Badı´a P, Izquierdo MS, Ferna´ndez-Palacios H, Herna´ndez AL. The n-3 highly unsaturated fatty acids requirements of gilthead seabream (Sparus aurata L.) larvae when using an appropriate DHA/EPA ratio in the diet. Aquaculture. 1998; 169: 9–23. 57. Collier TK, Chiang WLM, Au WTD, Rainbow PS. Biomarkers currently used in environmental monitor- ing. Ecological biomarkers: indicators of ecotoxicological effects. 2012; 385–410. 58. Menoyo D, Izquierdo MS, Robaina L, Gine´s R, Lopez-Bote CJ, Bautista JM. Adaptation of lipid metabo- lism, tissue composition and flesh quality in gilthead sea bream (Sparus aurata) to the replacement of dietary fish oil by linseed and soya bean oils. Br J Nutr. 2004; 92: 41–52. https://doi.org/10.1079/ BJN20041165 PMID: 15230986 59. Panserat S, Kolditz C, Richard N, Plagnes-Juan E, Piumi F, Esquerre´ D, et al. Hepatic gene expression profiles in juvenile rainbow trout (Oncorhynchus mykiss) fed fishmeal or fish oil-free diets. Br J Nutr. 2008; 100: 953–967. https://doi.org/10.1017/S0007114508981411 PMID: 18439330 60. Morais S, Pratoomyot J, Taggart JB, Bron JE, Guy DR, Bell JG, et al. Genotype-specificresponses in Atlantic salmon (Salmo salar) subject todietary fish oil replacement by vegetable oil: a liver transcrip- tomic analysis. BMC. Genomics. 2011a; 12: 1. 61. References Morais S, Silva T, Cordeiro O, Rodrigues P, Guy DR, Bron JE, et al. Effects of genotype and dietary fish oil replacement with vegetable oil on the intestinal transcriptome and proteome of Atlantic salmon (Salmo salar). BMC Genomics. 2012; 13: 1. 62. Peng M, Xu W, Mai KS, Zhou HH, Zhang YJ, Liufu ZG, et al. Growth performance, lipid deposition and hepatic lipid metabolism related gene expression in juvenile turbot (Scophthalmus maximus L.) fed diets with various fish oil substitution levels by soybean oil. Aquaculture. 2014; 433: 442–449. 63. Zheng JL, Luo Z, Zhuo MQ, Pan YX, Song YF, Hu W, et al. Dietary L-carnitine supplementation increases lipid deposition in the liver and muscle of yellow catfish (Pelteobagrus fulvidraco) through changes in lipid metabolism. Br J Nutr. 2014; 112: 698–708. https://doi.org/10.1017/ S0007114514001378 PMID: 24933091 64. Liang XF, Ogata HY, Oku H. Effect of dietary fatty acids on lipoprotein lipase gene expression in the liver and visceral adipose tissue of fed and starved red sea bream Pagrus major. Comp Biochem Phy- siol A. 2002; 132: 913–919. 65. Stubhaug I, Froyland L, Torstensen BE. Beta-oxidation capacity of red and white muscle and liver in Atlantic salmon (Salmo salar L.)–effects of increasing dietary rapeseed oil and olive oil to replace cape- lin oil. Lipids, 2005; 40: 39–47. PMID: 15825829 66. Jordal AEO, Torstensen BE, Tsoi S, Tocher DR, Lall SP, Douglas SE.Dietaryrapeseed oil affects the expression of genes involved in hepatic lipid metabolism in Atlantic salmon (Salmo salar L.). J Nutr. 2005; 135: 2355–2361. PMID: 16177195 67. Turchini GM, Mentasti T, Frøyland L, Orban E, Caprino F, Moretti VM, et al. Effects of alternative dietary lipid sources on performance, tissue chemical composition, mitochondrial fatty acid oxidation capabili- ties and sensory characteristics in brown trout (Salmo trutta L.). Aquaculture. 2003; 225: 251–267. 68. Monroig O´ , Navarro JC, Tocher DR. Long-chain poly unsaturated fatty acids in fish: recent advances on desaturases and elongases involved in their biosynthesis. Ava. Nutr. Acuı´cola. 2011; 11: 257– 283.2011, 69. Zuo RT, Mai KS, Xu W, Dong X, Ai QH. Molecular cloning, tissue distribution and nutritional regulation of a Δ6-fatty acyl desaturase-like enzyme in large yellow croaker (Larimichthys crocea). Aquacult Res. 2014. 70. Gregory MK, Collins RO, Tocher DR, James MJ, Turchini GM. Nutritional regulation of long-chain PUFA biosynthetic genes in rainbow trout (Oncorhynchus mykiss). Br J Nutr, 2016; 115: 1721–1729. https://doi.org/10.1017/S0007114516000830 PMID: 26987422 71. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 References Castro C, Corraze G, Firmino-Dio´genes A, Larroquet L, Panserat S, Oliva-Teles A. Regulation of glu- cose and lipid metabolism by dietary carbohydrate levels and lipid sources in gilthead sea bream juve- niles. Br J Nutr. 2016; 116: 19–34. https://doi.org/10.1017/S000711451600163X PMID: 27160810 72. Levent G, Ali A, Ahmet A, Polat EC, Aytac¸ C¸ , Ayşe E, et al. Oxidative stress and antioxidant defense in patients with chronic hepatitis C patients before and after pegylated interferon alfa-2b plus ribavirin ther- apy. J Transl Med. 2006; 4: 1. 73. Koruk M, Taysi S, Savas MC, Yilmaz O, Akcay F, Karakok M. Oxidative stress and enzymatic antioxi- dant status in patients with nonalcoholic steatohepatitis. Ann Clin Lab Sci. 2004; 34: 57–62. PMID: 15038668 74. Lu KL, Xu WN, Li XF, Liu WB, Wang LN, Zhang CN. Hepatic triacylglycerol secretion, lipid transport and tissue lipid uptake in blunt snout bream (Megalobrama amblycephala) fed high-fat diet. Aquaculture. 2013; 408: 160–168. 19 / 20 PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 Dietary DHA/EPA ratio impact on black seabream 75. Lu KL, Xu WN, Wang LN, Zhang DD, Zhang CN, Liu WB. Hepatic β-oxidation and regulation of carnitine palmitoyltransferase (CPT) I in blunt snout bream Megalobrama amblycephala fed a high fat diet. PloS one. 2014; 9: e93135. https://doi.org/10.1371/journal.pone.0093135 PMID: 24676148 76. DiMascio P, Murphy ME, Sies H. Antioxidant defense systems: the role of the carotenoids, tocopherols and thiols. Am J Clin Nutr 1991; 53(Suppl): 194S–200S. 77. Gamboa-Go´mez C, Salgado LM, Gonza´lez-Gallardo A, Ramos-Go´mez M, Loarca-Piña G, Reynoso- Camacho R. Consumption of Ocimum sanctum L. and Citrus paradisi infusions modulates lipid metabo- lism and insulin resistance in obese rats. Food Funct, 2014; 5: 927–935. https://doi.org/10.1039/ c3fo60604j PMID: 24584283 78. van der Leij FR, Bloks VW, Grefhorst A, Hoekstra J, Gerding A, Kooi K, et al. Gene expression profiling in livers of mice after acute inhibition of β-oxidation. Genomics. 2007; 90: 680–689. https://doi.org/10. 1016/j.ygeno.2007.08.004 PMID: 17933490 79. Yan J, Liao K, Wang T, Mai KS, Xu W, Ai QH. Dietary Lipid Levels Influence Lipid Deposition in the Liver of Large Yellow Croaker (Larimichthys crocea) by Regulating Lipoprotein Receptors, Fatty Acid Uptake and Triacylglycerol Synthesis and Catabolism at the Transcriptional Level. PloS one. 2015; 10: e0129937. https://doi.org/10.1371/journal.pone.0129937 PMID: 26114429 80. Wang X, Li Y, Hou C, Gao Y, Wang Y. Physiological and molecular changes in large yellow croaker (Pseudosciaena crocea R.) with high-fat diet-induced fatty liver disease. Aquacult Res. PLOS ONE \| https://doi.org/10.1371/journal.pone.0176216 April 21, 2017 References 2015; 46: 272– 282. 81. Li M, Yu N, Qin JG, Li E, Du Z, Chen L. Effects of ammonia stress, dietary linseed oil and Edwardsiella ictaluri challenge on juvenile darkbarbel catfish Pelteobagrus vachelli. Fish Shellfish Immunol. 2014; 38: 158–165. https://doi.org/10.1016/j.fsi.2014.03.015 PMID: 24657724 82. Zhou QC, Jin M, Elmada ZC, Liang X, Mai K S. Growth, immune response and resistance to Aeromo- nas hydrophila of juvenile yellow catfish, Pelteobagrus fulvidraco, fed diets with different arginine levels. Aquaculture. 2015; 437: 84–91. 83. Weber RE, Wells RM. Hemoglobin structure and function. In Lung biology in health and disease. Mar- cel Dekker, Inc. 1989. 84. Minghetti M, Leaver MJ, Tocher DR. Transcriptional control mechanisms of genes of lipid and fatty acid metabolism in the Atlantic salmon (Salmo salar L.) established cell line, SHK-1. BBA-Mol Cell Bioll. 2011; 1811: 194–202. 85. Richard N, Mourente G, Kaushik S, Corraze G. Replacement of a large portion of fish oil by vegetable oils does not affect lipogenesis, lipid transport and tissue lipid uptake in European seabass (Dicen- trarchus labrax L.). Aquaculture, 2006a; 261: 1077–1087. 86. Richard N, Kaushik S, Larroquet L, Panserat S, Corraze G. Replacing dietary fish oil by vegetable oils has little effect on lipogenesis, lipid transport and tissue lipid uptake in rainbow trout (Oncorhynchus mykiss). Br J Nutr. 2006b; 96: 299–309. 87. Morais S, Pratoomyot J, Torstensen BE, Taggart JB, Guy DR, Bell JG, et al. Diet x genotype interac- tions in hepatic cholesterol and lipoprotein metabolism in Atlantic salmon (Salmo salar) in response to replacement of dietary fish oil with vegetable oil. Br J Nutr. 2011b; 106: 1457–1469. 88. Brufau G, Canela M., Rafecas M. Phytosterols:physiologic and metabolic aspects related to choles- terol-lowering properties. Nutr Res. 2008; 28: 217–225. https://doi.org/10.1016/j.nutres.2008.02.003 PMID: 19083411 89. MacKay DS, Jones PJH. Phyto sterols in human nutrition: type, formulation, delivery, and physiological function. Eur J Lipid Sci Tech. 2011; 113: 1427–1432. 90. Gilman CI, Leusch FD, Breckenridge WC, MacLatchy DL. Effects of a phytosterol mixture on male fish plasma lipoprotein fractions and testis P450scc activity. Gen Comp Endocrinol. 2003; 130: 172–184. PMID: 12568795 20 / 20
https://openalex.org/W4387730608	https://www.nature.com/articles/s42005-023-01407-6.pdf	English	null	General construction scheme for geometrically nontrivial flat band models	Communications physics	2,023	cc-by	9,528	1 Department of Physics, Ajou University, Suwon 16499, Korea. 2 Department of Applied Physics, The University of Tokyo, Tokyo 113-8656, Japan. 3 Research Center for Novel Epitaxial Quantum Architectures, Department of Physics, Seoul National University, Seoul 08826, Korea. ✉email: jwrhim@ajou.ac.kr COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys General construction scheme for geometrically nontrivial ﬂat band models Then it was shown that if the n0-th band is ﬂat, one can always ﬁnd a linear combination of the Bloch wave functions resulting in the CLS of the form: χR ¼ cχ ∑ k2BZ ∑ R0 ∑ Q q¼1 αkvn0;k;qeikðR0RÞ R0; q ; ð3Þ ð3Þ d2 ¼ 1 hψ1jψ2i 2; ð1Þ ð1Þ where cχ is the normalization constant and αk is a mixing coef- ﬁcient of the linear combination40. It is important to note that αkvn0;k;q is a ﬁnite sum of exponential factors eik⋅R so that the range of R0 in (3) with the nonzero coefﬁcient of R0; q is ﬁnite. If αkvn0;k;q ¼ 0 at k = k0 for all kinds of αk satisfying the above properties, we call the band the SFB because vn0;k;q becomes discontinuous at k0 in this case. From (3), one can note that the constants in front of each exponential factor of αkvn0;k;q becomes the amplitude of the CLS. which is positive-valued and ranging from 0 to 137–39. If a ﬂat band has a band-touching point with another parabolic band and the maximum value of the quantum distance, denoted by dmax, between eigenvectors around the touching point, is nonzero, we call it a singular ﬂat band (SFB)40. The SFB hosts non- contractible loop states featuring exotic topological properties in real space41,42. The Landau level structure of the SFB is shown to be anomalously spread into the band gap region32,33, and the maximum quantum distance determines the magnitude of the Landau level spreading. Moreover, if we introduce an interface in the middle of an SFB system by applying different electric potentials, an interface mode always appears, and the maximum quantum distance determines its effective mass43. We construct a ﬂat band Hamiltonian from a CLS arbitrarily designed on a given lattice. This part corresponds to the third and fourth stages of the construction scheme sketched in Fig. 1. By using the correspondence between the CLS and Bloch eigenvector in (3), one can obtain αkvn0;k;q in the form of the ﬁnite sum of exponential factors from the designed CLS. Then, by normalizing αkvn0;k;q, one can have the ﬂat band’s eigenvector vn0;k;q corresponding to the CLS. Our purpose is to ﬁnd a tight- binding Hamiltonian of the form q Diverse unconventional phenomena characterized by quantum distance are expected to occur in the SFB systems. General construction scheme for geometrically nontrivial ﬂat band models Hyeongseop Kim 1, Chang-geun Oh2 & Jun-Won Rhim 1,3✉ Hyeongseop Kim 1, Chang-geun Oh2 & Jun-Won Rhim 1,3✉ A singular ﬂat band (SFB), a distinct class of the ﬂat band, has been shown to exhibit various intriguing material properties characterized by the quantum distance. We present a general construction scheme for a tight-binding model hosting an SFB, where the quantum distance proﬁle can be controlled. We ﬁrst introduce how to build a compact localized state (CLS), endowing the ﬂat band with a band-touching point and a speciﬁc value of the maximum quantum distance. Then, we develop a scheme designing a tight-binding Hamiltonian hosting an SFB starting from the obtained CLS, with the desired hopping range and symmetries. We propose several simple SFB models on the square and kagome lattices. Finally, we establish a bulk-boundary correspondence between the maximum quantum distance and the boundary modes for the open boundary condition, which can be used to detect the quantum distance via the electronic structure of the boundary states. 1 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 W d function of the n-th band with momentum k is given by W hen a band has a macroscopic degeneracy, we call it a ﬂat band1,2. Flat band systems have received great attention because their van Hove singularity is expected to stabilize various many-body states when the Coulomb interaction is introduced. Examples of such correlated states induced by ﬂat bands are unconventional superconductivity3–11, ferromagnetism12–18, Wigner crystal19–21, and fractional Chern insulator22–31. Recently, it was revealed that the ﬂat band could be nontrivial from the perspective of geometric notions, such as the quantum distance, quantum metric, and cross-gap Berry connection32–36. The quantum distance is related to the resem- blance between two quantum states deﬁned by jψn;ki ¼ 1ﬃﬃﬃﬃ N p ∑ R ∑ Q q¼1 eikRvn;k;q R; q ; ð2Þ ð2Þ where N is the number of unit cells in the system, R represents the position vectors of the unit cells, R; q corresponds to the q-th orbital among Q orbitals in a unit cell, and vn,k,q is the q-th component of the eigenvector vn,k of the Q × Q Bloch Hamiltonian53. General construction scheme for geometrically nontrivial ﬂat band models First, we ﬁnd two vectors with complex components to yield the desired dmax. Then, we build a compact localized state (CLS) from the two vectors in the second and third steps. Finally, we obtain an SFB Hamiltonian from the CLS using the general ﬂat band construction scheme. A tight-binding Hamiltonian hosting a SFB with dmax Fig. 1 The SFB construction scheme overview. A scheme for the construction of a tight-binding model hosting a singular ﬂat band (SFB) characterized by the maximum quantum distance (dmax). First, we ﬁnd two vectors with complex components to yield the desired dmax. Then, we build a compact localized state (CLS) from the two vectors in the second and third steps. Finally, we obtain an SFB Hamiltonian from the CLS using the general ﬂat band construction scheme. Fig. 1 The SFB construction scheme overview. A scheme for the Fig. 1 The SFB construction scheme overview. A scheme for the construction of a tight-binding model hosting a singular ﬂat band (SFB) characterized by the maximum quantum distance (dmax). First, we ﬁnd two vectors with complex components to yield the desired dmax. Then, we build a compact localized state (CLS) from the two vectors in the second and third steps. Finally, we obtain an SFB Hamiltonian from the CLS using the general ﬂat band construction scheme. General construction scheme for geometrically nontrivial ﬂat band models However, we lack good tight-binding models hosting the SFB where one can control the quantum distance, although numerous ﬂat band construction methods have been developed44–52. This paper suggests a general construction scheme for the tight-binding Hamiltonians with an SFB and the controllable maximum quantum distance. The construction process’s essential part is designing a compact localized state (CLS), which gives the desired maximum quantum distance. The CLS is a characteristic eigen- state of the ﬂat band, which has ﬁnite amplitudes only inside a ﬁnite region in real space40. The CLS can be transformed into the Bloch eigenstate, and any Hamiltonian having this as one of the eigenstates must host a ﬂat band40. Among inﬁnitely many possible tight-binding Hamiltonians for a given CLS, one can choose several ones by implementing the wanted symmetries and hopping range into the construction scheme. Using the con- struction scheme, we suggest several simple tight-binding models hosting an SFB characterized by the maximum quantum distance on the square and kagome lattices. Using the obtained tight- binding models, we propose a bulk-boundary correspondence of the ﬂat band system from the maximum quantum distance to address the question of how to measure the maximum quantum distance in experiments. The previous work established the bulk- interface correspondence for the interface between two domains with different electric potentials in the same SFB system, where the maximum quantum distance of the bulk determines the interface mode’s effective mass43. We show that the same cor- respondence applies to open boundaries if a boundary mode exists. Hlattice ij ðkÞ ¼ ∑ ΔR tijðΔRÞeikΔR; ð4Þ ð4Þ which satisﬁes Hlattice ij ðkÞ Eflat h i αkvn0;k ¼ 0; ð5Þ ð5Þ where Eﬂat is the ﬂat band’s energy and vn0;k is a column vector with components vn0;k;q. Here, tij (ΔR) represents the hopping parameter between the ithe and jth orbitals in unit cells separated by ΔR ¼ ∑d ν¼1 nνaν, where nν is an integer, d is spatial Find ∝q1u1 + q2u2 CLS : αkvk Regularization Flat band construction scheme in Sec. 2 A tight-binding Hamiltonian hosting a SFB with dmax & u1 u2 d2 max = 1 −\|u1 ⋅u2\|2 1 −(Reu1 ⋅u2)2 satisfying Singular flat band vk Fig. 1 The SFB construction scheme overview. A scheme for the construction of a tight-binding model hosting a singular ﬂat band (SFB) characterized by the maximum quantum distance (dmax). Results We use a bar notation for the complex conjugate such that tn1;n2 ¼ nν ij ¼ tn1;n2 ¼ nν ij and eν ¼ ðeνÞ. Then, the matrix element of the tight-binding Hamiltonian is rewritten as dimension, and aν is the primitive vector. For convenience, we denote tn1;n2 ¼ nν ij tijðΔRÞ and eν eikaν. We use a bar notation for the complex conjugate such that tn1;n2 ¼ nν ij ¼ tn1;n2 ¼ nν ij and eν ¼ ðeνÞ. Then, the matrix element of the tight-binding Hamiltonian is rewritten as HCB1 ¼ 2ð1 þ cos kyÞ a b ð1 þ e1Þð1 þ e2Þ a b ð1 þ e1Þð1 þ e2Þ 2a2 b2 ð1 þ cos kxÞ ! ; ð11Þ ð11Þ where we further assume that a and b are real constants and t0;0 11 ¼ 2 for convenience. This Hamiltonian yields a zero-energy ﬂat band and lower parabolic with a singular band-touching point at k = (π, π) as plotted in Fig. 2b. In fact, this band-crossing is already designed at the construction stage of the CLS in (7) by assigning a simultaneous zero of all the components of αkvn0;k at k = (π, π). HLattice ij ðkÞ ¼ ∑ n1;n2 ¼ nν ∑ ij tn1;n2 ¼ nν ij Y ν0 enν0 ν0 : ð6Þ ð6Þ Here, the hopping parameters tn1;n2 ¼ nν ij can be considered complex unknowns determined by the matrix equation in (5). One can encode some wanted hopping range and symmetries by manipulating the number of unknown hopping parameters and setting relations between them, respectively. Noting that αkvn0;k ¼ ∑n1;n2;;nνcn1;n2;;nν Q ν0enν0 ν0 as described above, the matrix equation (5) leads to a system of linear equations obtained from the coefﬁcients of the independent exponential factors. Here, the hopping parameters tn1;n2 ¼ nν ij can be considered complex unknowns determined by the matrix equation in (5). One can encode some wanted hopping range and symmetries by manipulating the number of unknown hopping parameters and setting relations between them, respectively. Noting that αkvn0;k ¼ ∑n1;n2;;nνcn1;n2;;nν Q ν0enν0 ν0 as described above, the matrix equation (5) leads to a system of linear equations obtained from the coefﬁcients of the independent exponential factors. Maximum quantum distance. Results In this section, we discuss how to endow the band-crossing of the ﬂat band with the wanted value of the maximum quantum distance dmax when we construct a ﬂat band model. Speciﬁcally, the quantum distance between two Bloch eigenstates with momenta k and k0 is denoted as dðk; k0Þ2 ¼ 1 jv k0 vkj2 and dmax is deﬁned as Let us consider a simple example, the ﬂat band Hamiltonian on the checkerboard lattice, which is illustrated in Fig. 2a. We design a CLS in the shape of a square represented by a gray region in Fig. 2a, having amplitudes a and b on the A and B sites, respectively. From the CLS, one can obtain the ﬂat band’s eigenvector αkvn0;k in momentum space such that the CLS’s d2 max ¼ lim rD!0 max dðk; k0Þ2 h i k;k02Dðk0Þ; ð12Þ ð12Þ where vk is the ﬂat band’s eigenvector and D (k0) is a closed disk with radius rD centered at the band-crossing point k032. In the previous study, dmax was proposed to measure the strength of the singularity at k0. The ﬁnite dmax also indicates the divergence of the Fubini-Study metric54 deﬁned by the real part of the quantum geometric tensor Qμν = 〈∂μψ∣∂νψ〉−〈∂μψ∣ψ〉〈ψ∣∂νψ〉. For the generic SFB Hamiltonian characterized by dmax, the quantum geometric tensor is evaluated as amplitude in the unit cell ΔR ¼ ∑d ν¼1 nνaν becomes the coefﬁcient of the exponential factor Q νeν ν. As a result, we have αkvn0;k ¼ a þ ae1 b þ be2 : ð7Þ ð7Þ The next step is to design the tight-binding Hamiltonian (6). We seek one with real-valued hopping parameters up to the next- nearest hopping range. Then, the matrix elements of HCB1 are of the form QSFBðkÞ ¼ d2 max 4m1m2EparaðkÞ2 k2 x kxky kxky k2 y ! Results General ﬂat band construction scheme. Since the key ingredient of the ﬂat band construction scheme is designing a CLS, we begin with a brief review of it. The general form of the Bloch wave COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys 2 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 A site B site Unit cell CLS (a) (b) Flat band 0 1 2 3 4 5 Fig. 2 The checkerboard lattice model CB1. a A red box represents the unit cell. The hopping amplitudes are 1 for the dashed lines along the y axis, −a/b for black solid lines along diagonal directions, and a2/b2 for the black solid lines along the x axis. The compact localized state (CLS) corresponding to the ﬂat band is drawn by a gray region. The CLS’s amplitudes are a at the A-sites and b at the B-sites. b The band structure of the checkerboard model for a = 1 and b = 2. A site B site Unit cell CLS (a) (b ) Flat band 0 1 2 3 4 5 Fig. 2 The checkerboard lattice model CB1. a A red box represents the unit cell. The hopping amplitudes are 1 for the dashed lines along the y axis, −a/b for black solid lines along diagonal directions, and a2/b2 for the black solid lines along the x axis. The compact localized state (CLS) corresponding to the ﬂat band is drawn by a gray region. The CLS’s amplitudes are a at the A-sites and b at the B-sites. b The band structure of the checkerboard model for a = 1 and b = 2. Fig. 2 The checkerboard lattice model CB1. a A red box represents the unit cell. The hopping amplitudes are 1 for the dashed lines along the y axis, −a/b for black solid lines along diagonal directions, and a2/b2 for the black solid lines along the x axis. The compact localized state (CLS) corresponding to the ﬂat band is drawn by a gray region. The CLS’s amplitudes are a at the A-sites and b at the B-sites. b The band structure of the checkerboard model for a = 1 and b = 2. dimension, and aν is the primitive vector. For convenience, we denote tn1;n2 ¼ nν ij tijðΔRÞ and eν eikaν. COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys Results ; ð13Þ ð13Þ where EparaðkÞ ¼ 2t1k2 x þ 2t2kxky þ ð2t3 þ t2 4=t1Þk2 y is the energy HCB1 11 ¼ t0;0 11 þ t0;1 11 e2 þ t0;1 11 e2; ð8Þ HCB1 12 ¼ t0;0 12 þ t1;0 12 e1 þ t0;1 12 e2 þ t1;1 12 e1e2; ð9Þ HCB1 22 ¼ t0;0 22 þ t1;0 22 e1 þ t1;0 22 e1; ð10Þ where EparaðkÞ ¼ 2t1k2 x þ 2t2kxky þ ð2t3 þ t2 4=t1Þk2 y is the energy ð8Þ parað Þ 1 x þ 2 x y þ ð 3 þ 4= 1Þ y gy of the parabolic band touching with the SFB, and m1 and m2 are the minimum and maximum effective mass it [See Supplemen- tary Note 1 for details]. While the real and imaginary parts of the quantum geometric tensor correspond to the quantum metric and the Berry curvature, respectively, the Berry curvature van- ishes for all momenta so that the Fubini-Study metric is identical to the quantum geometric tensor in the SFB. Since Epara (k) ∝k2, where k ¼ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ k2 x þ k2 y q , every element of the Fubini-Study metric parað Þ 1 x 2 x y ð 3 4= 1Þ y gy of the parabolic band touching with the SFB, and m1 and m2 are the minimum and maximum effective mass it [See Supplemen- tary Note 1 for details]. While the real and imaginary parts of the quantum geometric tensor correspond to the quantum metric and the Berry curvature, respectively, the Berry curvature van- ishes for all momenta so that the Fubini-Study metric is identical to the quantum geometric tensor in the SFB. Since Epara (k) ∝k2, where k ¼ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ k2 x þ k2 y q , every element of the Fubini-Study metric ð9Þ HCB1 12 ¼ t0;0 12 þ t1;0 12 e1 þ t0;1 12 e2 þ t1;1 12 e1e2; ð10Þ HCB1 22 ¼ t0;0 22 þ t1;0 22 e1 þ t1;0 22 e1; ð10Þ From the ﬂat band condition (5) and by enforcing the hermicity, one can ﬁnd relationships between the tight-binding parameters, which lead to the following form of the Hamiltonian: 3 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 A site B site CLS (a) A site B site CLS (a) (b) Flat band Fig. 3 The checkerboard lattice model CB2. Results However, the CLS should consist of different complex amplitudes in at least two atomic sites for generic ﬂat bands with 0 < dmax < 1. The tight- binding Hamiltonian stabilizing such a CLS usually requires complex hopping parameters. Moreover, it is shown in Methods that we need more than two exponential factors for at least one component of αkvk. This implies that we usually need hopping processes between atoms at a longer distance than the nearest- neighbor ones. One can design vk of the ﬂat band to have a speciﬁc value of dmax by manipulating the form of the linear expansion of αkvk around the band-crossing point. Denoting qμ = kμ−k0,μ, where k0 is the band-crossing point, the eigenvector αkvk can be written as Let us consider the checkerboard lattice example again. We assume that the touching point is at k = (0, 0). First, to obtain a model with dmax ¼ 1, we can choose u1 = (i, 0)T and u2 = (0, −i)T in (14), using the formula (15). Then, we apply the regularization ik1 ! 1 eik1 and ik2 ! 1 eik2 to obtain the CLS’s Fourier transform. Second, on the other hand, one can let the CLS have dmax ¼ 1= ﬃﬃﬃ 2 p by choosing u1 ¼ ði; 1ÞT= ﬃﬃﬃ 2 p and u2 = (0, −i)T. In this case, an example of the regularization gives αkvk ¼ ð1 eik1; 1 þ i ieik1 eik2Þ T. The CLS corresponding to this eigenvector is drawn in Fig. 3a. An example of the ﬂat band tight-binding Hamiltonian obtained from this choice of the CLS is given by αkvk ’ q1u1 þ q2u2; ð14Þ ð14Þ in the vicinity of k0 up to the linear order of q. Here, u1 and u2 are Q × 1 constant normalized vectors. Then, one can show that d2 max ¼ 1 ju 1 u2j2 1 ðReu 1 u2Þ2 : ð15Þ ð15Þ αkvk ¼ ð1 eik1; 1 þ i ieik1 eik2Þ T. The CLS corresponding to this eigenvector is drawn in Fig. 3a. An example of the ﬂat band tight-binding Hamiltonian obtained from this choice of the CLS is given by See Supplementary Note 2 for the detailed derivations. By using this relationship, one can choose two constant vectors u1 and u2, giving the desired value of dmax. Results Then, performing a regulariza- tion of (14) by applying transformations, such as qi ! sin qi and qi ! 1 eiqi, one can obtain αkvk, the Fourier transform of a CLS, in the form of a ﬁnite sum of exponential factors eν and eν. In this stage, corresponding to the ﬁrst to third steps in Fig. 1, one can control the size of the CLS, which is closely related to the hopping range of the tight-binding model obtained from this CLS. While the regularization process contains a large degree of arbitrariness for the ﬁnal tight-binding model, this arbitrariness can be reduced quite much if we select CLSs with a size as small as possible. Once we obtain αkvk, the tight-binding Hamiltonian with the desired dmax can be built by using the construction scheme in the previous section. HCB2 ¼ v2v 2 v1v 2 v2v 1 v1v 1 ; ð16Þ ð16Þ where v1 ¼ 1 eik1 and v2 ¼ 1 þ i ieik1 eik2. The band structure of this model is shown in Fig. 3b. One can note that the band has nonzero slopes at X and M points due to the broken time-reversal, mirror, and inversion symmetries. As discussed above, the CLS contains both the real and imaginary amplitudes, and the Hamiltonian possesses imaginary hopping processes in the dmax ¼ 1= ﬃﬃﬃ 2 p case. Flat band models characterized by the quantum distance. We ﬁrst construct a simple tight-binding model hosting an SFB characterized by dmax in the kagome lattice as shown in Fig. 4a, b. When we consider only the nearest-neighbor hopping processes in the kagome lattice, which is the most popular case, the ﬂat band already has a quadratic band-touching, but the corre- sponding dmax is ﬁxed to 132. We generalize this conventional kagome lattice model so that dmax can vary by adding some next- nearest-neighbor hopping processes. From the dmax-formula (15), one can note that dmax can be less than one and larger than zero only when u 1 u2 is not real or pure imaginary. Namely, u 1;mu2;m should be imaginary at least for one m, where ui,m is the m-th component of ui. Let us denote such an index m by m0. Results a The checkerboard ﬂat band model with the maximum quantum distance dmax ¼ 1= ﬃﬃﬃ 2 p . A red box represents the unit cell. The hopping parameters are given below the lattice structure. For the complex hopping processes, the hopping direction is represented by the arrow. The gray region stands for the compact localized state (CLS). b The band structure of the checkerboard model CB2. (a) Fig. 3 The checkerboard lattice model CB2. a The checkerboard ﬂat band model with the maximum quantum distance dmax ¼ 1= ﬃﬃﬃ 2 p . A red box represents the unit cell. The hopping parameters are given below the lattice structure. For the complex hopping processes, the hopping direction is represented by the arrow. The gray region stands for the compact localized state (CLS). b The band structure of the checkerboard model CB2. diverges algebraically (~1/k2) approaching the band-crossing point at k = 0. This singularity disappears if dmax becomes zero because the Fubini-Study metric simply vanishes in this case. Note that if there is a singularity at the touching point, the quantum distance between the Bloch eigenstates can remain ﬁnite even if the momenta of them are very close to k032. For the well- known SFB models, such as the kagome and checkerboard lattice models, dmax is found to be unity. While 0 ≤dmax ≤1 in general32, there have been almost no examples of the tight-binding models hosting dmax smaller than 1. diverges algebraically (~1/k2) approaching the band-crossing point at k = 0. This singularity disappears if dmax becomes zero because the Fubini-Study metric simply vanishes in this case. Note that if there is a singularity at the touching point, the quantum distance between the Bloch eigenstates can remain ﬁnite even if the momenta of them are very close to k032. For the well- known SFB models, such as the kagome and checkerboard lattice models, dmax is found to be unity. While 0 ≤dmax ≤1 in general32, there have been almost no examples of the tight-binding models hosting dmax smaller than 1. corresponding to the SFB with 0 < dmax < 1 cannot be constructed only with the real amplitudes. Note that the CLS of the ﬂat band of the kagome lattice can be represented by only real amplitudes because the corresponding dmax is unity. COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys Results Then, the m0-th component of αkvk, given by αkvkjm0 ¼ u1;m0q1 þ u2;m0q2, must be regularized into a form, where the coefﬁcients of the exponential factors contain both the real and imaginary values. This implies that the CLS 4 4 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 Unit Cell CLS (a) ei ei ei ei e-i e-i e-i e-i e-i e-i iei iei (b) (c) (d) (e) (f) dmax = (2α2 + 3)/(6α2 + 3) Lattice model α = 0 dmax = 1 α = 0.766 dmax = 0.8 α = 3.464 dmax = 0.6 ϕA ϕB ϕC ϕA ϕB Fig. 4 The kagome lattice model characterized by the quantum distance. a The nearest and the next-nearest-neighbor hopping processes are denoted by the black solid and green dashed lines, respectively. The compact localized state (CLS) corresponding to the ﬂat band of this model is represented by the gray region. b The phase parts of the hopping parameters are highlighted. The magnetic ﬂuxes for the complex hopping parameters are given by ϕA = π/ 2−θ, ϕB = θ, and ϕC = −π. c–e Band dispersions for α = 0, α = 0.766, and α = 3.464, where θ ¼ cos1ð1= ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p Þ. f The maximum quantum distance dmax as a function of α. The formula (17) drawn by a black curve is compared with the numerically calculated dmax from the lattice model, represented by circles. Fig. 4 The kagome lattice model characterized by the quantum distance. a The nearest and the next-nearest-neighbor hopping processes are denoted by the black solid and green dashed lines, respectively. The compact localized state (CLS) corresponding to the ﬂat band of this model is represented by the gray region. b The phase parts of the hopping parameters are highlighted. The magnetic ﬂuxes for the complex hopping parameters are given by ϕA = π/ 2−θ, ϕB = θ, and ϕC = −π. c–e Band dispersions for α = 0, α = 0.766, and α = 3.464, where θ ¼ cos1ð1= ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p Þ. f The maximum quantum distance dmax as a function of α. The formula (17) drawn by a black curve is compared with the numerically calculated dmax from the lattice model, represented by circles. Results We begin with two vectors u1 = c1 (−i, −2α, −i)T and u2 = c2 (0, −i−α, −i)T, where c1 ¼ ð2 þ 4α2Þ1=2 and c2 ¼ ð2 þ α2Þ1=2. This set of vectors yields We begin with two vectors u1 = c1 (−i, −2α, −i)T and u2 = c2 (0, −i−α, −i)T, where c1 ¼ ð2 þ 4α2Þ1=2 and c2 ¼ ð2 þ α2Þ1=2. This set of vectors yields g4 ¼ tð1 þ e1Þ, t ¼ eiθ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p , and θ ¼ cos1ð1= ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p Þ. Note that when α = 0, where dmax ¼ 1, the model reduces to the kagome lattice model with only nearest-neighbor hopping processes. As the parameter α grows, the nearest-neighbor hopping parameters become complex-valued, and the next- nearest-neighbor hopping processes are developed as represented by green dashed lines in Fig. 4a. One can assign threading magnetic ﬂuxes corresponding to the complex hopping para- meters as illustrated in Fig. 4b, similar to the Haldane model in graphene. In Fig. 4c–e, we plot band dispersions for various values of α, where we have a zero-energy ﬂat band at the bottom. g4 ¼ tð1 þ e1Þ, t ¼ eiθ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p , and θ ¼ cos1ð1= ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ α2 p Þ. Note that when α = 0, where dmax ¼ 1, the model reduces to the kagome lattice model with only nearest-neighbor hopping processes. As the parameter α grows, the nearest-neighbor hopping parameters become complex-valued, and the next- nearest-neighbor hopping processes are developed as represented by green dashed lines in Fig. 4a. One can assign threading magnetic ﬂuxes corresponding to the complex hopping para- meters as illustrated in Fig. 4b, similar to the Haldane model in graphene. In Fig. 4c–e, we plot band dispersions for various values of α, where we have a zero-energy ﬂat band at the bottom. dmax ¼ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 3 þ 2α2 3 þ 6α2 r ; ð17Þ ð17Þ ð17Þ where α can take any real number from −∞to ∞. dmax of the constructed SFB model can take values from 1= ﬃﬃﬃ 3 p to 1. Then, we regularize the linearized vector vfb = u1k1 + u2k2 to where α can take any real number from −∞to ∞. dmax of the constructed SFB model can take values from 1= ﬃﬃﬃ 3 p to 1. Results Then, we regularize the linearized vector vfb = u1k1 + u2k2 to vfb ¼ 1 e1 1 þ iαe1 þ e2 iαe3 e1 e2 0 B @ 1 C A; ð18Þ ð18Þ gy Fig. 4c) is the well-known band diagram of the kagome lattice with the nearest-neighbor hopping processes. If α is nonzero, the Dirac point is gapped out due to the broken C6 symmetry, but the quadratic band-crossing at the Γ point is maintained. We calculate dmax of this model directly using (12) and check that the continuum formula (17) works well as shown in Fig. 4f. where e3 ¼ e1e2. The CLS corresponding to this eigenvector of the ﬂat band is drawn in Fig. 4a by the gray region. From this choice of the CLS, we construct a tight-binding Hamiltonian as follows: We also construct an SFB tight-binding model in the square lattice bilayer, where one can adjust dmax. The lattice structure is illustrated in Fig. 5a, b, and its band dispersion is plotted in Fig. 5c. As in the kagome lattice case, the construction scheme starts by setting two constant vectors. Our choice is u1 = c (iα+γ, −α−iγ)T, and u2 ¼ u1, where c ¼ ð2α2 þ 2γ2Þ1=2. One can show HkagðkÞ ¼ g1 g 2 g 3 g2 2 g 4 g3 g4 g1 0 B @ 1 C A; ð19Þ HkagðkÞ ¼ g1 g 2 g 3 g2 2 g 4 g3 g4 g1 0 B @ 1 C A; ð19Þ ð19Þ where g1 = 2∣t∣2, g2 ¼ tð1 þ e3Þ, g3 ¼ t 1 þ e2 þ iαtðe1 þ e3Þ, where g1 = 2∣t∣2, g2 ¼ tð1 þ e3Þ, g3 ¼ t 1 þ e2 þ iαtðe1 þ e3Þ, COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphy 5 5 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 i(γ2 −α2) −α 4αγ −α2 −γ2 x y z (a) (b) E/(α2 + γ2) (c) (d) Lattice model dmax = 1 −4γ2/(1 + γ2)2 Fig. 5 The square lattice bilayer model characterized by the quantum distance. We plot the interlayer and intralayer hopping processes in a and b, respectively. c We plot the band structure, where the energy is scaled by α2 + γ2. The relation between the maximum quantum distance dmax and the band parameter γ is presented in d. Results The results from the lattice model (red circles) are compared with the analytic formula (20) drawn by the solid curve (b) Fig. 5 The square lattice bilayer model characterized by the quantum distance. We plot the interlayer and intralayer hopping processes in a and b, respectively. c We plot the band structure, where the energy is scaled by α2 + γ2. The relation between the maximum quantum distance dmax and the band parameter γ is presented in d. The results from the lattice model (red circles) are compared with the analytic formula (20) drawn by the solid curve. that dmax calculated from these vectors is given by was developed43. Here, a speciﬁc type of interface is considered, which is generated between two domains of an SFB system with different onsite potentials UR and UL. Note that the two domains are characterized by the same geometric quantity dmax, unlike the topological bulk-boundary correspondence, where the boundary is formed between two regions with different topological char- acters. In the case of the SFB systems, an interface state is guar- anteed to exist if the value of dmax is nonzero, and the corresponding band dispersion around the band-crossing point is given by dmax ¼ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 4α2γ2 ðα2 þ γ2Þ2 s ; ð20Þ ð20Þ where α and γ can take any real values. dmax can vary from 0 to 1. If α or γ is zero As shown in Fig. 5d, dmax of the constructed SFB model can take values from 0 to 1. Then, we regularize a vector vfb = u1k1 + u2k2 to vfb ¼ iγð1 e1e2Þ αðe1 e2Þ iαð1 e1e2Þ þ γðe1 e2Þ : ð21Þ ð21Þ EIFðkÞ d2 max 2mb k2 þ U0; ð23Þ ð23Þ From this choice of the CLS, we construct a tight-binding Hamiltonian as follows: where k and mb are the crystal momentum and the bulk mass along the direction of the interface, respectively, and U0 ¼ minðUR; ULÞ. This formula implies that the effective mass of the interface mode is m ¼ mb=d2 max. HsqðkÞ ¼ jf 2j2 f 3 f 3 jf 1j2 ! ; ð22Þ ð22Þ γ (1−e1e2) −α (e1−e2), f2 = iα (1−e1e2) + γ (e1−e2 where f1 = −iγ (1−e1e2) −α (e1−e2), f2 = iα (1−e1e2) + γ (e1−e2) and f 3 ¼ f 1f 2. COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 h The dmax-dependence of the effective mass of the boundary modes of the square lattice bilayer around ky = 0. The general continuum formula (23) of the effective mass is drawn by red circles. We ﬁrst consider the kagome lattice model. We note that boundary modes exist for the ribbon geometry of this system illustrated in Fig. 6a, which respects the translational symmetry along ð1=2; ﬃﬃﬃ 3 p =2Þ while terminated along the x axis. The width W of the kagome ribbon is deﬁned as the number of the unit cells along the x axis. For example, the width of the kagome ribbon shown in Fig. 6a is 4. We plot the band dispersions of the kagome ribbons with W = 20 for dmax ¼ 1 and dmax ¼ 0:8 in Fig. 6b, c, respectively. The red and blue lines represent the boundary modes stemming from the band-crossing point at ky = 0. While the band dispersions of the left (near ky ≈0)- and right (near ky ≈2π)-localized modes in Fig. 6b are precisely the same for the dmax ¼ 1 case, it is not for 0 < dmax < 1 case due to the broken time-reversal symmetry. For this reason, we distinguish the left- and right-localized modes by the red and blue colors in Fig. 6c. We check that the blue and red curves, although they look asymmetric with respect to ky = 0, they follow the same parabolic equation (23) in the vicinity of the touching point at ky = 0 [See Supplementary Fig. 2 in Supplementary Note 4]. We numerically calculate the effective mass of the boundary modes from the kagome lattice model and compare it with the analytic result of the effective mass m ¼ mb=d2 max in (23). As plotted in Fig. 6d, the formula (23) describes the numerical results perfectly for any values of dmax. Second, we also investigate the edge state of the square lattice bilayer ribbon shown in Fig. 6e. As in the kagome model, the width W of this system is deﬁned as the number of unit cells along the x axis. In Fig. 6f, g, we plot the band structures of the square lattice bilayer ribbon with W = 20. The red curves, which are doubly degenerate, correspond to the boundary modes. COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 x y E 6 3 0 0 2 ky 0 2 ky 2.5 2.0 1.5 1.0 0.6 0.7 0.8 0.9 1.0 dmax (a) (b) (c) (d) dmax=1 dmax=0.8 0 ky 0 ky 1 0 2 3 0 1 dmax E 6 3 0 log10 (e) (f) (g) (h) dmax=1 dmax=0.6 Lattice Continuum Lattice Continuum Fig. 6 The Bulk-boundary correspondence of singular ﬂat band systems. a The lattice structure of the kagome lattice model. The unit cell is indicated by the red box. b, c are the band structures of the kagome lattice model with W = 20 for various values of the maximum quantum distance dmax. d We plot the effective mass of the boundary modes of the kagome lattice model around ky = 0 as a function of dmax and compare it with the continuum result (red circles) in (23). e The lattice structure of the square lattice bilayer model, where the red box represents the unit cell. f, g We plot the band spectra of the square lattice bilayer model with W = 20 for various values of dmax. h The dmax-dependence of the effective mass of the boundary modes of the square lattice bilayer around ky = 0. The general continuum formula (23) of the effective mass is drawn by red circles. x y (a) (e) 0 2 ky 2.5 2.0 1.5 1.0 0.6 0.7 0.8 0.9 1.0 dmax (c) (d) dmax=0.8 Lattice Continuum x y (a) ky 0 ky (g) dmax=0.6 (e) 1 0 2 3 0 1 dmax log10 (h) Lattice Continuum dmax Fig. 6 The Bulk-boundary correspondence of singular ﬂat band systems. a The lattice structure of the kagome lattice model. The unit cell is indicated by the red box. b, c are the band structures of the kagome lattice model with W = 20 for various values of the maximum quantum distance dmax. d We plot the effective mass of the boundary modes of the kagome lattice model around ky = 0 as a function of dmax and compare it with the continuum result (red circles) in (23). e The lattice structure of the square lattice bilayer model, where the red box represents the unit cell. f, g We plot the band spectra of the square lattice bilayer model with W = 20 for various values of dmax. Results When α = γ or α and γ are zero, dmax ¼ 1. As parameters α and γ grow, interlayer and intralayer hopping appear, and if α ≠γ, the complex-value hopping process is developed as represented by blue arrow in Fig. 5a. Unlike kagome lattice model, this model has an isotropic band dispersion. As shown in Fig. 5c, the ﬂat band is ﬁxed at the zero energy, and the parabolic band is scaled with α2 + γ2. Figure 5d shows dmax of this model as a function of γ. One calculated by the continuum formula (15) complies with the numerical results from the lattice model evaluated directly from (12). b= max Now, we examine the formula (23) for the ﬁnite systems satisfying the open boundary condition. In the previous work, (23) could be obtained by presuming an exponentially decaying edge mode, and the existence of such a state was guaranteed for the speciﬁc interface of the step-like potential. While the open boundaries are naturally induced when we prepare a sample, the application of the step-like potential is not usually straightforward in experiments. Therefore, it is worthwhile to investigate the bulk-boundary correspondence for the open boundary systems. In the case of the open boundary, the bulk-boundary correspon- dence states that if edge-localized modes exist, their energy spectrum is given by (23). Note that the edge modes are not guaranteed to appear within the open boundary condition. For example, the modiﬁed Lieb lattice model yields an interface mode when there is a chemical potential difference over the system as studied in reference [43], while we do not have an edge mode under the open boundary condition [See Supplementary Fig. 1 in Supplementary Note 3]. f 2. When α = γ or α and γ are zero, dmax ¼ d l d l h Bulk-boundary correspondence. The bulk-boundary corre- spondence is the essential idea of the topological analysis of materials55–62. Based on this, one can detect the topological information of the bulk by probing the electronic structure of the boundary states. Recently, a new kind of bulk-interface corre- spondence from the quantum distance for the ﬂat band systems COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys 6 ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 We conﬁrm that the effective mass of the boundary modes obeys the continuum formula (23) well as plotted in Fig. 6h. square lattice bilayer, where one can control dmax. We note that complex and long-range (at least the next-nearest ones) hopping amplitudes are necessary to change dmax between 0 and 1. This implies that the candidate materials hosting an SFB with 0 < dmax < 1 could be found among the materials with strong spin- orbit coupling. We believe that our construction scheme could inspire the material search for geometrically nontrivial ﬂat band systems. If we extend the category of the materials to the artiﬁcial systems, our lattice models with the ﬁne-tuned complex hopping parameters are expected to be realized in the synthetic dimensions63–69 and circuit lattices70,71. Then, we propose a bulk-boundary correspondence between the bulk number dmax and the shape of the low-energy dispersion of the boundary modes within the open boundary condition. The information of dmax is embedded in the effective mass of the band dispersion of the edge states. This correspondence provides us with a tool to detect dmax from the spectroscopy of the ﬁnite SFB systems. Notably, the bulk-boundary correspondence is obtained from the continuum Hamiltonian around the band-crossing point. This implies that even if the ﬂat band obtained from our construction scheme is slightly deformed in real systems, one can investigate the geometric properties of the SFB. It is worthwhile to remark that the dmax-driven bulk-boundary correspondence is dis- tinguished from the conventional one in that the former predicts the effective mass of the interface mode instead of its existence. Method d A condition for the CLS to have a noninteger dmax. In this section, we show that at least one component of αkvk, the Fourier transform of the CLS, should contain more than two different exponential factors eiðmq1þnq2Þ. Here, qi is the momentum with respect to the band-crossing point, and m and n are integer numbers. To this end, we verify that if all the components of αkvk have two or less than two exponential factors, dmax of the cor- responding ﬂat band is one or zero. The q-th component of such an eigenvector can be written as COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys References 112, 070403 (2014). 9. Volovik, G. E. Graphite, graphene, and the ﬂat band superconductivity. JETP Lett. 107, 516–517 (2018). 44. Mizoguchi, T. & Udagawa, M. Flat-band engineering in tight-binding models: beyond the nearest-neighbor hopping. Phys. Rev. B 99, 235118 (2019). 10. Aoki, H. Theoretical possibilities for ﬂat band superconductivity. J. Supercond. Nov. Magn. 33, 2341–2346 (2020). 45. Călugăru, D. et al. General construction and topological classiﬁcation of crystalline ﬂat bands. Nat. Phys. 18, 185–189 (2022). 45. Călugăru, D. et al. General construction and topological classiﬁcation of t lli ﬂt b d N t Ph 18 185 189 (2022) crystalline ﬂat bands. Nat. Phys. 18, 185–189 (2022). 11. Kononov, A. et al. Superconductivity in type-ii weyl-semimetal wte2 induced by a normal metal contact. J. Appl. Phys. 129, 113903 (2021). 46. Mizoguchi, T. & Hatsugai, Y. Systematic construction of topological ﬂat-band models by molecular-orbital representation. Phys. Rev. B 101, 235125 (2020). 12. Mielke, A. Ferromagnetism in the hubbard model and hund’s rule. Phys. Lett. A 174, 443–448 (1993). 47. Graf, A. & Piéchon, F. Designing ﬂat-band tight-binding models with tunable multifold band touching points. Phys. Rev. B 104, 195128 (2021). 13. Tasaki, H. From nagaoka’s ferromagnetism to ﬂat-band ferromagnetism and beyond: an introduction to ferromagnetism in the hubbard model. Prog. Theor. Phys. 99, 489–548 (1998). g y 48. Hwang, Y., Jung, J., Rhim, J.-W. & Yang, B.-J. Wave-function geometry of b d i i t i t di i Ph R B 103 L241102 (2021) 48. Hwang, Y., Jung, J., Rhim, J.-W. & Yang, B.-J. Wave-function geometry of band crossing points in two dimensions. Phys. Rev. B 103, L241102 (2021). g p y ( ) 49. Hwang, Y., Rhim, J.-W. & Yang, B.-J. Flat bands with band crossings enforced by symmetry representation. Phys. Rev. B 104, L081104 (2021). y 14. Mielke, A. Stability of ferromagnetism in hubbard models with degenerate single-particle ground states. J. Phys. A 32, 8411 (1999). 50. Hwang, Y., Rhim, J.-W. & Yang, B.-J. General construction of ﬂat bands with and without band crossings based on wave function singularity. Phys. Rev. B 104, 085144 (2021). 15. Hase, I., Yanagisawa, T., Aiura, Y. & Kawashima, K. Possibility of ﬂat-band ferromagnetism in hole-doped pyrochlore oxides Sn2Nb2O7 and S2Ta2O7. Phys. Rev. Lett. 120, 196401 (2018). 51. Maimaiti, W., Flach, S. & Andreanov, A. Universal d= 1 ﬂat band generator from compact localized states. Phys. Rev. B 99, 125129 (2019). 16. References 1. Leykam, D., Andreanov, A. & Flach, S. Artiﬁcial ﬂat band systems: from lattice models to experiments. Adv. Phys.: X 3, 1473052 (2018). 1. Leykam, D., Andreanov, A. & Flach, S. Artiﬁcial ﬂat band systems: from lattice models to experiments. Adv. Phys.: X 3, 1473052 (2018). 37. Bužek, V. & Hillery, M. Quantum copying: beyond the no-cloning theorem. Phys. Rev. A 54, 1844 (1996). 2. Rhim, J.-W. & Yang, B.-J. Singular ﬂat bands. Adv. Phys.: X 6, 1901606 (2021). l k h f d h ddl d h 2. Rhim, J.-W. & Yang, B.-J. Singular ﬂat bands. Adv. Phys.: X 6, 1901606 (2021). 38. Dodonov, V., Man’Ko, O., Man’Ko, V. & Wünsche, A. Hilbert-schmidt distance and non-classicality of states in quantum optics. J. Mod. Opt. 47, 633–654 (2000). 3. Volovik, G. The fermi condensate near the saddle point and in the vortex core. JETP Lett. 59, 830 (1994). 4. Cao, Y. et al. Unconventional superconductivity in magic-angle graphene superlattices. Nature 556, 43–50 (2018). 39. Wilczek, F. & Shapere, A. Geometric phases in physics, vol. 5 (World Scientiﬁc, 1989). p 5. Liu, X. et al. Spectroscopy of a tunable moiré system with a correlated and topological ﬂat band. Nat. Commun. 12, 1–7 (2021). 5. Liu, X. et al. Spectroscopy of a tunable moiré system 40. Rhim, J.-W. & Yang, B.-J. Classiﬁcation of ﬂat bands according to the band- crossing singularity of bloch wave functions. Phys. Rev. B 99, 045107 (2019). topological ﬂat band. Nat. Commun. 12, 1–7 (2021) 6. Balents, L., Dean, C. R., Efetov, D. K. & Young, A. F. Superconductivity and strong correlations in moiré ﬂat bands. Nat. Phys. 16, 725–733 (2020). g g y y 41. Bergman, D. L., Wu, C. & Balents, L. Band touching from real-space topology in frustrated hopping models. Phys. Rev. B 78, 125104 (2008). 7. Peri, V., Song, Z.-D., Bernevig, B. A. & Huber, S. D. Fragile topology and ﬂat- band superconductivity in the strong-coupling regime. Phys. Rev. Lett. 126, 027002 (2021). 42. Ma, J. et al. Direct observation of ﬂatband loop states arising from nontrivial real-space topology. Phys. Rev. Lett. 124, 183901 (2020). 43. Oh, C.-g, Cho, D., Park, S. Y. & Rhim, J.-W. Bulk-interface correspondence from quantum distance in ﬂat band systems. Communications Physics 5, 320 (2022). 8. Yudin, D. et al. Fermi condensation near van hove singularities within the hubbard model on the triangular lattice. Phys. Rev. Lett. Code availability Code availability Code sharing is not applicable. y Code sharing is not applicable. 33. Hwang, Y., Rhim, J.-W. & Yang, B.-J. Geometric characterization of anomalous landau levels of isolated ﬂat bands. Nat. Commun. 12, 1–9 (2021) Received: 9 May 2023; Accepted: 28 September 2023; 34. Peotta, S. & Törmä, P. Superﬂuidity in topologically nontrivial ﬂat bands. Nat. Commun. 6, 8944 (2015). 35. Törmä, P., Peotta, S. & Bernevig, B. A. Superconductivity, superﬂuidity and quantum geometry in twisted multilayer systems. Nat. Rev. Phys. 4, 528–542 (2022). 36. Piéchon, F., Raoux, A., Fuchs, J.-N. & Montambaux, G. Geometric orbital susceptibility: quantum metric without berry curvature. Phys. Rev. B 94, 134423 (2016). Data availability Data availability Data sharing is not applicable. Data sharing is not applicable. 31. Bergholtz, E. J. & Liu, Z. Topological ﬂat band models and fractional chern insulators. Int. J. Mod. Phys. B 27, 1330017 (2013). y 32. Rhim, J.-W., Kim, K. & Yang, B.-J. Quantum distance and anomalous landau levels of ﬂat bands. Nature 584, 59–63 (2020). COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 touching point, the coefﬁcients satisfy 22. Wang, l touching point, the coefﬁcients satisfy 22. Wang, F. & Ran, Y. Nearly ﬂat band with chern number c= 2 on the dice lattice. Phys. Rev. B 84, 241103 (2011). Am1;n1 þ Am2;n2 ¼ 0: ð25Þ Am1;n1 þ Am2;n2 ¼ 0: ð25Þ y 23. Tang, E., Mei, J.-W. & Wen, X.-G. High-temperature fractional quantum hall states. Phys. Rev. Lett. 106, 236802 (2011). As a result, the linear expansion of αk becomes y 24. Sun, K., Gu, Z., Katsura, H. & Sarma, S. D. Nearly ﬂatbands with nontrivial topology. Phys. Rev. Lett. 106, 236803 (2011). αkvkjq iAm1;n1 ðm1 m2Þq1 ðn1 n2Þq1 ; ð26Þ p gy y 25. Neupert, T., Santos, L., Chamon, C. & Mudry, C. Fractional quantum hall states at zero magnetic ﬁeld. Phys. Rev. Lett. 106, 236804 (2011). p gy y 25. Neupert, T., Santos, L., Chamon, C. & Mudry, C. Fractional quantum hall q 1; 1 1 1 leading to u 1;qu2;q ¼ jAm1;n1j2ðm1 m2Þðn1 n2Þ, where ui,q is the q-th component of ui deﬁned in (14). Therefore, u 1 u2 ¼ ∑qu 1;qu2;q is a real number, which proves the statement at the beginning of this section. Namely, we need at least three different exponential factors in at least one component of αk. 25. Neupert, T., Santos, L., Chamon, C. & Mudry, C. Fractional quantum hall states at zero magnetic ﬁeld. Phys. Rev. Lett. 106, 236804 (2011). states at zero magnetic ﬁeld. Phys. Rev. Lett. 106, 236804 (20 26. Sheng, D., Gu, Z.-C., Sun, K. & Sheng, L. Fractional quantum hall effect in the absence of landau levels. Nat. Commun. 2, 1–5 (2011). absence of landau levels. Nat. Commun. 2, 1–5 (2011) 27. Regnault, N. & Bernevig, B. A. Fractional chern insulator. Phys. Rev. X 1, 021014 (2011). 28. Weeks, C. & Franz, M. Flat bands with nontrivial topology in three dimensions. Phys. Rev. B 85, 041104 (2012). dimensions. Phys. Rev. B 85, 041104 (2012). 29. Yang, S., Gu, Z.-C., Sun, K. & Sarma, S. D. Topological ﬂat band models with arbitrary chern numbers. Phys. Rev. B 86, 241112 (2012). 30. Liu, Z., Bergholtz, E. J., Fan, H. & Läuchli, A. M. Fractional chern insulators in topological ﬂat bands with higher chern number. Phys. Rev. Lett. 109, 186805 (2012). Discussion In summary, we propose a construction scheme for tight-binding Hamiltonians hosting a ﬂat band whose band-touching point is characterized by dmax, the maximum value of the quantum dis- tance between Bloch eigenstates around the touching point. Based on the scheme, we built several ﬂat band tight-binding models with simple hopping structures in the kagome lattice and the αkvkjq ¼ Am1;n1eiðm1q1þn1q2Þ þ Am2;n2eiðm2q1þn2q2Þ: ð24Þ ð24Þ Since we assume that the ﬂat band is singular at the band- 7 ARTICLE Reprints and permission information is available at http://www.nature.com/reprints 67. Balčytis, A. et al. Synthetic dimension band structures on a si cmos photonic platform. Sci. Adv. 8, eabk0468 (2022). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 68. Ozawa, T., Price, H. M., Goldman, N., Zilberberg, O. & Carusotto, I. Synthetic dimensions in integrated photonics: from optical isolation to four- dimensional quantum hall physics. Phys. Rev. A 93, 043827 (2016). Ozawa, T. & Carusotto, I. Synthetic dimensions with magnetic 69. Ozawa, T. & Carusotto, I. Synthetic dimensions with magnetic ﬁelds and local interactions in photonic lattices. Phys. Rev. Lett. 118, 013601 (2017). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. 70. Albert, V. V., Glazman, L. I. & Jiang, L. Topological properties of linear circuit lattices. Phys. Rev. Lett. 114, 173902 (2015). 71. Kim, Y. et al. Realization of non-hermitian hopf bundle matter. arXiv https:// arxiv.org/abs/2303.13721 (2023). Author contributions 58. Hatsugai, Y. Chern number and edge states in the integer quantum hall effect. Phys. Rev. Lett. 71, 3697 (1993). H.K. and C.-g.O. contributed equally to this work. H.K. developed the general ﬂat band construction scheme. C.-g.O. performed the theoretical analysis of the bulk-boundary correspondence. J.W.R. supervised the project. y 59. Kitaev, A. Periodic table for topological insulators and superconductors. In: AIP conference proceedings, vol. 1134, 22–30 (American Institute of Physics, 2009). 60. Fukui, T., Shiozaki, K., Fujiwara, T. & Fujimoto, S. Bulk-edge correspondence for chern topological phases: a viewpoint from a generalized index theorem. J. Phys. Soc. Jpn. 81, 114602 (2012). References You, J.-Y., Gu, B. & Su, G. Flat band and hole-induced ferromagnetism in a novel carbon monolayer. Sci. Rep. 9, 1–7 (2019). 52. Huda, M. N., Kezilebieke, S. & Liljeroth, P. Designer ﬂat bands in quasi-one- dimensional atomic lattices. Phys. Rev. Res. 2, 043426 (2020). 17. Saito, Y. et al. Hofstadter subband ferromagnetism and symmetry-broken chern insulators in twisted bilayer graphene. Nat. Phys. 17, 478–481 (2021). dimensional atomic lattices. Phys. Rev. Res. 2, 043426 (2020) y 53. Rhim, J.-W., Behrends, J. & Bardarson, J. H. Bulk-boundary correspondence from the intercellular zak phase. Phys. Rev. B 95, 035421 (2017). 18. Sharpe, A. L. et al. Emergent ferromagnetism near three-quarters ﬁlling in twisted bilayer graphene. Science 365, 605–608 (2019). from the intercellular zak phase. Phys. Rev. B 95, 035421 (20 54. Provost, J. P. & Vallee, G. Riemannian structure on manifolds of quantum states. Commun. Math. Phys. 76, 289 (1980). 19. Wu, C., Bergman, D., Balents, L. & Sarma, S. D. Flat bands and wigner crystallization in the honeycomb optical lattice. Phys. Rev. Lett. 99, 070401 (2007). 55. Kane, C. L. & Mele, E. J. Quantum spin hall effect in graphene. Phys. Rev. Lett. 95, 226801 (2005). 20. Chen, Y. et al. Ferromagnetism and wigner crystallization in kagome graphene and related structures. Phys. Rev. B 98, 035135 (2018). 56. Kane, C. L. & Mele, E. J. Z 2 topological order and the quantum spin hall effect. Phys. Rev. Lett. 95, 146802 (2005). 21. Jaworowski, B. et al. Wigner crystallization in topological ﬂat bands. New J. Phys. 20, 063023 (2018). 57. Bernevig, B. A. & Zhang, S.-C. Quantum spin hall effect. Phys. Rev. Lett. 96, 106802 (2006). 8 COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys ARTICLE COMMUNICATIONS PHYSICS \| https://doi.org/10.1038/s42005-023-01407-6 Additional information 62. Rhim, J.-W., Bardarson, J. H. & Slager, R.-J. Uniﬁed bulk-boundary correspondence for band insulators. Phys. Rev. B 97, 115143 (2018). Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s42005-023-01407-6. 63. Celi, A. et al. Synthetic gauge ﬁelds in synthetic dimensions. Phys. Rev. Lett. 112, 043001 (2014). Correspondence and requests for materials should be addressed to Jun-Won Rhim. 64. Ozawa, T. & Price, H. M. Topological quantum matter in synthetic dimensions. Nat. Rev. Phys. 1, 349–357 (2019). Peer review information Communications Physics thanks Alexei Andreanov, Aleksandra Maluckov, and the other, anonymous, reviewer (s) for their contribution to the peer review of this work. A peer review ﬁle is available. y 65. Yuan, L., Lin, Q., Xiao, M. & Fan, S. Synthetic dimension in photonics. Optica 5, 1396–1405 (2018). 66. Dutt, A. et al. Experimental band structure spectroscopy along a synthetic dimension. Nat. Commun. 10, 3122 (2019). Reprints and permission information is available at http://www.nature.com/reprints Competing interests p g The authors declare no competing interests. 61. Mong, R. S. & Shivamoggi, V. Edge states and the bulk-boundary correspondence in dirac hamiltonians. Phys. Rev. B 83, 125109 (2011). Acknowledgements This work was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korean government (MSIT) (Grant no. 2021R1A2C1010572 and 2021R1A5A1032996 and 2022M3H3A106307411). This work was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korean government (MSIT) (Grant no. 2021R1A2C1010572 and 2021R1A5A1032996 and 2022M3H3A106307411). © The Author(s) 2023 COMMUNICATIONS PHYSICS \| (2023) 6:305 \| https://doi.org/10.1038/s42005-023-01407-6 \| www.nature.com/commsphys 9
https://openalex.org/W2576795123	https://revistas.ubiobio.cl/index.php/MCT/article/download/2666/2247	English	null	Modeling and simulation of an industrial indirect solar dryer for Iroko wood (Chlorophora excelsa) in a tropical environment	Maderas. Ciencia y tecnología	2,017	cc-by	8,134	1LERMaB, Post-Doctoral position, 27 rue Philippe Séguin, P.O. Box 1041, F-88051 Epinal, France. 2University of Lorraine, LERMaB, Faculty of Sciences and Technologies, Aiguillettes Campus, P.O. Box 70239-54506 VlN, Nancy, France. 3University of Lorraine, LERMaB, ENSTIB, 27 rue Philippe Séguin, P.O. Box 1041, F-88051 Epinal, France. 4Higher Teacher’s Training College, Applied Physic Laboratory, P.O.Box 47, Yaoundé, Cameroon. ♠ Corresponding author: simotagne2002@yahoo.fr Received: 17.01.2016 Accepted: 20.11.2016 MODELING AND SIMULATION OF AN INDUSTRIAL INDIRECT SOLAR DRYER FOR IROKO WOOD (Chlorophora excelsa) IN A TROPICAL ENVIRONMENT Merlin Simo-Tagne1,♠, André Zoulalian2, Romain Remond3, Yann Rogaume3, Beguidé Bonoma4 ABSTRACT This paper presents a modeling of an instrumental indirect solar wood dryer less expensive functioning in a Cameroonian climate applied to the climate of Yaoundé. The dryer is easy to build and electric energy is only used for the fan. Applications are done on Iroko wood (Chlorophora excelsa), a tropical wood 50mm thick most utilized in Africa. A satisfactory agreement between experimental and numerical results was found. Influences of thickness, wood initial water content and airflow rate were studied. Keywords: Drying curves, numerical simulation, solar greenhouse dryer, tropical climate, tropical woods. Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 DOI: 10.4067/S0718-221X2017005000009 ISSN impresa 0717-3644 ISSN online 0718-221X MODELING AND SIMULATION OF AN INDUSTRIAL INDIRECT SOLAR DRYER FOR IROKO WOOD (Chlorophora excelsa) IN A TROPICAL ENVIRONMENT Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 DOI: 10.4067/S0718-221X2017005000009 ISSN impresa ISSN online INTRODUCTION In Cameroon, forest exploitation gives provides considerable to the state. In addition, natural forests represent an important area. Iroko (Chlorophora excelsa) is prohibited exporting in the form of rough lumber. Then, it is very important to develop the first transformations in the country. For this reason, the reflections to optimize the drying are necessary in function of the local realities. Locally, after sawing, natural drying is used in the majority to dry wood. This method is known to have a lot of drying time and to destroy more boards than solar drying in an appropriate dryer. Solar drying is a method recommended where sustainable development is essential to reduce the environment impact of the industrial activities. Cameroon has good solar potential with an average incident solar energy from 4,5 kWh/m2 in the south and 5,8kWh/m2 daily in the great north (Ayangma et al. 2008). For example, the study of Ayangma et al. (2008) using 20 years data and applied in Garoua, a city of the great north region of 95 Universidad del B í o - B í o Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Cameroon gives an estimation near 5,743 kWh/(m2.day). Another works in the literature such as Lealea and Tchinda (2013) and Njomo and Wald (2006) present some estimations of solar irradiation of many Cameroonian towns in general. Figure 1 presents the position of Cameroon’s incident solar energy with some others developing countries. In the literature, we have solar dryers with the wall in glass or with energy storage (Awadalla et al. 2004, Bauer 2003, Bekkioui 2009, Bekkioui et al. 2009, Bekkioui et al. 2011, Bentayeb et al. 2008a, Bentayeb et al. 2008b, Luna 2008, Luna et al. 2010). These dryers are very expensive for the level of development of many people coming in tropical region. Mathematical modeling is an essential tool that facilitates the understanding of the physical phenomena occurring in the product to dry during solar drying. It is possible to numerically model a dryer in order to give the characteristics to impose during the construction and to satisfy the needs of a population. Figure 1. Global solar irradiation on a horizontal plane of some developing countries (Weiss and Buchinger 2003) Figure 1. INTRODUCTION Global solar irradiation on a horizontal plane of some developing countries (Weiss and Buchinger 2003) In this work, a modeling and a numerical simulation of the instrumental solar dryer of tropical woods coming from Cameroon are done. The price of construction of this simple dryer is low. The environment of the drying is the one of Yaoundé. Experiments are done from 22th November to 12th December 2004. Before, we present a model for the desorption isotherms based on experimental values presented in the literature. Modeling of the desorption isotherms We have used experimental values (Jannot et al. 2006) to propose a relation for the isotherms of desorption. For the equilibrium moisture content (Xeq,-) as a function of temperature (T, K) and fractional relative humidity (HR,-), we obtained: 1 2 2 1 (1 )(1 ) m eq b X HR X b HR b HR b HR = − − + (1a) (1b) 4 2 7,33 10 0,286; 0,994 m X x T R − = − + = 1 2 2 1 (1 )(1 ) m eq b X HR X b HR b HR b HR = − − + (1a) (1b) 4 2 7,33 10 0,286; 0,994 m X x T R − = − + = (1a) (1b) (1b) Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 (1c) 2 1 2135,87 27,827exp ; R =0,324 b RT −   =     (1d) 2 2 2308,798 1,931exp ; R =0,919 b RT −   =     (1c) (1d) Figure 2a illustrates the equilibrium moisture content of the desorption summarized in Eq (1a). The fractional moisture content at the fiber saturation point (XPSF,-) is obtained using Eq (1a) with a value of 1 for relative humidity. This is presented in Eq (2) and plotted as a function of temperature in Figure 2b. 1 2 2 1 (1 )(1 ) m fsp b X X b b b = − − + (2) 1 2 2 1 (1 )(1 ) m fsp b X X b b b = − − + (2) (a) (b) Figure 2. (a) Experimental and theoretical (Dent’s model) isotherms of desorption of Iroko wood and (b) evolution of the fiber saturation point versus temperature. (b) (a) (b) (a) Figure 2. (a) Experimental and theoretical (Dent’s model) isotherms of desorption of Iroko wood and (b) evolution of the fiber saturation point versus temperature. Solar data The drying was done in Yaoundé. After using the data given in Kemajou et al. (2012), we obtained Equations (3a) and (3b) of the correlations of experimental values of diffuse and global insulation on a horizontal plane in Yaoundé respectively. Geographical coordinates of Yaoundé are: Latitude: 3,87°N; Longitude: 11,52°E; Altitude: 720m (Afungchui and Neba-Ngwa 2013). D=0,12152t4-5,44546t3+76,93216t2-349,09338t +357,73432 ; R2=0,99967 (3a) Gt=0,51681t4-24,19623t3+386,86802t2-2427,06266t+5207,34134 ; R2=0,99882 (3b) Gt=0,51681t4-24,19623t3+386,86802t2-2427,06266t+5207,34134 ; R2=0,99882 (3b) (3b) Using the relationships given in Appendix A, we have deduced the global irradiations on a slope plane at 10° (3c) and on the wall at 90° (3d). 97 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Universidad del B í o - B í o Gt(10°)=-0,02411t5+1,86434t4-53,12658t3+685,0667t2-3906,6173t+8037,91894; R2=0,99 (3c) ( Gt(Wall)=0,12073t4-5,32896t3+74,29878t2-332,08327t+312,788 ; R2=0,9913 (3d) (3c) Gt(10°)=-0,02411t5+1,86434t4-53,12658t3+685,0667t2-3906,6173t+8037,91894; R2=0,99 (3c) ( G (W ll) 0 12073t4 5 32896t3+74 29878t2 332 08327t+312 788 R2 0 9913 (3d) (3c) (3c) (3d) Where t is the time of the day in hour. Between 6am and 6pm, Gt* is given by (3c) and (3d) on the slope and the wall respectively. At night, Gt* is equal to zero. Figure 3 presents the diffuse and global solar irradiation on horizontal plane experimentally obtained by Kemajou et al. (2012) and estimations of theses physical parameters on inclined plane with 10° angle. Figure 3. Global irradiations on a slope plane at 10° angle and on the wall (Kemajou et al. 2012). Figure 3. Global irradiations on a slope plane at 10° angle and on the wall (Kemajou et al. 2012). The wall and the slope are made in polyethylene. We have 3,7kg and 5,28m2 respectively the mass and surface of the east side wall. 4,623kg and 6,6m2 are respectively the mass and surface Solar dryer design The fan is the type SK012/4EHBWC. The electrical characteristics are 0,37kW power, 1350tr/min rate of rotation, 3,6A intensity, 240V tension and 50Hz frequency. The dryer dimensions are 3,10m length 2,40m width and 2,75m west height. The roof at the slope of 10 degrees with horizontal. The floor is well insulated to reduce the heat losses. Black body situated at the top of the dryer is made in a steel sheet (aluminum) with an inclination 10 degrees with horizontal, smooth 5e/10e. Another physical parameters of the steel sheet are 0,5mm thickness, 2m length, 1m width and 2,7kg mass. The air volume in the dryer is equal to the dryer volume with the reduction of the volume of all other component located in the dryer. The obtained air volume is equal to 13m3. Layout of the dryer is presented in Figure 4. Figure 4. Schematic representation of the indirect solar dryer. Figure 4. Schematic representation of the indirect solar dryer. The wall and the slope are made in polyethylene. We have 3,7kg and 5,28m2 respectively the mass and surface of the east side wall. 4,623kg and 6,6m2 are respectively the mass and surface 98 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. of the west side wall. The north and south sides have the mass and surface respectively equal to 5,374kg and 7,673m2.The mass and surface of the slope are 5,292kg and 7,555m2 respectively. of the west side wall. The north and south sides have the mass and surface respectively equal to 5,374kg and 7,673m2.The mass and surface of the slope are 5,292kg and 7,555m2 respectively. Modeling of the Dryer Hypotheses taken are: Each component inside the dryer is homogeneous; Each component inside the dryer is homogeneous; The thermophysical characteristics of the air are only influenced by the temperature. The hermophysical characteristics of the black body and the wall are constant; The thermophysical characteristics of the air are only influenced by the temperature. The hermophysical characteristics of the black body and the wall are constant; The entire wall has the same global solar radiation; We neglected the variation between the days of the global solar radiation; Natural convection is neglected; The floor of the dryer is supposed adiabatic; Transfers by conduction are neglected on all components of the dryer. Wood stack characteristics We have 14 layers with 4 boards by layer. Two consecutive layers are separated by the sticks and each stick dimensions are 220cm length and 3,5cm thickness. Each board in the wood stack has 5cm thickness, 220cm length and 38,5cm width. Thus, the wood stack porosity is equal to 0,412. The total board surface and the total wood stack volume are respectively equal to 97,16m2 and 2,156m3. Using the wood stack characteristics and the dryer dimensions, we obtained a fill factor (FF) equal to 0,195. Model equations Mass transfer on the wood stack: We have used a purely diffusive mass transfer model such as Ananías et al. (2009, 2011), Bekkioui (2009), Bekkioui et al. (2009, 2011) and Bentayeb et al. (2008a, 2008b): 0 ( ) b eq dx m KS X X dt − = − 0 ( ) b eq dx m KS X X dt − = − (4) (4) 0 0 p (1 )V 1108,26Kg m ρ ε = − = and 2 2 ( ) 97,16 b c pc t S LpN lN e m = + = . We (Gérard et al. 1998). 3 0 514,1 / Kg m ρ = Applied on temperate and Chilean woods, the literature proposes the variation of the global mass transfer coefficient K in function of the air temperature, air velocity, air relative humidity, fiber saturation point and equilibrium water content given on Equation (5) (Alvear et al. 2003, Ananías et al. 2009, Ananías et al. 2011, Bekkioui 2009, Bekkioui et al. 2009, Bekkioui et al. 2011, Bentayeb et al. 2008a, Bentayeb et al. 2008b): 99 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Universidad del B í o - B í o 0 0 0 0 int 1 1 exp exp exp p a a fsp eq C C HR a e b V K T T X X −       − = + −         −       (5) (5) a0=0,2265ms/kg; b0=268,9m2/kg; c0=2543,6K; p=2,7158 a0=0,2265ms/kg; b0=268,9m2/kg; c0=2543,6K; p=2,7158 In this relation, e is used expressed in mm (Chrusciel et al. 1999). Simo-Tagne et al. (2016) show that it is possible to use the same influences of the global mass transfer coefficient in the case of tropical woods. To estimate the external air temperature, we have used the relation (Benlahmidi 2013): min max max min 12 2 2 12 a a a a aext T T T T t T Cos π + − −   = +     (6) (6) We have Tamax =303,75K and Tamin =292,95K. Initially, we have taken the temperature of all the component of the dryer equal to the value min max 2 a a T T + . 2 2 Mass transfer of air The global mass balance of water in the dryer is given by Eq(7) taken from Bekkioui (2009), Bekkioui et al. (2009, 2011) and Bentayeb et al. (2008a, 2008b). (7) 0 ( ) (1 ) 0 s s E a p dY dX G Y Y m V dt dt ρ ε − + + − = (7) With ( (1 ) ) a a dryer p m V V ρ ε = − − G=0,05kg/s. ( (1 ) ) a a dryer p m V V ρ ε = − − ( (1 ) ) a dryer p V V ε − − (8) g/s. (8) G=0,05kg/s. Air relative humidity (HR,-) is given by: ve humidity (HR,-) is given by: Air relative humidity (HR,-) is given by: 0,622 atm s vsat s P Y HR P Y = + (9a) 0,622 atm s vsat s P Y HR P Y = + (9a) , vsat s Patm=101325-12z+5,2x10-4z2 (Galilée 2005) (9b) Patm=101325-12z+5,2x10-4z2 (Galilée 2005) (9b) Patm=101325-12z+5,2x10-4z2 (Galilée 2005) Patm=101325-12z+5,2x10-4z2 (Galilée 2005) Patm=101325-12z+5,2x10-4z2 (Galilée 2005) (9b) with z=720m for Yaoundé (Afungchui and Neba-Ngwa 2013). with z=720m for Yaoundé (Afungchui and Neba-Ngwa 2013). with z=720m for Yaoundé (Afungchui and Neba-Ngwa 2013). with z=720m for Yaoundé (Afungchui and Neba-Ngwa 2013). 5 5120 1,013125 10 exp 13,7 vsat a P x T   = −     ,(Nadeau and Puiggali 1995) (9c) (9c) Thermal transfer on the black body The rate of accumulation of thermal energy in the black body situated on the top of the drying chamber is equal to the sum of the rate of solar energy transferred by the roof and accumulated on the black body from solar radiation, the rate of thermal energy transfer by convection between the inside air and the black body, the rates of thermal energy exchanged by radiation between the roof and the black body, between the black body and the wall and between the black body and the wood stack respectively. Thus we have: (12) 4 4 4 4 to to 4 4 (10 ) ( ) S ( ) S ( ) S ( ) to to pto p to t to to to a to p to p pl pl to to pl to to b to p dT m C S G h S T T F T T F T T dt F T T τ α σ σ σ − − − = ° − − − − − − − − (12) Thermal transfer of the roof slope in polyethylene The rate of accumulation of thermal energy in the roof is equal to the sum of the rate of solar energy accumulated in the roof from solar radiation, the rate of thermal energy exchanged by radiation between the roof and the black body, and between the roof and the sky, the rate of thermal energy transferred by convection between the inside air and the roof, also between the outside air and the roof. Thus we have: (13) 4 4 4 4 p p ciel 1 (10 ) ( ) S F ( ) ( ) 2 1 ( ) 2 p p pp p p t p p to to p ciel p p vint p a p vext p aext dT m C S G S F T T T T S h T T dt S h T T α σ σ − − = ° + − + − − − − − (13) Thermal transfer on the wood stack We have neglected the exchange between the wood stack and the floor, also between the wood stack and the cover. Thus the rate of accumulation of thermal energy in the product is equal to the sum of the rate of thermal energy gain from the product due to sensible and latent heat, the rate of thermal energy received from air by the product due to convection and the rate of radiation between black body and wood stack. Thus we have: 100 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. ( ) ( ) ( ) ( ) 4 4 0 b pb b b eq b b a b bb bb b to b dT m C K L E S X X h S T T S F T T dt σ − = − + − + − + − (10) ( ) ( ) ( ) ( ) 4 4 0 b pb b b eq b b a b bb bb b to b dT m C K L E S X X h S T T S F T T dt σ − = − + − + − + − (10) (10) Thermal transfer of the air The rate of accumulation of thermal energy in the air inside the dryer is written as the sum of the rate of solar energy accumulated inside dryer from solar radiation through the roof and the wall respectively, the rate of thermal energy change in the air chamber due to inflow and outflow of the air in the chamber, the rate of thermal energy loss from air inside due to latent heat, the rates of thermal energy exchanged by convection between inside air and the roof, between inside air and the wall, between inside air and the black body and between inside air and wood stack respectively. Thus we have: (11) ( ) 1 1 pa 1 (10 ) (1 ) (Wall) GC ( ) ( ) ( ) ( ) ( ) ( ) a a pa p p t p p t aext a E S ci p a p cil pl a pl cto to a to cb b a b dT m C S G S G T T GL Y Y h S T T dt h S T T h S T T h S T T α α = − ° + − + − + − − − − − − − − − (11) Thermal transfer on the wall of the dryer The rate of accumulation of thermal energy in the wall (polyethylene cover) is equal to the sum of the rate of solar energy accumulated in the wall from solar radiation, the rate of thermal energy transfer by convection between the inside air and the wall, also between the outside air and the wall, the rate of thermal energy exchanged by radiation between the wall and the black body, and between the wall and the sky. Thus we have: 101 Universidad del B í o - B í o Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 (14) 4 4 pl pl to 4 4 pl pl 1 1 ( ) ( ) ( ) S F ( ) 2 2 S F ( ) pl pl ppl pl pl t pl cint pl a pl cext pl aext to pl ciel ciel pl dT m C S G Wall S h T T S h T T T T dt T T α σ σ − − = − − − − + − + − (14) The relations below are used in the program. Form factors are obtained after calculations with the dimensions of the components of the dryer. 1,5 0,0552 ciel aext T T = (15) 5,67 3,86 cext ext h V = + (16) 1,5 0,0552 ciel aext T T = (15) 5,67 3,86 cext ext h V = + (16) 5,67 3,86 cext ext h V = + (16) Vext=1,3m/s (meteorological data), Vint=1,5m/s;hciL=7w/(m2K);hci=8w/(m2K); hcto=8w/(m2K); 0,8 0,33 int ; 0,023 ; ; 2 u air h b u e r e h t h air N V D h N R P R D e D u λ = = = = (17a) (17a) We have Re=5687,36; Dh=0,07m; Pr=0,708; 0,026248 / ( ) air w mK λ = ; hcb=8w/(m2K). Thus we have hb =7,77w/(m2K). The thermophysical correlations on the wood and on the component of the dryer are taken in the literature Simo-Tagne (2014) and Simpson and TenWolde (1999). Those of air, black body (in aluminum) and wall (in polyethylene) are taken in the literature (Jannot 2011, Lienhard IV and Lienhard V 2011, Nadeau and Puiggali 1995). Method of Resolution and Experimental Drying Process Method of resolution Method of Resolution and Experimental Drying Process Method of Resolution and Experimental Drying Process Thermal transfer on the wall of the dryer We have: Cpto=900J/(kg.K); Cpp=CppL=2300J/(kg.K); 1835 0,734( 273,15) pa a C T = − − (17b) 1000(3335 2,91 ) a L T = − (17c) 3 1170,4 10 exp( 14 ) b E x X = − (17d) 4 6 8 ( 6,191 10 2,36 10 1,33 10 ) 1 po pw pb b C XC C X x x T x X X − − − + = + − + − + (17e) 3 0,1031 3,867 10 ; 4190 / (Kg.K) po b pw C x T C J − = + = (17f) 0,05; 0,91; 0,95; p pL to p α α α τ = = = = (17b) 1000(3335 2,91 ) a L T = − (17c) 3 1170,4 10 exp( 14 ) b E x X = − (17d) 4 6 8 ( 6,191 10 2,36 10 1,33 10 ) 1 po pw pb b C XC C X x x T x X X − − − + = + − + − + (17e) 3 0,1031 3,867 10 ; 4190 / (Kg.K) po b pw C x T C J − = + = (17f) (17f) Second term of second member of (17e) is equal to zero in a non-hygroscopic domain, domain that is limited by the water content at the fiber saturation points. The geometric factor is obtained using the method developed in Clark and Korybalsky (1974). This method is also used by Bekkioui (2009), Bekkioui et al. (2009, 2011) and Bentayeb et al. (2008). We have obtained: 102 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. Ftop=0,537; Fpciel=0,8; Fpto=0,81; FpLto=0,75; Ftob=0,3; Fbbb=0,323. The parameter used to compare experimental and numerical results is given by the average relative error: exp 1 exp 100 (%) 2 i i i i N th r i th X X E X X N = − = + ∑ (18) (18) Method of Resolution and Experimental Drying Process Method of resolution Finite difference method (Gonçalves 2005) is used to resolve equation relative to the mass transfer on the air (Eq. 7). We obtained: , 1 , 1 , 1 ( ) s to a o to t E t s t a Y m m X X GY t Y m G t + − + ∆ = + ∆ (19) (19) Known air humidity of the air at the time t0, ( ,s to Y ) we deducted the air humidity at the time 1 0 , 1 ( ) s t t t t Y = + ∆ . For all the others relations, we have used Runge Kutta method in the order 4 (Gonçalves 2005). Method of resolution For example, Eq.10 is treated such as the following: 4 4 b b a b bb bb b ( , ) ( ) ( ) h S (T T ) S F ( ) n bn b b eq to b A t T K L E S X X T T σ − = − + − + − + − (20a) (20a) ( , ) ( , ) n bn n bn o pb A t T f t T m C = ( , ) ( , ) n bn n bn o pb A t T f t T m C = (20b) (20b) (20b) Given Tbo we have: Given Tbo we have: Given Tbo we have: 1 n bn (t ,T ) k tf = ∆ 1 2 n bn (t ,T ) 2 2 k t k tf ∆ = ∆ + + 2 3 n bn (t ,T ) 2 2 k t k tf ∆ = ∆ + + 4 n bn 3 (t ,T ) k tf t k = ∆ + ∆ + 1 1 2 3 4 1 (k 2 2 ) 6 n n b b T T k k k + = + + + + Given Tbo we have: 1 n bn (t ,T ) k tf = ∆ 1 2 n bn (t ,T ) 2 2 k t k tf ∆ = ∆ + + 2 3 n bn (t ,T ) 2 2 k t k tf ∆ = ∆ + + 4 n bn 3 (t ,T ) k tf t k = ∆ + ∆ + 1 1 2 3 4 1 (k 2 2 ) 6 n n b b T T k k k + = + + + + 1 n bn (t ,T ) k tf = ∆ 1 n bn (t ,T ) k tf = ∆ (20c) (20d) 4 n bn 3 (t ,T ) k tf t k = ∆ + ∆ + (20f) 1 1 2 3 4 1 (k 2 2 ) 6 n n b b T T k k k + = + + + + (20g) (20f) (20g) Universidad del B í o - B í o Maderas. Experimental drying process The mission of promoting local materials of Cameroon (MIPROMALO) had built and experienced the solar dryer studied in this paper. After having positioned the wood stack, the fan is commanded ON. The air relative humidity inside is controlled after each two hours during the day. If the air relative humidity is more greater than 0,7 (near 0,9), the air inside is replaced by fresh outside air. When the air relative humidity inside is lower than 0,7 (near 0,3), the fresh air replaces the air inside the dryer. At night, inside air is exchanged each 1hr in order to avoid atmospheric saturation. Method of resolution Ciencia y tecnología 19(1): 95 - 112, 2017 Fortran language in the version 77 was used to generate our results and the step time used was 15s. We recorded our results each 1hr drying time. RESULTS AND DISCUSSION Known that initial drying time 0 corresponds to 00pm (local time in Yaoundé). Figure 5 shows a satisfactory agreement between theoretical and experimental measurements. We have a average relative error Er equal to 4,49%. From initial water content equal to 0,4kg/kg with the thickness equal to 50mm, 20days are necessary to dry iroko wood to 0,15kg/kg, Figure 5. Figure 6 shows the influence of the environment on the water content of our wood. During the night, equilibrium water content is great and it is weak at 12 o’clock. Then, the drying of wood stack is weak during the night. It is clear that in this condition, the wood stack dried until the equilibrium value of 0,11kg/kg after 768h drying time, Figure 6. Figure 5. Experimental and theoretical water content evolution versus drying time, 50mm. Figure 5. Experimental and theoretical water content evolution versus drying time, 50mm. 104 Figure 6. Theoretical water content and equilibrium water content evolutions versus drying time, 104 Figure 6. Theoretical water content and equilibrium water content evolutions versus drying time, Figure 6. Theoretical water content and equilibrium water content evolutions versus drying time, Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. Modeling and simulation of an..: Simo-Tagne et al. Modeling and simulation of an..: Simo-Tagne et al. 50mm. If we want to use this solar dryer to dry our wood in order to use in a house in Yaoundé, knowing that the means values in the Yaoundé house are HR=0,727 and T=24,8°C (Kameni et al. 2014), relation (1a) gives equilibrium water content equal to 0,144kg/kg for our wood. In the same experimental process and climatic conditions, our indirect solar dryer gives these equilibrium values after 18days. Figure 7 shows the time variations of temperatures relative to the wood stack (red), to the interior air (blue) and to the black body (green) during 528hrs drying time. Curves are presented 528hrs of the drying process. We notice that the curves are similar evolutions than the temperature of the air outside the dryer. Black body absorbs most energy necessary to facilitate drying process during the night. Figure 8 shows that air relative humidity vary with the time. When air temperature increases, air relative humidity decreases. Also, air temperature inside the dryer is greater than air exterior with a difference equal to 10°C. Figure 7. Predicted temperatures of wood stack, black body and air interior during the process, 50mm. Figure 7. Predicted temperatures of wood stack, black body and air interior during the process, 50mm. Figure 8. Theoretical relative humidity, outside air temperature and inside air temperature, 50mm. ure 8. Theoretical relative humidity, outside air temperature and inside air temperature, 50mm. Figure 9 presents the variation of average water content of wood stack with the drying time in function to the wood thickness numerically obtained. When wood thickness decreases, drying process is rapid. When the thickness changes, final water content is near than equilibrium water content. 105 Figure 9. Predicted water content versus drying time and board thickness, 50mm. Figure 9. Predicted water content versus drying time and board thickness, 50mm. Universidad del B í o - B í o Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Figure 10a shows that the drying kinetic is fast when initial water content is important because it is easy to dry free water. Also, final water content of all curves at different initial water content is near to equilibrium water content. It is clear that, if initial water content or thicknesses of wood are different, final water content of each plank is different. 50mm. Thus, it is important to do a pretreatment to the wood stack in order to homogenize initial water content of the stack. Also, it is possible to do this homogenization at the end of the drying process with an increase of interior air relative humidity. Figure 10b shows that when air flow rate increases, drying kinetic decreases. Increase of air flow rate helps to decrease the value of air temperature because communication between air drying and the components of the dryer is reduced. Figure 11 presents the consequence of two types of ventilation of the dryer. Open night conduct (in the night, we remove moist air in the dryer) helps to increase drying kinetic, compared to the case where in the night fan is stopped. When moist air is not remove in the night, it is possible that interior air humidity becomes more than the board humidity at the surface. Thus, the drying kinetic becomes very lowest comparing by the case where ventilation is not stopped. It is clear that the drying duration is great in the case where ventilation is stopped during the night. (a) (b) Figure 10. Predicted water content versus drying time with 50mm of thickness. (a) Influence of initial water content, (b) influence of air flow rate. (b) (a) (b) (a) Figure 10. Predicted water content versus drying time with 50mm of thickness. (a) Influence of initial water content, (b) influence of air flow rate. Figure 11. Predicted water content in function of the ventilation, 50mm. Figure 11. Predicted water content in function of the ventilation, 50mm. 106 106 Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Modeling and simulation of an..: Simo-Tagne et al. Modeling and simulation of an..: Simo-Tagne et al. CONCLUSIONS The model developed gives a satisfactory agreement in comparison with experimental results. The influences of initial wood water content, airflow velocity and board thickness on the drying process are conformed. Mathematical model proposed can be used to design the performing dryer in order to build in tropical region many indirect solar dryers less expensive. During drying process, it is important to homogenize initial water content and the board thickness in order to obtain a same drying kinetic of each board. At night, it is economically important to stop the fan and open the drying chamber when inside air is near saturation. In the future, it will interesting to discuss which terms are important to carefully characterize in this model. ACKNOWLEDGEMENTS The principal author acknowledges the International Tropical Timber Organization (ITTO) for financially supporting a part of this work (ITTO Ref. Number: 012/15A). The authors are grateful to the administration and all staff of the mission of promoting local materials of Cameroon (MIPROMALO) for the experimental data given and all explanations obtained in order to put it in our program. We acknowledge the two anonymous reviewers for all suggestions given to ameliorate the presentation of this paper. REFERENCES Afungchui, D.; Neba-Ngwa, R. 2013. Global solar radiation of some regions of Cameroon using the linear Angstrom and non-linear polynomial relations (part I) model development. International Journal of Renewable Energy Research 3(4):984-992. Afungchui, D.; Neba-Ngwa, R. 2013. Global solar radiation of some regions of Cameroon using the linear Angstrom and non-linear polynomial relations (part I) model development. International Journal of Renewable Energy Research 3(4):984-992. Alvear, M.; Broche, W.; Salinas, C.; Ananias, R.A. 2003. Drying kinetic of Chilean coigüe: Study of the global drying coefficient. 8th International IUFRO Wood Drying Conference: 383-387. Ananías, R.A.; Chrusciel, L.; Zoulalian, A.; Salinas-Lira, C.; Mougel, E. 2011. Overall mass transfer coefficient for wood drying curves predictions. Mass Transfer in Multiphase Systems and its Applications. [on line] Prof. Mohamed El-Amin (Ed.), ISBN: 978-953-307-215-9, InTech. Available from: <http://www.intechopen.com/books/mass-transfer-in-multiphase-systems-and-its-applications/ overall-masstransfer-coefficient-for-wood-drying-curves-predictions> Ananías, R.A.; Mougel, E.; Zoulalian, A. 2009. Introducing an overall mass-transfer coefficient for prediction of drying curves at low temperature drying rates. Wood Science and Technology 43(1):43- 56. Awadalla, H.S.F.; El-Dib, A.F.; Mohamad, M.A.; Reuss, M.; Hussein, H.M.S. 2004. Mathematical modelling and experimental verification of wood drying process. Energy Conversion and Management (45):197-207. Ayangma, F.; Nkeng, G.E. ; Bonoma, D.B.; Nganhou, J. 2008. Evaluation du potentiel en énergie solaire au Cameroun : Cas du Nord Cameroun. African Journal of Science and Technology, Science and Engineering series 9(2):32-40. Bauer, K. 2003. Development and optimization of a low temperature drying schedule for Eucalyptus grandis (Hill) ex maiden in a solar assisted timber dryer. Ph.D. Thesis, Institut für Agrartechnik in den Tropen und Subtropen, Universität Hohenheim, Germany. 107 Universidad del B í o - B í o Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Bekkioui, N. 2009. Séchage solaire du bois: modélisation simplifiée du séchage solaire d’une pile de bois dans un séchoir solaire à parois vitrées. Ph.D.Thesis, University of Mohammed V-AGDAL, Morocco. Bekkioui, N.; Hakam, A.; Zoulalian, A.; Sesbou, A.; Kortbi, M.E. 2011. Solar drying of pin lumber: Verification of a mathematical model. Maderas. Ciencia y Tecnologia 13(1):29-40. Bekkioui, N.; Hakam, A.; Zoulalian, A.; Sesbou, A.; Kortbi, M.E. 2011. Solar drying of pin lumber: Verification of a mathematical model. Maderas. Ciencia y Tecnologia 13(1):29-40. Bekkioui, N.; Zoulalian, A.; Hakam, A.; Bentayeb, F.; Sesbou, F. 2009. Modelling of a solar wood dryer with glazed walls. Maderas. Ciencia y Tecnologia 11(3): 191-205. Benlahmidi, S. 2013. Etude du séchage convectif par l’énergie solaire des produits rouges. Ph.D.Thesis. University of Mohamed Khider-Biskra. REFERENCES Njomo, D.; Wald, L. 2006. Solar irradiation retrieval in Cameroon from meteosat satellite imagery using Helio_2 method. ISESCO Science and Technology Vision 2(1):19-24. Simo-Tagne, M. 2014. Numerical study of heat and mass transfer during the thermal drying of tropical woods. International Journal of Thermal and Environment Engineering 8(2):9-15. Simo-Tagne, M.; Monkam, L.; Rémond, R.; Zoulalian, A.; Rogaume, Y.; Beguide-Bonoma. 2016. Experimental determination of the global mass transfer coefficient of the tropical woods in order to deduce the drying curves at the lower temperature. International Journal of Thermal and Environment Engineering 12(1):9-14. Simo-Tagne, M.; Monkam, L.; Rémond, R.; Zoulalian, A.; Rogaume, Y.; Beguide-Bonoma. 2016. Experimental determination of the global mass transfer coefficient of the tropical woods in order to deduce the drying curves at the lower temperature. International Journal of Thermal and Environment Engineering 12(1):9-14. Simpson, W.T.; TenWolde, A. 1999. Physical properties and Moisture Relations of Wood. Chapter 3 from Forest Products Laboratory. Wood handbook. Weiss, W.; Buchinger, J. 2003. Solar drying. AEEE INTEC. Weiss, W.; Buchinger, J. 2003. Solar drying. AEEE INTEC. REFERENCES Algeria. Bentayeb, F.; Bekkioui, N.; Camacho, E. F. 2008b. Simulation of a solar dryer functioning in a Moroccan climate. VI Minsk International Seminar. Heat Pipes, Heat Pumps, Refrigerators : 214-219. Bentayeb, F.; Bekkioui, N.; Zeghmati, B. 2008a. Modelling and simulation of a wood solar dryer in a Moroccan climate. Renewable Energy (33):501-506. Chrusciel, L.; Mougel, E.; Zoulalian, A.; Meunier, T. 1999. Characterisation of water transfer in a low temperature convective wood drier: influence of the operating parameters on the mass transfer coefficient. Holz Roh Werkstoff (57):439-445. Clark, J.A.; Korybalski, M.E. 1974. Algebraic methods for the calculation of radiation exchange in an enclosure. Warme-und Stoffubertragung 7:31-44. Galilée, L.P. 2005. L’air humide Cours de climatisation, (Chapter 1). BTS Cours FEE 1ère Année. Gérard, J.; Kouassi, A.E.; Daigremont, C.; Détienne, P.; Fouquet, D.; Vernay, M. 1998. Synthèse sur les caractéristiques technologiques de référence des principaux bois commerciaux africains. Série FORAFRI, Document 11, CF, CIRAD, CIFOR. Gonçalves, E. 2005. Résolution numérique, discrétisation des EDP et EDO. Institut National Polytechnique de Grenoble. Jannot, Y. 2011. Thermique solaire. Cours de transfert thermique. Jannot, Y.; Kanmogne, A.; Talla, A.; Monkam, L. 2006. Experimental determination and modelling of water desorption isotherms of tropical woods: afzelia, ebony, iroko, moabi and obeche. Holz als Roh-und Werkstoff (64):121-124. Kameni, N.M.; Tchinda, R.; Orosa, J.A.; Roshan, G. 2014. Study of dioxide carbon concentration and indoor àir quality in some buildings in the equatorial region of Cameroon (Yaoundé). Iranian Journal of Health Sciences 2(2):1-15. Kemajou, A.; Mba, L.; Pako-Mbou, G. 2012. Energy efficiency in air-conditioned buildings of the tropical humid climate. IJRRAS 11(2):235-240. Lealea, T.; Tchinda, R. 2013. Estimation of diffuse solar radiation in the South of Cameroon. Journal of Energy Technologies and Policy 3(6):32-42. Lienhard IV, J.H.; Lienhard V, J.H. 2011. A heat transfer textbook. Fourth edition, Cambridge Massachusetts. Luna, D.L. 2008. Modélisation et conception préliminaire d’un séchoir solaire pour bois de pin avec stockage d’énergie. Ph.D. Thesis, ENSAM, France. Luna, D.L.; Nadeau, J.P.; Jannot, Y. 2010. Model and simulation of a solar kiln with energy storage. Renewable Energy 36(11):2533-2542. 108 Modeling and simulation of an..: Simo-Tagne et al. Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 Nadeau, J.P.; Puiggali, J.R. 1995. Séchage, des processus physiques aux procédés industriels. Paris, New York, Londres, Tec and Doc. Njomo, D.; Wald, L. 2006. Solar irradiation retrieval in Cameroon from meteosat satellite imagery using Helio_2 method. ISESCO Science and Technology Vision 2(1):19-24. (Jannot 2011) (Jannot 2011) St=Gt-D; Gt, St and D are global, direct and diffuse insolation on horizontal plane. (A1) For inclined plane i, we have: (A1) For inclined plane i, we have: ( , ) ( , ) ( , ) ( , ) G i S i D i R i γ γ γ γ = + + (A1) ( , ) ( , ) ( , ) ( , ) G i S i D i R i γ γ γ γ = + + ( , ) ( , ) ( , ) ( , ) G i S i D i R i γ γ γ γ = + + [ ( , ) cos( )sin( )cos( ) s sin( ) S S i h i a h γ γ = − + [ ] ( , ) 1 cos( ) 2 G R i i γ ρ = − [ ] ( , ) 1 cos( ) 2 D D i i γ = + [ ] ( , ) cos( )sin( )cos( ) sin( )cos( ) sin( ) S S i h i a h i h γ γ = − + (A2) (A3) (A4) [ ] ( , ) 1 cos( ) 2 G R i i γ ρ = − [ ] ( , ) 1 cos( ) 2 D D i i γ = + [ ] ( , ) cos( )sin( )cos( ) sin( )cos( ) sin( ) S S i h i a h i h γ γ = − + (A2) (A2) [ ] ( , ) 1 cos( ) 2 G R i i γ ρ = − [ ] ( , ) 1 cos( ) 2 D D i i γ = + [ ] ( , ) 1 cos( ) 2 D D i i γ = + (A3) ρ is the albedo sin(h)=sin(Lat)sin(δ)+cos(Lat)cos(δ)cos(w) (A5) sin(h)=sin(Lat)sin(δ)+cos(Lat)cos(δ)cos(w) sin(h)=sin(Lat)sin(δ)+cos(Lat)cos(δ)cos(w) (A5) Universidad del B í o - B í o Maderas. Ciencia y tecnología 19(1): 95 - 112, 2017 )s cos( si in( ) n( ) cos( ) w a h δ = (A6) (A7) TS is Solar time expressed in h. (Jannot 2011) (A8) δ=23,45ºsin[0,98º(j+284] j day number in the year ; 15 ref l long TS TL C ET − = − + + (A9) TL is the legal time (h), ET is Equation of the time (h) lref and long are reference of longitude and longitude of the site (°) respectively. We have used C=1h and lref=0° for Yaoundé. Nomenclature Cpa : mass heat of the air (J/(kg.K)); Cpto : mass heat of the black body (J/(kg.K)); Cpb : mass heat of the wood (J/(kg.K)); Cpp : mass heat of the roof slope (polyethylene) (J/(kg.K)); Cppl : mass heat of the wall (polyethylene) (J/(kg.K)); D : diffuse solar irradiation on a horizontal plane (w/m2); : derivative with respect of the drying time (s-1); e : thickness of the plank (mm); Eb : desorption heat (J/kg); Er : average relative error (%); et : stick thickness (m); Fbb-b : geometric factor black body-wood(-); FF : volume of the wood stack divided by the volume of the dryer chamber (-); d dt 110 Modeling and simulation of an..: Simo-Tagne et al. Maderas. Nomenclature Ciencia y tecnología 19(1): 95 - 112, 2017 Fpl-ciel : geometric factor between wall and sky (-); Fpl-to : geometric factor between wall and black body (-); Fp-ciel : geometric factor between roof slope and sky (-); Fto-b : geometric factor between black body and wood (-) G : mass flow rate of dry air (kg/s); Gt : global solar radiation on a horizontal plane (w/m2) Gt * (10°) : global solar radiation on the roof (w/m2); Gt * (Wall) : global solar radiation on the wall (w/m2); hb : convective transfer between wood and air (w/(m2K hcb : interior convective coefficient between the wood s hcext : convective coefficient between exterior air and wa hci : interior convective coefficient between the roof slo hcil : interior convective coefficient between the wall an hcto : interior convective coefficient between the black b HR : air relative humidity (-); hvext : convective coefficient between exterior air and the hvint : convective coefficient between interior air and the K : mass global transfer coefficient (kg/(m2s)); l : width of each plank (m); L : latent heat of evaporation (J/kg); Lp : length of each plank (m); Pr : Prandtl number (-); Pvsat : pressure of vapor in saturation (Pa); R : perfect gas constant (8.314J/(mol.K)); r : regression of determination (-); Sb : total exchange surface of the wood (m2); Sbb : black body surface (m2); Sp : surface of the roof slope (m2); Spl : surface of the wall (m2); t : drying time (s); Ta : interior air temperature (K); Taext : temperature of the air exterior of the dryer (K); Tamax : maximum air temperature (K); Tamin : minimum air temperature (K); Tb : wood temperature (K); Tciel : temperature of the sky (K); Tpl : temperature of the wall (K); Tto : black body temperature (K); Vext : ambient air velocity (m/s); Vint : air velocity inside dryer(m/s); Vdryer : volume of the dryer (m3) ; Vp : wood stack volume (m3); X : water content of the wood stack in dry basis (kg/k Xeq : equilibrium water content in dry basis (kg/kg); Xfsp : fractional moisture content in dry basis at the fiber Xm : fractional moisture content in dry basis at the mon kg); Fpl-ciel : geometric factor between wall and sky (-); Fpl-to : geometric factor between wall and black body (-); Fp-ciel : geometric factor between roof slope and sky (-); Fto-b : geometric factor between black body and wood (-); G : mass flow rate of dry air (kg/s); Gt : global solar radiation on a horizontal plane (w/m2), Gt * (10°) : global solar radiation on the roof (w/m2); Gt * (Wall) : global solar radiation on the wall (w/m2); hb : convective transfer between wood and air (w/(m2K)); hcb : interior convective coefficient between the wood stack and the air (w/(m2K)); hcext : convective coefficient between exterior air and wall (w/(m2K)); hci : interior convective coefficient between the roof slope and the air (w/(m2K)); hcil : interior convective coefficient between the wall and the air (w/(m2K)); hcto : interior convective coefficient between the black body and the air (w/(m2K)); HR : air relative humidity (-); hvext : convective coefficient between exterior air and the roof slope (w/(m2K)); hvint : convective coefficient between interior air and the roof slope (w/(m2K)); K : mass global transfer coefficient (kg/(m2s)); l : width of each plank (m); L : latent heat of evaporation (J/kg); Lp : length of each plank (m); Pr : Prandtl number (-); Pvsat : pressure of vapor in saturation (Pa); R : perfect gas constant (8.314J/(mol.K)); r : regression of determination (-); Sb : total exchange surface of the wood (m2); Sbb : black body surface (m2); Sp : surface of the roof slope (m2); Spl : surface of the wall (m2); t : drying time (s); Ta : interior air temperature (K); Taext : temperature of the air exterior of the dryer (K); Tamax : maximum air temperature (K); Tamin : minimum air temperature (K); Tb : wood temperature (K); Tciel : temperature of the sky (K); Tpl : temperature of the wall (K); Tto : black body temperature (K); Vext : ambient air velocity (m/s); Vint : air velocity inside dryer(m/s); Vdryer : volume of the dryer (m3) ; Vp : wood stack volume (m3); X : water content of the wood stack in dry basis (kg/kg); Xeq : equilibrium water content in dry basis (kg/kg); Xfsp : fractional moisture content in dry basis at the fiber saturation point(kg/kg); Xm : fractional moisture content in dry basis at the monolayer saturation point (kg/ kg); 111 Universidad del B í o - B í o Maderas. Nomenclature Ciencia y tecnología 19(1): 95 - 112, 2017 i exp X : experimental water content of the wood stack in dry basic number i (kg/kg); i th X : theoretical water content of the wood stack in dry basic number i(kg/kg); YE : inlet air humidity of the dryer (kg/kg); YS : inside air humidity of the dryer (kg/kg); z : altitude of the dryer (the one of the town)(m); αp : absorptivity of the roof (-); αpL : absorptivity of the wall (-); αto : absorptivity of the black body (-); τp : transmittivity of the roof and wall (-); λair : thermal conductivity of the air(W/(m.K)); μair : dynamical viscosity of the air(Pa.s); ε : wood stack porosity (-); ρa : air density (kg/m3) ; ρo : wood dry density (kg/m3); σ : Stefan-Boltzmann constant value ∆t : time step (s); \| \| : symbol of the absolute value. 8 2 4 (5,67 10 / (m ); x W K − 112
https://openalex.org/W3182507118	https://periodicos.saude.sp.gov.br/RIAL/article/download/35002/33497	Portuguese	null	O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw	Revista do Instituto Adolfo Lutz	2,002	cc-by	2,299	1 Instituto Adolfo Lutz Central – São Paulo, S.P. Seção de Microbiologia Alimentar * Endereço para correspondência: Av. Dr. Arnaldo, 355 – 01246-902 – São Paulo – SP Ristori, C. A., Pereira, M Rev. Inst. Adolfo Lutz, Rev. Inst. Adolfo Lutz, 61(2):131-133, 2002 Ristori, C. A., Pereira, M Rev. Inst. Adolfo Lutz, Rev. Inst. Adolfo Lutz, 61(2):131-133, 2002 eta moída frente a contaminação in vitro com Salmonella Rubislaw COMUNICAÇÃO CIENTÍFICA/ BRIEF COMUNICATION Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 Rubislaw, obtida de pimenta, por ter sido este sorotipo o mais freqüentemente isolado em amostras deste produto, analisadas pela Seção de Microbiologia Alimentar, do Instituto Adolfo Lutz Central. A cultura da cepa, em fase estacionária, foi obtida em caldo BHI (Brain Heart Infusion) a 35ºC por 18-24h (cerca de 109 UFC/mL), A composição do óleo essencial de pimenta é complexa5. É importante ressaltar que os níveis dos compostos que possuem ação antibacteriana podem variar em função da espécie de pimenta, das condições de cultivo da mesma e da forma de extração, o que afeta diretamente a atividade inibitória do óleo essencial2. Os componentes dos óleos essenciais exercem atividade antibacteriana por 1) interferência na dupla camada fosfolipídica da parede celular por aumento da permeabilidade e perda dos constituintes celulares; 2) alteração de uma variedade de sistemas enzimáticos, incluindo aqueles envolvidos na produção de energia celular e síntese de componentes estruturais e 3) inativação ou destruição do material genético8. Foram realizados dois experimentos, baseados nos testes de Beuchat1, um com meios de cultura sólidos e outro, em caldo. Beuchat testou a sensibilidade do Vibrio parahaemolyticus frente especiarias secas adicionadas em meios de cultura nas concentrações finais de 0.1-1.0% e seus óleos essenciais nas concentrações finais de 10 e 100 µg/mL. De acordo com os resultados apresentados no referido trabalho, foram escolhidas as concentrações mais altas de 1% e 100 µg. No presente estudo, o primeiro experimento foi realizado preparando-se dois meios de cultura, um adicionado de 1% de pimenta do reino preta moída de três marcas diferentes (A, B e C) e o outro com 1% de óleo essencial de pimenta do reino preta diluído em etanol absoluto. O meio de cultura base foi: 3 g de caldo tripticaseína de soja (TSB) em pó e 1g de ágar para solidificação em 100mL de água destilada. O controle negativo dos meios foi realizado sem adição do óleo essencial ou pimenta, isto é, apenas o meio base sem adição. Após a esterilização em autoclave a 121ºC por 15 minutos, o meio foi distribuído em placas de Petri (15-18mL/placa). Foram feitas diluições decimais até 10-7 da cultura de S. Rubislaw, em fase estacionária, em água peptonada tamponada a 1%. Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev Inst Adolfo Lutz 62(2): 131-133 2002 sólido preparado anteriormente. As placas foram incubadas em estufa a 35ºC e as Unidades Formadoras de Colônias (UFC) contadas após 24 e 48h de incubação. As especiarias e seus derivados têm sido usados no preparo de alimentos há milhares de anos, conferindo-lhes sabor e aroma diferenciados. O Brasil é o segundo maior produtor de pimenta no mundo5, sendo que um dos componentes que contribuem para aumentar o seu valor como condimento é o seu óleo essencial. A ação inibitória das especiarias e seus extratos nos diferentes microrganismos tem sido relatada em diversos estudos1,2,4,8,9. No segundo experimento, o óleo essencial foi diluído em etanol (0,1mL de óleo para 10mL de etanol) e uma alíquota de 0,1mL dessa solução foi adicionada em 100mL de TSB resultando em uma concentração final de 100µg de óleo por mL. Como controle, o etanol puro foi adicionado ao TSB. Alíquotas de 1mL da cultura de S. Rubislaw diluída até 10-3 foram inoculadas em TSB adicionado do óleo e no TSB controle. Após incubação do caldo a 35ºC por 18/24h, foi retirada uma alíquota de 1mL para ser realizada a contagem padrão em placas, de acordo com SWANSON et al.11. O interesse renovado no uso de especiarias como agentes antibacterianos é atribuído basicamente a duas razões: (1) a segurança dos aditivos químicos é constantemente questionada havendo uma tendência ao uso de substâncias naturais de plantas; (2) a redução do sal ou do açúcar em alimentos por razões dietéticas tende a aumentar o uso de outros temperos6. O objetivo deste trabalho foi avaliar a ação antibacteriana da pimenta do reino preta moída (Piper Nigrum L.) e do seu óleo essencial frente a uma cepa de Salmonella Rubislaw. Cada experimento foi repetido três vezes. Os resultados não demonstraram diferença significativa entre os meios controles e os meios testes, mostrando assim que a pimenta do reino moída e o seu óleo essencial não foram eficientes para inibir o desenvolvimento de S. Rubislaw, quando em concentrações de 1% (experimento 1 – Tabela 1) e 100µg/mL (experimento 2 – Tabela 2). Nesse estudo foi utilizada cepa de S. Tabela 1 – Contagem padrão (UFC/mL) e número (em Log10) de S. Rubislaw em meio de cultura adicionado de três tipos de pimenta do reino preta moída e de seu óleo essencial na concentração final de 1%. O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. The effect in vitro of ground black pepper on contamination with Salmonella Rubislaw Christiane A. RISTORI1* Marco A. dos S. PEREIRA1 Dilma S. GELLI1 RIALA6/929 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 RESUMO. As especiarias e seus derivados têm sido usados no preparo de alimentos há milhares de anos, conferindo-lhes sabor e aroma diferenciados. A ação inibitória das especiarias e seus extratos nos diferentes microrganismos tem sido relatada. O objetivo deste trabalho foi avaliar a ação antibacteriana da pimenta do reino preta moída (Piper nigrum L.) e de seu óleo essencial frente a uma cepa de Salmonella Rubislaw. Os efeitos da pimenta e seu óleo adicionados em meios de cultura foram avaliados em dois experimentos. No primeiro foram preparados dois tipos de meio de cultura sólidos, um com 1% de pimenta do reino e o outro com 1% de óleo essencial diluído em etanol, adicionados de caldo tripticase de soja (TSB) e ágar. As semeaduras foram feitas em superfície de diluições decimais (10-1 até 10-7) da cepa de S. Rubislaw, em fase estacionária. Após a incubação a 35ºC durante 24 a 48h, procedeu-se a contagem das colônias. No segundo experimento, 1mL da diluição 10-3 da mesma cepa, em fase estacionária, foi adicionada a 100mL de TSB com o óleo diluído em etanol, para uma concentração final de 100µg/mL. Após 24h de incubação a 35ºC foram realizadas contagens padrão em placas. Os resultados obtidos nos experimentos não demonstraram efeito inibitório da pimenta do reino preta moída e de seu óleo essencial na cepa estudada. PALAVRAS-CHAVES. pimenta do reino preta moída, Salmonella Rubislaw, atividade antibacteriana 131 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 Gelli, D. S The effect in vitro of ground black pepper on contamination with salmonella rubislaw . Rev. Inst. Adolfo Lutz, 62(2):131-133, 2002 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S The effect in vitro of ground black pepper on contamination with salmonella rubislaw . Rev. Inst. Adolfo Lutz, 62(2):131-133, 2002 ABSTRACT. Spices and their derivatives have been used in food housekeeping for thousands of years, bringing about differentiated flavor and aroma. Inhibitory action of the spices and their extracts on the different microorganisms has been reported. The objective of this work was to evaluate the antibacterial action of the ground black pepper (Piper nigrum L.) and of its essential oil on Salmonella Rubislaw strain culture. The effect of the pepper and its oil addition in culture medium was evaluated in two experiments. In the first one two types of solid culture medium had been prepared, one with 1% of ground black pepper and the other with 1% ethanol oil solution, added to broth tripticase soy (TSB) and agar. Sowings of dilutions (10-1 up to 10-7) of the Salmonella strain, in stationary phase had been made in agar surface. After incubation at 35ºC for 24/48h, standard count method was carried out. In a second experiment, 1mL of the same strain dilution 10-3, in stationary phase, was added to 100ml of TSB with the oil diluted in ethanol, to a final concentration of 100µg/mL. Standard count was carried out after 24h of incubation at 35ºC. The results gotten in the experiments had not demonstrated inhibitory effect of the ground black pepper and its essential oil for the studied strain. KEY WORDS: ground black pepper, Salmonella Rubislaw, antibacterial activity KEY WORDS: ground black pepper, Salmonella Rubislaw, antibacterial activity KEY WORDS: ground black pepper, Salmonella Rubislaw, antibacterial activity Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 De cada diluição 0,1mL foi semeado em superfície de placas de Petri (7 placas) contendo o meio Alguns fatores citados por diferentes autores podem explicar os resultados obtidos no presente trabalho: a quantidade de óleo essencial presente na pimenta pode variar2; o óleo essencial pode não se solubilizar de forma adequada7, a variação na composição do óleo em função das condições de cultivo10. Além disso, estudos têm demonstrado que, de maneira geral, as bactérias Gram-negativas são mais resistentes à ação do óleo essencial3,4,9. Observa-se que nesse estudo foi testada apenas uma cepa e sorotipo de Salmonella e que pode existir diferença no comportamento de outras cepas, assim como de sorotipos, no que se refere à sensibilidade frente a estes compostos naturais. Tabela 1 – Contagem padrão (UFC/mL) e número (em Log10) de S. Rubislaw em meio de cultura adicionado de três tipos de pimenta do reino preta moída e de seu óleo essencial na concentração final de 1%. MEIOS DE CULTURA SÓLIDOS CEPA TSB controle TSB + pimenta TSB + pimenta TSB + pimenta TSB + óleo (Amostra A) (Amostra B) (Amostra C) essencial S. Rubislaw 1,6 x109 8,8x108 2,2x108 3,9x108 3,0x109 UFC/mL (Log 10) (9,20) (8,94) (8,34) (8,59) (9,48) 132 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S O efeito da pimenta do reino preta moída frente a contaminação in vitro com Salmonella Rubislaw. Rev. Inst. Adolfo Lutz, 62(2): 131-133, 2002 Tabela 2 - Contagem padrão (UFC/mL) e número (em Log10) de S. Rubislaw em meio de cultura adicionado de óleo essencial de pimenta do reino moída preta em concentração de 100µg/mL. Tabela 2 - Contagem padrão (UFC/mL) e número (em Log10) de S. Rubislaw em meio de cultura adicionado de óleo essencial de pimenta do reino moída preta em concentração de 100µg/mL. g p ( ) ( g ) pimenta do reino moída preta em concentração de 100µg/mL. MEIOS DE CULTURA LÍQUIDOS CEPA TSB controle TSB + óleo essencial S. Rubislaw 9,0x108 1,1x109 UFC/mL (Log 10) (8,95) (9,04) RIALA6/929 Ristori, C. A., Pereira, M. A.dos S.,. Gelli, D. S The effect in vitro of ground black pepper on contamination with salmonella rubislaw . Rev. Inst. Adolfo Lutz, 62(2):131-133, 2002 pimenta do reino moída preta em concentração de 100µg/mL. MEIOS DE CULTURA LÍQUIDOS CEPA TSB controle TSB + óleo essencial S. Rubislaw 9,0x108 1,1x109 UFC/mL (Log 10) (8,95) (9,04) RIALA6/929 Ristori, C. A., Pereira, M. A.dos S.,. 7. Juven, B.J. et al. Factors that interact with the antibacterial action of thyme essential oil and its active constituents. J. Appl. Bacteriol. 76:626-631,1994. Recebido em 27/02/2002; Aprovado em 05/12/2002 REFERÊNCIAS 7. Juven, B.J. et al. Factors that interact with the antibacterial action of thyme essential oil and its active constituents. J. Appl. Bacteriol. 76:626-631,1994. 8. Kim, J.M. et al. Antibacterial Activity of Carvacrol, Citral, and Geraniol against Salmonella typhimurium in Culture Medium and on Fish Cubes. J. Food Sci. 60(6):1364-1368,1995. 1. Beuchat, L.R. Sensitivity of Vibrio parahaemolyticus to spices and organic acids. J. Food Sci. 41:899-902,1976. 2. Deans,S.G.; Ritchie, G. Antibacterial properties of plant essential oils. Int. J. Food Microbiol. 5:165-180,1987. 9. Ouattara, B. et al. Antibacterial activity of selected fatty acids and essential oils against six meat spoilage organisms. Int. J. Food Microbiol. 37:155-162,1997. 3. Dorantes, L. et al. Inhibition of growth of some foodborne pathogenic bacteria by Capsicum annum extracts. Int. J. Food Microbiol. 57:125- 128,2000. 10. Sivropoulou, A. et al. Antimicrobial and cytotoxic activities of Origanum essential oils. J. Agric. Food Chem. 44:1202-1205,1996. 4. Farag, R.S. et al. Antimicrobial Activity of Some Egyptian Spice Essential Oils. J. Food Protect. 52 (9):665-667, 1989. 11. Swanson, K.M.J. et al. “Colony count methods”. In: Vanderzant, C. & Splittstoesser, D.F., ed. Compendium of Methods for the Microbiological Examination of Foods, Washington, D.C.; 1992, p.75-95. 5. Ferreira, S.R.S. et al. Supercritical fluid extraction of black pepper (Piper nigrun L.) essential oil. J. Superc. Fuids 14:235-245, 1999. 6. Ismaiel, A.; Pierson, M.D. Inhibition of Growth and Germination of C. botulinum 33A, 40B, and 1623E by Essential Oil of Spices. J. Food Sci. 55(6): 1676-1678,1990. Recebido em 27/02/2002; Aprovado em 05/12/2002 133
https://openalex.org/W2767120054	https://www.mdpi.com/2073-4360/9/10/522/pdf?version=1508332400	English	null	Evaluation of Thermal and Thermomechanical Behaviour of Bio-Based Polyamide 11 Based Composites Reinforced with Lignocellulosic Fibres	Polymers	2,017	cc-by	12,066	Received: 14 September 2017; Accepted: 16 October 2017; Published: 18 October 2017 Abstract: In this work, polyamide 11 (PA11) and stone ground wood ﬁbres (SGW) were used, as an alternative to non-bio-based polymer matrices and reinforcements, to obtain short ﬁbre reinforced composites. The impact of the reinforcement on the thermal degradation, thermal transitions and microstructure of PA11-based composites were studied. Natural ﬁbres have lower degradation temperatures than PA11, thus, composites showed lower onset degradation temperatures than PA11, as well. The thermal transition and the semi-crystalline structure of the composites were similar to PA11. On the other hand, when SGW was submitted to an annealing treatment, the composites prepared with these ﬁbres increased its crystallinity, with increasing ﬁbre contents, compared to PA11. The differences between the glass transition temperatures of annealed and untreated composites decreased with the ﬁbre contents. Thus, the ﬁbres had a higher impact in the composites mechanical behaviour than on the mobility of the amorphous phase. The crystalline structure of PA11 and PA11-SGW composites, after annealing, was transformed to α’ more stable phase, without any negative impact on the properties of the ﬁbres. Keywords: polyamide 11; lignocellulosic fibres; thermomechanical behaviour; annealing; microstructure polymers polymers Evaluation of Thermal and Thermomechanical Behaviour of Bio-Based Polyamide 11 Based Composites Reinforced with Lignocellulosic Fibres Helena Oliver-Ortega 1 ID , José Alberto Méndez 1, Pere Mutjé 1, Quim Tarrés 1,* ID , Francesc Xavier Espinach 2 and Mònica Ardanuy 3 ID 1 Group LEPAMAP, Department of Chemical Engineering, University of Girona, C/M.Aurèlia Capmany, 61, 17003 Girona, Spain; helena.oliver@udg.edu (H.O.-O.); jalberto.mendez@udg.edu (J.A.M.); pere.mutje@udg.edu (P.M.) 1 Group LEPAMAP, Department of Chemical Engineering, University of Girona, C/M.Aurèlia Capmany, 61, 17003 Girona, Spain; helena.oliver@udg.edu (H.O.-O.); jalberto.mendez@udg.edu (J.A.M.); pere.mutje@udg.edu (P.M.) p j g 2 Design, Development and Product Innovation, Dpt. Organization, Business Management and Product Design, University of Girona, C/M.Aurèlia Capmany, 61, 17003 Girona, Spain; francisco.espinach@udg.edu 2 Design, Development and Product Innovation, Dpt. Organization, Business Management and Product Design, University of Girona, C/M.Aurèlia Capmany, 61, 17003 Girona, Spain; francisco.espinach@udg 2 Design, Development and Product Innovation, Dpt. Organization, Business Management and Product Design, University of Girona, C/M.Aurèlia Capmany, 61, 17003 Girona, Spain; francisco.espinach@udg.edu 3 Departament de Ciència dels Materials i Enginyeria Metal lúrgica Secció Enginyeria Tèxtil g , y , / p y, , , p ; p g 3 Departament de Ciència dels Materials i Enginyeria Metal.lúrgica, Secció Enginyeria Tèxtil, Universitat Politècnica de Catalunya, C/Colom, 11, 08222 Terrassa, Barcelona, Spain; monica.ardanuy@upc.edu y p * Correspondence: joaquimagusti.tarres@udg.edu; Tel.: +34-690-754-563 y p * Correspondence: joaquimagusti.tarres@udg.edu; Tel.: +34-690-754-563 Received: 14 September 2017; Accepted: 16 October 2017; Published: 18 October 2017 1. Introduction Composites are produced to obtain new materials, designed to be used for a speciﬁc applications, with comparatively better properties than its phases [1]. In this sense, the literature shows that polymeric matrices with comparatively poor mechanical properties, such as polyoleﬁn, have been successfully reinforced to obtain competitive materials. One example of these composites are glass ﬁbre reinforced polypropylene (PP) materials, currently produced and used at industrial level [2]. Nonetheless, despite the high mechanical performance of these composites, there are health problems associated with the manipulation of glass ﬁbres [3] and its poor recyclability. This justiﬁes the search for more healthy and environmentally friend substitutive reinforcements. Moreover, the UE has ﬁxed for 2025–2030 some recyclability goals [4] which are not possible to achieve using glass ﬁbres as reinforcements. Cellulosic or lignocellulosic reinforcing ﬁbres have become a recyclable alternative to glass ﬁbres. These ﬁbres have been extensively studied since the 1980s showing good intrinsic properties Polymers 2017, 9, 522; doi:10.3390/polym9100522 www.mdpi.com/journal/polymers Polymers 2017, 9, 522 2 of 17 such as comparatively high tensile strengths [5]. Moreover, cellulosic ﬁbres are fully bio-based and biodegradable [6]. Besides, the use of natural ﬁbres accomplishes objectives proposed by green chemistry and sustainable production [7,8]. In addition, their use and manipulation is healthier. Since the 1980s, the environmental awareness of the society has increased noticeably. The use of non-renewable resources and the high impact of oil-based polymers on the environment drive the research of more environmentally friendly alternatives to such resources. One of the proposed challenges is reducing the oil-based polymer dependence and replacing these polymers by bio-based or biodegradable ones [9], with lower environmental ﬁngerprints [10]. Besides, the use of biodegradable polymers can also promote waste reduction, diminishing its environmental impact [11]. Therefore, the composite materials researchers have increased their interest on these materials as possible replacements to oil-based matrices [12]. Despite all the aforementioned advantages, cellulosic ﬁbres have low thermal degradation temperatures, being this its main drawback that limits their use as reinforcement for polymer-based composites [12]. The degradation temperatures of cellulosic ﬁbres are between 200 ◦C and 250 ◦C, and, therefore, the matrices must be able to be processed at such temperatures. Polyamide 11 (PA11), also called nylon 11, is a bio-based polyamide obtained from castor oil. Moreover, it is a non-biodegradable polymer matrix allowing its use for long-time applications, such as those needed in the construction or automotive sectors [12,13]. 1. Introduction Besides, its melting temperature, around 190 ◦C, is low enough to avoid the thermal degradation of cellulosic ﬁbres during composites manufacturing. Furthermore, PA11 is a recyclable thermoplastic matrix with mechanical properties similar to polypropylene. Thus, PA11 is a promising green alternative to polypropylene [14,15]. There are some studies in the about different nanomaterial reinforced PA11 nanocomposites [16,17]. More recently, there are works devoted to cellulosic fibres reinforced PA11 composites [15,18,19]. The analysis of the tensile and ﬂexural properties of these cellulosic composites showed signiﬁcant increases of its modulus and strength together with low reductions of its strains at break [15,19], compared to other ﬁbre, like glass ﬁbres (GF), based composites [14]. However, although the thermal properties and the structure of PA11-based composites reinforced with several micro- or nano-reinforcements have been extensively studied, there are only few studies about the effect of cellulosic ﬁbres in a PA11 matrix [15]. In this study, Stone Groundwood (SGW) ﬁbres from softwood were used as reinforcement for PA11 biocomposites. SGW ﬁbres were obtained by high yield mechanical processes (98%) and, therefore, the chemical composition of such ﬁbres and the wood from which they come were the same [19]. Moreover, SGW is produced in a sustainable way due to its use in the papermaking industry. The effect of these lignocellulosic ﬁbres in the thermal stability, thermal transitions and thermomechanical behaviour of PA11 composites were analysed. In addition, an annealing treatment was applied to the ﬁbres to study the impact of such treatment in the thermal properties of the composites. 2.4. Composite Characterization Data were collected on the 2θ range from 5◦ p q y p p g −40 ◦C to 120 ◦C with a heating rate of 3 ◦C/min and the analysis was performed in air atmosphere. X-ray diffraction measurements were performed on a Bruker D8 Advance diffractometer (Bruker, −40 ◦C to 120 ◦C with a heating rate of 3 ◦C/min and the analysis was performed in air atmo X-ray diffraction measurements were performed on a Bruker D8 Advance diffractometer (Bruker, Madrid, Spain) with a Cu-Kα radiation (λ = 0.15406 nm). Data were collected on the 2θ range from 5◦ to 40◦operating at 40 KV and 40 mA. A Fourier Transformed infrared spectroscopy (FT-IR) using a Bruker Alpha FT-IR spectrometer (FT-IR) (Bruker, Madrid, Spain) was performed in the PA11 and PA11 composites. 2.3. Annealing Treatment The annealing treatment, based previous publications, was performed following the conditions that led to a maximum crystallinity enhancement [20,21]. This treatment consisted of heating the matrix or the composites in an oven at 165 ◦C for 1 h followed by cooling them at room temperature (23 ◦C). 2.1. Materials Rilsan®BMNO TLD polyamide 11, kindly supplied by Arkema S.A. (Colombes, France), with a density of 1.03 g/cm3 and a melt ﬂow index (MFI) of 11 cc/10 min measured at 235 ◦C/2.16 kg, was used as polymer matrix. Stone Groundwood (SGW), a mechanically deﬁbrated pulp from softwood (Pinus radiata), provided by Zubialde S.A. (Aizarnazabal, Spain), was used as reinforcement. The length and diameter distributions of SGW ﬁbres were studied in a previous work [19]. In the mentioned study, the density of ﬁbre was determined to be 1.40 g/cm3. Polymers 2017, 9, 522 3 of 17 2.2. Composite Compounding Composites reinforced with 10%, 20% and 50% w/w ﬁbre content were produced in a Gelimat Kinetic Mixer (model G5S, Draiswerke, Mahaw, New Jersey, NJ, USA). The ﬁbres and the polymer matrix were ﬁrstly added and premixed at low speed (300 rpm) and then the speed was increased up to 2500 rpm. When the mixture reached 200 ◦C, the material was discharged, cooled and pelletized with a knives mill. To produce the specimens, the composites were mould injected in a Meteor-40 injection machine (Mateu & Solé, Barcelona, Spain). The processing temperature proﬁle was 170–185–200 ◦C and the pressures were modiﬁed regarding the ﬁbre content to a maximum of 75 bars for the volumetric phase and 30 bars for the pressure maintenance phase. The samples were stored in a climatic chamber at 23 ◦C and 50% RH before their analysis, according to ASTM D618 standard speciﬁcations. 2.4. Composite Characterization Thermogravimetrical analysis (TGA) was performed in a Mettler Toledo SDTA 851 thermobalance (Mettler Toledo, L’Hospitalet de Llobregat, Spain). Samples were heated from 30 to 700 ◦C at a heating rate of 10 ◦C/min under nitrogen atmosphere at a ﬂow rate of 40 mL/min. Differential Scanning Calorimetry analysis (DSC) was performed using a Mettler Toledo DSC822e calorimeter (Mettler Toledo, L’Hospitalet de Llobregat, Spain) following ASTM E 1269.01 standard speciﬁcation. The samples were initially heated from 40 to 210 ◦C to erase their thermal history. Afterwards, the samples were cooled and heated again using the same temperature range. All runs were performed at heating or cooling rates of 10 ◦C/min under 40 mL/min ﬂow of nitrogen atmosphere. g p Dynamic mechanical thermal analysis (DMTA) was carried out in a Mettler Toledo DMA/SDTA 861 (Mettler Toledo, L’Hospitalet de Llobregat, Spain) using dual cantilever conﬁguration. Specimens of 65 × 13 × 3 mm3 were cut from the ﬂexural specimens obtained following the ASTM D3641. The tests were performed at a frequency of 1 Hz and a preload of 3 N. The temperature range was −40 ◦C to 120 ◦C with a heating rate of 3 ◦C/min and the analysis was performed in air atmosphere. X-ray diffraction measurements were performed on a Bruker D8 Advance diffractometer (Bruker, Madrid, Spain) with a Cu-Kα radiation (λ = 0.15406 nm). Data were collected on the 2θ range from 5◦ to 40◦operating at 40 KV and 40 mA. Dynamic mechanical thermal analysis (DMTA) was carried out in a Mettler Toledo DMA/SDTA 861 (Mettler Toledo, L’Hospitalet de Llobregat, Spain) using dual cantilever conﬁguration. Specimens of 65 × 13 × 3 mm3 were cut from the ﬂexural specimens obtained following the ASTM D3641. The tests were performed at a frequency of 1 Hz and a preload of 3 N. The temperature range was of 65 × 13 × 3 mm were cut from the ﬂexural specimens obtained following the ASTM D3641. The tests were performed at a frequency of 1 Hz and a preload of 3 N. The temperature range was −40 ◦C to 120 ◦C with a heating rate of 3 ◦C/min and the analysis was performed in air atmosphere. X-ray diffraction measurements were performed on a Bruker D8 Advance diffractometer (Bruker, Madrid, Spain) with a Cu-Kα radiation (λ = 0.15406 nm). 3.1. Thermal Stability of the Composites Temperatures (°C) PA11 PA11 + 20% SGW PA11 + 50% SGW T5% 409 322 307 T10% 417 386 336 T 439 451 461 Temperatures (◦C) PA11 PA11 + 20% SGW PA11 + 50% SGW T5% 409 322 307 T10% 417 386 336 Tmax 439 451 461 Although the degradation of PA11-SGW composites seemed to start around 200 °C, the weight loss did not surpass 5% until 300 °C. This first decomposition step was related to the reinforcement degradation, and gained importance when the fibre contents were increased. During this step, the degradation of the O-glucosidic bonds of the cellulose and hemicelluloses occurred [22,23]. The degradation of the other component of the fibres, lignin and extractives occurred in a broader temperature range from around 200 °C to 900 °C [24,25]. SGW fibres showed high lignin contents [19,26], thus, a high degradation range was expected. It was considered that the slight differences obtained in the fibres length, produced by attrition phenomena during composite preparation of the fibres [14,19], had little influence in its decomposition. Moreover, it was possible to measure the onset temperature for the 95% weight loss (T95%) for the matrix (560 °C) but it was impossible for the composites. An inflection point in the curve, where a second degradation step started, was observed at around 350 °C and 375 °C for PA11 + 20% SGW and PA11 + 50% SGW, respectively. This was Although the degradation of PA11-SGW composites seemed to start around 200 ◦C, the weight loss did not surpass 5% until 300 ◦C. This ﬁrst decomposition step was related to the reinforcement degradation, and gained importance when the ﬁbre contents were increased. During this step, the degradation of the O-glucosidic bonds of the cellulose and hemicelluloses occurred [22,23]. The degradation of the other component of the ﬁbres, lignin and extractives occurred in a broader temperature range from around 200 ◦C to 900 ◦C [24,25]. SGW ﬁbres showed high lignin contents [19,26], thus, a high degradation range was expected. It was considered that the slight differences obtained in the ﬁbres length, produced by attrition phenomena during composite preparation of the ﬁbres [14,19], had little inﬂuence in its decomposition. Moreover, it was possible to measure the onset temperature for the 95% weight loss (T95%) for the matrix (560 ◦C) but it was impossible for the composites. 3.1. Thermal Stability of the Composites As mentioned in the Introduction, the comparatively low degradation temperature of natural ﬁbres limits its use with a broad set of polymeric matrices. The matrices must have melting temperatures in the order of 200 ◦C. Moreover, the ﬁbre degradation could have a negative effect on the thermal degradation of the polymer matrix. The TGA proﬁles and the ﬁrst derivate of the PA11 and the composites reinforced with a 20% and 50% w/w of SGW are shown in Figure 1. Polymers 2017, 9, 522 Polymers 2017 9 522 4 of 17 4 of 18 Figure 1. TGA curves and first derivate of the TGA curve for neat PA11 and PA11-SGW composites. Figure 1. TGA curves and ﬁrst derivate of the TGA curve for neat PA11 and PA11-SGW composites. Figure 1. TGA curves and first derivate of the TGA curve for neat PA11 and PA11-SGW composites. Figure 1. TGA curves and ﬁrst derivate of the TGA curve for neat PA11 and PA11-SGW composites. As shown in Figure 1, neat PA11 presents a TGA trend with one main decomposition step starting around 400 °C. PA11 composites started degrading before, and presented two main decomposition steps, as was expected for cellulose reinforced composite materials. Table 1 shows the onset temperatures for the 5% and 10% of weight loss (T5% and T10%) in PA11 and PA11-SGW composites As shown in Figure 1, neat PA11 presents a TGA trend with one main decomposition step starting around 400 ◦C. PA11 composites started degrading before, and presented two main decomposition steps, as was expected for cellulose reinforced composite materials. Table 1 shows the onset temperatures for the 5% and 10% of weight loss (T5% and T10%) in PA11 and PA11-SGW composites. o posi es Table 1. Onset temperatures for the 5% and 10% of weight loss and temperature in the maximum Table 1. Onset temperatures for the 5% and 10% of weight loss and temperature in the maximum decomposition rate. posi es Table 1. Onset temperatures for the 5% and 10% of weight loss and temperature in the maximum Table 1. Onset temperatures for the 5% and 10% of weight loss and temperature in the maximum decomposition rate. mposition rate. 3.1. Thermal Stability of the Composites An inﬂection point in the curve, where a second degradation step started, was observed at around 350 ◦C and 375 ◦C for PA11 + 20% SGW and PA11 + 50% SGW, respectively. This was related with the degradation of the polymeric phase. This inﬂection point in 5 of 17 5 of 18 Polymers 2017, 9, 522 Polymers 2017, 9, 522 the curve indicated that the matrix started degrading before the ﬁbre degradation was ﬁnished [22]. The ﬁrst derivative of the TGA curve was performed to evaluate this effect (Figure 1). As expected, only one peak was obtained for the PA11 matrix, while two peaks, corresponding with the degradation of the ﬁbres and polymer matrix, were found in PA11-SGW composites. Moreover, these peaks appeared overlapped indicating that there was a range of temperatures where ﬁbres and matrix degraded simultaneously. related with the degradation of the polymeric phase. This inflection point in the curve indicated that the matrix started degrading before the fibre degradation was finished [22]. The first derivative of the TGA curve was performed to evaluate this effect (Figure 1). As expected, only one peak was obtained for the PA11 matrix, while two peaks, corresponding with the degradation of the fibres and polymer matrix, were found in PA11-SGW composites. Moreover, these peaks appeared overlapped indicating that there was a range of temperatures where fibres and matrix degraded simultaneously. On the other hand a shift of the maximum temperature of the decomposition step (T ) of the On the other hand, a shift of the maximum temperature of the decomposition step (Tmax) of the polymer from 439 ◦C for pure PA11 to 451 ◦C and 461 ◦C for PA11 + 20% SGW and PA11 + 50% SGW composites, respectively, was observed. From these results, it was concluded that, although the cellulose ﬁbres had a negative effect on the onset decomposition temperature of the composites, their presence contributed to thermally stabilize the composite once the degradation started. The literature shows similar thermal stabilizations in the case of cellulose nanoﬁber (CNF) reinforced PA11 [27] and other thermoplastic matrices reinforced with lignocellulosic ﬁbres [28]. This phenomenon has been explained as an inhibiting effect of the char obtained from the ﬁbres decomposition to the diffusion of volatile and radical compounds implied in PA11 decomposition. 3.1. Thermal Stability of the Composites On the other hand, a shift of the maximum temperature of the decomposition step (Tmax) of the polymer from 439 °C for pure PA11 to 451 °C and 461 °C for PA11 + 20% SGW and PA11 + 50% SGW composites, respectively, was observed. From these results, it was concluded that, although the cellulose fibres had a negative effect on the onset decomposition temperature of the composites, their presence contributed to thermally stabilize the composite once the degradation started. The literature shows similar thermal stabilizations in the case of cellulose nanofiber (CNF) reinforced PA11 [27] and other thermoplastic matrices reinforced with lignocellulosic fibres [28]. This phenomenon has been explained as an inhibiting effect of the char obtained from the fibres decomposition to the diffusion of volatile and radical compounds implied in PA11 decomposition. Another important effect of the fibre addition was related to the residue found at 700 °C It was Another important effect of the ﬁbre addition was related to the residue found at 700 ◦C. It was observed that the addition of ﬁbre enhanced its content from 3.4%, obtained in monolithic PA11, up to 20.5% when 50% SGW ﬁbres were added to the composite material. This increase on the residue was related with the extractives and inorganic compounds contents in the lignocellulosic reinforcement, and lignin molecules which did not totally degrade [24,25,29]. Finally, from these TGA curves, it was observed that the degradation of the ﬁbres occurred at higher temperatures than the processing temperatures of the corresponding composites, allowing the preparation of these composites. Another important effect of the fibre addition was related to the residue found at 700 C. It was observed that the addition of fibre enhanced its content from 3.4%, obtained in monolithic PA11, up to 20.5% when 50% SGW fibres were added to the composite material. This increase on the residue was related with the extractives and inorganic compounds contents in the lignocellulosic reinforcement, and lignin molecules which did not totally degrade [24,25,29]. Finally, from these TGA curves, it was observed that the degradation of the fibres occurred at higher temperatures than the processing temperatures of the corresponding composites, allowing the preparation of these composites. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites The melting and crystallization behaviours of the polymer matrix and the composites were studied by means of DSC. The thermographs obtained during the second heating, after erasing the thermal history, are shown in Figure 2. 3.2. Thermal Transitions and Structure of the Composites The melting and crystallization behaviours of the polymer matrix and the composites were studied by means of DSC. The thermographs obtained during the second heating, after erasing the thermal history are shown in Figure 2 Figure 2. DSC thermographs of the melting point of PA11 and PA11 composites. Figure 2. DSC thermographs of the melting point of PA11 and PA11 composites. Figure 2. DSC thermographs of the melting point of PA11 and PA11 composites. Figure 2. DSC thermographs of the melting point of PA11 and PA11 composites. It was observed that PA11 presented a main peak at around 189 °C, preceded by a shoulder with a maximum at around 181 °C as a rearrangement in the structure [30]. PA11 has different crystalline structures which can be transformed from one to another depending on temperature, It was observed that PA11 presented a main peak at around 189 ◦C, preceded by a shoulder with a maximum at around 181 ◦C as a rearrangement in the structure [30]. PA11 has different crystalline structures which can be transformed from one to another depending on temperature, 6 of 17 Polymers 2017, 9, 522 cooling conditions and pressure, among others [31–33]. This shoulder peak corresponded to the melt-crystallization process of the γ phase to the α’ crystalline form of PA11 [30,34]. The secondary peak seemed to decrease and the main peak became broader when the ﬁbre content was increased in the composite material. This effect has also been found when CNF were used as reinforcement, indicating that the ﬁbres promoted the α’ crystalline form instead of the γ form observed in the neat PA11 [27,34]. On the other hand, it was found that the melting temperature (Tm) was not affected by SGW presence (189 ◦C for all the studied materials). p The degree of crystallinity of the polymer matrix was calculated as the ratio between the enthalpy of main melting endotherm and the theoretical enthalpy for a fully crystalline polymer matrix (∆ˆH m = 226.4 J/g) [27,35]. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites The obtained degree of crystallinities for the composites materials were 26.4%, 26.4% and 27.2% for PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW, respectively. These values were very similar to the pure matrix (26.7%), indicating no signiﬁcant effect of the ﬁbres on the degree of crystallinity of the PA11 matrix. In the literature, a slight increase (around 3%) in the degree of crystallinity for CNF reinforced PA11 composites with contents lower than 5% has been reported [27]. However, for higher contents of CNF, the crystallinity decreased. This could be related with a disruption of the PA11 structure by the effect of the CNF [27]. Moreover, the chemical surface’s composition of SGW ﬁbres is different to CNF studied in the literature, mainly due to the presence of lignin in the ﬁbre surface. This could inhibit the nucleating effect of the ﬁbres, as it also inhibited the interactions between the polymer matrix and the ﬁbre [15,19,29]. On the other hand, cellulose nanocrystals (CNC) and modiﬁed CNC also obtained no signiﬁcant changes in crystallinity when they were used as reinforcement in PA11 [36]. This is the opposite effect to that observed in other polymer matrices, where the cellulosic ﬁbres acted as nucleating agents, increasing the polymer crystallinity [22,37,38]. Nonetheless, these matrices had weaker interactions with cellulose ﬁbres than PA11. The capacity of this polymer to establish H-bonds with cellulose [19] can explain this disruption in the polymer matrix, and hence no considerable nucleating effect was observed. Moreover, this capacity can inhibit the formation of the γ crystalline phase on the composites. Concerning the crystallization behaviour, a peak with a maximum crystallization temperature (Tc) at 164 ◦C was observed for the PA11 and the composites. The crystallinity of PA11 matrix was also measured as the ratio between the enthalpy of the polymer crystallization and the theoretical value of the crystalline polymer matrix. The obtained values were not different from those obtained during the melting except for the PA11 + 10% SGW composite, where a lower degree of crystallinity (22.8%) was measured. The results allowed concluding that, as found in the literature for other cellulose reinforced PA11 and PP composites, the presence of SGW ﬁbres did not affect the main transition temperatures of the crystalline polymer phase [22,33,37]. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites The measured values of the Tg obtained from the tan δ curve shifted from 53.1 ◦C for neat PA11 ◦C to 50.0 ◦C, 51.0 ◦C and 53.2 ◦C for the PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW, respectively. No considerable differences were observed in the Tg of the composites by the effect of SGW content. This was related with the same crystallinity values of the materials. Thus, no changes were observed in the amorphous phase of the PA11 matrix in the composite materials. It was observed that tan δ decreased with SGW contents as a consequence of the enhancement of the loss moduli. As expected, higher values of storage moduli were obtained for the composite materials due to the stiffening effect of the reinforcing material [39,40]. For all materials, a slight decrease was observed when the temperature was raised in the analysis below 20 °C, as a result of a slight mobility gain in the polymer chain, and a drastic drop was observed when the Tg was overpass. Once the temperature was over the Tg, the storage modulus values were really low, indicating a high mobility of the polymer molecules corresponding to the amorphous phase of the PA11. However, the presence of SGW fibres stiffened the material, achieving higher values of storage modulus due to the higher stiffness of cellulosic fibres and counteracting the reduction of the modulus when the Tg was exceeded. It must be noted that the presence of lignocellulosic fibres in the composite material significantly reduced the mobility of the polymer chains for temperatures higher than its Tg. The influence of SGW fibres is clearly observed in Figure 4, where the modulus of PA11 + 50% SGW at 80 °C is 10 times higher than the PA11 modulus at the same temperature and slightly higher than the PA11 matrix at 20 °C. As expected, higher values of storage moduli were obtained for the composite materials due to the stiffening effect of the reinforcing material [39,40]. For all materials, a slight decrease was observed when the temperature was raised in the analysis below 20 ◦C, as a result of a slight mobility gain in the polymer chain, and a drastic drop was observed when the Tg was overpass. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites The DMTA thermograms of the PA11 composite materials was performed to observe other processes in which the material loss energy take place, as well as to understand the behaviour of the stiffness of the materials with the temperature. In this sense, a softening is usually experimented when thermoplastic materials overpass the glass transition (Tg) as the amorphous chains of the polymer suffers an important gain of mobility. As shown in Figure 3, the evolution of the loss modulus (E”) of pure PA11 and the composites with respect to the temperature showed a unique transition in the studied range, related with the Tg of the polymer matrix. Polymers 2017, 9, 522 Polymers 2017 9 522 7 of 17 7 of 18 Figure 3. Loss Modulus of the PA11, PA11 + 20% SGW and PA11 + 50% SGW with respect to the temperature. Figure 3. Loss Modulus of the PA11, PA11 + 20% SGW and PA11 + 50% SGW with respect to the temperature. Figure 3. Loss Modulus of the PA11, PA11 + 20% SGW and PA11 + 50% SGW with respect to the temperature. Figure 3. Loss Modulus of the PA11, PA11 + 20% SGW and PA11 + 50% SGW with respect to the temperature. Figure 4 represents the evolution of the values of storage modulus (E’) and tan δ of PA11 and the composites with the temperature. The measured values of the Tg obtained from the tan δ curve shifted from 53.1 °C for neat PA11 °C to 50.0 °C, 51.0 °C and 53.2 °C for the PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW, respectively. No considerable differences were observed in the Tg of the composites by the effect of SGW content. This was related with the same crystallinity values of the materials. Thus, no changes were observed in the amorphous phase of the PA11 matrix in the composite materials. It was observed that tan δ decreased with SGW contents as a consequence of the enhancement of the loss moduli. As e pected highe alues of sto age oduli e e obtai ed fo the co posite ate ials due to Figure 4 represents the evolution of the values of storage modulus (E’) and tan δ of PA11 and the composites with the temperature. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites Once the temperature was over the Tg, the storage modulus values were really low, indicating a high mobility of the polymer molecules corresponding to the amorphous phase of the PA11. However, the presence of SGW ﬁbres stiffened the material, achieving higher values of storage modulus due to the higher stiffness of cellulosic ﬁbres and counteracting the reduction of the modulus when the Tg was exceeded. It must be noted that the presence of lignocellulosic ﬁbres in the composite material signiﬁcantly reduced the mobility of the polymer chains for temperatures higher than its Tg. The inﬂuence of SGW ﬁbres is clearly observed in Figure 4, where the modulus of PA11 + 50% SGW at 80 ◦C is 10 times higher than the PA11 modulus at the same temperature and slightly higher than the PA11 matrix at 20 ◦C. 8 of 17 Polymers 2017, 9, 522 Figure 4. Storage modulus and tan δ results of PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW. 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 -40 -20 0 20 40 60 80 100 120 tan δ E' (MPa) Temperature (ºC) PA11 PA11+10%SGW PA11+20%SGW PA11+50%SGW Figure 4. Storage modulus and tan δ results of PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW. Figure 4. Storage modulus and tan δ results of PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW. Figure 4. Storage modulus and tan δ results of PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW. The crystalline structure of PA11 and PA11-SGW composites was analysed using an X-ray diffractometer (Figure 5). It is known that PA11 shows polymorphism that highly influences its properties [41,42]. The DSC study showed two different structures during the second melting (γ and α’ forms), for PA11 and the composite with 10% of SGW. However, it must be noted that these structures were obtained after a melting process and controlled crystallization with a cooling rate of 10 °C/min. The obtained samples produced by injection-moulding were not cooled under the same conditions and can have a different structure. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites X-ray diffractograms of PA11 and PA11-SGW composites. Figure 5. X-ray diffractograms of PA11 and PA11-SGW composites. Figure 5. X-ray diffractograms of PA11 and PA11-SGW composites. The broad peak of PA11 and composites related with the PA11 δ’ crystalline form was similar to that obtained in the literature for this PA11 structure [17]. Despite this, two more peaks at 22.3° and 15° of 2θ appeared for PA11 + 20% SGW and PA11 + 50% SGW composites and were related with cellulose [29]. These peaks can only be appreciated in the composites with higher fibre contents due to the low crystallinity of the fibre: 48.5% measured with the X-ray diffractometer and calculated as the ratio between the intensity at the peak at 22°–23° and the minimum at 15°–18°, as has been described in the literature [45]. This low crystallinity of the SGW fibres is in agreement with the values obtained for untreated pine fibres [29]. These results differ from those found on the literature for PA11 nanocomposites, where, usually, the α’ structure is observed [21,36,44]. However, the reinforcement content with respect to the matrix for these nanocomposites was significantly lower. Moreover, the formation of the different phases is strongly influenced by the cooling process, which could be different than the performed in this work [35,44]. h l l h h h d ff l d f h d ff The broad peak of PA11 and composites related with the PA11 δ’ crystalline form was similar to that obtained in the literature for this PA11 structure [17]. Despite this, two more peaks at 22.3◦ and 15◦of 2θ appeared for PA11 + 20% SGW and PA11 + 50% SGW composites and were related with cellulose [29]. These peaks can only be appreciated in the composites with higher ﬁbre contents due to the low crystallinity of the ﬁbre: 48.5% measured with the X-ray diffractometer and calculated as the ratio between the intensity at the peak at 22◦–23◦and the minimum at 15◦–18◦, as has been described in the literature [45]. This low crystallinity of the SGW ﬁbres is in agreement with the values obtained for untreated pine ﬁbres [29]. These results differ from those found on the literature for PA11 nanocomposites, where, usually, the α’ structure is observed [21,36,44]. However, the reinforcement content with respect to the matrix for these nanocomposites was signiﬁcantly lower. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites Moreover, the formation of the different phases is strongly inﬂuenced by the cooling process, which could be different than the performed in this work [35,44]. Nonetheless, although γ phase is difficult to identify in the diffractograms at room temperature, FT-IR can serve as a complementary technique to determine the presence of this crystalline phase [39,45]. Figure 6 shows the FT-IR normalized for PA11 and PA11-SGW composites. There are some differences in the main bands observed regarding the crystal phase. Nevertheless, usually the difference in the wavelength is really small. One of the most differentiated bands in the FT-IR profiles was in the fingerprint zone of the FT-IR spectrum. The bands comprised in the range 500– 800 cm−1 were related with amide V and VI amide bands (marked in the Figure 6). When the γ phase was present, a shoulder peak was observed in the 721 cm−1 peak, while in other forms a double peak at 721 and 686 cm−1 was shown. In PA11 and PA11 + 10% SGW a broad peak was observed while a double peak 721 and 681 cm−1 was appreciated in 20% and 50% of SGW fibre reinforced composites. Moreover, a small peak was observed around 627 cm−1 in the FT-IR of the PA11 and PA11 + 10% p Nonetheless, although γ phase is difﬁcult to identify in the diffractograms at room temperature, FT-IR can serve as a complementary technique to determine the presence of this crystalline phase [39,45]. Figure 6 shows the FT-IR normalized for PA11 and PA11-SGW composites. There are some differences in the main bands observed regarding the crystal phase. Nevertheless, usually the difference in the wavelength is really small. One of the most differentiated bands in the FT-IR proﬁles was in the ﬁngerprint zone of the FT-IR spectrum. The bands comprised in the range 500–800 cm−1 were related with amide V and VI amide bands (marked in the Figure 6). When the γ phase was present, a shoulder peak was observed in the 721 cm−1 peak, while in other forms a double peak at 721 and 686 cm−1 was shown. In PA11 and PA11 + 10% SGW a broad peak was observed while a double peak 721 and 681 cm−1 was appreciated in 20% and 50% of SGW ﬁbre reinforced composites. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites As can be seen in Figure 5, the samples after injection-moulding process showed a broad peak at 2θ = 21°, corresponding to the δ’ phase produced from quenching from the melt [35,43,44], which is similar to the process produced during injection-moulding. The controlled crystallization of the DSC led to obtaining the α’ crystalline form that was produced by melt crystallization [44]. Moreover, a small content of γ phase is typically obtained during this analysis and is transformed to the α’ more stable form [21]. Nonetheless, the γ phase or was impossible to be observed at room temperature [43] or its content was low and was difficult to identify. Thus, its formation simultaneously with the δ’ phase was not observed in the X-ray diffractograms. The crystalline structure of PA11 and PA11-SGW composites was analysed using an X-ray diffractometer (Figure 5). It is known that PA11 shows polymorphism that highly inﬂuences its properties [41,42]. The DSC study showed two different structures during the second melting (γ and α’ forms), for PA11 and the composite with 10% of SGW. However, it must be noted that these structures were obtained after a melting process and controlled crystallization with a cooling rate of 10 ◦C/min. The obtained samples produced by injection-moulding were not cooled under the same conditions and can have a different structure. As can be seen in Figure 5, the samples after injection-moulding process showed a broad peak at 2θ = 21◦, corresponding to the δ’ phase produced from quenching from the melt [35,43,44], which is similar to the process produced during injection-moulding. The controlled crystallization of the DSC led to obtaining the α’ crystalline form that was produced by melt crystallization [44]. Moreover, a small content of γ phase is typically obtained during this analysis and is transformed to the α’ more stable form [21]. Nonetheless, the γ phase or was impossible to be observed at room temperature [43] or its content was low and was difﬁcult to identify. Thus, its formation simultaneously with the δ’ phase was not observed in the X-ray diffractograms. Polymers 2017, 9, 522 Polymers 2017 9 522 9 of 17 9 of 18 9 of 17 9 of 18 Figure 5. X-ray diffractograms of PA11 and PA11-SGW composites. 5 10 15 20 25 30 35 40 Intensity 2θ (º) PA11 PA11+10%SGW PA11+20%SGW PA11+50%SGW δ' PA11 21° Cellulose 23° Cellulose 15° Figure 5. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions Thermal annealing treatment is used in polymers to obtain a specific crystallinity form, to increase their crystallinity reducing the softening effect when the Tg is overpassed or as a useful form to remove residual stresses, which could appear during the extrusion, injection or other processing procedures [17,42,46]. Thermal annealing has growing importance in PA11 thermal studies because of its piezoelectric and ferroelectric properties. However, these studies usually involved the use of nanomaterials in the case of composite materials [21,34]. Thermal annealing treatment is used in polymers to obtain a speciﬁc crystallinity form, to increase their crystallinity reducing the softening effect when the Tg is overpassed or as a useful form to remove residual stresses, which could appear during the extrusion, injection or other processing procedures [17,42,46]. Thermal annealing has growing importance in PA11 thermal studies because of its piezoelectric and ferroelectric properties. However, these studies usually involved the use of nanomaterials in the case of composite materials [21,34]. In this case, the objective of the annealing was to enhance the crystallinity of the samples to study the influence of SGW fibres during this process. As mentioned above, no considerable increase of the crystallinity was obtained in the DSC of PA11-SGW composites while a nucleating effect of the fibre was observed for PP-SGW composites [23]. As mentioned above, using CNF at low reinforcement contents increased the crystallinity of PA11. Otherwise, high contents of fibre seemed to inhibit the crystalline production, probably due to the H-bonds established between the fibres and the matrix. Crystalline structures in PA11 have a large dependence of the H-bond orientation [32]. In the literature, it can be found that, after an annealing process, the PA11 structure shows higher crystallinity [17,35]. However, the fibres and their capacity to interact with PA11 can impact this In this case, the objective of the annealing was to enhance the crystallinity of the samples to study the inﬂuence of SGW ﬁbres during this process. As mentioned above, no considerable increase of the crystallinity was obtained in the DSC of PA11-SGW composites while a nucleating effect of the ﬁbre was observed for PP-SGW composites [23]. As mentioned above, using CNF at low reinforcement contents increased the crystallinity of PA11. Otherwise, high contents of ﬁbre seemed to inhibit the crystalline production, probably due to the H-bonds established between the ﬁbres and the matrix. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions Crystalline structures in PA11 have a large dependence of the H-bond orientation [32]. In the literature, it can be found that, after an annealing process, the PA11 structure shows higher crystallinity [17,35]. However, the ﬁbres and their capacity to interact with PA11 can impact this process. process. As explained before, the PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW p y p p As explained before, the PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW samples were annealed at 165 ◦C for 1 h and studied using DSC, DMTA and X-Ray diffraction. process. As explained before, the PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW As explained before, the PA11, PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW samples were annealed at 165 ◦C for 1 h and studied using DSC, DMTA and X-Ray diffraction. samples were annealed at 165 °C for 1 h and studied using DSC, DMTA and X-Ray diffraction. The DSC results showed an increase of the degree of the crystallinity in the first heat by the effect of the annealing, which increased considerably when the fibre content was augmented. In particular, the crystallinity of PA11 after the annealing treatment increased from 26.7% to 28.9%, and, in the composites, from 26.4%, 26.4% and 27.2% to 27.5%, 30.7% and 40.5% for PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW, respectively. It seems that SGW fibres can act as nucleating agent during the annealing process. Trans-crystallinity process between fibres and matrix are highly dependent on the characteristics of the fibre, matrix and their interactions [47]. PA11 The DSC results showed an increase of the degree of the crystallinity in the ﬁrst heat by the effect of the annealing, which increased considerably when the ﬁbre content was augmented. In particular, the crystallinity of PA11 after the annealing treatment increased from 26.7% to 28.9%, and, in the composites, from 26.4%, 26.4% and 27.2% to 27.5%, 30.7% and 40.5% for PA11 + 10% SGW, PA11 + 20% SGW and PA11 + 50% SGW, respectively. It seems that SGW ﬁbres can act as nucleating agent during the annealing process. Trans-crystallinity process between ﬁbres and matrix are highly dependent on the characteristics of the ﬁbre, matrix and their interactions [47]. 3.2. Thermal Transitions and Structure of the Composites 3 2 Thermal Transitions and Structure of the Composites Moreover, a small peak was observed around 627 cm−1 in the FT-IR of the PA11 and PA11 + 10% SGW composites and the band of 581 shifted to higher wavelengths, usually found in in the γ phase [45]. Polymers 2017, 9, 522 SGW composites 10 of 17 γ phase Figure 6. FT-IR PA11 and PA11-SGW composites. Figure 6. FT-IR PA11 and PA11-SGW composites. Figure 6. FT-IR PA11 and PA11-SGW composites. Figure 6. FT-IR PA11 and PA11-SGW composites. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions PA11 requires higher time and temperature than PP due to the high intermolecular interactions, which can be produced between PA11 and ﬁbres, explaining the increments produced during the annealing. The highest crystallinity 11 of 17 11 of 18 hi h Polymers 2017, 9, 522 Polymers 2017, 9, 522 was observed for the highest ﬁbre content, with a difference in the crystallinity up to 10%, compared with the neat annealed PA11. be produced between PA11 and fibres, explaining the increments produced during the annealing. The highest crystallinity was observed for the highest fibre content, with a difference in the crystallinity up to 10% compared with the neat annealed PA11 Furthermore, a secondary crystal formation was observed in all the samples after annealing (Figure 7). PA11 showed a small peak around the temperature of the annealing treatment (165 ◦C) corresponding to the polymer chains crystallized during the annealing. In the composites, this small peak increased and was shifted to higher temperatures, indicating higher Tm of the crystals produced during the annealing process. In the case of PA11 + 50% SGW, the secondary crystal formation merged with the peak of the melting temperature at 189 ◦C. This phenomenon can be related with a nucleating effect of the ﬁbre which has been observed in other polymers [22]. Moreover, the enhancement of the Tm of the crystals obtained after the annealing was also observed in the literature when the annealing time was increased for PA11 matrix [31]. The addition of SGW shortened the annealing treatment to achieve similar crystallinities. crystallinity up to 10%, compared with the neat annealed PA11. Furthermore, a secondary crystal formation was observed in all the samples after annealing (Figure 7). PA11 showed a small peak around the temperature of the annealing treatment (165 °C) corresponding to the polymer chains crystallized during the annealing. In the composites, this small peak increased and was shifted to higher temperatures, indicating higher Tm of the crystals produced during the annealing process. In the case of PA11 + 50% SGW, the secondary crystal formation merged with the peak of the melting temperature at 189 °C. This phenomenon can be related with a nucleating effect of the fibre which has been observed in other polymers [22]. Moreover, the enhancement of the Tm of the crystals obtained after the annealing was also observed in the literature when the annealing time was increased for PA11 matrix [31]. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions The addition of SGW shortened the annealing treatment to achieve similar crystallinities. Figure 7. DSC thermographs of the firsts heating comparing annealed and not annealed samples. Figure 7. DSC thermographs of the ﬁrsts heating comparing annealed and not annealed samples. Figure 7. DSC thermographs of the firsts heating comparing annealed and not annealed samples. Figure 7. DSC thermographs of the ﬁrsts heating comparing annealed and not annealed samples. Although an improvement in the crystallinity due the secondary crystal growth obtained during annealing was observed, only slight differences in the Tm values were found by the effect of the annealing treatment. Although an improvement in the crystallinity due the secondary crystal growth obtained during annealing was observed, only slight differences in the Tm values were found by the effect of the annealing treatment. A change in the Tg was expected, as this transition temperature is directly related with the amorphous phase in the material. After the annealing, the amorphous phase of the polymer matrix was reduced, thus the crystalline phase was increased, diminishing the mobility of the chains. This can imply a higher Tg temperature. DMTA was performed in the annealed PA11 and their composites and the results were compared to the untreated ones. A change in the Tg was expected, as this transition temperature is directly related with the amorphous phase in the material. After the annealing, the amorphous phase of the polymer matrix was reduced, thus the crystalline phase was increased, diminishing the mobility of the chains. This can imply a higher Tg temperature. DMTA was performed in the annealed PA11 and their composites and the results were compared to the untreated ones. In the loss modulus (Figure 8), the annealed samples achieved higher values due to the higher crystalline phase in the polymer. Nonetheless, this increment in the modulus was reduced as the fibre content increased. Moreover, again, a unique transition corresponding to the Tg was observed In the loss modulus (Figure 8), the annealed samples achieved higher values due to the higher crystalline phase in the polymer. Nonetheless, this increment in the modulus was reduced as the ﬁbre content increased. Moreover, again, a unique transition corresponding to the Tg was observed in the studied range. However, the peaks were shifted to higher temperatures, except for the PA11 + 50% SGW. Polymers 2017, 9, 522 in the studied ran 12 of 17 e PA11 + Figure 8. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions Loss modulus obtained by DMTA of the treated and untreated samples. Figure 8. Loss modulus obtained by DMTA of the treated and untreated samples. in the studied range. However, the peaks were shifted to higher temperatures, except for the PA11 + 50% SGW. Figure 8. Loss modulus obtained by DMTA of the treated and untreated samples. d to higher temperatures, except for the PA11 + in the studied range. However, the peaks were shifte 50% SGW. Figure 8. Loss modulus obtained by DMTA of the treated and untreated samples. Figure 8. Loss modulus obtained by DMTA of the treated and untreated samples. Figure 8 Loss modulus obtained by DMTA of the treated and untreated samples As aforementioned, the tan δ was used to determine the Tg of the composite materials. Figure 9 shows the Tg obtained for untreated and treated samples. 57 00 t t d l As aforementioned, the tan δ was used to determine the Tg of the composite materials. Figure 9 shows the Tg obtained for untreated and treated samples. As aforementioned, the tan δ was used to determine the Tg of the composite materials. Figure 9 shows the Tg obtained for untreated and treated samples. Figure 9. Tg measured values for treated and untreated samples. The Tg temperatures increased for all the samples, except for the PA11 + 50% SGW, where 45.00 47.00 49.00 51.00 53.00 55.00 PA11 PA11+10%SGW PA11+20%SGW PA11+50%SGW Temperature (ºC) treated samples Figure 9. Tg measured values for treated and untreated samples. 45.00 47.00 49.00 51.00 53.00 55.00 57.00 PA11 PA11+10%SGW PA11+20%SGW PA11+50%SGW Temperature (ºC) untreated samples treated samples Figure 9. Tg measured values for treated and untreated samples. ratures increased for all the samples except for the PA11 Figure 9. Tg measured values for treated and untreated samples. Figure 9. Tg measured values for treated and untreated samples. similar Tg values were found for the untreated and annealed samples. On the other hand, the effect of the annealing in the Tg decreased as the fibre contents increased. This can be due to higher impact of the stiffness of the fibres in the modulus and in the reduction of the chain mobility than to the increase of crystallinity produced by the annealing treatment. The Tg temperatures increased for all the samples, except for the PA11 + 50% SGW, where similar Tg values were found for the untreated and annealed samples. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions On the other hand, the effect of the annealing in the Tg decreased as the fibre contents increased. This can be due to higher impact The Tg temperatures increased for all the samples, except for the PA11 + 50% SGW, where similar Tg values were found for the untreated and annealed samples. On the other hand, the effect of the annealing in the Tg decreased as the ﬁbre contents increased. This can be due to higher impact of the stiffness of the ﬁbres in the modulus and in the reduction of the chain mobility than to the increase of crystallinity produced by the annealing treatment. 13 of 17 Polymers 2017, 9, 522 Polymers 2017, 9, 522 Figure 10 shows the results obtained for the evolution of the storage modulus with the temperature analysed by DMTA. Higher moduli were obtained in the annealed samples due to the higher crystalline phase in the polymer. Nevertheless, as was observed, this effect was higher in the neat PA11 than in PA11-SGW composites. y Figure 10 shows the results obtained for the evolution of the storage modulus with the temperature analysed by DMTA. Higher moduli were obtained in the annealed samples due to the higher crystalline phase in the polymer. Nevertheless, as was observed, this effect was higher in the neat PA11 than in PA11-SGW composites Figure 10. Storage modulus of treated and untreated samples obtained by DMTA. Figure 10. Storage modulus of treated and untreated samples obtained by DMTA. Figure 10. Storage modulus of treated and untreated samples obtained by DMTA. Figure 10. Storage modulus of treated and untreated samples obtained by DMTA. Finally, the annealed samples were also studied by X-ray diffraction. As shown, the α’ form was observed in the neat PA11 and PA11-based composites (Figure 11). The annealing treatment produced a change in the polymer structure from the δ’ obtained after the injection-moulding process to the triclinic α’ structure, which is more thermodynamically stable [17,35]. The broad peak of the δ’ phase observed at 2θ = 21° was transformed in two well defined peaks at room temperature which shifted to 20.3° and 22.7°. The γ phase can be also obtained but is not usually observed under Finally, the annealed samples were also studied by X-ray diffraction. As shown, the α’ form was observed in the neat PA11 and PA11-based composites (Figure 11). 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions The annealing treatment produced a change in the polymer structure from the δ’ obtained after the injection-moulding process to the triclinic α’ structure, which is more thermodynamically stable [17,35]. The broad peak of the δ’ phase observed at 2θ = 21◦was transformed in two well deﬁned peaks at room temperature which shifted to 20.3◦and 22.7◦. The γ phase can be also obtained but is not usually observed under 100 ◦C [43]. 100 °C [43]. Moreover, in the FT-IR of the PA11 annealed sample (Figure 12), the peaks shifted to α’ reported wavelengths [45] and their presence can be discarded or really reduced. In the composite materials, when the fibre content were augmented, the intensity of the second peak intensity corresponding to α’ form did the same, and exceeded the intensity of the first peak in the PA11 + 20% SGW and PA11 + 50% SGW composites. This effect, can be related with the overlap of this second peak with the cellulose peak and also with higher crystallinity [21] in the material which was d h h b d d h SC Moreover, in the FT-IR of the PA11 annealed sample (Figure 12), the peaks shifted to α’ reported wavelengths [45] and their presence can be discarded or really reduced. In the composite materials, when the ﬁbre content were augmented, the intensity of the second peak intensity corresponding to α’ form did the same, and exceeded the intensity of the ﬁrst peak in the PA11 + 20% SGW and PA11 + 50% SGW composites. This effect, can be related with the overlap of this second peak with the cellulose peak and also with higher crystallinity [21] in the material which was in accordance with the obtained data in the DSC. 14 of 17 14 of 18 14 of 18 Polymers 2017, 9, 522 Polymers 2017 9 522 Polymers 2017, 9, 522 14 of 17 14 of 18 14 of 18 Figure 11. X-ray diffractograms of annealed samples. Figure 12. FT-IR of PA11 and annealed PA11. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass T 5 10 15 20 25 30 35 40 Intensity 2θ (º) PA11 annealed PA11+10%SGW annealed PA11+20%SGW annealed PA11+50%SGW annealed Figure 11. X-ray diffractograms of annealed samples. Figure 11. X-ray diffractograms of annealed samples. 4. Conclusions A thermal and structural characterization of PA11 and its composites reinforced with SGW was performed by means of TGA, DSC, DMTA and X-ray diffractionto determine the effect of the lignocellulosic reinforcement on the thermal transitions and morphology of the polymer matrix. A lower onset degradation temperature of PA11 by the effect of the ﬁbres was found; however, once the decomposition temperature was started, the cellulosic ﬁbres contributed to thermally stabilize the composites. No differences in Tm and Tc, the main melting and crystallization peaks, and the crystallinity degree (26–27%) were found by the incorporation of SGW. Nonetheless, ﬁbres promoted the α’ crystalline form instead of the both γ and α’ forms observed in the neat PA11. DMTA results revealed no considerable differences in the Tg of the composites with the ﬁbre content attributed to the formation of an amorphous phase with more restricted mobility caused by the presence of the reinforcement. An improvement on the storage modulus was observed when increasing the ﬁbre content throughout all the measured temperature range. g p g Some samples were subjected to an annealing treatment during 1 h at 165 ◦C. The effect of the annealing on the materials was analysed by means of DSC and DMTA. The matrix crystallinity increased after the treatment. Moreover, higher crystal growth was also observed when the ﬁbre content was augmented in the materials, achieving crystallinities increments around 13% for PA11 + 50% SGW. A slight increase of the Tg was observed, which can be related with the different crystal growths during the annealing treatment. A lower effect was obtained when the ﬁbre content increased, probably due to the higher melting temperature of these crystals, similar to the PA11 melting point in the case of PA11 + 50% SGW composite. Slightly higher storage modulus values were obtained in the annealed samples, although the effect of the annealing in the stiffness seemed to decrease when the reinforcement content were increased. Finally, a different structure was observed before and after annealing in all samples. X-ray diffractograms showed a δ’ phase in the samples after the injection moulding process. However, in the annealed sample, this phase was transformed to a more stable α’ form. Moreover, FT-IR detected a small part of γ phase in neat PA11 and low ﬁbre contents, which was not shown after the annealing. 3.3. Effect of Annealing on the Structure and Thermal Transitions 3.3. Effect of Annealing on the Structure and Thermal Transitions Figure 12. FT-IR of PA11 and annealed PA11. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass Tg. 5 10 15 20 25 30 35 40 Intensity 2θ (º) PA11 annealed PA11+10%SGW annealed PA11+20%SGW annealed PA11+50%SGW annealed Figure 12. FT-IR of PA11 and annealed PA11. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass T Figure 11. X-ray diffractograms of annealed samples. 5 10 15 20 25 30 35 40 Intensity 2θ (º) PA11 annealed PA11+10%SGW annealed PA11+20%SGW annealed PA11+50%SGW annealed Figure 11. X-ray diffractograms of annealed samples. Figure 11. X-ray diffractograms of annealed samples. 5 10 15 20 25 30 35 40 Intensity 2θ (º) PA11 annealed PA11+10%SGW annealed PA11+20%SGW annealed PA11+50%SGW annealed Figure 11. X-ray diffractograms of annealed samples. Figure 11. X-ray diffractograms of annealed samples. Figure 11. X-ray diffractograms of annealed samples. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. Figure 12. FT-IR of PA11 and annealed PA11. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass Tg. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass Tg. The use of an annealing treatment has interest in lignocellulosic reinforced PA11 materials to increase the stiffness and the mechanical properties of the composite materials if working temperatures do not overpass Tg. Polymers 2017, 9, 522 15 of 17 15 of 17 4. Conclusions Acknowledgments: The authors hereby thank Arkema for kindly supplying the polyamide 11 that was used in this work, and Patrick Dang (Arkema France) and Pep Català (Arkema Spain) for the technical information provided. The authors are also grateful to Luis Angel Granda for the technical support and helpful discussion. Author Contributions: Helena Oliver-Ortega performed the experimental part and wrote the ﬁrst version of the paper. José Alberto Méndez and Mònica Ardanuy conceived and designed the experiments. Francesc Xavier Espinach and Quim Tarrés analysed and represented the data. Pere Mutjé guided the project. All the authors contributed to write and correct the paper. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Lubin, G. Handbook of Composites; Lubin, G., Ed.; Springer: Boston, MA, USA, 1982; ISBN 978-1-4615-7141-4. 1. Lubin, G. Handbook of Composites; Lubin, G., Ed.; Springer: Boston, MA, USA, 1982; ISBN 978-1-4615-7141-4. 1. Lubin, G. Handbook of Composites; Lubin, G., Ed.; Springer: Boston, MA, USA, 1982; ISBN 978-1-4615-7141-4. 2 Bodros E ; Pillin I ; Montrelay N ; Baley C Could biopolymers reinforced by randomly scattered ﬂax ﬁbre 1. Lubin, G. Handbook of Composites; Lubin, G., Ed.; Springer: Boston, MA, USA, 1982; ISBN 978-1-4615-7141-4. 2. Bodros, E.; Pillin, I.; Montrelay, N.; Baley, C. Could biopolymers reinforced by randomly scattered ﬂax ﬁbre be used in structural applications? Compos. Sci. Technol. 2007, 67, 462–470. [CrossRef] , f p ; , , ; p g , , , ; 2. Bodros, E.; Pillin, I.; Montrelay, N.; Baley, C. Could biopolymers reinforced by randomly scattered ﬂax ﬁbre be used in structural applications? Compos. Sci. Technol. 2007, 67, 462–470. [CrossRef] 2. Bodros, E.; Pillin, I.; Montrelay, N.; Baley, C. Could biopolymers reinforced by randomly scattered ﬂax ﬁbre be used in structural applications? Compos. Sci. Technol. 2007, 67, 462–470. [CrossRef] . Martino, L.; Basilissi, L.; Farina, H.; Ortenzi, M.A.; Zini, E.; Di Silvestro, G.; Scandola, M. Bio-based polyam 11: Synthesis, rheology and solid-state properties of star structures. Eur. Polym. J. 2014, 59, 69–77. [Cross y gy p p y 5. Lefeuvre, A.; Bourmaud, A.; Morvan, C.; Baley, C. Elementary ﬂax ﬁbre tensile properties: Correlation between stress-strain behaviour and ﬁbre composition. Ind. Crops Prod. 2014, 52, 762–769. [CrossRef] 6. Bledzki, A.K.; Gassan, J. Composites reinforced with cellulose based ﬁbres. Prog. Polym. Sci. 1999, 24, 221–274. [CrossRef] 7. Anastas, P.T.; Zimmerman, J.B. Design through the 12 principles of green engineering. IEEE Eng. Manag. Rev. 2007, 35, 16. [CrossRef] Polymers 2017, 9, 522 16 of 17 16 of 17 8. Anastas, P.T.; Kirchhoff, M.M. Origins, current status, and future challenges of green chemistry. Acc. Chem. Res. 2002, 35, 686–694. [CrossRef] [PubMed] 9. Mohanty, A.K.; Misra, M.; Drzal, L.T. Sustainable Bio-Composites from renewable resources: Opportunities and challenges in the green materials world. J. Polym. Environ. 2002, 10, 19–26. [CrossRef] 10. Winnacker, M.; Rieger, B. Biobased Polyamides: Recent Advances in Basic and Applied Research. Macromol. Rapid Commun. 2016, 37, 1391–1413. [CrossRef] [PubMed] 11. Heitzmann, M.T.; Veidt, M.; Ng, C.T.; Lindenberger, B.; Hou, M.; Truss, R.; Liew, C.K. Single- plant biocomposite from ricinus communis: Preparation, properties and environmental performance. J. Polym. Environ. References 2013, 21, 366–374. [CrossRef] 12. Ashori, A. Wood–plastic composites as promising green-composites for automotive industries! Bioresour. Technol. 2008, 99, 4661–4667. [CrossRef] [PubMed] 13. Bourmaud, A.; Le Duigou, A.; Gourier, C.; Baley, C. Inﬂuence of processing temperature on mechanical performance of unidirectional polyamide 11-ﬂax ﬁbre composites. Ind. Crops Prod. 2016, 84, 151–165. [CrossRef] 14. Oliver-Ortega, H.; Granda, L.A.; Espinach, F.X.; Delgado-Aguilar, M.; Duran, J.; Mutjé, P. Stiffness of bio-based polyamide 11 reinforced with softwood stone ground-wood ﬁbres as an alternative to polypropylene-glass ﬁbre composites. Eur. Polym. J. 2016, 84, 481–489. [CrossRef] 15. Zierdt, P.; Theumer, T.; Kulkarni, G.; Däumlich, V.; Klehm, J.; Hirsch, U.; Weber, A. Sustainable wood-plastic composites from bio-based polyamide 11 and chemically modiﬁed beech ﬁbers. Sustain. Mater. Technol. 2015, 6, 6–14. [CrossRef] 16. Hu, Y.; Shen, L.; Yang, H.; Wang, M.; Liu, T.; Liang, T.; Zhang, J. Nanoindentation studies on Nylon 11/clay nanocomposites. Polym. Test. 2006, 25, 492–497. [CrossRef] 17. Zhang, X.; Yang, G.; Lin, J. Crystallization behavior of nylon 11/montmorillonite nanocomposites under annealing. J. Appl. Polym. Sci. 2006, 102, 5483–5489. [CrossRef] 18. Le Duigou, A.; Bourmaud, A.; Gourier, C.; Baley, C. Multi-scale shear properties of ﬂax ﬁbre reinforced polyamide 11 biocomposites. Compos. Part A Appl. Sci. Manuf. 2016, 85, 123–129. [CrossRef] 19. Oliver-Ortega, H.; Granda, L.A.; Espinach, F.X.; Méndez, J.A.; Julian, F.; Mutjé, P. Tensile properties and micromechanical analysis of stone groundwood from softwood reinforced bio-based polyamide11 composites. Compos. Sci. Technol. 2016, 132, 123–130. [CrossRef] p p 20. Zhang, Q.; Mo, Z.; Liu, S.; Zhang, H. Inﬂuence of annealing on structure of Nylon 11. Macromolecules 2000, 33, 5999–6005. [CrossRef] 21. Panaitescu, D.M.; Gabor, R.A.; Frone, A.N.; Vasile, E. Inﬂuence of Thermal Treatment on Mechanical and Morphological Characteristics of Polyamide 11/Cellulose Nanoﬁber Nanocomposites. J. Nanomater. 2015, 2015, 1–11. [CrossRef] 22. Granda, L.A.; Méndez, J.A.; Espinach, F.X.; Puig, J.; Delgado-Aguilar, M.; Mutjé, P. Polypropylene reinforced with semi-chemical ﬁbres of Leucaena collinsii: Thermal properties. Compos. Part B Eng. 2016, 94, 75–81. [CrossRef] 23. López, J.P.; Gironés, J.; Méndez, J.A.; El Mansouri, N.E.; Llop, M.; Mutjé, P.; Vilaseca, F. Stone-ground wood pulp-reinforced polypropylene composites: Water uptake and thermal properties. BioResources 2012, 7, 5478–5487. [CrossRef] 4. Yang, H.; Yan, R.; Chen, H.; Lee, D.H.; Zheng, C. Characteristics of hemicellulose, cellulose and lig pyrolysis. Fuel 2007, 86, 1781–1788. [CrossRef] 25. Azwa, Z.N.; Yousif, B.F.; Manalo, A.C.; Karunasena, W. A review on the degradability of polymeric composites based on natural ﬁbres. Mater. Des. References 2013, 47, 424–442. [CrossRef] 26. Lopez, J.P.; Mendez, J.A.; Espinach, F.X.; Julian, F.; Mutjé, P.; Vilaseca, F. Tensile strength characteristics of polypropylene composites reinforced with stone groundwood ﬁbres from softwood. BioResources 2012, 7, 3188–3200. [CrossRef] 27. Panaitescu, D.M.; Frone, A.N.; Nicolae, C. Micro- and nano-mechanical characterization of polyamide 11 and its composites containing cellulose nanoﬁbers. Eur. Polym. J. 2013, 49, 3857–3866. [CrossRef] 28. Ardanuy, M.; Antunes, M.; Velasco, J.I. Vegetable ﬁbres from agricultural residues as thermo-mechanical reinforcement in recycled polypropylene-based green foams. Waste Manag. 2012, 32, 256–263. [CrossRef] [PubMed] 17 of 17 Polymers 2017, 9, 522 17 of 17 29. Aydemir, D.; Kiziltas, A.; Erbas Kiziltas, E.; Gardner, D.J.; Gunduz, G. Heat treated wood-nylon 6 composites. Compos. Part B Eng. 2015, 68, 414–423. [CrossRef] 30. Stoclet, G.; Sclavons, M.; Devaux, J. Relations between structure and property of polyamide 11 nanocomposites based on raw clays elaborated by water-assisted extrusion. J. Appl. Polym. Sci. 2013, 127, 4809–4824. [CrossRef] 31. Castagnet, S.; Thilly, L. High-pressure dependence of structural evolution in polyamide 11 during annealing. J. Polym. Sci. Part B Polym. Phys. 2009, 47, 2015–2025. [CrossRef] 32. Stempﬂe, F.; Ortmann, P.; Mecking, S. Long-chain aliphatic polymers to bridge the gap between semicrystalline polyoleﬁns and traditional polycondensates. Chem. Rev. 2016, 116, 4597–4641. [CrossRef] [PubMed] 33. Mancic, L.; Osman, R.F.M.; Costa, A.M.L.M.; d’Almeida, J.R.M.; Marinkovic, B.A.; Rizzo, F.C. Thermal and mechanical properties of polyamide 11 based composites reinforced with surface modiﬁed titanate nanotubes. Mater. Des. 2015, 83, 459–467. [CrossRef] 34. Mago, G.; Kalyon, D.M.; Fisher, F.T. Nanocomposites of polyamide-11 and carbon nanostructures: Development of microstructure and ultimate properties following solution processing. J. Polym. Sci. Part B Polym. Phys. 2011, 49, 1311–1321. [CrossRef] 35. Pepin, J.; Miri, V.; Lefebvre, J.-M. New Insights into the Brill Transition in Polyamide 11 and Polya Macromolecules 2016, 49, 564–573. [CrossRef] 36. Peng, S.X.; Shrestha, S.; Youngblood, J.P. Crystal structure transformation and induction of shear banding in Polyamide 11 by surface modiﬁed Cellulose Nanocrystals. Polymer 2017, 114, 88–102. [CrossRef] 37. Reixach, R.; Puig, J.; Méndez, J.A.; Gironès, J. Orange Wood Fiber Reinforced Polypropylene Composites: Thermal Properties. BioResources 2015, 10, 2156–2166. [CrossRef] 38. Joseph, P.V.; Joseph, K.; Thomas, S.; Pillai, C.K.S.; Prasad, V.S.; Groeninckx, G.; Sarkissova, M. The thermal and crystallisation studies of short sisal ﬁbre reinforced polypropylene composites. Compos. Part A Appl. Sci. Manuf. 2003, 34, 253–266. [CrossRef] 39. Mancic, L.; Pontón, P.I.; Letichevsky, S.; Costa, A.M.; Marinkovic, B.A.; Rizzo, F.C. References Application of silane grafted titanate nanotubes in reinforcing of polyamide 11 composites. Compos. Part B Eng. 2016, 93, 153–162. [CrossRef] 40. Semba, T.; Ito, A.; Kitagawa, K.; Nakatani, T.; Yano, H.; Sato, A. Thermoplastic composites of polyamide-12 reinforced by cellulose nanoﬁbers with cationic surface modiﬁcation. J. Appl. Polym. Sci. 2014, 131. [CrossRef] 40. Semba, T.; Ito, A.; Kitagawa, K.; Nakatani, T.; Yano, H.; Sato, A. Thermoplastic composites of polyamide-12 reinforced by cellulose nanoﬁbers with cationic surface modiﬁcation. J. Appl. Polym. Sci. 2014, 131. [CrossRef] 41. Rhee, S.; White, J.L. Crystalline structure and morphology of biaxially oriented polyamide-11 ﬁlms. J. Polym. g p p p y reinforced by cellulose nanoﬁbers with cationic surface modiﬁcation. J. Appl. Polym. Sci. 2014, 131. [CrossRef] 41. Rhee, S.; White, J.L. Crystalline structure and morphology of biaxially oriented polyamide-11 ﬁlms. J. Polym. Sci. Part B Polym. Phys. 2002, 40, 2624–2640. [CrossRef] 41. Rhee, S.; White, J.L. Crystalline structure and morphology of biaxially oriented polyamide-11 ﬁlms. J. Polym. Sci. Part B Polym. Phys. 2002, 40, 2624–2640. [CrossRef] 42. Frübing, P.; Kremmer, A.; Gerhard-Multhaupt, R.; Spanoudaki, A.; Pissis, P. Relaxation processes at the glass transition in polyamide 11: From rigidity to viscoelasticity. J. Chem. Phys. 2006, 125. [CrossRef] [PubMed] 42. Frübing, P.; Kremmer, A.; Gerhard-Multhaupt, R.; Spanoudaki, A.; Pissis, P. Relaxation processes at the glass transition in polyamide 11: From rigidity to viscoelasticity. J. Chem. Phys. 2006, 125. [CrossRef] [PubMed] 43. Nair, S.S.; Ramesh, C.; Tashiro, K. Crystalline phases in nylon-11: Studies using HTWAXS and HTFTIR. 43. Nair, S.S.; Ramesh, C.; Tashiro, K. Crystalline phases in nylon-11: Studies using HTWAXS and HTFTIR. Macromolecules 2006, 39, 2841–2848. [CrossRef] 44. Naffakh, M.; Shuttleworth, P.S.; Ellis, G. Bio-based polymer nanocomposites based on nylon 11 and WS 2 inorganic nanotubes. RSC Adv. 2015, 5, 17879–17887. [CrossRef] 45. Nair, S.S.; Ramesh, C.; Tashiro, K. Polymorphism in nylon-11: Characterization using HTWAXS and HTFTIR. Macromol. Symp. 2006, 242, 216–226. [CrossRef] 46. Pérez-Fonseca, A.A.; Robledo-Ortíz, J.R.; González-Núñez, R.; Rodrigue, D. Effect of thermal annealing on the mechanical and thermal properties of polylactic acid-cellulosic ﬁber biocomposites. J. Appl. Polym. Sci. 2016, 133, 1–10. [CrossRef] 47. Quan, H.; Li, Z.M.; Yang, M.B.; Huang, R. On transcrystallinity in semi-crystalline polymer composites. Compos. Sci. Technol. 2005, 65, 999–1021. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2088549918	https://discovery.ucl.ac.uk/1394338/1/journal.pone.0054754.pdf	English	null	In Vivo Cell Reprogramming towards Pluripotency by Virus-Free Overexpression of Defined Factors	PloS one	2,013	cc-by	8,235	Introduction multiple rounds of gene transfer vectors, growth factors, antibiotics and other cell media cocktails to promote reprogramming and select for pluripotency. All of these are considered major culprits for the potential risks associated with the ensuing cells as recent studies investigating the genomic integrity of iPS have alluded to [14,15,16]. In terms of iPS generation using non-viral gene transfer vectors, plasmid DNA [4,5,6] or RNA [7,10] delivery using liposomes or electroporation have been reported. Compared to viruses, episomal vectors are generally considered safer, however transduction and reprogramming efficiencies are much lower [13]. Alternatively, Warren et al. reported somatic cell reprogramming in vitro by direct delivery of synthetic mRNAs [7]. Although this methodology offers significantly higher reprogram- ming efficiency, high RNA dosages, multiple rounds of trans- fection and complex cell culturing protocols are still needed [13]. Forced reprogramming of somatic cells into a pluripotent, stem cell-like state by the ectopic expression of specific transcription factors results in the generation of induced pluripotent stem (iPS) cells. Such transcription factor cell reprogramming has been achieved today by viral [1,2,3] and non-viral [4,5,6,7] gene transfer, protein cytoplasmic translocation [8,9], miRNA [10] and is changing the landscape in developmental biology, can potentially resolve all ethical concerns about the use of embryonic stem cells and open further opportunities for regenerative medicine. The original discovery by Yamanaka and colleagues that the in vitro expression of four transcription factors, Oct3/4, Klf4, Sox2, c-Myc (OKSM) was capable to revert fully differen- tiated mouse and human skin fibroblasts into iPS cells [1,11] constitutes the most widely used transcription-based reprogram- ming technology today. Due to the paradigm-shifting nature of transcription-induced reprogramming to pluripotency there is still limited understanding of the exact mechanisms and pathways implicated in induced cell reprogramming, and the exact features of reprogrammed cells [17,18]. Morever, the majority of experimental evidence today is based on the concept of extraction and in vitro manipulation of the somatic cells to be reprogrammed, leading to the array of caveats mentioned above that make clinical translation of iPS cells seem distant [19,20,21]. In the present work, we hypothesized that in vivo forced expression of the OKSM transcription factors by non-viral transient over-expression within living tissue can induce The initial reports of transcription-mediated somatic cell reprogramming involved the use of retroviruses to stably transduce skin fibroblasts with defined transcription factors [1,2,11]. Received October 11, 2012; Accepted December 18, 2012; Published January 23, 2013 In the present work, we hypothesized that in vivo forced expression of the OKSM transcription factors by non-viral transient over-expression within living tissue can induce Ac¸elya Yilmazer, Irene de La´zaro, Cyrill Bussy, Kostas Kostarelos* Ac¸elya Yilmazer, Irene de La´zaro, Cyrill Bussy, Kostas Kostarelos* Nanomedicine Lab, UCL School of Pharmacy, UCL School of Life & Medical Sciences, University College London, Brunswick Squar PLOS ONE \| www.plosone.org January 2013 \| Volume 8 \| Issue 1 \| e54754 In Vivo Cell Reprogramming towards Pluripotency by Virus-Free Overexpression of Defined Factors Ac¸elya Yilmazer, Irene de La´zaro, Cyrill Bussy, Kostas Kostarelos* Nanomedicine Lab, UCL School of Pharmacy, UCL School of Life & Medical Sciences, University College London, Brunswick Square, London, United Kingdom Abstract The ability to induce the reprogramming of somatic mammalian cells to a pluripotent state by the forced expression of specific transcription factors has helped redefine the rules of cell fate and plasticity, as well as open possibilities for disease modeling, drug screening and regenerative medicine. Here, we hypothesized that the non-viral forced expression of the four originally discovered defined factors (OKSM) in adult mice could result in in vivo reprogramming of cells in the transfected tissue in situ. We show that a single hydrodynamic tail-vein (HTV) injection of two plasmids encoding for Oct3/4, Sox2, Klf4 and c-Myc respectively, are highly expressed in the liver tissue of Balb/C adult mice. Hallmark pluripotency markers were upregulated within 24–48 h after injection, followed by down-regulation of all major hepatocellular markers. Generation of transcriptionally reprogrammed cells in vivo was further confirmed by positive staining of liver tissue sections for all major pluripotency markers in Balb/C mice and the Nanog-GFP reporter transgenic strain (TNG-A) with concomitant upregulation of GFP expression in situ. No signs of physiological or anatomical abnormalities or teratoma formation were observed in the liver examined up to 120 days. These findings indicate that virus-free expression of OKSM factors in vivo can transcriptionally reprogram cells in situ rapidly, efficiently and transiently, absent of host tissue damage or teratoma formation. Citation: Yilmazer A, de La´zaro I, Bussy C, Kostarelos K (2013) In Vivo Cell Reprogramming towards Pluripotency by Virus-Free Overexpression of Defined Factors. PLoS ONE 8(1): e54754. doi:10.1371/journal.pone.0054754 Editor: Knut Stieger Justus Liebig University Giessen Germany Editor: Knut Stieger, Justus-Liebig-University Giessen, Germany Editor: Knut Stieger, Justus-Liebig-University Giessen, Germany Received October 11, 2012; Accepted December 18, 2012; Published January 23, 2013 Received October 11, 2012; Accepted December 18, 2012; Published January 23, 2013 Received October 11, 2012; Accepted December 18, 2012; Published January 23, 2013 Received October 11, 2012; Accepted December 18, 2012; Published January 23, 2013 mazer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2013 Yilmazer et al. This is an open-access article distributed under the terms of the Creative Commons Attribut unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was partially supported by the Engineering and Physical Sciences Research Council (EPSRC, Swindon, United Kingdom) and the European Commission FP-7 Programs NANONEUROHOP (PIEF-GA-2010-276051) and ANTICARB (HEALTH-2008-20157). AY was supported by a full-time PhD scholarship from the Ministry of Education (Turkey) and IL acknowledges partial support from the British Council and the CAIXA Foundation (Spain). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: k.kostarelos@ucl.ac.uk multiple rounds of gene transfer vectors, growth factors, antibiotics and other cell media cocktails to promote reprogramming and select for pluripotency. All of these are considered major culprits for the potential risks associated with the ensuing cells as recent studies investigating the genomic integrity of iPS have alluded to [14,15,16]. In terms of iPS generation using non-viral gene transfer vectors, plasmid DNA [4,5,6] or RNA [7,10] delivery using liposomes or electroporation have been reported. Compared to viruses, episomal vectors are generally considered safer, however transduction and reprogramming efficiencies are much lower [13]. Alternatively, Warren et al. reported somatic cell reprogramming in vitro by direct delivery of synthetic mRNAs [7]. Although this methodology offers significantly higher reprogram- ming efficiency, high RNA dosages, multiple rounds of trans- fection and complex cell culturing protocols are still needed [13]. Due to the paradigm-shifting nature of transcription-induced reprogramming to pluripotency there is still limited understanding of the exact mechanisms and pathways implicated in induced cell reprogramming, and the exact features of reprogrammed cells [17,18]. Morever, the majority of experimental evidence today is based on the concept of extraction and in vitro manipulation of the somatic cells to be reprogrammed, leading to the array of caveats mentioned above that make clinical translation of iPS cells seem distant [19,20,21]. Isolation of Hepatocyte Population Mice livers were perfused as previously described [36] with some modifications. In brief, livers were first perfused with Ca2+ and Mg2+ free Hank’s buffered salt solution (Sigma-Aldrich, UK) and then with liver digest medium (Gibco, UK) at 37uC. After digestion, liver was washed with Hepatocyte Wash Medium Figure 1. In vivo overexpression of OKSM transcription factors in adult mouse liver. Balb/C mice were HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline and at days 2, 4, 8, 12, 24, RT-qPCR analysis of hepatocytes was performed to determine the relative gene expression of: (a) transfected transcription factors (OKSM) and (b) endogenous pluripotency markers. All gene expression levels were normalized to HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (c) flow cytometry analysis of OCT3/4 positive and NANOG positive cells in liver extracts; (d) relative gene expression of hepatocyte markers as determined by RT-qPCR. All gene expression levels were normalized to saline HTV-injected group ( p,0.05 indicates statistically significant differences between the expression levels for hepatocyte markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison). doi:10.1371/journal.pone.0054754.g001 PLOS ONE \| www plosone org 2 January 2013 \| Volume 8 \| Issue 1 \| e54754 Figure 1. In vivo overexpression of OKSM transcription factors in adult mouse liver. Balb/C mice were HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline and at days 2, 4, 8, 12, 24, RT-qPCR analysis of hepatocytes was performed to determine the relative gene expression of: (a) transfected transcription factors (OKSM) and (b) endogenous pluripotency markers. All gene expression levels were normalized to HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (c) flow cytometry analysis of OCT3/4 positive and NANOG positive cells in liver extracts; (d) relative gene expression of hepatocyte markers as determined by RT-qPCR. Introduction This methodology of gene transfer is still today the most popular way to reprogram animal and human somatic cells despite the risks from insertional mutagenesis, stable transduction and long-term gene expression of known proto-oncogenes [12,13]. Moreover, the vast majority of current methodologies to generate iPS cells involve use of long-term culture conditions and treatment of cells with January 2013 \| Volume 8 \| Issue 1 \| e54754 January 2013 \| Volume 8 \| Issue 1 \| e54754 1 PLOS ONE \| www.plosone.org In Vivo Cell Reprogramming in Mouse Tissue Animals and Hydrodynamic Tail Vein (HTV) Injection of Plasmids Animals and Hydrodynamic Tail Vein (HTV) Injection of Plasmids cell reprogramming towards pluripotency. In order to test this hypothesis we chose the most naive, non-viral gene transfer technology available today: large-volume, rapid hydrodynamic tail vein (HTV) injection of plasmid DNA [22,23] encoding the originally proposed OKSM factors. This gene transfer methodol- ogy circumvents most complications or risks associated with viral gene transfer vectors, as has been previously shown in numerous preclinical [24,25] and clinical [26,27] studies allowing un- precedented levels of exogenous gene expression in hepatocytes. All experiments were performed with prior approval from the UK Home Office under a Home Office project license (PPL 80/ 2296). Female Balb/C mice, 6 weeks old were purchased from Harlan, UK. TNG-A mice which carry the eGFP reporter inserted into the Nanog locus [28], were a kind gift from The Wellcome Trust Centre for Stem Cell Research, University of Cambridge, UK. TNG-A mice and WT controls, were of 129 background and were bred and genotyped at the UCL. Mice were allowed one week to acclimatize prior to use. Mice (4 animals/ group)were warmed in a 37uC heating chamber, anesthetized with isofluorane and were injected via tail vein in 5–7 sec with 1.5 ml of 0.9% saline including 75 mg of pCX-OKS-2A with and without 75 mg of pCX-cMyc or 150 mg pCAG-GFP plasmids or no plasmid. Mice were culled at different time points, such as 2, 4, 8, 12, 24, 50, 120 days after HTV injections. Plasmids Reprogramming plasmids pCX-OKS-2A encoding OCT3/4, KLF4, SOX2; pCX-cMyc encoding CMYC and pCAG-GFP encoding eGFP under the control of CAG promoter (as previously described by Okita et al. [4]) were obtained from Addgene (USA) as bacterial stabs. Research grade plasmid production was performed from these stabs (Plasmid Factory, Germany). pGFP- Luc plasmid (Clontech, USA) encodes for the eGFP and Luc transgenes under the control of a CMV promoter. Vector maps of all plasmids used in this work are included in Figure S5. Isolation of Hepatocyte Population All gene expression levels were normalized to saline HTV-injected group ( p,0.05 indicates statistically significant differences between the expression levels for hepatocyte markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison). doi:10.1371/journal.pone.0054754.g001 January 2013 \| Volume 8 \| Issue 1 \| e54754 PLOS ONE \| www.plosone.org 2 In Vivo Cell Reprogramming in Mouse Tissue Figure 2. OKSM and OKS factor overexpression with dose-response in adult mouse liver. Balb/C mice HTV injected with 0.9% saline alone, pCX-OKS-2A with (OSKM) and without (OKS) pCX-cMyc in 0.9% saline, or pCAG-GFP in 0.9% saline, at the indicated doses. On day 4, RT-qPCR analysis of hepatocyte extracts was performed. (a) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes were determined for plasmid dose-escalation (total plasmid dose 50, 75 and 100 mg/animal). All gene expression levels were normalized to the HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (b) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes with and without inclusion of cMyc. All gene expression levels were normalized to HTV-injected saline group (p,0.01 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and OKS injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison). doi:10.1371/journal.pone.0054754.g002 Figure 2. OKSM and OKS factor overexpression with dose-response in adult mouse liver. Balb/C mice HTV injected with 0.9% saline alone, pCX-OKS-2A with (OSKM) and without (OKS) pCX-cMyc in 0.9% saline, or pCAG-GFP in 0.9% saline, at the indicated doses. On day 4, RT-qPCR analysis of hepatocyte extracts was performed. (a) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes were determined for plasmid dose-escalation (total plasmid dose 50, 75 and 100 mg/animal). All gene expression levels were normalized to the HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (b) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes with and without inclusion of cMyc. All gene expression levels were normalized to HTV-injected saline group (*p,0 01 indicates statistically significant differences between the expression levels of pluripotency Figure 2. Flow Cytometry Flow Cytometry Cell density was adjusted to 16107 cells/ml and 100 ml suspensions were aliquoted in microfuge tubes. BD Mouse Pluripotent Stem Cell Transcription Factor Analysis Kit (BD Biosciences, UK) were used to analyze OCT3/4-positive or NANOG-positive cells. In brief, cells were firstly fixed with BD Cytofix fixation buffer and permeabilized with 1X BD Perm/ Wash buffer. Then cells were incubated with either anti-mouse OCT4-PerCP-Cy5.5 or anti-mouse NANOG-PE for 30 min according to manufacturer’s instructions. Negative and isotype controls were included in each experiment. Stained cells were then analyzed by CyAnTM ADP High-Performance Research Flow Isolation of Hepatocyte Population OKSM and OKS factor overexpression with dose-response in adult mouse liver. Balb/C mice HTV injected with 0.9% saline alone, pCX-OKS-2A with (OSKM) and without (OKS) pCX-cMyc in 0.9% saline, or pCAG-GFP in 0.9% saline, at the indicated doses. On day 4, RT-qPCR analysis of hepatocyte extracts was performed. (a) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes were determined for plasmid dose-escalation (total plasmid dose 50, 75 and 100 mg/animal). All gene expression levels were normalized to the HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV-injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (b) Expression levels of the injected reprogramming transcription factors and endogenous pluripotency genes with and without inclusion of cMyc. All gene expression levels were normalized to HTV-injected saline group (*p,0.01 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and OKS injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison). doi:10.1371/journal.pone.0054754.g002 Rad, UK) with the following protocol: 95uC for 3 min, 1 cycle; 95uC for 10 sec, 60uC for 30 sec, – repeated for 40 cycles. b-actin was used as a housekeeping gene and gene expression levels were normalized to saline groups. (HWM, Gibco, UK) and cell suspension was passed through a 100 ml cell strainer (VWR, UK) at 4uC. Cells were centrifuged at 50 g for 5 min to separate parenchymal cells (PC including hepatocytes) which were collected in pellet and non-parenchymal cells (NPC including Kupffer cells and epithelial cells) which stayed in supernatant. The hepatocyte fraction was collected after washing twice with HWM. RNA Isolation and Reverse Transcription-quantitative PCR (RT-qPCR) Analysis Following liver perfusion and isolation of hepatocytes, RNeasy Blood and Tissue Kit (Qiagen, UK) were used to isolate total RNA. cDNA synthesis from 1 mg of RNA sample was performed by iScript cDNA synthesis kit (Bio-Rad, UK) according to manufacturer’s instructions. Two microliters of each cDNA sample was used to perform RT-qPCR reactions with iO SYBR Green Supermix (Bio-Rad, UK). Primer sequences are shown in Table S1. Samples were run on CFX-96 Real Time System (Bio- January 2013 \| Volume 8 \| Issue 1 \| e54754 PLOS ONE \| www.plosone.org 3 In Vivo Cell Reprogramming in Mouse Tissue Figure 3. In vivo cell reprogramming on adult mouse liver tissue by immunohistochemistry. Balb/C mice HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline, or 150 mg of pCAG-GFP in 0.9% saline and at day 4, livers were collected and frozen tissue sections were stained with anti-OCT4, anti-SOX2 or anti-NANOG antibodies to assess immunoreactivity, or BCIP/NBT to determine ALP activity in the tissue (40x). Scale bars represent 100 mm. doi:10.1371/journal.pone.0054754.g003 Figure 3. In vivo cell reprogramming on adult mouse liver tissue by immunohistochemistry. Balb/C mice HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline, or 150 mg of pCAG-GFP in 0.9% saline and at day 4, livers were collected and frozen tissue sections were stained with anti-OCT4, anti-SOX2 or anti-NANOG antibodies to assess immunoreactivity, or BCIP/NBT to determine ALP activity in the tissue (40x). Scale bars represent 100 mm. doi:10 1371/journal pone 0054754 g003 Cytometer (DakoCytomation, USA) at the Institute of Child Health, University College London, UK. tome (Leica Microsystems, CM3050S), air-dried for 1hour at room temperature, before storage at 220uC. Before staining, liver sections were post-fixed with methanol, pre-cooled at 220uC, for 10 min at 220uC, then air-dried for 15 min and finally washed twice with PBS for 5 min. For Oct4, Sox2 and Nanog, we used a conventional immunostaining protocol that consisted of 1 h incubation in blocking buffer (5% goat serum-0.1% Triton in PBS pH 7.3) at room temperature, followed by two washing steps with PBS (1%BSA- 0.1% Triton, pH 7.3) before overnight incubation at +4uC with the different primary antibodies [(rabbit pAb anti- OCT4 (ab19857,3 mg/ml, Abcam, UK)/rabbit pAb anti-SOX2 (ab97959, 1 mg/ml, Abcam, UK)/rabbit polyclonal anti-NANOG (ab80892, 1 mg/ml, Abcam, UK]). RNA Isolation and Reverse Transcription-quantitative PCR (RT-qPCR) Analysis Next day sections were washed (2 min each) with PBS and incubated (1.5 hours at room temperature) with the secondary antibody (goat polyclonal anti- Quantification of Serum Cytokine Levels Blood was collected from mice at different time points. Serum levels of ALT, AST, ALP, GLDH were determined by Diagnostic Laboratories in the Royal Veterinary College, London, UK. Immuno-histochemistry (IHC) of Mouse Livers The staining procedure was performed according to the supplier recommendations, using either the provided secondary biotinylated anti-mouse IgG revealed by fluorescein-tagged Avidin (1/250, Vector Laboratories) or a goat polyclonal anti-mouse IgG labeled with Cy3 (1/250, Jackson ImmunoResearch Laboratories Inc.). Then liver sections were washed with PBS and mounted in DAPI and antifade containing medium (Vectashield mounting medium, Vector Laboratories, UK). Slides were visualised under epi-fluorescence microscope (Zeiss Axio Observer). Paraformalde- hyde fixed samples were paraffin-embedded and liver sections were stained with Haematoxylin and Eosin (H&E) by Diagnostic Laboratories in the Royal Veterinary College, London, UK. Random images were captured by light microscopy for different treatment groups. rabbit IgG labeled with Cy3, 1/250, Jackson ImmunoResearch Laboratories Inc.). For SSEA1 immunostain, a mouse monoclonal antibody (ab16285, 20 mg/ml, Abcam,UK) was used, which make it more suitable to use a specific immunodetection kit to localize mouse antibodies on mouse tissues (vector MOM immunodection kit, Vector Laboratories). The endogenous liver avidin/biotin were blocked with Avidin/blocking Kit (Vector Laboratories, UK). The staining procedure was performed according to the supplier recommendations, using either the provided secondary biotinylated anti-mouse IgG revealed by fluorescein-tagged Avidin (1/250, Vector Laboratories) or a goat polyclonal anti-mouse IgG labeled with Cy3 (1/250, Jackson ImmunoResearch Laboratories Inc.). Then liver sections were washed with PBS and mounted in DAPI and antifade containing medium (Vectashield mounting medium, Vector Laboratories, UK). Slides were visualised under epi-fluorescence microscope (Zeiss Axio Observer). Paraformalde- hyde fixed samples were paraffin-embedded and liver sections were stained with Haematoxylin and Eosin (H&E) by Diagnostic Laboratories in the Royal Veterinary College, London, UK. Random images were captured by light microscopy for different treatment groups. for 30 min with the BCIP/NBT liquid substrate system. Color development was stopped by rinsing with water. In order to determine ALP activity in the tissue, frozen liver sections were prepared as described above and incubated with BCIP/NBT liquid substrate system for 30 min. Sections were washed with water and mounted. Random images were captured by light microscopy (106) for saline, GFP or OKSM treatment groups. Statistical Analysis Experiments were performed with at least four animals per group. Statistical analysis was performed by analysis of variance and Tukey’s pairwise comparison using SPSS software, version 16.0. Immuno-histochemistry (IHC) of Mouse Livers Livers were perfused with 10 mL HBSS, pre-warmed at 37uC, then immediately immersed into isopentane, pre-cooled in liquid nitrogen or into stabilized 4% paraformaldehyde for fixation. Frozen livers were stored at 280uC until further processing. Fourteen micron thick sections were prepared on a Cryomicro- PLOS ONE \| www.plosone.org January 2013 \| Volume 8 \| Issue 1 \| e54754 4 In Vivo Cell Reprogramming in Mouse Tissue Figure 4. In vivo cell reprogramming in TNG-A mice. TNG-A mice were HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline and at days 2, 4, RT-qPCR analysis of hepatocytes was performed to determine the relative gene expression of: (a) transfected transcription factors (OKSM) and (b) endogenous pluripotency markers. All gene expression levels were normalized to HTV-injected saline group (p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV- injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (c) flow cytometry analysis of GFP positive cells in liver extracts; (d) liver tissue frozen and sectioned to image GFP-positive cells with fluorescence microscopy at day 4 (10x). doi:10.1371/journal.pone.0054754.g004 Figure 4. In vivo cell reprogramming in TNG-A mice. TNG-A mice were HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline and at days 2, 4, RT-qPCR analysis of hepatocytes was performed to determine the relative gene expression of: (a) transfected transcription factors (OKSM) and (b) endogenous pluripotency markers. All gene expression levels were normalized to HTV-injected saline group (*p,0.05 indicates statistically significant differences between the expression levels of pluripotency markers in the OKSM and saline HTV- injected groups, obtained by the analysis of variance and Tukey’s pairwise comparison); (c) flow cytometry analysis of GFP positive cells in liver extracts; (d) liver tissue frozen and sectioned to image GFP-positive cells with fluorescence microscopy at day 4 (10x). doi:10.1371/journal.pone.0054754.g004 rabbit IgG labeled with Cy3, 1/250, Jackson ImmunoResearch Laboratories Inc.). For SSEA1 immunostain, a mouse monoclonal antibody (ab16285, 20 mg/ml, Abcam,UK) was used, which make it more suitable to use a specific immunodetection kit to localize mouse antibodies on mouse tissues (vector MOM immunodection kit, Vector Laboratories). The endogenous liver avidin/biotin were blocked with Avidin/blocking Kit (Vector Laboratories, UK). Periodic Acid Schiff (PAS) Staining of Liver Sections Periodic Acid Schiff (PAS) Staining of Liver Sections Liver samples were fixed in paraformaldehyde, paraffin- embedded and liver sections were stained with PAS stain (Sigma, UK). Random images were captured by light microscopy for saline or OKSM treatment groups. A significant increase in the gene expression of all transduced transcription factors was observed on day 2 that decreased over time (Figure 1a). At the same time endogenous pluripotency markers were upregulated, reached peak values on day 4 and decreased to background levels from day 8 onward (Figure 1b). Protein expression in the hepatocyte extracts from the transduced Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections Results Mice were injected by HTV with an equimolar mix of two plasmids, pCX-OKS-2A and pCX-cMyc, encoding for the OKS and M reprogramming factors respectively. HTV injection of plasmid DNA has been known to result in high levels of gene expression in hepatocytes [22,23] and primary hepatocytes from OKSM-injected animals were extracted at different time points here. RT-qPCR and flow cytometry were used to analyze the expression levels for various reprogramming (Oct3/4, Sox2, c-Myc), pluripotency (Nanog, Ecat1, Rex1, Cripto, Gdf3 and endogenous Oct3/4, Klf,4 or Sox2) and hepatocyte markers (Alb, Trf, Aat) directly after extraction of the hepatocytes at different time points after HTV injection. Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections ALP activity staining was performed using the BCIP/NBT liquid substrate system (Sigma UK). Methanol-fixed cell cultures were first washed with HEPES buffer and then incubated at 37uC January 2013 \| Volume 8 \| Issue 1 \| e54754 PLOS ONE \| www.plosone.org 5 January 2013 \| Volume 8 \| Issue 1 \| e54754 In Vivo Cell Reprogramming in Mouse Tissue Figure 5. The effect of in vivo cell reprogramming on hepatotoxicity and liver damage. Balb/C mice HTV injected with either 75 mg of pCX- OKS-2A and 75 mg pCX-cMyc in 0.9% saline or 0.9% saline only. On days 2, 4, 8, 12, 50 and 120 livers and sera were isolated. (a) H&E staining of liver sections; (b) levels of liver enzymes and (c) albumin were analyzed; (d) liver sections were PAS stained to determine glycogen storage levels. Representative images were captured with light microscopy (10x). Scale bars represent 100 mm. doi:10.1371/journal.pone.0054754.g005 Figure 5. The effect of in vivo cell reprogramming on hepatotoxicity and liver damage. Balb/C mice HTV injected with either 75 mg of pCX- OKS-2A and 75 mg pCX-cMyc in 0.9% saline or 0.9% saline only. On days 2, 4, 8, 12, 50 and 120 livers and sera were isolated. (a) H&E staining of liver sections; (b) levels of liver enzymes and (c) albumin were analyzed; (d) liver sections were PAS stained to determine glycogen storage levels. Representative images were captured with light microscopy (10x). Scale bars represent 100 mm. doi:10.1371/journal.pone.0054754.g005 eGFP) plasmid DNA at an escalated dose regime (Figure 2a). A sharp increase in the levels of gene expression profile of the transduced reprogramming and other endogenous pluripotent factors on day 4 after administration was observed only in the case of OKSM plasmid DNA injections, and the levels of up-regulation reached a plateau at 75 mg/animal. c-Myc is a known oncogene and its use in cell reprogramming (in particular for in vivo strategies) is preferable to be avoided. However, a decrease in the efficiency of reprogramming has been previously described in the absence of this factor [30]. The effect of the presence or absence of c-Myc in the reprogramming cocktail was also investigated in the present study (Figure 2b). Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections In vivo cell reprogramming could be achieved without c-Myc, however the levels of pluripotency markers Nanog and Ecat1 observed were significantly lower compared to use of the OKSM cocktail, in line with the previous reports omitting c-Myc from in vitro cell reprogramming. liver tissue using flow cytometry indicated that on day 1 only OCT3/4 was expressed, whereas by day 4 both OCT3/4 and NANOG positive cells were detected (Figure 1c). Examination of the expression levels of hepatocyte markers by RT-qPCR indicated that on day 2 they were similar to those of animals HTV injected with saline, while on day 4 significant down- regulation in hepatocyte marker expression was considered as another indication of cell reprogramming, manifested by the de- differentiation of hepatocytes (Figure 1d). Later than day 8, hepatocyte markers returned back to control levels, while on day 24 increased levels of these markers were obtained. liver tissue using flow cytometry indicated that on day 1 only OCT3/4 was expressed, whereas by day 4 both OCT3/4 and NANOG positive cells were detected (Figure 1c). Examination of the expression levels of hepatocyte markers by RT-qPCR indicated that on day 2 they were similar to those of animals HTV injected with saline, while on day 4 significant down- regulation in hepatocyte marker expression was considered as another indication of cell reprogramming, manifested by the de- differentiation of hepatocytes (Figure 1d). Later than day 8, hepatocyte markers returned back to control levels, while on day 24 increased levels of these markers were obtained. Transgene expression by HTV injection was monitored separately by injection of control plasmids encoding for GFP- luciferase (pCMV-GFP-luc) and eGFP (pCAG-eGFP). Figure S1 confirms that hepatocytes were the targeted cell population in the liver following HTV injection as previously reported. The kinetics of transgene expression in the liver following HTV injection with the OKSM plasmids indicated that from day 9 onward, pluripotency factors began to be down-regulated reaching naive levels by day 22 (Figure S2). Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections H&E staining of liver tissues indicated that HTV injection of OKSM plasmids did not lead to any tissue damage beyond day 2, nor development of teratomas at later time points, with all liver sections and animals exhibiting healthy structural morphology and behavior with no signs of dysplasia (Figure 5a). Serum levels were also analyzed for liver enzymes over the same period, with no aberrant change in the levels of ALT, AST and GLDH between the saline-injected and OKSM-injected groups (Figure 5b). Albumin levels and glycogen staining of liver sections (Figure 5c & d) further confirmed no hepatic structural or functional abnormality throughout the course of the study for any of the animals. These findings in conjunction with the gene and protein expression analyses of the liver tissue (Figure 1) suggested that transcriptional cell reprogramming was occurring in vivo rapidly after HTV-injection of the OKSM factors, similar to recent direct reprogramming (transdifferentiation) studies [29,30,31]. We spec- ulate that the endogenous tissue milieu (growth factors, homeo- static signaling cues) leads to rapid loss of the transcriptionally reprogrammed cells into phenotypically normal, functioning hepatocytes. However, much more work using lineage-tracing techniques is needed to elucidate the mechanisms and extent of in vivo cell reprogramming to a pluripotent state. Further progress from this work will be the isolation and characterization of the transcriptionally reprogrammed cells from the primary hepatocyte extracts. Determination of the quality and level of functional pluripotent capacity of those cells, their genomic stability and epigenetic character, along with all the comparative studies currently undertaken by numerous laboratories to define the similarities and differences between iPS and ES cells will be needed. Also, the mechanistic basis of direct in vivo induced reprogramming of somatic cells to a pluripotent state and the dissection of the reasons behind the rapid kinetics observed, the extent and possible implications of transient cell reprogramming to ground-state pluripotency that could possibly take place transiently in situ, and the mechanisms by which the tissue microenvironment drives rapid re-differentiation in the host cells will have to be determined. Enhancement of the level of in vivo cell reprogramming to a pluripotent state can also be envisaged by improvements in gene transfer methodology, dosing regimen and vector construct design. Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections Even though upregulation of eGFP and pluripotency markers in the TNG-A mice strongly suggest that HTV injection of OKSM plasmids can result in adult somatic cell reprogramming toward pluripotency in vivo, further studies are needed to de- termine the level of pluripotency achieved, similar to previous Nanog-activity studies using ES cells [34], as well as more detailed characterization of the extracted reprogrammed cells in relation to the loss of hepatocyte phenotype. Moreover, the in vivo cell reprogramming towards pluripotency in fully developed mammals shown here is also in agreement with recently reported cell reprogramming for non-mammalian tissue [35] and did not lead to any structural or functional side effects in the liver, nor did it lead to any manifestation of carcinogenesis or teratoma formation. Further progress from this work will be the isolation and characterization of the transcriptionally reprogrammed cells from the primary hepatocyte extracts. Determination of the quality and level of functional pluripotent capacity of those cells, their genomic stability and epigenetic character, along with all the comparative studies currently undertaken by numerous laboratories to define the similarities and differences between iPS and ES cells will be needed. Also, the mechanistic basis of direct in vivo induced reprogramming of somatic cells to a pluripotent state and the dissection of the reasons behind the rapid kinetics observed, the extent and possible implications of transient cell reprogramming to ground-state pluripotency that could possibly take place transiently in situ, and the mechanisms by which the tissue microenvironment drives rapid re-differentiation in the host cells will have to be determined. g y py ( g ) One of the key concerns by induction of reprogramming towards pluripotency in vivo may be the spontaneous occurrence of teratomas within the tissues where reprogrammed cells are generated [12,19]. To address this, animals HTV-injected with the reprogramming factors were kept for a period of 120 days and at frequent intervals (days 2, 4, 8, 12, 50 and 120) different groups were analyzed heamatologically and histologically (Figure 5a). HTV injection of plasmid DNA can result in moderate tissue damage manifested by increased serum levels of liver enzymes at early time points [26], which was also observed here at day 2 for both saline and OKSM injected groups (Figure 5a & S4). Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections Finally, it can be envisioned that improved (and less invasive) protocols and technologies for the possible generation of in vivo induced pluripotent cells may lead to a general methodology for the extraction of autologous cells from tissue (liver or other) rapidly and with minimization of risks (mutagenesis, media contamina- tion, xenobiotic reagents, etc) from culturing conditions and protocols. Overall, this study offers the basis of a rapid, virus-free, efficient, and in the absence of culturing complications in vivo platform to study reprogramming mechanisms in situ using adult, fully developed, somatic tissue. Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections The effect of plasmid dose was then studied by HTV injections of OKSM against control (pCAG- To further interrogate the occurrence of in vivo induced cell reprogramming in the liver by the forced expression of the OKSM transcription factors, tissue sections from transfected Balb/C mice were directly stained immunohistochemically (IHC) at day 4 (post- January 2013 \| Volume 8 \| Issue 1 \| e54754 PLOS ONE \| www.plosone.org January 2013 \| Volume 8 \| Issue 1 \| e54754 6 In Vivo Cell Reprogramming in Mouse Tissue oocyte nuclear transfer methodologies [32]. Zhou et al. have previously reported rapid and highly efficient in vivo conversion of pancreatic exocrine cells directly into insulin secreting b-cells without reverting to a pluripotent state (i.e. by transdifferentiation) using adenovirus-mediated transcription factor (not OKSM) overexpression [33]. Very recently, three studies have also offered proof-of-concept evidence that transcriptionally-induced, direct transdifferentiation from fibroblasts to cardiomyocyte-like cells in infracted heart tissue in situ can offer therapeutic benefits [29,30,31]. In all such previous work direct conversion to different lineages is reported without reprogramming towards a plutipotent state, shown to take place rapidly and efficiently (compared to in vitro transdifferentiation methodologies). From our studies herein, in vivo cell reprogramming by overexpression of OKSM transcription factors is also shown to be rapid and efficient enough to be detected in situ, however transiently maintained. According to our semi-quantitative analysis based on FACS and IHC of liver sections from transfected Balb/C and TNG-A mice, it took place efficiently in the order of 5–15% of the total hepatocyte population. Even though upregulation of eGFP and pluripotency markers in the TNG-A mice strongly suggest that HTV injection of OKSM plasmids can result in adult somatic cell reprogramming toward pluripotency in vivo, further studies are needed to de- termine the level of pluripotency achieved, similar to previous Nanog-activity studies using ES cells [34], as well as more detailed characterization of the extracted reprogrammed cells in relation to the loss of hepatocyte phenotype. Moreover, the in vivo cell reprogramming towards pluripotency in fully developed mammals shown here is also in agreement with recently reported cell reprogramming for non-mammalian tissue [35] and did not lead to any structural or functional side effects in the liver, nor did it lead to any manifestation of carcinogenesis or teratoma formation. HTV) for different pluripotency markers (Oct3/4, Sox2, Nanog and ALP). January 2013 \| Volume 8 \| Issue 1 \| e54754 Alkaline Phosphatase (ALP) Staining of Cell Cultures and Liver Sections Figure 3 shows that distinctive IHC-positive cells were obtained for all four markers only in the case of animals injected with the OKSM plasmids compared to saline or GFP plasmid HTV-injected groups. Positive staining for Nanog was obtained reproducibly in all liver sections in all OKSM-transfected animals indicating the presence of transcriptionally reprogrammed cells throughout the liver tissue (Figure S3). Considering the critical role of Nanog in the control of pluripotency, we performed a separate experiment using the transgenic strain TNG-A that carries the eGFP reporter inserted into the Nanog locus [28]. HTV injection of the OKSM plasmids in TNG-A mice showed enhanced gene expression in both reprogramming and endoge- nous pluripotency markers on days 2 and 4 post-injection (Figure 4a & b). Hepatocyte extracts from OKSM injected TNG-A mice on day 4 were analyzed by flow cytometry for eGFP protein expression, revealing that 6–15% of the total cell population were eGFP positive (Figure 4c). Lastly, further confirmation of the generation of reprogrammed cells in the liver was offered by the presence of eGFP-positive cells in frozen tissue sections imaged by fluorescence microscopy (Figure 4d). oocyte nuclear transfer methodologies [32]. Zhou et al. have previously reported rapid and highly efficient in vivo conversion of pancreatic exocrine cells directly into insulin secreting b-cells without reverting to a pluripotent state (i.e. by transdifferentiation) using adenovirus-mediated transcription factor (not OKSM) overexpression [33]. Very recently, three studies have also offered proof-of-concept evidence that transcriptionally-induced, direct transdifferentiation from fibroblasts to cardiomyocyte-like cells in infracted heart tissue in situ can offer therapeutic benefits [29,30,31]. In all such previous work direct conversion to different lineages is reported without reprogramming towards a plutipotent state, shown to take place rapidly and efficiently (compared to in vitro transdifferentiation methodologies). From our studies herein, in vivo cell reprogramming by overexpression of OKSM transcription factors is also shown to be rapid and efficient enough to be detected in situ, however transiently maintained. According to our semi-quantitative analysis based on FACS and IHC of liver sections from transfected Balb/C and TNG-A mice, it took place efficiently in the order of 5–15% of the total hepatocyte population. PLOS ONE \| www.plosone.org References 1. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, et al. (2007) Induction of Pluripotent Stem Cells from Adult Human Fibroblasts by Defined Factors. Cell 131: 861–872. 16. Lister R, Pelizzola M, Kida YS, Hawkins RD, Nery JR, et al. (2011) Hotspots of aberrant epigenomic reprogramming in human induced pluripotent stem cells. Nature 471: 68–73. 17. Jaenisch R, Young R (2008) Stem Cells, the Molecular Circuitry of Pluripotency and Nuclear Reprogramming. Cell 132: 567–582. 2. Yu J, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J, Frane JL, et al. (2007) Induced Pluripotent Stem Cell Lines Derived from Human Somatic Cells. Science 318: 1917–1920. 18. Plath K, Lowry WE (2011) Progress in understanding reprogramming to the induced pluripotent state. Nat Rev Genet 12: 253–265. 3. Stadtfeld M, Nagaya M, Utikal J, Weir G, Hochedlinger K (2008) Induced Pluripotent Stem Cells Generated Without Viral Integration. Science 322: 945– 949. 19. Miura K, Okada Y, Aoi T, Okada A, Takahashi K, et al. (2009) Variation in the safety of induced pluripotent stem cell lines. Nat Biotech 27: 743–745. 20. Dolgin E (2011) Flaw in induced-stem-cell model. Nature 470: 1 4. Okita K, Nakagawa M, Hyenjong H, Ichisaka T, Yamanaka S (2008) Generation of Mouse Induced Pluripotent Stem Cells Without Viral Vectors. Science 322: 949–953. 21. Pera M (2011) The dark side of pluripotency. Nature 471: 46–47. 22. Liu F, Song YK, Liu D (1999) Hydrodynamics-based transfection 22. Liu F, Song YK, Liu D (1999) Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther 6: 1258–1266. systemic administration of plasmid DNA. Gene Ther 6: 1258–12 5. Gonzalez F, Barragan Monasterio M, Tiscornia G, Montserrat Pulido N, Vassena R, et al. (2009) Generation of mouse-induced pluripotent stem cells by transient expression of a single nonviral polycistronic vector. Proc Natl Acad Sci USA 106: 8918–8922. 23. Zhang G, Budker V, Wolff JA (1999) High Levels of Foreign Gene Expression in Hepatocytes after Tail Vein Injections of Naked Plasmid DNA. Human Gene Therapy 10: 1735–1737. 24. Andrianaivo F, Lecocq M, Wattiaux-De Coninck S, Wattiaux R, Jadot M (2004) Hydrodynamics-based transfection of the liver: entrance into hepatocytes of DNA that causes expression takes place very early after injection. The Journal of Gene Medicine 6: 877–883. 6. Woltjen K, Michael IP, Mohseni P, Desai R, Mileikovsky M, et al. (2009) piggyBac transposition reprograms fibroblasts to induced pluripotent stem cells. Nature 458: 766–770. 7. Supporting Information Figure S1 Transfection after HTV injection. Balb/C mice were HTV injected with pCMV?GFP-Luc in 0.9% saline and hepatocytes were isolated after 24 h and analyzed for (a) luciferase activity by luciferase assay (b) GFP and Luc gene expression by real-time PCR. Balb/C mice were HTV injected with pCAG?GFP in 0.9% saline and hepatocytes were isolated after 24 h and analyzed for (c) eGFP gene expression by real-time PCR (d) transfection efficiency by FACS. (e) Liver samples were frozen and sectioned to image transfected hepatocytes under (i) differential interference contrast (DIC) illumination or (ii) blue light excitation (10x). (TIF) Figure S1 Transfection after HTV injection. Balb/C mice were HTV injected with pCMV?GFP-Luc in 0.9% saline and hepatocytes were isolated after 24 h and analyzed for (a) luciferase activity by luciferase assay (b) GFP and Luc gene expression by real-time PCR. Balb/C mice were HTV injected with pCAG?GFP in 0.9% saline and hepatocytes were isolated after 24 h and analyzed for (c) eGFP gene expression by real-time PCR (d) transfection efficiency by FACS. (e) Liver samples were frozen and sectioned to image transfected hepatocytes under (i) differential interference contrast (DIC) illumination or (ii) blue light excitation (10x). (TIF) Figure S5 Plasmid DNA maps used in this study. (a) pCX- OKS-2A, (b) pCX-c-Myc, (c) pCAG-GFP and (d) pGFP-Luc plasmids. ( ) Table S1 Primer seqeunces used in this study. (TIF) Figure S2 Flow cytometry analysis of in vivo-reprogrammed hepatocyte extracts. Balb/C mice HTV injected with 0.9% saline alone, 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline, or 150 mg of pCAG-GFP in 0.9% saline. On days 1, 4, 9, 11 and 22, hepatocytes were isolated and stained for OCT3/4, SOX2 and NANOG. (TIF) Figure S3 NANOG immunofluorescence staining of different liver sections after HTV injection of reprogramming plasmids. Balb/C mice HTV injected with 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline. On day 4, liver tissue was collected and frozen tissue sections were immunostained with an anti-NANOG antibody. Scale bars represent 100 mm. In Vivo Cell Reprogramming in Mouse Tissue Figure S4 The effect of HTV injection of plasmids on liver damage at early time points. Balb/C mice HTV injected with either 75 mg of pCX-OKS-2A and 75 mg pCX-cMyc in 0.9% saline or 0.9% saline only. On days 2, 4, 8, 12 sera were isolated and analyzed for the levels of liver enzymes. (TIF) References Warren L, Manos PD, Ahfeldt T, Loh Y-H, Li H, et al. (2010) Highly Efficient Reprogramming to Pluripotency and Directed Differentiation of Human Cells with Synthetic Modified mRNA. Cell stem cell 7: 618–630. 25. Herweijer H, Wolff JA (2006) Gene therapy progress and prospects: Hydrodynamic gene delivery. Gene Ther 14: 99–107. 26. Suda T, Liu D (2007) Hydrodynamic Gene Delivery: Its Principles and Applications. Mol Ther 15: 2063–2069. 8. Kim D, Kim C-H, Moon J-I, Chung Y-G, Chang M-Y, et al. (2009) Generation of Human Induced Pluripotent Stem Cells by Direct Delivery of Reprogram- ming Proteins. Cell stem cell 4: 472–476. 27. Khorsandi SE, Bachellier P, Weber JC, Greget M, Jaeck D, et al. (2008) Minimally invasive and selective hydrodynamic gene therapy of liver segments in the pig and human. Cancer Gene Ther 15: 225–230. g 9. Zhou H, Wu S, Joo JY, Zhu S, Han DW, et al. (2009) Generation of Induced Pluripotent Stem Cells Using Recombinant Proteins. Cell stem cell 4: 381–384. 28. Chambers I, Silva J, Colby D, Nichols J, Nijmeijer B, et al. (2007) Nanog safeguards pluripotency and mediates germline development. Nature 450: 1230– 1234. 10. Anokye-Danso F, Trivedi Chinmay M, Juhr D, Gupta M, Cui Z, et al. (2011) Highly Efficient miRNA-Mediated Reprogramming of Mouse and Human Somatic Cells to Pluripotency. Cell stem cell 8: 376–388. 29. Jayawardena TM, Egemnazarov B, Finch EA, Zhang L, Payne JA, et al. (2012) MicroRNA-Mediated In Vitro and In Vivo Direct Reprogramming of Cardiac Fibroblasts to Cardiomyocytes/Novelty and Significance. Circ Res 110: 1465– 1473. 11. Takahashi K, Yamanaka S (2006) Induction of Pluripotent Stem Cells from Mouse Embryonic and Adult Fibroblast Cultures by Defined Factors. Cell 126: 663–676. 12. Ben-David U, Benvenisty N (2011) The tumorigenicity of human embryonic and induced pluripotent stem cells. Nat Rev Cancer 11: 268–277. 30. Qian L, Huang Y, Spencer CI, Foley A, Vedantham V, et al. (2012) In vivo reprogramming of murine cardiac fibroblasts into induced cardiomyocytes. Nature 485: 593–598. 13. Gonzalez F, Boue S, Belmonte JCI (2011) Methods for making induced pluripotent stem cells: reprogramming a la carte. Nat Rev Genet 12: 231–242. 31. Song K, Nam Y-J, Luo X, Qi X, Tan W, et al. (2012) Heart repair by reprogramming non-myocytes with cardiac transcription factors. Nature 485: 599–604. 14. Gore A, Li Z, Fung H-L, Young JE, Agarwal S, et al. (2011) Somatic coding mutations in human induced pluripotent stem cells. Author Contributions Conceived and designed the experiments: KK AY. Performed the experiments: AY IL CB. Analyzed the data: AY IL CB KK. Contributed reagents/materials/analysis tools: AY IL CB KK. Wrote the paper: AY IL CB KK. Conceived and designed the experiments: KK AY. Performed the experiments: AY IL CB. Analyzed the data: AY IL CB KK. Contributed reagents/materials/analysis tools: AY IL CB KK. Wrote the paper: AY IL CB KK. Conceived and designed the experiments: KK AY. Performed the experiments: AY IL CB. Analyzed the data: AY IL CB KK. Contributed reagents/materials/analysis tools: AY IL CB KK. Wrote the paper: AY IL CB KK. Discussion This study provides previously unreported evidence that direct in vivo transcriptional cell reprogramming towards the pluripotent state in adult mammalian somatic (hepatocyte) cells is possible using non-viral transfection of plasmids encoding the OKSM (or OKS) transcription factors. This occurs following rapid kinetics (within 24–48 h) that are, in principle, reminiscent of the kinetics described for somatic cell reprogramming using the egg and January 2013 \| Volume 8 \| Issue 1 \| e54754 7 PLOS ONE \| www.plosone.org Acknowledgments The authors would like to thank Prof.A.Smith of The Wellcome Trust Centre for Stem Cell Research, University of Cambridge, UK for the kind gift of the TGN-A cells. KK would also like to acknowledge Prof.J.Gurdon and his team at the Wellcome Trust/Cancer Research UK Gurdon Institute, University of Cambridge for the useful discussions of the data included in this work. 35. Vivien C, Scerbo P, Girardot F, Le Blay K, Demeneix BA, et al. (2012) Non- viral Expression of Mouse Oct4, Sox2, and Klf4 Transcription Factors Efficiently Reprograms Tadpole Muscle Fibers in Vivo. J Biol Chem 287: 7427–7435. 34. Kalmar T, Lim C, Hayward P, Mun˜oz-Descalzo S, Nichols J, et al. (2009) Regulated Fluctuations in Nanog Expression Mediate Cell Fate Decisions in Embryonic Stem Cells. PLoS Biol 7. 36. ten Hagen TLM, Van Vianen W, Bakker-Woudenberg IAJM (1996) Isolation and characterization of murine Kupffer cells and splenic macrophages. J Immunol Meth 193: 81–91. 33. Zhou Q, Brown J, Kanarek A, Rajagopal J, Melton DA (2008) In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455: 627– 632. In Vivo Cell Reprogramming in Mouse Tissue References Nature 471: 63–67. 15. Hussein SM, Batada NN, Vuoristo S, Ching RW, Autio R, et al. (2011) Copy number variation and selection during reprogramming to pluripotency. Nature 471: 58–62. 32. Jullien J, Pasque V, Halley-Stott RP, Miyamoto K, Gurdon JB (2011) Mechanisms of nuclear reprogramming by eggs and oocytes: a deterministic process? Nat Rev Mol Cell Biol 12: 453–459. PLOS ONE \| www.plosone.org January 2013 \| Volume 8 \| Issue 1 \| e54754 January 2013 \| Volume 8 \| Issue 1 \| e54754 8 In Vivo Cell Reprogramming in Mouse Tissue In Vivo Cell Reprogramming in Mouse Tissue PLOS ONE \| www.plosone.org January 2013 \| Volume 8 \| Issue 1 \| e54754 PLOS ONE \| www.plosone.org 9 9
W4229026845.txt	https://www.mdpi.com/2075-4442/10/5/87/pdf?version=1651752804	en		Electrical Field Strength in Rough Infinite Line Contact Elastohydrodynamic Conjunctions	Lubricants	2,022	cc-by	8,228	lubricants Article Electrical Field Strength in Rough Infinite Line Contact Elastohydrodynamic Conjunctions Samuel A. Morris 1 , Michael Leighton 2 1 2 * and Nicholas J. Morris 1, * Wolfson School of Mechanical, Manufacturing and Electrical Engineering, Loughborough University, Loughborough LE11 3TU, UK; s.a.morris-17@student.lboro.ac.uk AVL List GmbH, Hans-List-Platz 1, 8020 Graz, Austria; michael.leighton@avl.com Correspondence: n.j.morris@lboro.ac.uk Abstract: Rolling element bearings are required to operate in a variety of use cases that determine voltage potentials will form between the rolling elements and races. When the electrical field strength causes the dielectric breakdown of the intermediary lubricant film electrical discharge can damage the bearing surfaces. To reduce the prevalence and severity of electrical discharge machining an improved understanding of the coupled electrical and mechanical behavior is necessary. This paper aims to improve understanding of the problem through a combined elastohydrodynamic and electrostatic numerical study of charged elastohydrodynamic conjunctions. The results show the effect of amplitude reduction means that for typical surface topographies found in EHL conjunctions the maximum field strength is adequately predicted by the elastohydrodynamic minimum film thickness and potential difference. The paper also indicates the width of the elevated electrical field strength region is dependent on EHL parameters which could have important implications on the magnitude of current density during dielectric breakdown. Keywords: tribology; EDM; electrical erosion; EHL; electrical field strength; roughness; bearing; wear Citation: Morris, S.A.; Leighton, M.; Morris, N.J. Electrical Field Strength in Rough Infinite Line Contact Elastohydrodynamic Conjunctions. Lubricants 2022, 10, 87. https:// doi.org/10.3390/lubricants10050087 Received: 23 February 2022 Accepted: 28 April 2022 Published: 5 May 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1. Introduction Electrical voltages form between bearing rolling elements in many applications such as electric vehicle powertrains [1], aircraft [2], rotorcraft [3], rolling stock axles [4], wind turbines [5] and spacecraft [6]. In the former common mode shaft voltages resulting from the inverter control systems lead to coronal discharge events at the bearing interfaces [1] and in space crafts a correlation between proximity to geomagnetic storms and electrical discharge initiated bearing failure has been observed [6]. In turbines, a tribostatic effect due to the lubricant flow over the bearing has resulted in charge build up between the rolling elements and race [7]. Electrical erosion and specifically electrical pitting otherwise known as Electrical Discharge Erosion (EDE) [8] (or electrical discharge machining wear EDM) are both common bearing failure modes. These events occur when a bearing ring to rolling element voltage exists and exceeds the ignition voltage [8]. When the ignition voltage is reached the electrical field strength in the lubricant film exceeds its dielectric breakdown strength [9]. An excessive current density discharge event can lead to localised high temperature regions on the surface, melt pools and in turn pit formation [8]. Electrical erosion causes frosting if the current is low and pitting if it is high. If the frosting damage continues, the damage increases until pits form. After pitting, the continued operation leads to fluting damage on the surface or surface fatigue of the race [2]. As well as damaging the surface the process can cause carbonisation and oxidation of the lubricant. An experimental study conducted by Joshi and Blennow [10] on the electrical properties of lubricants showed that the number and duration of current discharge events was dependant on the voltage, shaft speed and bearing load. Bechev et al. [11] conducted an analysis of lubricant properties for use in electric motor bearings. They identified differences in breakdown voltage of the lubricant measured ex-situ and the breakdown voltage Lubricants 2022, 10, 87. https://doi.org/10.3390/lubricants10050087 https://www.mdpi.com/journal/lubricants Lubricants 2022, 10, x FOR PEER REVIEW Lubricants 2022, 10, 87 2 of 17 2 of 16 analysis of lubricant properties for use in electric motor bearings. They identified differences in breakdown voltage of the lubricant measured ex-situ and the breakdown voltage of the elastohydrodynamic film within the bearing assembly. The difference in idealised of the elastohydrodynamic film within the bearing assembly. The difference in idealised contact geometry of the ex-situ measurements andand pressure-dielectric strength relation of of contact geometry of the ex-situ measurements pressure-dielectric strength relation the the lubricant were suggested as possible reasons for the differences. lubricant were suggested as possible reasons for the differences. Within inverter fed fed variable speed e-motors, which are are prevalent in electric vehicles, Within inverter variable speed e-motors, which prevalent in electric vehicles, the the charge build-up has been investigated [12–14]. Link [14] gave the main causes of static charge build-up has been investigated [12–14]. Link [14] gave the main causes of static charge accumulation as; as; motor magnetic asymmetry, power charge accumulation motor magnetic asymmetry, powersupply supplyasymmetry asymmetryand andvoltvoltage agetransients transients(dV/dt). (dV/dt). With Withrecent recenttrends trendstowards towardshigh high switching frequencies and voltswitching frequencies and voltages in agesautomotive in automotive inverter systems, fuelled by the drive efficiency, automotive e-mo- are inverter systems, fuelled by the drive for for efficiency, automotive e-motors torsset aretoset to become much more vulnerable to EDE failures. Bearing failures are already become much more vulnerable to EDE failures. Bearing failures are already the main the failure main failure of electric of electric drivesdrives [15,16].[15,16]. OneOne other technological trend maymay alsoalso introduce additional importance to electroother technological trend introduce additional importance to electrostatic discharge, the the selection of lubricants. Traditionally, e-motor bearings have been static discharge, selection of lubricants. Traditionally, e-motor bearings have been greased for for lifetime useuse but but as e-motor speeds increase, in pursuit of higher efficiencies, greased lifetime as e-motor speeds increase, in pursuit of higher efficiencies, forced oil lubrication is becoming more common. ThisThis is sometimes shared with the the rotor forced oil lubrication is becoming more common. is sometimes shared with rotor cooling system. The demand for environmentally sustainable lubricant alternatives is cooling system. The demand for environmentally sustainable lubricant alternativesalso is also driving a growing market in water-based lubricants, manymany of which have the benefits of driving a growing market in water-based lubricants, of which have the benefits veryoflow viscosity and good capacity but have quite different electricalelectrical conductivity very low viscosity andheat good heat capacity but have quite different conduc[17].Capacitive bearing models have beenhave created bycreated Schneider et al. [18] among others. tivity [17].Capacitive bearing models been by Schneider et al. [18] among Theothers. tribological and electrical the bearing showed there is little The tribological andanalysis electricalofanalysis of theconjunction bearing conjunction showed there is effect of the EHL Petrusevich spike [19]spike and minimum film thickness portion ofportion the hertzlittle effect of the EHL Petrusevich [19] and minimum film thickness of the on thecapacitance. bearing capacitance. Thisisfinding is in agreement with a significant ian hertzian region onregion the bearing This finding in agreement with a significant body bodywhich of work has capacitive used the capacitive of the Elasto-Hydrodynamic Lubriof work haswhich used the propertiesproperties of the Elasto-Hydrodynamic Lubrication cation (EHL) film lubricant film as atomethod to its measure its thickness, with the methodology (EHL) lubricant as a method measure thickness, with the methodology being being pioneered in 1960′s the early 1960’sAn [20,21]. An experimental study conducted by Sunahara pioneered in the early [20,21]. experimental study conducted by Sunahara et al. al. [22] investigated the location of electrical discharge within the hertzian contact [22]etinvestigated the location of electrical discharge eventsevents within the hertzian contact patch. They used a modification of the interferometry technique pioneered by Gohar patch. They used a modification of the interferometry technique pioneered by Gohar andand Cameron to capture light emission produced during a discharge event. It was Cameron [23][23] to capture thethe light emission produced during a discharge event. It was observed that the discharge locations were found in the highlighted regions shown observed that the discharge locations were found in the highlighted regions shown in Fig- in 1. ureFigure 1. Figure 1. Location of discharge events shown by Sunahara al. [22] in their paper “Preliminary Figure 1. Location of discharge events shown by Sunahara et al.et[22] in their paper “Preliminary measurements of electrical micropitting in grease-lubricated point contacts” in Tribology transacmeasurements of electrical micropitting in grease-lubricated point contacts” in Tribology transactions, copyright © Society of Tribiologists andand Lubrication Engineers, www.stle.org, reprinted by by tions, copyright © Society of Tribiologists Lubrication Engineers, www.stle.org, reprinted permission of Taylor & Francis Ltd. (Oxfordshire, UK), http://www.tandfonline.com (accessed on on permission of Taylor & Francis Ltd. (Oxfordshire, UK), http://www.tandfonline.com (accessed 20 January 2022) on behalf of Society of Tribiologists andand Lubrication Engineers, www.stle.org. 20 January 2022) on behalf of Society of Tribiologists Lubrication Engineers, www.stle.org. TheThe image shown in Figure 1 suggests thatthat the the location of discharge is associated image shown in Figure 1 suggests location of discharge is associated with the regions of minimum film thickness, high-curvature, and high-pressure. In the with the regions of minimum film thickness, high-curvature, and high-pressure. In the region of interest discussed in Figure 1 it 1was Johns-Rahnejat andand Gohar [24][24] andand SafaSafa andand region of interest discussed in Figure it was Johns-Rahnejat Gohar Gohar [25][25] whowho firstfirst managed to experimentally resolve the the Petrusevich pressure spike in in Gohar managed to experimentally resolve Petrusevich pressure spike point and line contacts respectively. It has also been shown however that microscale surface roughness can, under certain conditions cause a significant effect on EHL film thickness Lubricants 2022, 10, 87 3 of 16 and pressure distributions [26,27]. Venner and Napel [26] showed a significant flattening (amplitude reduction) of the microscale roughness in the EHL contact patch. The flattened asperities were accompanied by localised pressure spikes. Morales-Espejel [27] showed a complementary ripple amplitude model capable of accurately conforming to experimental micro EHL measurements. Experimental studies conducted by Becker and Abanteriba [2] on aircraft bearings proposed electrical discharge erosion took place between the asperities of the contiguous surfaces. At the same time there is an increased interest on the influence of lubricant physical properties on electrical discharge wear phenomena in rolling element bearings [28]. Despite the effect of amplitude reduction, it has been shown that the lambda ratio (oil film thickness over composite rms roughness) [29] plays an important role in predicting electrical discharge erosion wear in rolling element bearings [30]. A significant body of work has been conducted to determine lubricant electrical properties, experimental association between the lambda ratio and electrical discharge erosion wear. In addition, a significant amount of work has been conducted to predict surface roughness behaviour within elastohydrodynamic conjunctions, however, to date a numerical investigation of electrical field strength to indicate the location of discharge events in rough elastohydrodynamic contacts has not previously been published in open literature. This paper uses numerical analyses of infinite elastohydrodynamic line contacts and two-dimensional electrostatic problem to investigate if EHL microgeometry, lubricant density and microscale roughness has a significant influence on electrical field strength within dielectric lubricant films. 2. Methodology To investigate the electrical field strength in elastohydrodynamic conjunctions a one-dimensional EHL model and two-dimensional electrostatic model are employed. The elastohydrodynamic problem is solved to determine the surface deflection, fluid film thickness and lubricant density. These properties are then passed as inputs to the electrostatic problem. 2.1. Contact Geometry and Mechanics For simplicity an infinite line contact problem considering a pair of bodies which are axially aligned, cylindrical, counter formal and loaded against each other are considered. The two bodies rotate about their respective axes of cylindrical symmetry. The one-dimensional assumption is suitable for a range of conjunctions such as roller bearings and spur gears as a cross section at any axial position has the same tribological conditions (in the absence of Lubricants 2022, 10, x FOR PEER REVIEW 4 of 17 misalignment, thermo-mechanical distortions and edge effects). A diagram representing the modeling approach is shown in Figure 2. Figure 2. Depiction of elastohydrodynamic model formulation. Figure 2. Depiction of elastohydrodynamic model formulation. The dimensionless film thickness equation that describes the lubricant filled gap between the surfaces can be written as: 𝑯𝒊 = 𝑯𝟎𝟎 + 𝕽𝒊 − 𝟏 𝒏 𝑲𝒊𝒋 𝑷𝒋 + 𝑿𝟐𝒊 (1) Lubricants 2022, 10, 87 4 of 16 The dimensionless film thickness equation that describes the lubricant filled gap between the surfaces can be written as: Hi = H00 + <i − 1 π n ∑ Kij Pj + j =1 Xi2 2 (1) where < is the surface roughness profile which is provided at the beginning of the results section. Due to the 2D model the roughness is in effect considered as infinitely wide ridges running perpendicular to the direction of entrainment. H00 is an integration constant which is updated during the iterative load convergence scheme described in the solution procedure section. The third term describes the elastic deformation of the bodies resulting from the fluid film hydrodynamic pressure (P). While the last term describes the combined curvature of the two surfaces based on the dimensionless X coordinate. The surface roughness is considered as stationary as this study considers only pure rolling kinematics. Where the dimensionless X coordinate is defined as X = xb , where x is the coordinate and b is the hertzian contact half width: r 8wR b= (2) πE0 where w is the dimensional load per unit length and E’ is the composite elastic modulus combining the elastic modulus (E) and Poisson’s ratio (υ) of the two bodies. 1 − υ12 1 − υ22 2 = + E0 E1 E2 (3) And R is the composite radius of curvature of the two surfaces in contact and can be written as: 1 1 1 = + (4) R R1 R2 The penultimate term in Equation (1) can be discretised as described by Venner and Napel [26] and the deflection kernel can be described as: Kij = 1 1 1 1 i−j+ dX ln i − j + dX − 1 − i − j − dX ln i − j − dX − 1 2 2 2 2 (5) The elastic deformation considered above assumes a state of plane strain as is typical for analyses of this type however recent studies have shown potential refinement of the approach using plane stress assumptions [31,32]. 2.2. Hydrodynamics A one-dimensional stationary form of Reynolds Equation (6) is discretised as described by Venner and Napel [26]. d dP d(ρH ) e − =0 (6) dX dX dX The dimensionless: Pressure P and Reynolds equation Coefficient e terms are fully described in the Nomenclature and numerical values given in Tables 1 and 2. The discretised form of the equation is given as: ei− 1 Pi−1 − ei− 1 + ei+ 1 Pi + ei+ 1 Pi+1 ρ Hi − ρi−1 Hi−1 2 2 2 2 =0 (7) − i 2 dX dX where dX is the dimensionless grid spacing parameter and for convenience dimensionless ρH 3 terms are defined as: e = ηλ , P = ( p/ph ) and H = bhR 2 . A second order central difference of the Poiseuille term and first order upstream difference for the wedge term are employed to discretise the Reynolds equation as described by [26,33] due to the formulations improved stability [34]. The boundary conditions, labeled (B.C.) in Figure 2, are both set using Lubricants 2022, 10, 87 5 of 16 dimensionless atmospheric pressure ( p atm /ph ). The Moe’s dimensionless groups [35,36] are used to characterise the EHL contact. r 0 4 η0 u s L = αE (8) E0 R w η0 us −1/2 (9) E0 R E0 R where α is the pressure viscosity coefficient and us is the sum of the contacting surface velocities u1 and u2 : u s = u1 + u2 (10) M= Table 1. Lubricant physical properties. Parameter Value Units α η0 z po εr C1 C2 22 × 10−9 4 × 10−4 0.67 1.98 × 108 2.4 0.59 × 109 1.34 Pa−1 Pa s − Pa − Pa − Table 2. Geometric, Load, Elastic and Kinematic properties of the conjunction. Parameter Value Units E0 us R w X ε0 226 1, 2, 4 & 8 1.41 × 10−2 1.57 × 106 −4 to 2 8.854 × 10−12 GPa m s−1 m Nm−1 − Fm−1 In this paper only pure rolling is consider and therefore u1 = u2 . The load balance equation is written as: n −1 1 π dX ∑ Pj + Pj+1 = (11) 2 2 j =1 2.3. Lubricant Physical Properties The dielectric lubricant constant is dependent on the dimensionless lubricant density, ρ, which varies with pressure as described by Dowson and Higginson [37]: ρ= C1 + C2 p C1 + p (12) In the current study an isothermal assumption is employed and the dimensionless viscosity is described using Roelands [38] pressure viscosity relation as: η=e αp0 p z z [(1+ po ) −1] (13) where z is Roelands pressure viscosity parameter, po is a constant and α is the pressure viscosity index. There are limitations in the accuracy of Equation (13) to match experimentally measured data of real fluids [39,40]. In this study however where the specific lubricant is undefined the relation is deemed fit for purpose. In practice thermo-viscous behaviour and conjunctional thermal behaviour is often of importance and this would be an interesting addition to the current work [41]. 𝜼=𝒆 Lubricants 2022, 10, 87 𝜶𝒑𝟎 𝒑 𝒛 𝟏 𝟏 𝒛 𝒑𝒐 (13) where z is Roelands pressure viscosity parameter, 𝑝 is a constant and α is the pressure viscosity index. There are limitations in the accuracy of Equation (13) to match ex6 of 16 perimentally measured data of real fluids [39,40]. In this study however where the specific lubricant is undefined the relation is deemed fit for purpose. In practice thermo-viscous behaviour and conjunctional thermal behaviour is often of importance and this would be an interesting addition to model the current work [41].constant of the lubricant is required. The For the electrostatic the dielectric For the electrostatic model the dielectric constant of the lubricant is required. dielectric constant is calculated using the Clausius-Mossotti equation [18,42,43] usingThe the dielectric constant is calculated using the Clausius-Mossotti equation [18,42,43] using dimensionless density calculated using Equation (12) which is determined during the the dimensionless density calculated using Equation (12) which is determined during the elastohydrodynamic calculations. elastohydrodynamic calculations. ε r + 2 + 2( ε r − 1) ρ ε = 𝜺𝒓 + 𝟐 + 𝟐 𝜺𝒓 − 𝟏 𝝆 (14) (14) 𝜺 = ε r + 2 − ( ε r − 1) ρ 𝜺𝒓 + 𝟐 − 𝜺𝒓 − 𝟏 𝝆 where ε r is the reference dielectric constant measured at the same temperature and pressure where 𝜀 is the reference dielectric constant measured at the same temperature and presas the reference density [44]. sure as the reference density [44]. 2.4. Electrostatics 2.4. Electrostatics This paper investigates the electrical field intensity in an electrostatic condition during This investigates the electrical field intensity anat electrostatic condition which thepaper two electrically conductive contiguous surfacesinare a fixed potential of (±dur1V). ing which the two electrically conductive contiguous surfaces are at a fixed potential These values were selected to be in the same order of magnitude as those observed of in (±1V). These values were selected however to be in the same ordervalues of magnitude thoseparticular observed various applications [7,13,45,46] the specific have noasother in various applications [7,13,45,46] thefrom specific values haveelastohydrodynamic no other particular significance. The surface geometry however is provided the converged significance. The surface geometry is provided from the converged elastohydrodynamic model as shown in Figure 3. model as shown in Figure 3. Figure 3. Depiction of electrostatic model formulation. Figure 3. Depiction of electrostatic model formulation. The lubricant film which separates the surfaces is considered as a perfect dielecThe lubricant which separates the surfaces is considered as a of perfect tric which forms afilm continuous film between the surfaces. The effect polardielectric additive which forms a continuous film between the surfaces. The effect of polar additive constitconstituents of the lubricant (particularly detergents [47]), particulates [48] and other conuents of the lubricant (particularly detergents [47]), particulates [48] and other contamitaminants such as water are ignored in this analysis despite their potential importance nants as in water are ignored in this analysis despite their potential importance when when such present practice. The generalised Poisson equation shown below in Equation (15) present inthe practice. The generalised equation shown below in Equation (15) dedescribes electrostatic behaviour Poisson of the elastohydrodynamic contact [49]: scribes the electrostatic behaviour of the elastohydrodynamic contact [49]: − ρ (r ) ∇·[ε(r )∇V (r )] = (15) ε0 where the charge density ρ, dielectric constant ε and voltage potential V are functions of spatial position described by position vector r. The dielectric constant is calculated at each node from Equation (14) using the lubricant densities calculated during the elastohydrodynamic analysis. Finally, ε 0 is the permittivity of free space and should not be confused with the dielectric constant ε. The two dimensional Poisson equation is necessary as from an electrostatic perspective it is not clear if the contact can be idealised as 1D as the curvature of the surfaces may have an effect on the field strength. The problem is 1D in terms of the EHL problem as the pressure is assumed constant across the thickness of the film and symmetric through the depth of the contact. A successive over relaxed finite difference scheme is employed to solve the generalised Poisson equation in two dimensions. Although direct inversion is possible, the coefficient matrix is large and sparse making successive over relaxation preferable considering the Lubricants 2022, 10, 87 7 of 16 computational resources available for this study. The dielectric constant is defined using a grid staggered from that of the voltage potential function as described by Nagel [50]. Ql,k 1 Rl,k = a1 Vl +1,k + a2 Vl,k+1 + a3 Vl −1,k + a4 Vl,k−1 + − Vl,k (16) a0 ε0 where the coefficients a0 , a1 , a2 , a3 , & a4 are described by Nagel [50] and are also provided for completeness in the Appendix A. n +1 n Vl,k = Vl,k + ξ Rl,k (17) where R is the residual and ξ is the damping factor. The field strength is found simply from the following definition. Typically, it is assumed the maximum electric field strength in an EHL contact is determined by the voltage potential across the contact and the minimum film thickness. The electrical field strength is therefore non-dimensionalised using the voltage potential V12 ) between the two surfaces and minimum film thickness (hmin ) as: E = −∇V hmin V12 (18) The model was constructed with the capability to use a refined grid spacing in comparison to the elastohydrodynamic model. To solve the electrostatic problem linear interpolation was used to transfer the lubricant density and surface geometry data between the two grids in the region −1.25< X <1.25. In addition, it was assumed the lubricant had a uniform density across the film at any X coordinate location. On the two conducting surfaces a Dirichlet boundary with fixed voltage potential is used, these form the top and bottom region of the solution domain. This boundary conditions are applied using the local charge, Q, in Equation (16). The local charge on bodies 1 and two is determined from the voltage potential as Q = Vε 0 . The solution domain is discretised and solved numerically using a successive over relaxation scheme. 2.5. Solution Procedure The elastohydrodynamic problem is initialised using a hertzian pressure distribution and an initial guess for the integration constant. Then the hydrodynamic problem is solved values in the range −4 < X < 2, and the deflection field updated. A multigrid methodology outlined by Venner and Napel [26], Venner [33] and Wijnant [51] with the most refined grid consisting of 3200 nodes was employed. The refined grid has a sub micrometre spacing comparable to the surface roughness measurement lateral resolution. The integration constant is updated as: ! n −1 π 1 new old H00 = H00 + c − dX ∑ P + Pj+1 (19) 2 2 j j =1 where c is a numerical damping factor. Once the recursive multigrid solution has converged. The converged geometry and lubricant density are used to solve the electrostatic problem. The density of the lubricant from the elastohydrodynamic model is used to calculate the dielectric constant (Equation (14)) at each coordinate location. The electrostatic problem is initialised assuming a linearly varying voltage between the contiguous surfaces after which Equations (16) and (17) are solved until convergence is met. 3. Results The lubricant physical properties and the key tribological properties are shown in Tables 1 and 2 respectively. The values are chosen for purposes of validation and that they are typical of the tribological conditions found in many machine elements such as spur gear pairs and roller bearings. In addition, the minimum film thicknesses are sufficient that the quantum mechanical tunnel mechanism that leads to conductive behaviour in Lubricants 2022, 10, 87 8 of 16 sub-nanometer thickness boundary films does not need to be considered [52,53]. For the electrostatic model it is considered R1 = R2 , E1 = E2 and ν1 = ν2 . 3.1. Smooth Surface EHL and Electrical Field Strength Results The elastohydrodynamic section of the model was validated against the data provided by Venner and Napel [26]. The predicted minimum and average film thicknesses for all conditions presented are within 3% of their published values as shown in Table 3. The results provide a good degree of confidence in the numerical elastohydrodynamic model for both minimum and average film thicknesses. Table 3. Elastohydrodynamic model validation. 2GPa Venner & Napel [26] Current Model us [ms−1 ] M L hmin [µm] hav [µm] hmin [µm] hav [µm] 8 4 2 1 49.2 69.6 98.4 139.1 15.7 13.2 11.1 9.35 0.65 0.39 0.24 0.15 0.73 0.44 0.27 0.17 0.64 0.38 0.24 0.15 0.71 0.43 0.26 0.16 The voltage potential for two of the cases investigated is shown in Figure 4 where the constant voltage potential regions of ±1V are defined by the boundary condition applied to the conductive bodies. The electric field strength results for a range of M and L Moe’s non-dimensional EHL group values are presented for 1 and 2 GPa maximum hertzian contact pressure conjunctions shown in Figures 5 and 6 respectively. The non-dimensional electric field strength is shown to have a maximum value of very close to unity at the location of the minimum film thickness. This is independent of the varying EHL behaviour characterised by the non-dimensional EHL groups. The electric field strength is therefore Lubricants 2022, 10, x FOR PEER REVIEW 9 of 17 well predicted for a smooth conjunction by hV12 . The field strength in the central film min thickness region is shown to consistently have a non-dimensional field strength value close to 0.9. (a) (b) Figure4.4.Electrical Electricalpotneital potneitalof ofsmooth smoothsurfaces surfacesEHL EHLfilm filmatat(a) (a)M12L15 M12L15and and(b) (b)M34L9. M34L9. Figure (a) Lubricants 2022, 10, 87 (b) 9 of 16 Figure 4. Electrical potneital of smooth surfaces EHL film at (a) M12L15 and (b) M34L9. (a) (b) (c) (d) Figure surface,11 GPa GPamax maxHertzian Hertziancontact contactpressure pressure EHL conFigure5.5.Electrical Electricalfield fieldstrength strength of of smooth smooth surface, EHL tacts (a) M12L15 (b) M17L13 (c) M24 L11 (d) M34L9. contacts (a) M12L15 (b) M17L13 (c) M24 L11 (d) M34L9. Theresults results shown 6 show thethe same magnitude of dimensionless electrical The shownininFigure Figure 6 show same magnitude of dimensionless electrical field strength at the minimum and central film thickness regions even at the increased field strength at the minimum and central film thickness regions even at the increased maximum contact contact pressure. indicates a minor effect of the film pressure induced maximum pressure.This This indicates a minor effect offluid the fluid film pressure induced change to its dielectric constant on the electrical field strength. Thermal effects resulting from flash temperature rise within the contact are ignored in this paper, however their significance on electrostatic behvaiour has not yet be evaluated. This provides an interesting further area of research. The electrical field strength clearly increases linearly with minimum film thickness. The cross section of electrical field strength at middle of the fluid film (as defined by half of the average film thickness) is presented in Figure 7. Interestingly, the width of the high field strength region located in the vicinity of the minimum film thickness location changes with M and L non-dimensional groups. The width of this region is of great interest to help determine the current density during electrical breakdown of the dielectric lubricant. Current density A m−2 has been reported to be a key parameter to limit damage caused by electrical discharge erosion [54] with Busse et al. [45] suggesting the peak current density during a dielectric breakdown event should be less than 0.8 A m−2 to not adversely affect bearing life. It is suggested that elastohydrodynamic micro geometry Lubricants 2022, 10, 87 10 of 16 Lubricants 2022, 10, x FOR PEER REVIEW 10 of 17 could have a role on the width of the breakdown region and therefore current density during discharge. The experimental research of Plazenet et al. [46] showed the maximum and average lateral dimension of pits formed through electrical discharge erosion are 4 µm and 1.7 to µmitsdiameter respectively. in strength. the same order as theeffects width resulting of change dielectric constant These on thedimensions electrical are field Thermal the minimum film thickness region rather than the hertzian contact half width for both the from flash temperature rise within the contact are ignored in this paper, however their conjunctions in this research and for similar bearings in electric vehicle applications [55]. significance on electrostatic behvaiour has not yet be evaluated. This provides an interestThe presented numerical results therefore highlight the importance of elastohydrodynamic ing further area on of research. micro geometry electrical discharge erosion wear behaviour. (a) (b) (c) (d) Figure surface,22GPa GPamax maxHertzian Hertziancontact contact pressure EHL conFigure6.6.Electrical Electricalfield field strength strength of of smooth smooth surface, pressure EHL tacts (a) M49 L15 (b) M68 L13 (c) M98 L11 (d) M139 L9. contacts (a) M49 L15 (b) M68 L13 (c) M98 L11 (d) M139 L9. electrical field strength increases linearly with minimum film thickness. 3.2. The Rough Surface EHL and Electricalclearly Field Strength Results The cross section of electrical field strength at middle the fluid (asroughness defined by half A measured rough surface was included to investigateofthe effect of film surface ofonthe averagefield filmstrength. thickness) is presented in Figuredisc 7. Interestingly, thefor width of the high electrical A Mini-Traction-Machine commonly used elastohydrodynamic experimentation was measured using a focus variation microscopy technique field strength region located in the vicinity of the minimum film thickness location (Aliconawith InfiniteFocus with ×100 magnification. The disc rootofmean square roughness changes M and LG4) non-dimensional groups. The width this region is of great interRq is 0.01 µm. The rough surface profile is shown in Figure 8. est to help determine the current density during electrical breakdown of the dielectric lubricant. Current density 𝐴 𝑚 has been reported to be a key parameter to limit damage caused by electrical discharge erosion [54] with Busse et al. [45] suggesting the peak current density during a dielectric breakdown event should be less than 0.8 𝐴 𝑚 to not adversely affect bearing life. It is suggested that elastohydrodynamic micro geometry could have a role on the width of the breakdown region and therefore current density during discharge. The experimental research of Plazenet et al. [46] showed the maximum Lubricants 2022, 10, x FOR PEER REVIEW Lubricants 2022, 10, 87 11 of 17 applications [55]. The presented numerical results therefore highlight the importance of 11 of 16 elastohydrodynamic micro geometry on electrical discharge erosion wear behaviour. (a) (b) Figure 7. Mid fluid film thickness electrical field strength in (a) 1GPa and (b) 2GPa max hertzian contact pressure conjunctions. 3.2. Rough Surface EHL and Electrical Field Strength Results A measured rough surface was included to investigate the effect of surface roughness on electrical field strength. A Mini-Traction-Machine disc commonly used for elastohy(a) drodynamic experimentation was measured using a (b) focus variation microscopy technique (Alicona InfiniteFocus G4) with ×100 magnification. The disc root mean square Figure filmthickness thicknesselectrical electrical field strength in1GPa (a) 1GPa(b) and (b) max 2GPahertzian max hertzian Figure7.7. Mid Mid fluid fluid film field strength 2GPa roughness Rq is 0.01 µm. The rough surface profileinis(a) shown and in Figure 8. contact pressure conjunctions. contact pressure conjunctions. 3.2. Rough Surface EHL and Electrical Field Strength Results A measured rough surface was included to investigate the effect of surface roughness on electrical field strength. A Mini-Traction-Machine disc commonly used for elastohydrodynamic experimentation was measured using a focus variation microscopy technique (Alicona InfiniteFocus G4) with ×100 magnification. The disc root mean square roughness Rq is 0.01 µm. The rough surface profile is shown in Figure 8. (a) (b) Figure8.8.(a) (a)Measured Measuredrough roughsurface surface profile and example pressure distributions for M34L9 Figure profile and (b)(b) example pressure distributions for M34L9 with with and without withoutroughness. roughness. and The included with in in thethe elastohydrodynamic simulations and and Theroughness roughnessprofile profilewas was included with elastohydrodynamic simulations the toto the electrostatic model. TheThe roughness profile is trunthe converged convergedsolution solutionprovided provided the electrostatic model. roughness profile is truncated at ± 0.05 µm to help stability of the EHL model. The results shown in Figures 9 and cated at ±0.05 µm to help stability of the EHL model. The results shown in Figures109 and show the the electrical fieldfield strength for thefor same as previously investigated (1 and (1 10 show electrical strength the conditions same conditions as previously investigated 2and GPa max hertzian contact pressure) but on this occasion with the inclusion of surface with the inclusion of sur(a) 2 GPa max hertzian contact pressure) but on this occasion (b) roughness. face roughness. roughness shownsurface in Figures 9 and toexample create small perturbations to thefor electriFigureThe 8. (a) Measuredisrough profile and10(b) pressure distributions M34L9 with cal field strength due to a combination of small changes in local film thickness and pressure and without roughness. perturbations causing changes in the fluid film dielectric constant. The magnitude of the perturbations in dimensional electrical field strength are shown to be small relative to the The roughness profile was included with in the elastohydrodynamic simulations and electrical field strength at the location of the minimum film thickness in Figure 11. the converged solution provided to the electrostatic model. The roughness profile is truncated at ±0.05 µm to help stability of the EHL model. The results shown in Figures 9 and 10 show the electrical field strength for the same conditions as previously investigated (1 and 2 GPa max hertzian contact pressure) but on this occasion with the inclusion of surface roughness. Lubricants 2022, 10, x FOR PEER REVIEW 12 of 17 Lubricants 2022, 10, x FOR PEER REVIEW 12 of 17 Lubricants 2022, 10, 87 12 of 16 (a) (a) (b) (b) (c) (d) (c) (d) Figure 9. Electrical field strength of rough surface 1 GPa max Hertzian contact pressure EHL contacts (a)9. (b) M17L13 (c)ofM24 L11 (d)surface M34L9. Figure Electrical field strength of rough GPaHertzian max Hertzian contact pressure EHL conFigure 9.M12L15 Electrical field strength rough surface 1 GPa1max contact pressure EHL contacts tacts (a) M12L15 (b) M17L13 (c) M24 L11 (d) M34L9. (a) M12L15 (b) M17L13 (c) M24 L11 (d) M34L9. (a) (a) Figure 10. Cont. (b) (b) Lubricants 2022, 10, x FOR PEER REVIEW 13 of 17 Lubricants 2022, 10, 87 13 of 16 (c) (d) Figure 10. Electrical field strength of rough surface 2 GPa max Hertzian contact pressure EHL contacts (a) M49L15 (b) M68L13 (c) M98 L11 (d) M139L9. The roughness is shown in Figures 9 and 10 to create small perturbations to the electrical field strength due to a combination of small changes in local film thickness and pressure (c)perturbations causing changes in the fluid film dielectric (d) constant. The magnitude of the perturbations in dimensional electrical field strength are shown to be small relative to Figure 10. fieldstrength strength of rough surface 2max GPaHertzian max Hertzian contact pressure Figure 10.Electrical Electrical field of rough surface 2 GPa contact pressure EHL contacts the electrical field strength at the location of the minimum film thickness in Figure 11. EHL contacts (a) M49L15 (b) M68L13 (c) M98 L11 (d) M139L9. (a) M49L15 (b) M68L13 (c) M98 L11 (d) M139L9. The roughness is shown in Figures 9 and 10 to create small perturbations to the electrical field strength due to a combination of small changes in local film thickness and pressure perturbations causing changes in the fluid film dielectric constant. The magnitude of the perturbations in dimensional electrical field strength are shown to be small relative to the electrical field strength at the location of the minimum film thickness in Figure 11. Figure Figure11. 11.Mid Midfluid fluidfilm filmthickness thicknessdimensionless dimensionlesselectrical electricalfield fieldstrength strengthwith withroughness. roughness. 4.4.Conclusions Conclusions The Thecombined combinedelectrostatic electrostaticand andelastohydrodynamic elastohydrodynamicnumerical numericalmodelling modellingapproach approach expounded expoundedin inthis thisresearch researchhas has advanced advancedthe theunderstanding understandingof ofelectrical electricalfield fieldstrength strength within within elastohydrodynamic elastohydrodynamicfilms. films.The Themodel modelshows, shows,for forthe thefirst firsttime, time,an anapproach approachto to predicting the the electrical electrical field strength of predicting strength in inelastohydrodynamic elastohydrodynamiccontacts. contacts.The Thekey keyfindings findings this research are: of this research are: Forrough, rough,full fullfilm, film,rolling rollingEHL EHLcontacts contactsamplitude amplitudereduction reductiondetermines determinesthat that hV12 •• For min Figurecan 11. Mid fluid film thickness dimensionless electrical field strength with roughness. stillbe beused usedto toapproximate approximatethe themaximum maximumelectric electric field field strength strength in in the the contact. contact. can still DimensionlessMoe’s Moe’sEHL EHLgroup groupparameters parameters(M (Mand andL) L)change changethe thewidth widthof ofthe thehigh high •• Dimensionless 4. Conclusions field strength region with implications for current density during arcing events. field strength region with implications for current density during arcing events. • The Changes in contact pressure and result dielectric strength of the lubricant only have combined electrostatic andaselastohydrodynamic numerical modelling approach a minor effect on the non dimensional electrical field strength in the conjunction. expounded in this research has advanced the understanding of electrical field strength findings of this paper will inform the tribological design of electro-mechanical withinThe elastohydrodynamic films. The model shows, for the first time, an approach to systems and the lubricants and bearings that are required for their operation predicting the electrical field strength in elastohydrodynamic contacts. without The keythe findings scourge of electrical discharge erosion. of this research are: • For rough, full film, rolling EHL contacts amplitude reduction determines that • can still be used to approximate the maximum electric field strength in the contact. Dimensionless Moe’s EHL group parameters (M and L) change the width of the high field strength region with implications for current density during arcing events. Lubricants 2022, 10, 87 14 of 16 Author Contributions: Conceptualization, M.L. and N.J.M.; Funding acquisition, M.L.; Investigation, S.A.M.; Methodology, N.J.M.; Software, S.A.M.; Supervision, M.L. and N.J.M.; Writing—original draft, N.J.M.; Writing—review & editing, S.A.M., M.L. and N.J.M. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Acknowledgments: The authors would like to extend their thanks and appreciation to AVL List GmbH for the financial support and advice that has helped enabled this research. Conflicts of Interest: The authors declare no conflict of interest. Nomenclature A a C1 & C2 E E E0 hmin h H H00 i j k K l ph po p P Q r < us V U w x X z α V12 Ampere Coefficients for Electrostatic calculations Density pressure relation coefficients Elastic modulus Dimensionless electric field strength Composite elastic modulus Minimum film thickness Film thickness Dimensionless film thickness H = bhR 2 Integration constant Grid index for elastohydrodynamic model in x direction Grid indexing for elastohydrodynamic model in x direction Grid index for electrostatic model in y direction Deflection kernel Grid index for electrostatic model in x direction Maximum hertzian pressure Roelands equation constant pressure Dimensionless pressure ( p/ph ) Local charge Position vector Dimensionless surface roughness sum of surface velocities Voltage potential Dielectric strength of fluid load per unit length coordinate Dimensionless coordinate ( x/b) Roelands pressure viscosity parameter Pressure viscosity coefficient Voltage potential between surface 1 and surface 2 e ε0 εr ε ρ ρ Reynold’s equation coefficient ηλ Permittivity of free space Dielectric constant at reference conditions Dielectric constant Electrical charge density dimensionless lubricant density λ velocity parameter 6 η0 ν ξ viscosity at reference pressure and temperature Poisson’s ratio Numerical damping factor ρH 3 η0 u s R 2 b3 p h Subscript 1,2 denotes a property of Surface 1 and surface 2 respectively Lubricants 2022, 10, 87 15 of 16 Appendix A Residual coefficients as described by Nagel [50]: a0 = ε l,k + ε l −1,k + ε l,k−1 + ε l −1,k−1 1 (ε + ε l,k−1 ) 2 l,k 1 a2 = (ε l −1,k + ε l,k ) 2 1 a3 = (ε l −1,k−1 + ε l −1,k ) 2 1 a4 = (ε l,k−1 + ε l −1,k−1 ) 2 a1 = (A1) (A2) (A3) (A4) (A5) References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. He, F.; Xie, G.; Luo, J. Electrical bearing failures in electric vehicles. Friction 2020, 8, 4–28. [CrossRef] Becker, A.; Abanteriba, S. Electric discharge damage in aircraft propulsion bearings. Proc. Inst. Mech. Eng. Part J J. Eng. Tribol. 2013, 228, 104–113. [CrossRef] Budinski, M.K. Failure analysis of a bearing in a helicopter turbine engine due to electrical discharge damage. Case Stud. Eng. Fail. Anal. 2014, 2, 127–137. [CrossRef] Symonds, N.; Corni, I.; Wood, R.; Wasenczuk, A.; Vincent, D. Observing early stage rail axle bearing damage. Eng. Fail. Anal. 2015, 56, 216–232. [CrossRef] Zhang, P.; Neti, P.; Salem, S. Electrical Discharge and Its Impact on Drivetrains of Wind Turbines. IEEE Trans. Ind. Appl. 2015, 51, 5352–5357. [CrossRef] Bialke, W. A discussion of friction anomaly signatures in response to electrical discharge in ball bearings. Aerosp. Mech. Symp. 2018, 44, 55–68. Prashad, H. Investigations of corrugated pattern on the surfaces of roller bearings operated under the influence of electrical fields. Lubr. Eng. 1988, 44, 710–717. ISO 15243:2017, Rolling Bearings—Damage and Failures—Terms, Characteristics and Causes, Int. Org. for Standardization, Standard. March 2017. Available online: https://www.iso.org/standard/59619.htmlBSIS15243:2017 (accessed on 2 April 2021). Prashad, H. Appearance of craters on track surface of rolling element bearings by spark erosion. Tribol. Int. 2001, 34, 39–47. [CrossRef] Joshi, A.; Blennow, J. Investigation of the static breakdown voltage of the lubricating film in a mechanical ball bearing. In Proceedings of the Nordic Insulation Symposium (No. 23), Trondheim, Norway, 9–12 June 2013. Beshev, D. Characterization of Electrical Lubricant Properties for Modelling of Electrical Drive Systems with Rolling Bearings; Forschungsvereinigung Antriebstechnik e.V. (FVA): Frankfurt, Germany, 2018; Available online: https://www.vdmashop.de/ FVA/Bearing-World-Journal---Volume-2---December-2017-Print-1552.html (accessed on 20 April 2021). Chen, S.; Lipo, T.; Fitzgerald, D. Modeling of motor bearing currents in PWM inverter drives. IEEE Trans. Ind. Appl. 1996, 32, 1365–1370. [CrossRef] Erdman, J.; Kerkman, R.; Schlegel, D.; Skibinski, G. Effect of PWM inverters on AC motor bearing currents and shaft voltages. IEEE Trans. Ind. Appl. 1996, 32, 250–259. [CrossRef] Link, P.J. Minimizing electric bearing currents in ASD systems. IEEE Ind. Appl. Mag. 1999, 5, 55–66. [CrossRef] Benbouzid, M. Bibliography on induction motors faults detection and diagnosis. IEEE Trans. Energy Convers. 1999, 14, 1065–1074. [CrossRef] Thorsen, O.; Dalva, M. A survey of faults on induction motors in offshore oil industry, petrochemical industry, gas ter-minals, and oil refineries. IEEE Trans. Ind. Appl. 1995, 31, 1186–1196. Available online: http://www.scopus.com/inward/record.url?eid= 2-s2.0-0029374064&partnerID=tZOtx3y1 (accessed on 1 December 2021). [CrossRef] Wijanarko, W.; Khanmohammadi, H.; Espallargas, N. Ionic Liquid Additives in Water-Based Lubricants for Bearing Steel—Effect of Electrical Conductivity and pH on Surface Chemistry, Friction and Wear. Front. Mech. Eng. 2022, 7, 756929. [CrossRef] Schneider, V.; Liu, H.-C.; Bader, N.; Furtmann, A.; Poll, G. Empirical formulae for the influence of real film thickness distribution on the capacitance of an EHL point contact and application to rolling bearings. Tribol. Int. 2021, 154, 106714. [CrossRef] Petrusevich, A.I. Fundamental conclusions from the contact-hydrodynamic theory of lubrication. Izv. Akad. Nauk. SSSR OTN 1951, 2, 209–223. Archard, J.F.; Kirk, M.T. Lubrication at point contacts. Proc. R. Soc. Lond. A 1961, 261, 532–550. Crook, A.W. The lubrication of rollers II. Film thickness with relation to viscosity and speed. Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Sci. 1961, 254, 223–236. Lubricants 2022, 10, 87 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 16 of 16 Sunahara, K.; Ishida, Y.; Yamashita, S.; Yamamoto, M.; Nishikawa, H.; Matsuda, K.; Kaneta, M. Preliminary Measurements of Electrical Micropitting in Grease-Lubricated Point Contacts. Tribol. Trans. 2011, 54, 730–735. [CrossRef] Gohar, R.; Cameron, A. Optical Measurement of Oil Film Thickness under Elasto-hydrodynamic Lubrication. Nature 1963, 200, 458–459. [CrossRef] Johns-Rahnejat, P.M.; Gohar, R. Measuring contact pressure distributions under elastohydrodynamic point contacts. Tribotest 1994, 1, 33–53. [CrossRef] Safa, M.M.A.; Gohar, R. Pressure Distribution Under a Ball Impacting a Thin Lubricant Layer. J. Tribol. 1986, 108, 372–376. [CrossRef] Venner, C.H.; Napel, W.E.T. Surface Roughness Effects in an EHL Line Contact. J. Tribol. 1992, 114, 616–622. [CrossRef] Morales-Espejel, G.E.; Quiñonez, A.F. On the complementary function amplitude for film thickness in Mi-cro-EHL. Tribol. Int. 2019, 131, 631–636. [CrossRef] Gonda, A.; Capan, R.; Bechev, D.; Sauer, B. The Influence of Lubricant Conductivity on Bearing Currents in the Case of Rolling Bearing Greases. Lubricants 2019, 7, 108. [CrossRef] Tallian, T.E. On Competing Failure Modes in Rolling Contact. ASLE Trans. 1967, 10, 418–439. [CrossRef] Busse, D.F.; Erdman, J.M.; Kerkman, R.J.; Schlegel, D.W.; Skibinski, G.L. The effects of PWM voltage source inverters on the mechanical performance of rolling bearings. IEEE Trans. Ind. Appl. 1997, 33, 567–576. [CrossRef] Xu, K.; Chu, N.R.; Jackson, R.L. An investigation of the elastic cylindrical line contact equations for plane strain and stress considering friction. Proc. Inst. Mech. Eng. Part J J. Eng. Tribol. 2021. [CrossRef] Sharma, A.; Jackson, R.L. A Finite Element Study of an Elasto-Plastic Disk or Cylindrical Contact Against a Rigid Flat in Plane Stress with Bilinear Hardening. Tribol. Lett. 2017, 65, 112. [CrossRef] Venner, C.H. Multilevel Solution of the EHL Line and Point-Contact Problems. Ph.D. Thesis, University of Twente, Enschede, The Netherlands, December 1991. Lubrecht, A.A.; Breukink, G.A.C.; Moes, H.; Ten Napel, W.E.; Bosma, R. Solving Reynolds’ equation for EHL line contacts by application of a multigrid method. In Proceedings of the Thirteenth Leeds-Lyon Symposium on Tribology, Leeds, UK, 1 October 1987; pp. 175–182. Moes, H. Discussion on Paper D1 by D. Dowson. Proc. Inst. Mech. Eng. 1966, 180, 244–245. Moes, H. Communications. In Proceedings of the Symposium on Elastohydrodynamic Lubrication, London, UK, 21–23 September 1965; pp. 244–245. Dowson, D.; Higginson, G.R. Elasto-Hydrodynamic Lubrication: The Fundamentals of Roller and Gear Lubrication; Pergamon Press: Oxford, UK, 1966. Roelands, C.J.A. Correlational Aspects of the Viscosity Temperature-Pressure Relationship of Lubricating Oils. Ph.D. Thesis, Technische Hogeschool Delft, Delft, The Netherlands, 1966. Bair, S. High Pressure Rheology for Quantitative Elastohydrodynamics; Elsevier: Amsterdam, The Netherlands, 2019. Bair, S. Roelands’ missing data. Proc. Inst. Mech. Eng. Part J J. Eng. Tribol. 2004, 218, 57–60. [CrossRef] Liu, H.; Zhang, B.; Bader, N.; Poll, G.; Venner, C. Influences of solid and lubricant thermal conductivity on traction in an EHL circular contact. Tribol. Int. 2020, 146, 106059. [CrossRef] Bondi, A.A. Physical Chemistry of Lubricating Oils; Book Division, Reinhold Publishing Corporation: New York, NY, USA, 1951. Jackson, J.D. Classical Electrodynamics, 3rd ed.; Wiley: New York, NY, USA, 1999. Carey, A.A. The Dielectric Constant of Lubrication Oils; Computational Systems Inc.: Knoxville, TN, USA, 1998. Busse, D.; Erdman, J.; Kerkman, R.; Schlegel, D.; Skibinski, G. Bearing currents and their relationship to PWM drives. IEEE Trans. Power Electron. 1997, 12, 243–252. [CrossRef] Plazenet, T.; Boileau, T.; Caironi, C.C.; Nahid-Mobarakeh, B. Influencing Parameters on Discharge Bearing Currents in Inverter-Fed Induction Motors. IEEE Trans. Energy Convers. 2021, 36, 940–949. [CrossRef] McFadden, C.; Hughes, K.; Raser, L.; Newcomb, T. Electrical Conductivity of New and Used Automatic Transmission Fluids. SAE Int. J. Fuels Lubr. 2016, 9, 519–526. [CrossRef] Zhang, J.-G.; Mei, C.-H.; Wen, X.-M. Dust effects on various lubricated sliding contacts. In Proceedings of the Thirty Fifth Meeting of the IEEE Holm Conference on Electrical Contacts, Chicago, IL, USA, 18–20 September 1989; pp. 35–42. Sakidu, M.N.O. Numerical Techniques in Electromagnetics, 2nd ed.; CRC Press: Boca Raton, FL, USA, 2001. Nagel, J.R. Solving the Generalized Poisson Equation Using the Finite-Difference Method (FDM); Lecture Notes; Department of Electrical and Computer Engineering, University of Utah: Salt Lake City, UT, USA, 2011. Wijnant, Y.H. Contact Dynamics in the Field of Elastohydrodynamic Lubrication; Universiteit Twente: Enschede, The Netherlands, 2003. Holm, R. Electric Contacts, Theory and Applications, 4th ed.; Springer: Berlin/Heidelberg, Germany, 1967. Konchits, V.V.; Korotkevich, S.V.; Kim, C.K. Thermal effects and contact conductivity under boundary lubrication. Lubr. Sci. 2000, 12, 145–168. [CrossRef] Bell, S.; Cookson, T.J.; Cope, S.A.; Epperly, R.A.; Fischer, A.; Schlegel, D.W.; Skibinski, G.L. Experience with variable-frequency drives and motor bearing reliability. IEEE Trans. Ind. Appl. 2001, 37, 1438–1446. [CrossRef] Questa, H.; Mohammadpour, M.; Theodossiades, S.; Garner, C.; Bewsher, S.; Offner, G. Tribo-dynamic analysis of high-speed roller bearings for electrified vehicle powertrains. Tribol. Int. 2021, 154, 106675. [CrossRef]
https://openalex.org/W2076154029	https://lup.lub.lu.se/search/files/4069010/2856888.pdf	English	null	Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment	PloS one	2,012	cc-by	13,565	Citation for published version (APA): Håkansson, Å., Bränning, C., Molin, G., Adawi, D., Hagslätt, M.-L., Jeppsson, B., Nyman, M., & Ahrné, S. (2012). Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment. PLoS ONE, 7(3), Article e33510. https://doi.org/10.1371/journal.pone.0033510 Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment. Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment. Håkansson, Åsa; Bränning, Camilla; Molin, Göran; Adawi, Diya; Hagslätt, Marie-Louise; Jeppsson, Bengt; Nyman, Margareta; Ahrné, Siv Published in: PLoS ONE DOI: 10.1371/journal.pone.0033510 Link to publication Citation for published version (APA): Håkansson, Å., Bränning, C., Molin, G., Adawi, D., Hagslätt, M.-L., Jeppsson, B., Nyman, M., & Ahrné, S. (2012). Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment. PLoS ONE, 7(3), Article e33510. https://doi.org/10.1371/journal.pone.0033510 Total number of authors: 8 Total number of authors: General rights U l th General rights Unless other specific re-use rights are stated the following general rights apply: Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. p g g g g pp y Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. g q g • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal Read more about Creative commons licenses: https://creativecommons.org/licenses/ Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. LUND UNIVERSITY PO Box 117 221 00 Lund +46 46-222 00 00 Abstract Long-term colonic inflammation promotes carcinogenesis and histological abnormalities of the liver, and colorectal tumours frequently arise in a background of dysplasia, a precursor of adenomas. Altered colonic microbiota with an increased proportion of bacteria with pro-inflammatory characteristics, have been implicated in neoplastic progression. The composition of the microbiota can be modified by dietary components such as probiotics, polyphenols and dietary fibres. In the present study, the influence of probiotics in combination with blueberry husks on colorectal carcinogenesis and subsequent liver damage was evaluated. Colorectal tumours were induced in rats by cyclic treatment with dextran sulphate sodium (DSS). Blueberry husks and a mixture of three probiotic strains (Bifidobacterium infantis DSM 15159, Lactobacillus gasseri, DSM 16737 and Lactobacillus plantarum DSM 15313) supplemented a basic diet fortified with oats. The condition of the rats was monitored using a disease activity index (DAI). A qualitative and quantitative histological judgement was performed on segments of distal colon and rectum and the caudate lobe of the liver. The formation of short-chain fatty acids, bacterial translocation, the inflammatory reaction and viable count of lactobacilli and Enterobaceriaceae were addressed. Blueberry husks with or without probiotics significantly decreased DAI, and significantly reduced the number of colonic ulcers and dysplastic lesions. With a decreased proportion of blueberry husk in the diet, the probiotic supplement was needed to achieve a significant decrease in numbers of dysplastic lesions. Probiotics decreased faecal viable count of Enterobacteriaceae and increased that of lactobacilli. Blueberry husks with or without probiotics lowered the proportion of butyric acid in distal colon, and decreased the haptoglobin levels. Probiotics mitigated hepatic injuries by decreasing parenchymal infiltration and the incidence of stasis and translocation. The results demonstrate a dietary option for use of blueberry husks and probiotics to delay colonic carcinogenesis and hepatic injuries in the rat model. Received October 5, 2011; Accepted February 9, 2012; Published March 23, 2012 ansson et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2012 Ha˚kansson et al. This is an open-access article distributed under the terms of the Creative Commons Attribut unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This study was funded by the Functional Food Science Centre at Lund University. Abstract The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: SA, GM and BJ are minority shareholders in the public company Probi AB (Lund, Sweden). This does not alter the authors’ adherence to all the PLoS ONE policies on sharing data and materials. * E-mail: asa.hakansson@appliednutrition.lth.se A˚ sa Ha˚kansson1, Camilla Bra¨nning2, Go¨ ran Molin1, Diya Adawi3, Marie-Louise Hagsla¨tt1, Bengt Jeppsson3, Margareta Nyman2, Siv Ahrne´ 1 1 Food Hygiene, Division of Applied Nutrition and Food Chemistry, Lund University, Lund, Sweden, 2 Division of Applied Nutrition and Food Chemistry, Lund University, Lund, Sweden, 3 Department of Surgery, Lund University, Ska˚ne University Hospital, Malmo¨, Sweden PLoS ONE \| www.plosone.org Citation: Ha˚kansson A˚, Bra¨nning C, Molin G, Adawi D, Hagsla¨tt M-L, et al. (2012) Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment. PLoS ONE 7(3): e33510. doi:10.1371/journal.pone.0033510 Blueberry Husks and Probiotics Attenuate Colorectal Inflammation and Oncogenesis, and Liver Injuries in Rats Exposed to Cycling DSS-Treatment A˚ sa Ha˚kansson1, Camilla Bra¨nning2, Go¨ ran Molin1, Diya Adawi3, Marie-Louise Hagsla¨tt1, Bengt Jeppsson3, Margareta Nyman2, Siv Ahrne´ 1 Introduction Composition of diets (g/kg dwb) given to rats in the following groups: Group 1, active control (C); Group 2, probiotics (P); Group 3, blueberry husks (2B); Group 4, blueberry husks and probiotics (2BP), reduced amount of blueberry husks (B), and reduced amount of blueberry husks and probiotics (BP). DSS is a non-genotoxic, sulphated, polysaccharide that nevertheless can induce experimental chronic colitis and colitis- associated neoplasia in animals. The histopathological changes show reminiscence of human UC [19]. During long-term DSS exposure, dysplasia and/or cancer occurs as dysplasia-associated lesions, which has similarities to the development of dysplasia and cancer development in humans with colitis [19]. However, the effect on the liver of long-term DSS-induced colitis is mainly unknown. The aim of the present study has been to evaluate the potential of blueberry husks and a probiotic mixture to attenuate inflammatory injuries in colon and liver and to mitigate colonic dysplasia development. It was supported by evidence that the faecal flora was influenced and linked to changed profiles of SCFA production, the hepatic damage was affected and carcinogenic progression was delayed. g g y g 2Corresponding to 25 g dietary fibre/kg diet (dwb). ( , , y) 5+Freeze-dried Bifidobacterium infantis DSM 15159 ( = CURE21; dose, 2?109 CFU/ d), Lactobacillus gasseri DSM 16737 ( = VPG44; dose, 1?109 CFU/d) and Lactobacillus plantarum DSM 15313 ( = HEAL19; dose, 3?109 CFU/d) were added to the diet. doi:10.1371/journal.pone.0033510.t001 doi:10.1371/journal.pone.0033510.t001 Ethics Statement The Ethics Committee for Animal Studies at Lund University approved the animal experiment (permit number and approval- ID: M25-06). Methods of the amount of dietary fibre (25 g dietary fibre/kg) comprised oat bran and the other half comprised blueberry husks. The amount of fibre was chosen to approximate a moderate fibre intake, corresponding to an equivalent dose of around 30 g fibre/d in humans. The dry matter content was adjusted with wheat starch, and the dietary fibre content was 17.1 g/100 g (dwb) in oat bran and 40.8 g/100 g (dwb) in blueberry husks (Table 1). Of the dietary fibre content in oat bran, 1.5 g/100 g (dwb) was Klason lignin, i.e. components not soluble in 12 M H2SO4, whereas the amount of Klason lignin in blueberry husk was 14.1 g/100 g (dwb). The non-starch polysaccharides consisted mainly of glucose (61%), xylose (19%), and arabinose (12%) in oat bran and glucose (39%), uronic acids (25%) and xylose (20%) in blueberry husks. After 7 days of adaptation to the diets, an experimental period of 6 months followed, when feed residues were collected daily. Introduction ment and modulation of mucosal immunity [7]. During inflammation in UC-patients, different members of the Enterobac- teriaceae family and different Clostridium species have been found to increase in accordance with a decrease in bifidobacteria and lactobacilli [8,9]. This change in the composition of the microbiota, leading to an imbalance between potentially beneficial and adverse bacteria, can contribute to the pathogenesis. Lipopolysaccharides (LPS) associated to the cell wall of Gram- negative bacteria, are highly inflammatory compounds. LPS are associated with disturbed mucosal integrity, and bacterial translocation from the GI- tract [10]. Translocated LPS can cause extensive damage to a variety of organs, including the liver [11]. Cancers may arise from sites of infection, chronic irritation and inflammation [1] and the degree and extent of inflammation during, for example, ulcerative colitis (UC), is a critical component of tumour development and progression [2]. UC-associated colorectal tumours frequently arise in a background of dysplasia and differ in pathogenesis and molecular features from sporadic colorectal cancer [3]. The presence of dysplasia-associated lesions is highly indicative for underlying or associated cancer [4]. Histological abnormalities of the liver of patients with chronic UC have been observed [5], and steatosis and primary sclerosing cholangitis are the most common lesions [6]. The cause of gastrointestinal tumours is implicating chronic inflammation in response to an adverse bacterial flora as a promotion of neoplastic progression, and the intestinal environ- ment is considered important in both colorectal cancer develop- Dietary-induced changes in the different populations of the intestinal microbiota can be achieved by use of dietary fibres and/ or probiotics. Probiotics may affect the intestinal flora but may also modulate immunological functions and affect intestinal perme- March 2012 \| Volume 7 \| Issue 3 \| e33510 1 March 2012 \| Volume 7 \| Issue 3 \| e33510 Blueberries and Probiotics Attenuate Oncogenesis Table 1. Diet composition. Component C P 2B 2BP B BP Oat bran 2911 2911 - - 1452 1452 Blueberry husks (B) - - 1221 1221 612 612 Basal mixture3 369.2 369.2 369.2 369.2 369.2 369.2 Wheat starch4 380 380 549 549 465 465 Probiotics (P)5 - + - + - + Composition of diets (g/kg dwb) given to rats in the following groups: Group 1, active control (C); Group 2, probiotics (P); Group 3, blueberry husks (2B); Group 4, blueberry husks and probiotics (2BP), reduced amount of blueberry husks (B), and reduced amount of blueberry husks and probiotics (BP). Introduction 1Corresponding to 50 g dietary fibre/kg diet (dwb). 2Corresponding to 25 g dietary fibre/kg diet (dwb). 3Containing (g/kg dwb) 160 casein, 1.2 DL-methionine, 50 maize oil, 48 mineral mixture (Containing (g kg21) 0.55 CuSO4?H2O, 2.0 ZnSO4?7H2O, 498 KH2PO4, 258 NaH2PO4?2H2O, 487 CaCO3, 0.1 KI, 86 MgSO4, 12 FeSO4?7H2O, 5 MnSO4?H2O, 0.03 CoCl?6H2O, 153 NaCl, 0.02 CrCl3?6H2O, 0.02 Na2Se), 8 vitamin mixture (Containing (g kg21) 0.62 menadion, 2.5 thiamin hydrochloride, 2.5 riboflavin, 1.25 pyridoxin hydrochloride, 6.25 calcium pantothenate, 6.25 nicotinic acid, 0.25 folic acid, 12.5 inositol, 1.25 p-aminobenzoic acid, 0.05 biotin, 0.00375 cyanocobalamin, 0.187 retinol palmitate, 0.00613 calciferol, 25 d-a- tocopheryl acetate, 941.25 maize starch), 2 choline chloride, 100 sucrose. 4Wheat starch (Cerestar, Krefeld, Germany). 5+Freeze-dried Bifidobacterium infantis DSM 15159 ( = CURE21; dose, 2?109 CFU/ d), Lactobacillus gasseri DSM 16737 ( = VPG44; dose, 1?109 CFU/d) and Lactobacillus plantarum DSM 15313 ( = HEAL19; dose, 3?109 CFU/d) were added to the diet. doi:10.1371/journal.pone.0033510.t001 ability which will have effects on bacterial translocation [12] and liver health. Microbial degradation in the hindgut of dietary carbohydrates escaping digestion, results in production of short- chain fatty acids (SCFAs), mainly acetic acid, propionic acid and butyric acid [13]. Butyrate is considered to be the preferred source of energy for colonocytes but to some extent also propionic acid can be utilized [14]. Fermentation of dietary oat-fibres results in elevated amounts of butyric acid [15], which has been suggested to mitigate colorectal cancer development [16]. Table 1. Diet composition. Table 1. Diet composition. Other dietary components, such as various phenolic com- pounds, may modulate the composition of the intestinal microflora [17]. Blueberries are rich in a variety of phenolics, which have been shown to inhibit colon cancer and cell proliferation, and induce apoptosis in vitro [18]. Composition of diets (g/kg dwb) given to rats in the following groups: Group 1, active control (C); Group 2, probiotics (P); Group 3, blueberry husks (2B); Group 4, blueberry husks and probiotics (2BP), reduced amount of blueberry husks (B), and reduced amount of blueberry husks and probiotics (BP). Composition of diets (g/kg dwb) given to rats in the following groups: Group 1, active control (C); Group 2, probiotics (P); Group 3, blueberry husks (2B); Group 4, blueberry husks and probiotics (2BP), reduced amount of blueberry husks (B), and reduced amount of blueberry husks and probiotics (BP). Animals and experimental design Female Sprague-Dawley rats (n = 48), purchased from Scan- bur (Sollentuna, Sweden), were housed four per cage at room temperature of 22uC with 12 h light/dark cycles and given free access to water, while feed intake was restricted to 92 g (dwb, dry weight basis) per cage and day. Animals were randomly divided into six groups with eight animals in each group, and given different diets according to Table 1. In the diets, oat bran and blueberry husks were the sources of dietary fibres (Table 1). The blueberry husks were derived from pressed wild low-bush blueberry of Vaccinium myrtillus L, and were freeze-dried before inclusion (Probi AB, Lund, Sweden), and the oat bran (Avena sativa L. cv. Sang) was supplied by Lantma¨nnen (Ja¨rna, Sweden). A mixture of freeze-dried Bifidobacterium infantis DSM 15159 ( = CURE21; dose, 2?109 CFU/d), Lactobacillus gasseri DSM 16737 ( = VPG44; dose, 1?109 CFU/d) and Lactobacillus plan- tarum DSM 15313 ( = HEAL19; dose, 3?109 CFU/d) was added to the diet for three of the groups (Table 1). Oat bran and blueberry husks were included at a level of 50 g dietary fibre/kg in the diets (dwb) for groups C (oat bran), 2B (blueberry husks) and 2BP (blueberry husks and probiotics) (Table 1). The soluble and insoluble dietary fibres in the raw materials were determined by a gravimetric method [20]. The composition of the fibre residues was analysed by gas-liquid chromatography (GLC) for the neutral sugars as their alditol acetates and spectrophotometrically for the uronic acids [21]. In the blueberry diets marked B and BP (P = probiotics; Table 1), half All groups were administered 4% (w/v) DSS (MW = 36,000– 50,000; ICN Biomedicals Inc., Aurora, OH) dissolved in drinking water for 7 days, followed by 10 days of tap water, and this cycle was then repeated 11 times. The DSS solution was changed daily. Rats were weighed before and after the adaptation period, as well as daily during the DSS consumption. Body weight change during the experimental period was calculated as gram per animal or as body weight change per kilo gram feed consumed and animal. An attempt was made to quantify the amount of drinking water and DSS load ingested by the rats. Blueberries and Probiotics Attenuate Oncogenesis Blueberries and Probiotics Attenuate Oncogenesis eosin staining. The degree of dysplasia in the colon was scored from normal to mucosa with mild dysplasia (with distorted crypts of increased length and orientation), and mucosa with severe dysplasia (with severe crypt distortion, atypic epithelial cells, reduction or loss of goblet cells, hyperchromatic cell nuclei and increased numbers of cell mitoses). Liver specimens were evaluated for the degree of steatosis according to Brunt et al., [25], where steatosis was scored as absent ( = 0), mild when present in ,1/3 of the hepatocytes ( = 1), moderate when present in 1/3–2/3 of the hepatocytes ( = 2), and severe when present in .2/3 of the hepatocytes ( = 3). The presence and location of infiltrating inflammatory cells and liver injury was also recorded. The degree of infiltrating inflammatory cells in steatotic and non-steatotic areas, vascular stasis and loss of liver parenchyma in either zonal or nonzonal distribution were measured using a semi-quantitative graded scale of 0 (absent), 1 (mild), 2 (moderate), and 3 (extensive) [26], to enable statistical evaluation. Bacterial translocation Samples from the caudate lobe of the liver were removed aseptically and frozen immediately at 270uC until determination. For analysis, the samples were thawed, placed in an ultrasonic bath (Millipore, Sundbyberg, Sweden) for 5 min and swirled for 2 min on a Chiltern. Viable counts were obtained from Violet-red bile glucose (VRBG) agar (Oxoid) that was incubated aerobically at 37uC for 24 h (Enterobacteriaceae count), brain heart infusion (BHI) agar (Difco, Detroit, MI) that was incubated aerobically and under anaerobic conditions, as described above, at 37uC for 72 h (total aerobic and anaerobic counts, respectively), and from Rogosa agar (Oxoid), incubated anaerobically at 37uC for 72 h (lactobacilli count). Results were expressed as incidences of positive cultures/group. Clinical scoring of colitis Disease severity was analysed in terms of disease activity index (DAI), calculated on the basis of weight loss, stool consistency, and bleeding. The scoring system has been validated [22] and shown to correlate histologically with pathological findings [23]. The DAI was assessed daily during DSS administration from day 0 to day 7 and scored on a scale of 0–4 for each clinical parameter and then averaged for each animal. Weight-loss, stool and bleeding scores were defined by modified scoring limits [24]. Colonies were randomly picked from the plates with positive cultures and identified by sequencing the 16 S ribosomal RNA gene. The partial 16 S rRNA gene sequences were searched against GenBank (National Centre for Biotechnology Information, Bethesda, MD) using the Basic Local Alignment Search Tool (BLAST) accessible from the homepage at the National Centre for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih. gov/). The compared sequence lengths were between 200–900 base pairs. Viable count of Enterobacteriaceae and lactobacilli in Faecal samples were thawed and homogenised in freezing medium, diluted (sodium chloride (Merck), 8.5 g/l; Bacteriological peptone (Oxoid, Unipath LTD Basingstoke, Hampshire, En- gland), 1 g/l; Tween 80 (Merck), 1 g/l; L-Cystine hydrochloride monohydrate (Merck), 0.2 g/l) and plated on Rogosa agar for lactobacilli count (Oxoid; incubated anaerobically [Gas Pack System, Gas Pack; Becton Dickenson Microbiology Systems, Cockeynsville, MD] at 37uC for 72 h) and violet red bile-glucose agar (VRBG) for Enterobacteriaceae count (Oxoid; incubated aerobically at 37uC for 24 h). The number of colonies formed on each plate was counted and corrected for the weight of the original faecal sample, and expressed as CFU/g faeces. Sampling Blood samples for analysis of haptoglobin were taken from the saphenous vein at the beginning of the study, and during cycle 1, 5 and 10. During each DSS cycle samples were taken on the seventh day of DSS administration and on the tenth day of the subsequent water period. At the same time faecal samples were collected for viable count. The animals were anaesthetised with Hypnorm (Division of Janssen-Cilag Ltd., Janssen Pharmaceutica, Beerse, Belgium), Dormicum (F. Hoffman-La Roche AG, Basel, Switzerland) and water (1:1:2) at a dose of 0.15 ml/100 g of body weight by a subcutaneous injection. Arterial blood was collected for analysis of SCFAs and liver specimens were obtained for bacterial translo- cation and liver histology. The entire colorectum from the colocaecal junction to the anal verge was excised and the luminal content of caecum and colon was gently removed for analysis of SCFAs and pH was measured in caecal content before storage at 240uC. The large bowels were macroscopically examined for gross lesions all of which were recorded, and then the colons were cut and fixed in 10% buffered formalin, for 24 h. Haptoglobin The concentration of serum haptoglobin was analysed using a manual microplate (96 microwell plates, NuncTM, Roskilde, Denmark) method. In this assay, serum was incubated with haemoglobin (Hb) (0.12 mg/ml bovine haemoglobin (Sigma Aldrich, St Louis, USA) in 0.15 M NaCl (Merck Schuchardt, Hohenbrunn, Germany)) leading to preserved peroxidase activity of the complex. The preserved activity, measured by the addition of a peroxidase substrate (chromogenic solution; 0.5 M citrate buffer pH 3.8 (0.5 M sodium citrate dihydrate (J.T Baker B.V., Deventer, Holland), 0.5 M citric acid-1-hydrate (Merck)), 1% Tween 20 (Merck), 20 mM phenol (International Biotechnologies Inc., Eastman Kodak Co. Rochester, NY), 0.39 mM dithioery- thritol (Sigma), 1.6 mM 4-aminoantipyrine (Sigma), 1 mM 8- anilino-1-naphthalene sulphonic acid (Sigma) and 1 ml 30% H2O2/0.7 ml solution (Merck), is directly proportional to the amount of haptoglobin in the samples. Absorbance was measured at 600 nm (SpectraMaxH M2 Multi-detection Microplate Reader, Molecular Devices, Sunnyvale, California) and the results were compared with a haptoglobin standard (2 mg/ml) (Tridelta Development Ltd, Maynouth County Kildare, Ireland). Animals and experimental design Drinking volumes were recorded every 24 h for each cage (four animals) and the DSS load per animal was calculated over the experimental period as: Total drinking water ml ð Þ\| DSS g ð Þ=100 ml ð Þ ð Þ= number of animals: Total drinking water ml ð Þ\| DSS g ð Þ=100 ml ð Þ ð Þ= number of animals: March 2012 \| Volume 7 \| Issue 3 \| e33510 PLoS ONE \| www PLoS ONE \| www.plosone.org 2 PLoS ONE \| www.plosone.org 16 S rDNA sequencing DAI scores, haptoglobin, number of dysplastic lesions, number of ulcers, scoring used for histopathologic evaluation of liver samples, lactobacilli and Enterobacteriaceae counts (Fig. 1, 2, 5, 6, 11, 12 and Table 2) were presented as medians with 25 and 75 percentiles. The statistics were conducted in SigmaStatH version 3.0 (SPSS Inc., Chicago, Ill., USA). Differences between all groups were evaluated by Kruskal-Wallis test one way ANOVA on ranks followed by all pairwise multiple comparison procedures (Student- Newman-Keuls method), if appropriate. The differences between treatment groups were assessed by a Mann-Whitney rank sum test. Incidence of steatosis, cellinfiltration, stasis, loss of parenchyma and translocation to the liver (Table 3, 4), as well as statistical comparison of the total received score compared to maximum score (Table 2) were calculated in QuickStat version 2.6 and evaluated by the Fisher exact test. For sequencing, primers ENV1 (59-AGA GTT TGA TII TGG CTC AG-39, Escherichia coli numbering 8–27) and ENV2 (59-CGG ITA CCT TGT TAC GAC TT-39, E. coli numbering 1511–1492) [28] were used for amplification of the 16 S rRNA genes. The PCR reaction mixture contained 0.2 mM of both primers, 5 ml of template DNA, 5 ml of 106PCR reaction buffer with 1.5 mM MgCl2 (Roche Diagnostics GmbH, Mannheim, Germany), 200 mM of each deoxyribonucleotide triphosphate, and 2.5 U of Taq DNA polymer- ase (Roche Diagnostics, Mannheim, Germany). Water was added to a final volume of 50 ml. PCR was performed in a PCR Mastercycle 5333 (Eppendorf) with the following profile: 1 cycle at 94uC for 3 min, followed by 30 cycles of 96uC for 15 s, 50uC for 30 s, and 72uC for 90 s, with an additional extension at 72uC for 10 min. The amplification products (5 ml) were checked by running the products on 1.5% (wt./vol.) agarose gel in 16 TBE buffer (89 mM Tris, 89 mM boric acid, 2.5 mM EDTA, pH 8.3), after ethidium bromide staining. Amplicons were sent to MWG (Biotech, Ebersberg, Germany) for single strand sequencing. 16 S rDNA sequences (mostly around 500 bp) were searched against Genbank (blastn) option at the homepage of the National Centre for Biotechnology (http://www.ncbi.nlm.nih.gov/BLAST/) [29] or aligned to 16 S rDNA encoding sequences downloaded from the Ribosomal Data Base (RDP-II) [30] for an approximate phylogenetic affiliation. Histological evaluation Specimens from the distal part of colon and from the liver were evaluated by light microscopy. Macroscopic abnormalities through the entire length of colon, and microscopic alterations, were evaluated by an experienced surgeon and pathologist respectively. The biopsies of the distal colon, taken at selected sampling sites (polyps or dysplastic lesions and surrounding mucosa), and the left lobe of the liver, were each fixed in neutral buffered formalin, followed by standard procedure for paraffin embedding. Serial sections were cut for each organ and stained with haematoxylin- Colonies were randomly picked from countable Rogosa agar plates in order to get an idea of the identity of the dominant faecal lactobacilli flora. Colonies from four different time-points were collected (before adaptation period, day 7 of DSS cycle 1, day 7 of DSS cycle 10, during surgery) and a total of 56 isolates were identified through RAPD band pattern comparison and nucleotide sequencing. Obtained sequences were at least 700 base-pair long and the results showed no less than 99% sequence similarity to their nearest database entries. PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 March 2012 \| Volume 7 \| Issue 3 \| e33510 3 Blueberries and Probiotics Attenuate Oncogenesis ethylbutyric acid (internal standard) were added to the serum samples and the SCFAs were protonised with hydrochloric acid. The caecal and colonic amounts of SCFAs (acetic, propionic, isobutyric, butyric, isovaleric, valeric, caproic and heptanoic acids) were analysed by a GLC method [32] with minor modifications. Water mixed with hydrochloric acid and 2-ethylbutyric acid was added to the faecal samples before homogenisation of the suspension. Randomly Amplified Polymorphic DNA (RAPD) analysis Randomly Amplified Polymorphic DNA (RAPD) analysis As template for the polymerase chain reaction, crude cell extract was prepared [27] and one microlitre of PCR template was used in the polymerase chain reaction (PCR) [27]. Agarose gel (Type III, High EEO, Sigma) electrophoresis was run, and the gels were stained with ethidium bromide and photographed under UV illumination. RAPD band comparison of isolates taken from the feed was used for identification of Lactobacillus plantarum HEAL19. Short-chain fatty acids Body weight differed between groups at the start. Compared with the C group (189.5 g/animal (181.5–191.0), the animals of the 2BP group and BP group had a slightly higher body weight, The short-chain fatty acids (SCFAs; acetic, propionic, isobu- tyric, butyric, isovaleric and valeric acids) were analysed in serum using GLC [31] with small modifications. Water and 2- Figure 1. Disease activity index. Disease activity index (DAI) during the 11 cycles of DSS administration. DAI scores are expressed as medians (25 and 75 percentiles). Significant differences are expressed versus the C group. Cycle 1: P,0.01 for groups 2B, 2BP and B; Cycle 2: P,0.01 for groups B and BP, *** P,0.001 for groups 2B and 2BP groups; Cycle 3: P,0.05 for group P, P,0.001 for group B group, *P,0.001 for groups 2B, 2BP and BP; Cycle 4: P,0.05 for group BP, *P,0.001 for groups 2B and 2BP; Cycle 5: P,0.01 for group B group, **P,0.001 for groups 2B, 2BP, and BP; Cycle 6: P,0.05 for group B, P,0.01 for group BP,P,0.001 for groups 2B and BP; Cycle 7: P,0.05 for group B, P,0.01 for groups 2B and B, P,0.001 for groups 2BP and BP; Cycle 8: P,0.05 for group BP, P,0.01 for groups 2B and B, P,0.001 for group 2BP; Cycle 9: P,0.05 for groups 2B, 2BP, B, BP; Cycle 10, P,0.05 for group B, P,0.01 for groups 2B and 2BP; Cycle 11: P,0.05 for groups 2B, 2BP, B, BP. doi:10.1371/journal.pone.0033510.g001 Figure 1. Disease activity index. Disease activity index (DAI) during the 11 cycles of DSS administration. DAI scores are expressed as medians (25 and 75 percentiles). Significant differences are expressed versus the C group. 16 S rDNA sequencing One-way ANOVA was used for individual means to assess the effect of dietary fibre or probiotics by using Tukey’ procedure (SCFAs in colon content and blood samples). When error variance was found to be heterogeneous, data was transformed by BoxCox- transformation before ANOVA. Values are presented as means and differences resulting in P#0.05 were considered significant. Short-chain fatty acids Cycle 1: P,0.01 for groups 2B, 2BP and B; Cycle 2: P,0.01 for groups B and BP, *** P,0.001 for groups 2B and 2BP groups; Cycle 3: P,0.05 for group P, P,0.001 for group B group, *P,0.001 for groups 2B, 2BP and BP; Cycle 4: P,0.05 for group BP, *P,0.001 for groups 2B and 2BP; Cycle 5: P,0.01 for group B group, **P,0.001 for groups 2B, 2BP, and BP; Cycle 6: P,0.05 for group B, P,0.01 for group BP,P,0.001 for groups 2B and BP; Cycle 7: P,0.05 for group B, P,0.01 for groups 2B and B, P,0.001 for groups 2BP and BP; Cycle 8: P,0.05 for group BP, P,0.01 for groups 2B and B, P,0.001 for group 2BP; Cycle 9: P,0.05 for groups 2B, 2BP, B, BP; Cycle 10, P,0.05 for group B, P,0.01 for groups 2B and 2BP; Cycle 11: P,0.05 for groups 2B, 2BP, B, BP. doi:10.1371/journal.pone.0033510.g001 March 2012 \| Volume 7 \| Issue 3 \| e33510 PLoS ONE \| www.plosone.org 4 Blueberries and Probiotics Attenuate Oncogenesis Blueberries and Probiotics Attenuate Oncogenesis Figure 2. Concentrations of haptoglobin. Haptoglobin concentrations (mg/ml) in blood during the 11 cycles of DSS administration. Concentrations are expressed as medians (25 and 75 percentiles). Significant differences are expressed versus the C group. Base line: P,0.05 for group B, P#0.01 for group 2B group, *P,0.001 for group 2BP. DSS cycle 1: P,0.05 for group BP. DSS cycle 5: P,0.05 for group 2B. DSS cycle 10: P,0.05 for group 2B. Pure water cycle (W1): *P,0.01 for group B; Water cycle 5 (W5) P,0.05 for group 2B group. doi:10.1371/journal.pone.0033510.g002 Figure 2. Concentrations of haptoglobin. Haptoglobin concentrations (mg/ml) in blood during the 11 cycles of DSS administration. Concentrations are expressed as medians (25 and 75 percentiles). Significant differences are expressed versus the C group. Base line: P,0.05 for group B, P#0.01 for group 2B group, *P,0.001 for group 2BP. DSS cycle 1: P,0.05 for group BP. DSS cycle 5: P,0.05 for group 2B. DSS cycle 10: P,0.05 for group 2B. Pure water cycle (W1): *P,0.01 for group B; Water cycle 5 (W5) P,0.05 for group 2B group. doi:10.1371/journal.pone.0033510.g002 197.5 g/animal (194.0–202.5) (P = 0.001) and 200.0 g/animal (190.0–206.5) (P = 0.038)), respectively. group C (Fig. 1). Groups supplemented with blueberry husks continued to have lower DAI than the C group through the different cycles. The P group only reached significant difference in the 3rd cycle (P = 0.05). Short-chain fatty acids The signs of colitis gradually disappeared during the first periods with pure water, but DAI gradually increased over time and did not revert between the cycles of DSS administration (Fig. 1). At the 11th cycle, a significantly lower DAI was found compared with group C for groups 2B (P = 0.015), 2BP (P = 0.021), B (P = 0.043) and BP (P = 0.050) (Fig. 1). All animals exhibited body weight loss and most of them showed more or less rectal bleeding and loose stool, and DAI increased significantly between the first and eleventh cycle of DSS administration for all groups (P,0.01). The mortality rate was 0%. During the study, the feed intake was similar for all groups, and all animals gained weight with time, and at the end there were no significant differences between the groups, irrespective of the approach of calculation, i.e. the body weight change (g/animal) was for the C group 173.5 g (167.0–217.5), P group 194.5 g (158.0–217.5), 2B group 173.5 g (149.0–211.5), 2BP group 181.5 g (173.5–204.0), B group 153.0 g (138.0–211.0), BP group 163.0 g (146.5–196.5), and body weight change in relation to amount of consumed food (g/kg feed/animal) was for the C group 56.0 g (53.9–70.2), P group 62.7 g (51.0–70.2), 2B group 54.2 g (46.6–66.1), 2BP group 55.0 g (52.6–61.8), B group (47.8 g (43.1– 65.9), BP group 54.3 g (48.8–65.5). Mean total consumption of DSS (in water) was 21 g/rat during the experimental period with no differences between the groups. Haptoglobin The haptoglobin level in blood at base line was significantly lower in groups B (P = 0.043), 2B (P = 0.001) and 2BP (P,0.001) compared with that of group C (Fig. 2). Over the experimental period there is a general pattern that group P has higher haptoglobin values than group C, while all groups supplemented Disease severity scoring After the first DSS cycle the DAI score was significantly lower in groups 2B (P = 0.007), 2BP (P = 0.002) and B (P = 0.005) than in Table 2. Evaluation of liver injury (Scoring values). Scoring values Groups Steatosis Cellinfiltration1 Cellinfiltration2 Stasis Loss of parenchyma C 10/24 (41.7%) 2/24 (8.3%) 16/24 (66.7%) 8/24 (38.1%) 6/18 (33.3%) P 20,5/24# (85.4%) 4/24 (16.7%) 16/24 (66.7%) 8/15 (53.3%) 4/24 (16.7%) 2B 5/24 (20.8%) 0/24 15/24 (62.5%) 12/24 (50.0%) 10/21 (47.6%) 2BP 15/24# (62.5%) 0/24 10/24 (41.7%) 8/24 (33.3%) 3/24 (12.5%) B 6/15 (40.0%) 1/15 (6.7%) 6/15* (40.0%) 0/12 0/12 BP 16.5/21# (78.6%) 1/18 (5.6%) 4/18 (22.2%) 0/18 0/18 Liver specimens were histologically evaluated for the degree of steatosis, infiltrating inflammatory cells in steatotic and non-steatotic areas, vascular stasis and loss of liver parenchyma. The status of the livers of different groups is expressed as scoring values (degree of scoring and received value/maximum value). Between brackets are the percentages of the values. 1Cellinfiltration around CV within steatotic areas. 2Cellinfiltration elsewhere in the parenchyma. Degree of scoring: * denotes P,0.05 compared with the C group. Statistical comparison of the total received score compared to maximum score: # denotes P,0.01 compared to rats fed diets without bacteria. doi:10.1371/journal.pone.0033510.t002 Table 2. Evaluation of liver injury (Scoring values). Table 2. Evaluation of liver injury (Scoring values). Liver specimens were histologically evaluated for the degree of steatosis, infiltrating inflammatory cells in steatotic and non-steatotic areas, vascular stasis and loss of liver parenchyma. The status of the livers of different groups is expressed as scoring values (degree of scoring and received value/maximum value). Between brackets are the percentages of the values. : denotes P,0.05 compared with the C group. ison of the total received score compared to maximum score: # denotes P,0.01 compared to rats fed diets without bacteria. al.pone.0033510.t002 March 2012 \| Volume 7 \| Issue 3 \| e33510 March 2012 \| Volume 7 \| Issue 3 \| e33510 March 2012 \| Volume 7 \| Issue 3 \| e33510 PLoS ONE \| www.plosone.org 5 Blueberries and Probiotics Attenuate Oncogenesis Table 3. Evaluation of liver injury (Incidence). Table 3. Evaluation of liver injury (Incidence). Disease severity scoring Incidence Groups Steatosis Cellinfiltration1 Cellinfiltration2 Stasis Loss of parenchyma C 8/8 1/8 8/8 4/7 3/6 P 8/8 2/8 8/8 4/5 2/8 2B 4/8 * 0/8 8/8 6/8 5/7 2BP 8/8 0/8 7/8 4/8 2/8 B 3/5 1/5 5/5 0/4 0/4 BP 7/7 1/6 4/6 0/6 * 0/6 Liver specimens were histologically evaluated for the incidence of steatosis, infiltrating inflammatory cells in steatotic and non-steatotic areas, vascular stasis and loss of liver parenchyma. The status of the livers of different groups is expressed as incidence of phenomena. 1Cellinfiltration around CV within steatotic areas. 2Cellinfiltration elsewhere in the parenchyma. Incidence: * denotes P,0.05 compared with the C group. doi:10.1371/journal.pone.0033510.t003 uated for the incidence of steatosis, infiltrating inflammatory cells in steatotic and non-steatotic areas, vascular stasis and loss of ers of different groups is expressed as incidence of phenomena. i mal loss, haemorrhage, and small inflammatory infiltrations in non-steatotic areas (Fig. 7). Displaced nucleus to the periphery of the hepatocytes was occasionally found in livers from groups P (Fig. 8), B and BP. The overall histological changes in the different groups were similar to those of group C, but could be more or less severe. Based on scoring it was seen that, the degree of parenchymal inflammatory infiltration in non-steatotic area was significantly higher in group C (Fig. 7) compared to group 2BP (P = 0.038) (Fig. 9), group B (P = 0.019) and group BP (P,0.001) (Table 2). Significant increases in the degree of steatosis compared to group C was found in group P (P = 0.005) (Fig. 8) and in group BP (P = 0.004) (Table 2). Comparison of the total received score compared to maximum score revealed a significant increase of steatosis in groups P, 2BP and BP compared to rats fed corresponding diets without bacteria (P,0.01) (Table 2). with blueberry husks have lower values (groups 2B, 2BP, B and BP; Fig. 2). A significant increase in haptoglobin levels from the base line to the end of the water period of cycle 10 was shown in all groups: C group (P = 0.01), P group (P,0.001), 2B group (P,0.001), 2BP group (P,0.001), B group (P = 0.002), BP group (P = 0.043) (Fig. 2). Faecal viable count of Enterobacteriacea and lactobacilli Faecal viable count of Enterobacteriacea and lactobacilli At the start (base line), the viable count of Enterobacteriaceae between groups did not show any significant differences. On the last day of the study the Enterobacteriaceae count was higher in group C (P,0.001) and in group B (P = 0.002), compared with their individual base line level, while this increase over time could not be seen in any of the other groups. Quantitatively, the number of lesions classified as low-grade dysplasia was significantly reduced in groups 2B, 2BP and BP (P = 0.05) compared to group C (Fig. 5). A total of 26 dysplastic lesions distributed in 5 animals were found in group C, while the corresponding figures for group 2B were 2 lesions in one animal; for group 2BP and BP, 1 lesion for each group. A similar pattern was found for colonic ulcers, i.e. 11 ulcers were found distributed in 6 animals of group C, while no ulcers were found in group 2B group (P = 0.01); in group 2BP, 2 ulcers were found in 2 animals (P = 0.05); in group B, 1 ulcer was found (P = 0.043) and none in group BP (P = 0.01) (Fig. 6). At the end of the study, the count of Enterobacteriaceae was significantly decreased in groups P (P = 0.003), 2B (P,0.001), 2BP (P = 0.001) and BP (P = 0.017) compared with group C (Fig. 11). A significant decrease in Enterobacteriaceae count was achieved by the addition of probiotics to the diets, i.e. group C versus P (P = 0.003); 2B versus 2BP (P = 0.05); B versus BP (P = 0.036)) (Fig. 11). g At the base line, faecal viable count of lactobacilli differed between groups C and 2B (P = 0.003) and between groups 2B and 2BP (P,0.001). From the start to the end of the study, only group 2B exhibited a decrease in lactobacilli count (P,0.001), while it increased in groups B and BP (P = 0.002 and P = 0.012, respectively). At the last day of the study and compared with group C, the viable count was higher in groups 2BP (P = 0.001), B (P = 0.003) and BP (P = 0.017), and lower in group 2B (P = 0.001) (Fig. 12). The addition of probiotics to the 2B diet resulted in an Histological and macroscopic alterations of colon Macroscopic examination of colon from each animal revealed visible thickening of the colon walls in all groups. Invaginations as a cause of polyps and dilated descending colon were occasionally seen in animals of groups C, B and BP. No polyps were found in the groups 2B and 2BP. Examination in the microscope shows that group C had colonic inflammation, mostly confined to the mucosa and submucosa, with loss of surface epithelium, inflammatory cell infiltrations, loss of goblet cells, crypt distortion and abscesses, mucosal ulceration and erosion, and accompanying submucosal edema (Fig. 3, 4). The diseased condition was found throughout the colon but was particularly prominent on the left side and in the transverse colon. Regenerative and hyperplastic epithelium, which morphologically mostly resembled low-grade dysplasia with some sections of high grade dysplasia and polyps diagnosed as adenocarcinomas, was observed (Fig. 3, 4). The overall histological changes in group P were similar to those of group C, while mucosal injuries were less severe in groups BP, B, 2BP and 2B. In groups 2B and 2BP, no sections of high grade dysplasia and adenomatous polyps were found. None of the animals in groups 2B, 2BP, B and BP showed gross mucosal ulceration. The bleeding per rectum was caused by small and focal erosions. The incidence of steatosis was in comparison with group C found to be significantly reduced in group 2B (P = 0.038) (Fig. 10, Table 3). Compared to group C, the incidence of stasis was decreased in group BP (P = 0.049) (Table 3), and so was also the incidence of translocation to the liver (P,0.05) (Table 4). Identification of faecal lactobacilli All identified isolates were designated to Lactobacillus. The different species found in the different animal groups at different time points are shown in Table 5. The probiotic supplement gradually changed the dominance of L. murinus. Isolates identified as L. plantarum and isolated from groups P, 2BP and BP were identified as the strain L. plantarum HEAL19 (Table 5). Histopathological evaluation of the liver frumenti), Lactobacillus antri (L. antri), Lactobacillus gasseri (L. gasseri), Micrococcus luteus (M. luteus), Clostridium ramosum (C. ramosum), Staphylococcus warneri (S. warneri), Leifsonia xyli subsp. cynodontis (L. xyli subsp. cynodontis), Enterococcus casseliflavus (E. casseliflavus), Clostridium perfringens (C. perfringens), Paenibacillus lautus (P. lautus), Kocuria rosea (K. rosea), Corynebacterium lipophiloflavum (C. lipophiloflavum), Clostridium subterminale (C. subterminale), Bacillus siralis (B. siralis). doi:10 1371/journal pone 0033510 t004 increased viable count of faecal lactobacilli (P = 0.001) whereas it decreased when added to the B diet (P = 0.036) (Fig. 12). 12% in proximal colon, 66%, 16% and 13% in distal colon, and 95%, 1% and 2% in aortic blood (Table 6). Generally, higher proportions of acetic acid (P,0.05) and lower proportions of butyric acid (P,0.05) was found in the hindgut of group 2B compared with groups C and B (Table 6). The proportions of propionic acid in aortic blood was significantly decreased in the P group compared with the C group (P,0.001) and in the 2BP group compared with the 2B group (P,0.01) (Table 6). Histopathological evaluation of the liver Incidence: * denotes P,0.05 compared with the C group. Lactobacillus animalis (L. animalis), Lactobacillus apodemi (L. apodemi), Kocuria rhizophila (K. rhizophila), Lactobacillis frumenti (L. frumenti), Lactobacillus antri (L. antri), Lactobacillus gasseri (L. gasseri), Micrococcus luteus (M. luteus), Clostridium ramosum (C. ramosum), Staphylococcus warneri (S. warneri), Leifsonia xyli subsp. cynodontis (L. xyli subsp. cynodontis), Enterococcus casseliflavus (E. casseliflavus), Clostridium perfringens (C. perfringens), Paenibacillus lautus (P. lautus), Kocuria rosea (K. rosea), Corynebacterium lipophiloflavum (C. lipophiloflavum), Clostridium subterminale (C. subterminale), Bacillus siralis (B. siralis). doi:10.1371/journal.pone.0033510.t004 Thirty-four isolates were subjected to 16 S rDNA sequencing and the similarity levels for each isolate are shown in the table. Translocation is mentioned as incidence of phenomena/total number of animals. Isolates were received from Rogosa agar/Brain Heart Infusion agar (incubated anaerobically)/Brain Heart Infusion agar (incubated aerobically). Between brackets are species with the same sequence similarity as the before mentioned. Incidence: * denotes P,0.05 compared with the C group. Lactobacillus animalis (L. animalis), Lactobacillus apodemi (L. apodemi), Kocuria rhizophila (K. rhizophila), Lactobacillis frumenti (L. frumenti), Lactobacillus antri (L. antri), Lactobacillus gasseri (L. gasseri), Micrococcus luteus (M. luteus), Clostridium ramosum (C. ramosum), Staphylococcus warneri (S. warneri), Leifsonia xyli subsp. cynodontis (L. xyli subsp. cynodontis), Enterococcus casseliflavus (E. casseliflavus), Clostridium perfringens (C. perfringens), Paenibacillus lautus (P. lautus), Kocuria rosea (K. rosea), Corynebacterium lipophiloflavum (C. lipophiloflavum), Clostridium subterminale (C. subterminale), Bacillus siralis (B. siralis). d i 0 3 /j l 0033 0 00 Thirty-four isolates were subjected to 16 S rDNA sequencing and the similarity levels for each isolate are shown in the table. Translocation is mentioned as incidence of phenomena/total number of animals. Isolates were received from Rogosa agar/Brain Heart Infusion agar (incubated anaerobically)/Brain Heart Infusion agar (incubated aerobically). Between brackets are species with the same sequence similarity as the before mentioned. Incidence: * denotes P,0.05 compared with the C group. Lactobacillus animalis (L. animalis), Lactobacillus apodemi (L. apodemi), Kocuria rhizophila (K. rhizophila), Lactobacillis frumenti (L. frumenti), Lactobacillus antri (L. antri), Lactobacillus gasseri (L. gasseri), Micrococcus luteus (M. luteus), Clostridium ramosum (C. ramosum), Staphylococcus warneri (S. warneri), Leifsonia xyli subsp. cynodontis (L. xyli subsp. cynodontis), Enterococcus casseliflavus (E. casseliflavus), Clostridium perfringens (C. perfringens), Paenibacillus lautus (P. lautus), Kocuria rosea (K. rosea), Corynebacterium lipophiloflavum (C. lipophiloflavum), Clostridium subterminale (C. subterminale), Bacillus siralis (B. siralis). doi:10.1371/journal.pone.0033510.t004 Incidence: * denotes P,0.05 compared with the C group. Lactobacillus animalis (L. animalis), Lactobacillus apodemi (L. apodemi), Kocuria rhizophila (K. rhizophila), Lactobacillis frumenti (L. Histopathological evaluation of the liver Livers from group C showed mild to moderate degrees of steatosis. Liver lobules had occasional focal areas with parenchy- PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 6 Blueberries and Probiotics Attenuate Oncogenesis Table 4. Incidence of translocations and identified isolates from the livers. Groups Translocation Identity of isolate Similarity to strain of known identity (%) C 3/8; 2/8; 4/8 L. animalis 99 (2 isolates) L. apodemi 99 K. rhizophila 99 L. frumenti (L. antri) 96 L. antri 99, 93 L. gasseri 100 M. luteus 100 C. ramosum 100 S. warneri 100 (2 isolates) P 0/8; 2/8; 1/8 C. ramosum 100 L. xyli subsp. cynodontis 99 2B 1/8; 1/8; 3/8 L. animalis (L. apodemi) 99 L. apodemi 99 (2 isolates), 100 E. casseliflavus 100 C. perfringens 99 L. oris (L. frumenti, L. antri) 97 L. gasseri 100 K. rhizophila 99 P. lautus 99 2BP 0/8; 3/8; 3/8 K. rosea (K. rhizophila) 99 K. rosea 99 M. luteus 100 S. warneri 100 L. animalis 98 B 0/6; 1/6; 4/6 C. lipophiloflavum 99 C. subterminale 98 C. perfringens 100 K. rhizophila 99 B. siralis 99 BP 0/8; 0/8; 0/8 * - - Thirty-four isolates were subjected to 16 S rDNA sequencing and the similarity levels for each isolate are shown in the table. Translocation is mentioned as incidence of phenomena/total number of animals. Isolates were received from Rogosa agar/Brain Heart Infusion agar (incubated anaerobically)/Brain Heart Infusion agar (incubated aerobically). Between brackets are species with the same sequence similarity as the before mentioned. Incidence: * denotes P,0.05 compared with the C group. Lactobacillus animalis (L. animalis), Lactobacillus apodemi (L. apodemi), Kocuria rhizophila (K. rhizophila), Lactobacillis frumenti (L. frumenti), Lactobacillus antri (L. antri), Lactobacillus gasseri (L. gasseri), Micrococcus luteus (M. luteus), Clostridium ramosum (C. ramosum), Staphylococcus warneri (S. warneri), Leifsonia xyli subsp. cynodontis (L. xyli subsp. cynodontis), Enterococcus casseliflavus (E. casseliflavus), Clostridium perfringens (C. perfringens), Paenibacillus lautus (P. lautus), Kocuria rosea (K. rosea), Corynebacterium lipophiloflavum (C lipophiloflavum) Clostridium subterminale (C subterminale) Bacillus siralis (B siralis) Thirty-four isolates were subjected to 16 S rDNA sequencing and the similarity levels for each isolate are shown in the table. Translocation is mentioned as incidence of phenomena/total number of animals. Isolates were received from Rogosa agar/Brain Heart Infusion agar (incubated anaerobically)/Brain Heart Infusion agar (incubated aerobically). Between brackets are species with the same sequence similarity as the before mentioned. PLoS ONE \| www.plosone.org Blueberries and Probiotics Attenuate Oncogenesis Blueberries and Probiotics Attenuate Oncogenesis Figure 3. Flat dysplasia. Colonic mucosa showing flat dysplasia. The sample was taken from an animal in group C. doi:10.1371/journal.pone.0033510.g003 Discussion It has previously been demonstrated that dysplasia and adenocarcinomas can be induced by cyclic administration of DSS [19]. DSS is not mutagenic [33], and the changes occurring depend on the mucosal inflammatory ulceration and regeneration, recurrence-remission cycles typical of clinical ulcerative colitis cases [34]. Developing therapeutic regimens to combat colorectal cancer without significant side effects is of great interest, and in the current study we used a similar model to determine the protective effect of blueberry husks with and without addition of a probiotic mixture to delay or prevent colon carcinogenesis, and pathological abnormalities of the liver. abnormalities of the liver. The clinical findings during the first cycle of DSS administration showed lower scoring values for groups 2B, 2BP and B compared with the C group (Fig. 1). A gradually increased inflammatory activity was recorded in all groups, indicated by a significant increase in the DAI from cycle 1 through 11 (Fig. 1). There is a well-characterised sequence of changes in the DSS model, beginning with increased mucosal permeability, and eventual ulceration followed by inflammation. After DSS withdrawal, these events are followed by epithelial restitution, cell migration and proliferation. The supposed impaired recovery periods between exposures to DSS may suggest a chronological sequence, where repeated uninhibited acute inflammatory responses with less extension of mucosal repair develop chronicity. At the last DSS cycle, the DAI were significantly lower in groups administered supplementary blueberry husks with or without probiotics (Fig. 1). High level of haptoglobin has been implicated in patients with colorectal cancer and development of hepatic metastases and it has been stated to be a sensitive indicator [35]. In the present study, a successive rise in the concentration of haptoglobin in serum was observed for all groups, from the beginning to the end of the study period. During the 10th cycle of DSS administration, group 2B showed significantly lower values than group C, but also group 2BP had a reduced level that was close to significance (P = 0.054) (Fig. 2). The clinical findings during the first cycle of DSS administration showed lower scoring values for groups 2B, 2BP and B compared with the C group (Fig. 1). A gradually increased inflammatory activity was recorded in all groups, indicated by a significant increase in the DAI from cycle 1 through 11 (Fig. 1). Discussion There is a well-characterised sequence of changes in the DSS model, beginning with increased mucosal permeability, and eventual ulceration followed by inflammation. After DSS withdrawal, these events are followed by epithelial restitution, cell migration and proliferation. The supposed impaired recovery periods between exposures to DSS may suggest a chronological sequence, where repeated uninhibited acute inflammatory responses with less extension of mucosal repair develop chronicity. At the last DSS cycle, the DAI were significantly lower in groups administered supplementary blueberry husks with or without probiotics (Fig. 1). Figure 3. Flat dysplasia. Colonic mucosa showing flat dysplasia. The sample was taken from an animal in group C. doi:10.1371/journal.pone.0033510.g003 were lower in the caecum of rats fed 2B (P,0.001) and in distal part of colon this was valid for both propionic and butyric acid (6.160.6 vs 11.363.2 mmol/g and 3.760.3 vs 1565.7 mmol/g; P = 0.045 and P,0.001, respectively). The butyric acid level in the distal part of colon was higher in rats fed 2BP than in rats fed only 2B (5.060.3 vs.3.760.3 mmol/g; P = 0.005). High level of haptoglobin has been implicated in patients with colorectal cancer and development of hepatic metastases and it has been stated to be a sensitive indicator [35]. In the present study, a successive rise in the concentration of haptoglobin in serum was observed for all groups, from the beginning to the end of the study period. During the 10th cycle of DSS administration, group 2B showed significantly lower values than group C, but also group 2BP had a reduced level that was close to significance (P = 0.054) (Fig. 2). Proportions of SCFAs in the hindgut and in blood In general total levels of SCFAs along the hindgut were lower in rats fed 2B than in rats fed C (P,0.05), due to a lower content of most individual acids. Compared with B the butyric acid levels The proportions of acetic acid, propionic acid and butyric acid were in average 65%, 14% and 14% in caecum, 70%, 13% and PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 7 Levels of SCFAs in aortic blood Acetic acid was the major acid in aortic blood in all groups (in mean 1152.6 mmol/l) followed by butyric acid (in mean 18.7 mmol/l) and propionic acid (in mean 8.8 mmol/l). The level of propionic acid was higher in group C than in groups 2B and B (P,0.05). The propionic acid levels were significantly lower in group P (8.060.4 mmol/l) compared with group C (11.260.2 mmol/l) and in group 2BP (8.360.3 mmol/l) compared with group 2B (9.560.4 mmol/l) (P,0.001 and P,0.05, respectively). The pathogenesis of colorectal carcinogenesis associated with colonic inflammation is believed to involve progression from inflamed and hyperplastic cryptal cells, through dysplasia, to adenoma and carcinoma [36]. Strong circumstantial evidence documents the validity of UC-associated dysplasia as a precursor lesion or marker of carcinomas, and it is likely that colorectal carcinomas evolve through stages of increasingly severe epithelial dysplasia before becoming invasive lesions [37]. According to the present criteria, microscopic findings of descending colon revealed histological abnormalities with low and high grade dysplasia and adenocarcinomatous polyps in groups C, P, B and BP (Fig. 3, 4), whereas only sections with low-grade dysplasia were found in groups 2B and 2BP. Macroscopic evaluation showed a signifi- cantly lower number of low-grade dysplastic lesions compared to group C in colon of groups 2B, 2BP and BP (Fig. 5), as well as fewer mucosal ulcers in groups 2B, 2BP, B and BP (Fig. 6). Figure 4. Dysplasia-associated lesion. Colonic mucosa showing a dysplastic lesion in association with inflammation. The sample was taken from an animal in group C. doi:10.1371/journal.pone.0033510.g004 Increased gut permeability can be explained by malignant tissue disruption of the bowel architecture [38], making gut-derived bacteria and toxins accessible to the liver via the portal circulation [39]. According to the present liver scoring, the degree of parenchymal infiltration was significantly higher in group C compared to groups 2BP (Fig. 7, 9), B, and BP (Table 2), and the incidence of translocation to the liver was lower in group BP (Table 4). In fact, no translocated bacteria were found in this group. Among the identified bacteria isolated from the liver, both Enterococcus and Clostridium spp. were found. Enterococcus spp. are important causes of human infections [40] and different species of Figure 4. Dysplasia-associated lesion. Colonic mucosa showing a dysplastic lesion in association with inflammation. The sample was taken from an animal in group C. Levels of SCFAs in aortic blood doi:10.1371/journal.pone.0033510.g009 Figure 7. Liver injury group C. Focal parenchymal loss and haemorrhage in the liver of an animal in group C. doi:10.1371/journal.pone.0033510.g007 be regarded as an advantage, since it has been demonstrated that TLR4 expression is up-regulated in colitis-associated cancer lesions from patients with UC [45]. In active UC patients, the numbers of faecal lactobacilli decrease, indicating that a reduction in intestinal Lactobacillus species may be a sign of mucosal inflammation [9]. In the present study, the viable count of faecal lactobacilli at the end of the study period and in comparison with group C, was higher in groups 2BP, B and BP, but lower in group 2B (Fig. 12). The addition of probiotics increased the viable count of lactobacilli in group 2BP compared to group 2B (Fig. 12). In contrast, a decrease was found between groups B and BP (Fig. 12). Changes in luminal concentrations of phenolics can affect the microbiota composition [17] and the growth of certain bacterial groups with pathogenic potential. For example, Enterobacteriaceae, Clostridium and Bacteroides can be repressed by phenolics, while some probiotics are relatively unaffected. This may favour intestinal establishment of some probiotics, inhibition of pathogens and improvement of the bacterial balance of the gut [17]. An be regarded as an advantage, since it has been demonstrated that TLR4 expression is up-regulated in colitis-associated cancer lesions from patients with UC [45]. In active UC patients, the numbers of faecal lactobacilli decrease, indicating that a reduction in intestinal Lactobacillus species may be a sign of mucosal inflammation [9]. In the present study, the viable count of faecal lactobacilli at the end of the study period and in comparison with group C, was higher in groups 2BP, B and BP, but lower in group 2B (Fig. 12). The addition of probiotics increased the viable count of lactobacilli in group 2BP compared to group 2B (Fig. 12). In contrast, a decrease was found between groups B and BP (Fig. 12). Changes in luminal concentrations of phenolics can affect the microbiota composition [17] and the growth of certain bacterial groups with pathogenic potential. For example, Enterobacteriaceae, Clostridium and Bacteroides can be repressed by phenolics, while some probiotics are relatively unaffected. This may favour intestinal establishment of some probiotics, inhibition of pathogens and improvement of the bacterial balance of the gut [17]. An Okayasu et al., [34] presented significant increases in Enterobac- teriaceae and Clostridium spp. Levels of SCFAs in aortic blood doi:10.1371/journal.pone.0033510.g004 PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 8 Blueberries and Probiotics Attenuate Oncogenesis Figure 5. Quantification of dysplastic lesions. Number of dysplastic lesions in colon and rectum classified as low grade dysplasia in different treatment groups. * denotes P,0.05 compared to group C. doi:10.1371/journal.pone.0033510.g005 Blueberries and Probiotics Attenuate Oncogenesis Figure 5. Quantification of dysplastic lesions. Number of dysplastic lesions in colon and rectum classified as low grade dysplasia in different treatment groups. * denotes P,0.05 compared to group C. doi:10.1371/journal.pone.0033510.g005 with a diagnosis of gastrointestinal or hematologic malignancy [42]. A gastrointestinal source of infection, particularly carcinoma of the colon or rectum or enterocolitis, was evident in most patients [42]. Fatty infiltration of hepatocytes has been reported in intestinal inflammation [6]. The incidence of steatosis in the present study was found to be significantly reduced in group 2B (Fig. 10, Table 3), compared to group C (Fig. 7), while the scoring of steatosis was increased in groups P (Fig. 8) and BP (Table 2). Clostridium have been implicated in the induction of intestinal inflammation, due to production of extracellular toxins, and they may be detrimental during increased permeability of the colonic mucosa [41]. It should be pointed out, that group 2BP was the only group where no Clostridium was found in the liver (Table 4). A significantly reduced degree of parenchymal cellular infiltration was also found in this group (Table 2). These findings may be significant because Clostridium bacteriemia have been evident in many patients Figure 6. Quantification of colonic and rectal ulcers. Number of ulcers of colonic and rectal mucosa in different treatment groups. ** denotes P#0.01 and * denotes P,0.05 compared to group C. doi:10.1371/journal.pone.0033510.g006 Figure 6. Quantification of colonic and rectal ulcers. Number of ulcers of colonic and rectal mucosa in different treatment groups. ** denotes P#0.01 and * denotes P,0.05 compared to group C. doi:10.1371/journal.pone.0033510.g006 March 2012 \| Volume 7 \| Issue 3 \| e33510 PLoS ONE \| www.plosone.org 9 Blueberries and Probiotics Attenuate Oncogenesis Blueberries and Probiotics Attenuate Oncogenesis Figure 7. Liver injury group C. Focal parenchymal loss and haemorrhage in the liver of an animal in group C. doi:10.1371/journal.pone.0033510.g007 Figure 9. Liver injury group 2BP. Parenchymal inflammatory infiltration in steatotic areas from an animal in group 2BP. doi:10.1371/journal.pone.0033510.g009 Figure 9. Liver injury group 2BP. Parenchymal inflammatory infiltration in steatotic areas from an animal in group 2BP. Levels of SCFAs in aortic blood in faeces during DSS administration and C. ramosum were particularly evident after repeated adminis- trations. Also, C. ramosum has been one of the most frequently isolated anaerobes from the inflamed mucosa of UC patients, against which an enhanced antibody response was found [43]. In a study by Sokol et al., [44], the biodiversity of the active microbiota was shown to be lower for UC patients than for healthy controls and Escherichia coli (or related Enterobacteriaceae) were significantly associated with UC. In the present study, the viable count of Enterobacteriaceae in faeces increased during the experimental period in group C but also in group B. However, at the end of the study, a significant decrease was obvious in all groups, except for group B, compared to group C (Fig. 11). Also, the addition of probiotics seems to suppress Enterobacteriacea (Fig. 11). The action of LPS from gram-negative bacteria is effectuated via the toll-like receptor 4 (TLR4) protein and a decrease in the load of Enterobacteriaceae must Figure 8. Liver injury group P. Hepatocytes showing displaced nucleus to the periphery, known as ballooning in an animal from group P. doi:10.1371/journal.pone.0033510.g008 Figure 10. Liver injury group 2B. Parenchymal loss and haemor- rhage in liver from an animal in group 2B. doi:10.1371/journal.pone.0033510.g010 Figure 8. Liver injury group P. Hepatocytes showing displaced nucleus to the periphery, known as ballooning in an animal from group P. doi:10.1371/journal.pone.0033510.g008 Figure 10. Liver injury group 2B. Parenchymal loss and haemor- rhage in liver from an animal in group 2B. doi:10.1371/journal.pone.0033510.g010 doi:10.1371/journal.pone.0033510.g008 March 2012 \| Volume 7 \| Issue 3 \| e33510 PLoS ONE \| www.plosone.org 10 Blueberries and Probiotics Attenuate Oncogenesis Figure 11. Faecal load of Enterobacteriaceae. Viable count of Enterobacteriaceae (log CFU/g faecal sample) P,0.05, P,0.01, P#0.001 compared to group C. By the addition of probiotics: group C vs group P P,0.01; group 2B vs group 2BP #P#0.05; B group vs BP group P,0.05. doi:10.1371/journal.pone.0033510.g011 Figure 11. Faecal load of Enterobacteriaceae. Viable count of Enterobacteriaceae (log CFU/g faecal sample) P,0.05, P,0.01, * P#0.001 compared to group C. By the addition of probiotics: group C vs group P *P,0.01; group 2B vs group 2BP #P#0.05; B group vs BP group P,0.05. doi:10.1371/journal.pone.0033510.g011 increase in numbers of E. coli and clostridia in patients with active UC, may contribute to pathogenesis [8]. In the present study, the added strain of L. Levels of SCFAs in aortic blood plantarum was isolated and identified from faeces of animals that had been supplemented with probiotics (Table 5). Since Clostridium was not found in the livers of group 2BP (Table 4), it can be hypothesised that L. plantarum had a beneficial effect that supplemented that of the blueberry husks. The results of Russel et al., [46] suggest that the protective anti-inflammatory effect of blueberry associated phenolics in colon, probably occur due to increase in numbers of E. coli and clostridia in patients with active UC, may contribute to pathogenesis [8]. In the present study, the added strain of L. plantarum was isolated and identified from faeces of animals that had been supplemented with probiotics (Table 5). Since Clostridium was not found in the livers of group 2BP (Table 4), it can be hypothesised that L. plantarum had a beneficial effect that supplemented that of the blueberry husks. The results of Russel et al., [46] suggest that the protective anti-inflammatory effect of blueberry associated phenolics in colon, probably occur due to microbial metabolism, which in turn is dependent on the individual composition of the microbiota. The formation of SCFAs is mostly regulated by substrate availability and composition of the microbiota. Lower proportions of butyrate and higher proportions of acetate have been found in patients with adenomatous polyps [47], suggesting an increased floral capacity to produce acetate and a decreased one in the capacity to form butyrate [47]. In fact, the capacity of stool- derived bacteria from patients with colonic adenomas and colon Figure 12. Faecal load of lactobacilli. Viable count of lactobacilli (log CFU/g faecal sample)P,0.05, P,0.01, P#0.001 compared to group C. A decrease was found in group 2B * P#0.001 compared to group C. By the addition of probiotics: 2B group vs 2BP group **P#0.001, and B vs BP group P,0.05. doi:10.1371/journal.pone.0033510.g012 Figure 12. Faecal load of lactobacilli. Viable count of lactobacilli (log CFU/g faecal sample)P,0.05, P,0.01, P#0.001 compared to group C. A decrease was found in group 2B * P#0.001 compared to group C. By the addition of probiotics: 2B group vs 2BP group **P#0.001, and B vs BP group P,0.05. doi:10.1371/journal.pone.0033510.g012 March 2012 \| Volume 7 \| Issue 3 \| e33510 11 11 Blueberries and Probiotics Attenuate Oncogenesis Table 6. SCFAs in colon content and blood samples. Levels of SCFAs in aortic blood C P 2B 2BP B BP Caecum Acetic 66.0a 64.9 75.6b 75.2 69.0a 69.6 Propionic 16.9a 16.1 15.1a,b 14.8 13.5b 12.9 Butyric 17.1a 19.1 9.2b 9.9 17.6a 17.5 Total (mmol/g) 82.0a 89.7 55.3b 59.8 75.3a 80.4 Proximal colon Acetic 69.5a 70.9 76.0b 77.8 70.9a 72.7 Propionic 16.1 13.6 15.2 13.5 13.4 11.8 Butyric 14.4a 15.5 8.8b 8.8 15.7a 15.5 Total (mmol/g) 74.3a 73.9 56.0b 63.7 72.8a,b 75.3 Distal colon Acetic 66.4a 66.0 73.1b 73.0 67.5a 69.5 Propionic 18.4 19.1 16.6 15.6 14.5 14.5 Butyric 15.3a 14.9 10.3b 11.5 18.0a 16.0 Total (mmol/g) 75.0a 75.8 38.6b 47.6 82.1a 70.5 Blood from aorta1 Acetic 97.0 97.5 97.9 97.4* 97.9 98.0 Propionic 1.1 0.7* 0.7 0.8 0.7 0.6 Butyric 1.9 1.8 1.5 1.8 1.4 1.4 Table 6. SCFAs in colon content and blood samples. Table 5. Identification to species level of faecal lactobacilli. Group Base line After DSS cycle 1 After DSS cycle 10 Termination C L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus L. murinus P L. johnsonii L. johnsonii L. reuteri L. johnsonii L. plantarum 2B L. murinus L. murinus L. reuteri L. murinus L. reuteri L. murinus L. murinus L. murinus L. murinus 2BP L. gasseri L. plantarum L. plantarum L. gasseri L. plantarum L. plantarum L. gasseri L. plantarum L. plantarum B L. murinus L. reuteri L. murinus L. johnsonii L. murinus L. murinus L. reuteri L. reuteri L. johnsonii L. murinus BP L. vaginalis L. plantarum L. murinus L. plantarum L. vaginalis L. plantarum L. vaginalis L. plantarum L. plantarum Identification of isolates from viable count of faeces from the different groups of rats. (L. = Lactobacillus). doi:10.1371/journal.pone.0033510.t005 Proportions (%) of acetic-, propionic- and butyric acids in the hindgut and blood of rats given diets supplemented with different amounts of blueberry husks (2B = 122 g blueberry/kg dwb and B = 61 g blueberry/kg dwb) and/or probiotics (P; 6?109 CFU per day). p p y Probiotics: Bifidobacterium infantis (CURE21), Lactobacillus gasseri VPG44, and Lactobacillus plantarum (HEAL19). Mean values of C, 2B and B only, i.e. those without any added probiotics, with unlike superscript letters were significantly different (P,0.05). Mean values were significantly different from those rats fed diets without bacteria: P,0.05, * P,0.01, *** P,0.001. cancer to produce butyrate was significantly reduced [48]. PLoS ONE \| www.plosone.org Levels of SCFAs in aortic blood In the present study, the proportion of butyric acid was decreased and that of acetic acid increased in caecum, proximal and distal colon of group 2B compared with groups C and B (Table 6), and the same trend was seen in the 2BP group (Table 6). The utilisation of butyrate, the major energy substrate for colonocytes, is signifi- cantly impaired during both severe colitis [49] and during hepatic cirrhosis [50]. Negative correlations between lipid peroxidation and butyric acid as well as between DAI and butyric acid have also been shown in DSS-induced acute colitis [24]. However, histological examination revealed less epithelial affection in groups 2B and 2BP compared with the other groups, which may explain the superior disposal of butyrate in these groups, as judged by the findings of lower proportions of butyric acid. The high content of uronic acids in blueberry husks could explain the increased proportion of acetic acid (Table 6). 12B (n = 7), B and BP (n = 6). doi:10.1371/journal.pone.0033510.t006 doi:10.1371/journal.pone.0033510.t006 observed along with an increase in the concentration of oat. This might be caused by the increased fatty acid and cholesterol synthesis, which is enhanced in the liver when feeding soluble oat fibre [53]. The increase in fatty acid synthesis was due to the increased flow of SCFAs in the liver and a positive correlation between plasma propionate concentrations and hepatic fatty acid synthesis was found [53]. However, no cirrhotic, necrotic or inflammatory changes were observed so, in this model, the livers were not otherwise affected [53]. In the present study, the liver was highly affected from chronic intestinal inflammation, which probably impairs normal function. The addition of probiotic bacteria caused a significant increase in total scoring of steatosis compared to those animals fed diets without bacteria (Table 2), but the concentration of propionic acid in the blood decreased. It can be speculated that the livers from groups P, 2BP, B and BP have better preserved capability of fatty acid synthesis, which may reflect an improved liver capacity. observed along with an increase in the concentration of oat. This might be caused by the increased fatty acid and cholesterol synthesis, which is enhanced in the liver when feeding soluble oat fibre [53]. The increase in fatty acid synthesis was due to the increased flow of SCFAs in the liver and a positive correlation between plasma propionate concentrations and hepatic fatty acid synthesis was found [53]. References Roediger WE (1982) Utilization of nutrients by isolated epithelial cells of the rat colon. Gastroenterology 83: 424–429. 15. Drzikova B, Dongowski G, Gebhardt E (2005) Dietary fibre-rich oat-based products affect serum lipids, microbiota, formation of short-chain fatty acids and steroids in rats. Br J Nutr 94: 1012–1025. 37. Sjo¨qvist U (2004) Dysplasia in ulcerative colitis-clinical consequences? Langenbecks Arch Surg 389: 354–360. 38. Lescut D, Colombel JF, Vincent P, Cortot A, Fournier L, et al. (1990) Bacterial translocation in colorectal cancers. Gastroenterol Clin Biol 14: 811–814. J 16. Scheppach W, Bartram HP, Richter F (1995) Role of short-chain fatty acids in the prevention of colorectal cancer. Eur J Cancer 31A: 1077–1080. 39. Masubuchi Y, Horie T (2004) Endotoxin-mediated disturbance of hepatic cytochrome P450 function and development of endotoxin tolerance in the rat model of dextran sulfate sodium-induced experimental colitis. Drug Metab Dispos 32: 437–441. 17. Lee HC, Jenner AM, Low CS, Lee YK (2006) Effect of tea phenolics and their aromatic fecal bacterial metabolites on intestinal microbiota. Res Microbiol 157: 876–884. 18. Yi WG, Fischer J, Krewer G, Akoh CC (2005) Phenolic compounds from blueberries can inhibit colon cancer cell proliferation and induce apoptosis. J Agric Food Chem 53: 7320–7329. 40. Lewis CM, Zervos MJ (1990) Clinical manifestations of enterococcal infection. Eur J Clin Microbiol Infect Dis 9: 111–117. 41. Garcı´a-Lafuente A, Antolı´n M, Guarner F, Crespo E, Salas A, et al. (1998) Derangement of mucosal barrier function by bacteria colonizing the rat colonic mucosa. Eur J Clin Invest 28: 1019–1026. 19. Cooper HS, Murthy S, Kido K, Yoshitake H, Flanigan A (2000) Dysplasia and cancer in the dextran sulfate sodium mouse colitis model. Relevance to colitis- associated neoplasia in the human: a study of histopathology, b-catenin and p53 expression and the role of inflammation. Carcinogenesis 21: 757–768. J 42. Myers G, Ngoi SS, Cennerazzo W, Harris L, DeCosse JJ (1992) Clostridial septicemia in an urban hospital. Surg Gynecol Obstet 174: 291–296. expression and the role of inflammation. Carcinogenesis 21: 757– 20. Asp NG, Johansson CG, Hallmer H, Siljestro¨m M (1983) Rapid enzymatic assay of insoluble and soluble dietary fiber. J Agric Food Chem 31: 476–482. ˚ 43. Matsuda H, Fujiyama Y, Andoh A, Ushijima T, Kajinami T, et al. (2000) Characterization of antibody responses against rectal mucosaassociated bacterial flora in patients with ulcerative colitis. J Gastroenterol Hepatol 15: 61–68. 21. References 26. Deutschman CS, Cereda M, Ochroch EA, Raj NR (2006) Sepsis-induced cholestasis, steatosis, hepatocellular injury, and impaired hepatocellular regeneration are enhanced in interleukin-6 2/2 mice. Crit Care Med 34: 2613–2620. , ( ) 2. Gupta RB, Harpaz N, Itzkowitz S, Hossain S, Matula S, et al. (2007) Histologic inflammation is a risk factor for progression to colorectal neoplasia in ulcerative colitis: a cohort study. Gastroenterology 133: 1099–1105. 2. Gupta RB, Harpaz N, Itzkowitz S, Hossain S, Matula S, et al. (2007) Histologic inflammation is a risk factor for progression to colorectal neoplasia in ulcerative colitis: a cohort study. Gastroenterology 133: 1099–1105. y gy 3. Eaden JA, Mayberry JF (2000) Colorectal cancer complicating ulcerative colitis: a review. Am J Gastroenterol 95: 2710–2719. 3. Eaden JA, Mayberry JF (2000) Colorectal cancer complicating ulcerative colitis: a review. Am J Gastroenterol 95: 2710–2719. 27. Quednau M, Ahrne´ S, Peterson A, Molin G (1998) Identification of clinically important species of Enterococcus within 1 day with randomly amplified polymorphic DNA (RAPD). Curr Microbiol 36: 332–336. 4. Pohl C, Hombach A, Kruis W (2000) Chronic inflammatory bowel disease and cancer. Hepatogastroenterology 47: 57–70. 28. Brosius J, Palmer ML, Kennedy PJ, Noller HF (1978) Complete nucleotide sequence of a 16 S ribosomal RNA gene from Escherichia coli. Proc Natl Acad Sci USA 75: 4801–4805. 5. Ludwig T, Barham S, La Russo NF, Elveback LR, Wiesner RH, et al. (1981) Morphological features of chronic hepatitis associated with primary sclerosing cholangitis and chronic ulcerative colitis. Hepatology 1: 632–640. 29. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215: 403–410. g p gy 6. Shepherd HA, Selby W, Chapman RW, Nolan D, Barbatis C, et al. (1983) Ulcerative colitis and persistent liver dysfunction. Q J Med 52: 503–513. Shepherd HA, Selby W, Chapman RW, Nolan D, Barbatis C, et Ulcerative colitis and persistent liver dysfunction. Q J Med 52: 503–5 30. Cole JR, Chai B, Marsh TL, Farris RJ, Wang Q, et al. (2003) The Ribosomal Database Project (RDP-II): previewing and a new autoaligner that allows regular updates and the new prokaryotic taxonomy. Nucleic Acids Res 31: 442–443. ˚ 30. Cole JR, Chai B, Marsh TL, Farris RJ, Wang Q, et al. (2003) The Ribosomal Database Project (RDP-II): previewing and a new autoaligner that allows regular 7. Vannucci L, Stepankova R, Grobarova V, Kozakova H, Rossmann P, et al. References (2009) Colorectal carcinoma: importance of colonic environment for anti-cancer response and systemic immunity. J Immunotoxicol 6: 217–226. j p g g g updates and the new prokaryotic taxonomy. Nucleic Acids Res 31: 442–443. ˚ 31. Zhao G, Liu J-f, Nyman M, Jo¨nsson JA˚ (2007) Determination of short-chain fatty acids in serum by hollow fiber supported liquid membrane extraction coupled with gas chromatography. J Chromatogr 846: 202–208. ˚ 8. Mylonaki M, Rayment NB, Rampton DS, Hudspith BN, Brostoff J (2005) Molecular characterization of rectal mucosa-associated bacterial flora in inflammatory bowel disease. Inflamm Bowel Dis 11: 481–487. 32. Zhao G, Nyman M, Jo¨nsson JA˚ (2006) Rapid determination of short-chain fatty acids in colonic contents and faeces of humans and rats by acidified water- extraction and direct-injection gas chromatography. Biomed Chromatogr 20: 674–682. 9. Bullock NR, Booth JC, Gibson GR (2004) Comparative composition of bacteria in the human intestinal microflora during remission and active ulcerative colitis. Curr Issues Intest Microbiol 5: 59–64. 10. De Souza DA, Greene LJ (2005) Intestinal permeability and systemic infections in critically ill patients: effect of glutamine. Crit Care Med 33: 1125–1135. 33. Nagoya T, Hattori Y, Kobayashi F (1981) Mutagenicity and cytogenicity studies of dextran sulfate. Pharmacometrics 22: 621–627. in critically ill patients: effect of glutamine. Crit Care Med 33: 1 11. Kono H, Asakawa M, Fujii H, Maki A, Amemiya H, et al. (2003) Edaravone, a novel free radical scavenger, prevents liver injury and mortality in rats administered endotoxin. J Pharmacol Exp Ther 307: 74–82. 34. Okayasu I, Hatakeyama S, Yamada M, Ohkusa T, Inagaki Y, et al. (1990) A novel method in the induction of reliable experimental acute and chronic ulcerative colitis in mice. Gastroenterology 98: 694–702. J p 12. De Brese M, Schrezenmeir J (2008) Probiotics, prebiotics, and synbiotics. Adv Biochem Engin Biotechnol 111: 1–66. 35. Ward AM, Cooper EH, Turner R, Anderson JA, Neville AM (1977) Acute- phase reactant protein profiles: an aid to monitoring large bowel cancer by CEA and serum enzymes. Br J Cancer 35: 170–178. 13. Ruppin H, Bar-Meir S, Soergel KH, Wood CM, Schmitt MG, Jr. (1980) Absorption of short-chain fatty acids by the colon. Gastroenterology 78: 1500–1507. 36. Riddel RH, Goldman H, Ransohof DF, Appleman HD, Fenoglio CM, et al. (1983) Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical application. Hum Pathol 14: 931–968. 14. Levels of SCFAs in aortic blood However, no cirrhotic, necrotic or inflammatory changes were observed so, in this model, the livers were not otherwise affected [53]. In the present study, the liver was highly affected from chronic intestinal inflammation, which probably impairs normal function. The addition of probiotic bacteria caused a significant increase in total scoring of steatosis compared to those animals fed diets without bacteria (Table 2), but the concentration of propionic acid in the blood decreased. It can be speculated that the livers from groups P, 2BP, B and BP have better preserved capability of fatty acid synthesis, which may reflect an improved liver capacity. The increased concentration of propionic acid in serum of patients with chronic hepatitis of various etiology at the stage of hepatic cirrhosis with hepatic encephalopathy syndrome, has been shown to correlate with the results of clinical and laboratory methods [51]. In the present study, the addition of probiotics to the diet, significantly decreased the level of propionic acid in aortic blood in group P and group 2BP, compared with groups C and 2B, respectively. When the proportion of propionic acid in aortic blood was taken into account, lower values were found in all treatment groups (Table 6). In conclusion, DAI was significantly decreased by blueberry husks administered solely or along with probiotics. The use of blueberry husks alone or in association with probiotics reduced the number of dysplastic lesions and mucosal ulcers. The probiotic mixture could decrease the load of faecal Enterobacteriaceae and increase that of lactobacilli. The colonic epithelium of groups given the high dose of blueberry, with or without probiotics, was Fat accumulation in the liver of hypercholesterolaemic rats fed oat fibres has been shown earlier [52] and the accumulation was March 2012 \| Volume 7 \| Issue 3 \| e33510 March 2012 \| Volume 7 \| Issue 3 \| e33510 12 Blueberries and Probiotics Attenuate Oncogenesis less affected. In the same groups, haptoglobin levels were decreased. Furthermore, the probiotic mixture seems to provide protection against hepatic damage. Our data indicates a therapeutic option for use of blueberry husks and probiotics to delay colonic carcinogenesis and the subsequent hepatic damage, at least in the applied animal model. declare, but Siv Ahrne´, Go¨ran Molin and Bengt Jeppsson are minority shareholders in Probi AB. A˚ sa Ha˚kansson and Camilla Bra¨nning performed the animal experiments and the experimental work together during their PhD-studies. Author Contributions Conceived and designed the experiments: A˚ H CB GM DA MLH BJ MN SA. Performed the experiments: A˚ H CB DA MLH. Analyzed the data: A˚ H CB GM DA MLH BJ MN SA. Contributed reagents/materials/analysis tools: A˚ H CB DA MLH. Wrote the paper: A˚ H CB MLH. Acknowledgments We thank Lantma¨nnen (Ja¨rna, Sweden) and Probi AB (Lund, Sweden) for kindly supplying the oat bran and the freeze-dried blueberry husks and probiotics, respectively. None of the authors has any conflicts of interest to 52. Ma¨lkki Y, Autio K, Ha¨nninen O, Myllyma¨ki O, Pelkonen K, et al. (1992) Oat bran concentrates: physical properties of b-glucan and hypocholesterolemic effects in rats. Cereal Chemistry 69: 647–653. 51. Minushkin ON, Ardatskaia MD, Elizarova NA (2003) Short-chain fatty acids and implications of their measurement in the diagnosis of encephalopathies. Klin Med (Mosk) 81: 55–59. 50. Onori L, Pimpo MT, Palumbo GC, Gili L, Marchetti G, et al. (2001) N- Butyrate rectal transport in cirrhotic patients. Dig Dis Sci 46: 2084–2088. 49. Vernia P, Marcheggiano A, Caprilli R, Frieri G, Corrao G, et al. (1995) Short- chain fatty acid topical treatment in distal ulcerative colitis. Aliment Pharmacol Ther 9: 309–313. 48. Clausen MR, Bonne´n H, Mortensen PB (1991) Colonic fermentation of dietary fibre to short chain fatty acids in patients with adenomatous polyps and colonic cancer. Gut 32: 923–928. 53. Illman RJ, Topping DL (1985) Effects of dietary oat bran on faecal steroid excretion, plasma volatile fatty acids and lipid synthesis in rats. Nutrition Research 5: 839–846. References Theander O, A˚ man P, Westerlund E, Andersson R, Pettersson D (1995) Total dietary fiber determined as neutral sugar residues, uronic acid residues, and Klason lignin (the Uppsala method): collaborative study. J AOAC Int 78: 1030–1044. 44. Sokol H, Lepage P, Seksik P, Dore´ J, Marteau P (2006) Temperature gradient gel electrophoresis of fecal 16 S rRNA reveals active Escherichia coli in the microbiota of patients with ulcerative colitis. J Clin Microbiol 44: 3172–3177. 22. Murthy S, Murthy NS, Coppola D, Wood DL (1997) The efficiency of BAY y 1015 in dextran sulfate model of mouse colitis. Inflamm Res 46: 224–233. 45. Fukata M, Chen A, Vamadevan AS, Cohen J, Breglio K, et al. (2007) Toll-like receptor-4 promotes the development of colitis-associated colorectal tumors. Gastroenterology 133: 1869–1881. 23. Cooper HS, Murthy SN, Shah RS, Sedergran DJ (1993) Clinicopathologic study of dextran sulphates sodium experimental murine colitis. Lab Invest 69: 238–249. 46. Russel WR, Labat A, Scobbie L, Duncan SH (2007) Availability of blueberry phenolics for microbial metabolism in the colon and the potential inflammatory implications. Mol Nutr Food Res 51: 726–731. 24. Ha˚kansson A, Bra¨nning C, Adawi D, Molin G, Nyman M, et al. (2009) Blueberry husks, rye bran and multi-strain probiotics affect the severity of colitis induced by dextran sulphate sodium. Scand J Gastroenterol 44: 1213–1225. 47. Weaver GA, Krause JA, Miller TL, Wolin MJ (1988) Short chain fatty acid distributions of enema samples from a sigmoidoscopy population: an association of high acetate and low butyrate ratios with adenomatous polyps and colon cancer. Gut 29: 1539–1543. 25. Brunt EM, Jannery CG, Di Biscegle AM, Neuschwander-Tetri BA, Bacon BR (1999) Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 94: 2467–2474. PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 March 2012 \| Volume 7 \| Issue 3 \| e33510 13 Blueberries and Probiotics Attenuate Oncogenesis Blueberries and Probiotics Attenuate Oncogenesis 48. Clausen MR, Bonne´n H, Mortensen PB (1991) Colonic fermentation of dietary fibre to short chain fatty acids in patients with adenomatous polyps and colonic cancer. Gut 32: 923–928. 49. Vernia P, Marcheggiano A, Caprilli R, Frieri G, Corrao G, et al. (1995) Short- chain fatty acid topical treatment in distal ulcerative colitis. Aliment Pharmacol Ther 9: 309–313. PLoS ONE \| www.plosone.org March 2012 \| Volume 7 \| Issue 3 \| e33510 14
https://openalex.org/W2105054343	https://www.scielo.br/j/bjps/a/cTTsh4FrLGQjdQ4g33pvLHw/?lang=en&format=pdf	English	null	Prevalence of hemoglobinopathies in school children: the importance of using confirmatory methods	Brazilian Journal of Pharmaceutical Sciences	2,015	cc-by	3,744	Cristiane Fernandes de Freitas Tavares, Jacqueline da Silva Guimarães, Ana Maria de Souza* Department of Clinical Analysis, Toxicology and Food Science, Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, SP, São Paulo, Brazil The hemoglobinopathies are included among the most common genetic diseases in the world. In Brazil, hemoglobinopathies are related to the diversity of racial backgrounds and the degree of interbreeding. The study focused on the prevalence of hemoglobinopathies using conventional and confirmatory laboratory tests in children from public schools in Ribeirão Preto-SP. The study involved the participation of 427 children between six and nine years of age. Hematologic evaluation, hemoglobin electrophoresis on cellulose acetate at alkaline pH, quantification of hemoglobin fractions by high performance liquid chromatography (HPLC) and detection of -α3.7 deletion for α thalassemia by polymerase chain reaction were performed. The results of hemoglobin electrophoresis on cellulose acetate and HPLC of the children studied showed the presence of 30 children (7%) with hemoglobinopathies. Eleven children presented results indicating suspicion of S/β-thalassemia; their parents and/or siblings were evaluated and confirmed the presence of only Hb S. The analysis of deletion -α3.7 to characterize α-thalassemias sampling performed on 207 participants identified 26 children (12.6%) with deletion -α3.7. Thus, 54 (12.6%) of the children studied present this genetic alteration. For the detection of α-thalassemias it is necessary to use confirmatory methods such as molecular analysis and evaluation of family members in doubtful cases to facilitate genetic counseling in families, in which deletion -α3.7 is more frequent in Brazil. Uniterms: Hemoglobinopathy. Hemoglobin electrophoresis. High performance liquid chromatography/ quantitative analysis. Polymerase chain reaction. As hemoglobinopatias estão incluídas nas doenças genéticas mais comuns no mundo. No Brasil, as hemoglobinopatias são relatadas pela diversidade racial e o grau de miscigenação. O estudo focou a prevalência das hemoglobinopatias usando métodos laboratoriais convencionais como a eletroforese de hemoglobina em acetato de celulose em pH alcalino e confirmatório por reação em cadeia de polimerase (PCR) em crianças de escolas públicas de Ribeirão Preto-SP. O estudo envolveu a participação de 427 crianças entre 6-9 anos de idade. Determinaram-se os valores hematológicos, efetuou-se eletroforese de hemoglobina em acetato de celulose em pH alcalino, quantificação das frações de hemoglobina por HPLC e a detecção da deleção -α3,7 pela PCR. Os resultados da eletroforese de hemoglobina em acetato de celulose e do HPLC, nas crianças estudadas, mostraram a presença de 30 crianças (7%) com hemoglobinopatias. Correspondence: A. M. de Souza. Laboratório de Hematologia. Departamento de Análises Clínicas, Toxicológicas e Bromatológicas. Faculdade de Ciências Farmacêuticas de Ribeirão Preto. Universidade de São Paulo. Av. do Café, s/n - Monte Alegre - Bloco M, 20 andar, Sala 053-A. 14040-903 - Ribeirão Preto, SP, Brasil. E-mail: amsouza@fcfrp.usp.br Article Brazilian Journal of Pharmaceutical Sciences vol. 51, n. 2, apr./jun., 2015 http://dx.doi.org/10.1590/S1984-82502015000200013 Brazilian Journal of Pharmaceutical Sciences vol. 51, n. 2, apr./jun., 2015 http://dx.doi.org/10.1590/S1984-82502015000200013 Unitermos: Hemoglobinopatias. Eletroforese de hemoglobina. Cromatografia líquida de alta eficiência/ análise quantitativa. Reação em cadeia de polimerase. Cristiane Fernandes de Freitas Tavares, Jacqueline da Silva Guimarães, Ana Maria de Souza Department of Clinical Analysis, Toxicology and Food Science, Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, SP, São Paulo, Brazil ristiane Fernandes de Freitas Tavares, Jacqueline da Silva Guimarães, Ana M INTRODUCTION causing α+-talassemia (Weatherall, 2001). Molecular analyses indicate a high prevalence of deletion -α3.7 in the state of São Paulo (Sonati et al., 1991; Borges et al., 2001; Oliveira et al., 2006). The hemoglobinopathies are a group of inherited diseases that are classified based on the presence of structurally abnormal hemoglobin (Hb) such as hemoglobins S, C, D and E, and/or one or more globin chain disabilities, known as thalassemias (Clark, Thein, 2004; Henderson et al., 2009). These pathologies are included among the most common genetic diseases in the world, with an estimated prevalence of 7% of the worldwide population (Melo-Reis et al., 2006; Manca, Masala, 2008). Conventional laboratory diagnosis for α-thalassemia can be characterized by precipitation of hemoglobin H, visualized by supravital staining. This diagnosis can be difficult when there is a concomitant hemoglobinopathy such as Hb S, Hb C, Hb E and β-thalassemia, because of the decrease or absence of Hb H and cellular inclusions (Chui et al., 2003).i In general, the most used method for the identification of hemoglobinopathies is hemoglobin electrophoresis on cellulose acetate at alkaline pH, however its sensitivity is limited, especially in the case of isoform co-migration, such as Hb S, Lepore and Hb D (Chinelato-Fernandes et al., 2003). Thus, confirmatory tests are required, such as acid agarose gel electrophoresis, although there is controversy regarding the distinction of hemoglobins D and G, and O and E (Ondei et al., 2007). High-performance liquid chromatography (HPLC) and molecular techniques such as polymerase chain reaction (PCR) and sequencing the gene are more accurate methods for the diagnosis of hemoglobinopathies (Bertholo, Moreira, 2006; Wenning et al., 2000; Bonini-Domingos, 2004). In Brazil, hemoglobinopathies are related to the diversity of racial backgrounds and the degree of interbreeding, and may be regional. The most frequent hemoglobin variants are S and C of African origin, however, due to interbreeding these hemoglobins came to be found in other ethnic groups (Aigner et al., 2006). The impact of population migration increases the combinations of abnormal hemoglobins and the variety of mutations (Henderson et al., 2009). In most abnormal hemoglobins a point mutation, in other words, a single amino acid substitution occurs. Over 1100 hemoglobin variants involving the chains α, β, δ and γ have been described (Wagner et al., 2005). In the heterozygous state, production of Hb A and variant hemoglobin occurs without severe clinical manifestations. INTRODUCTION In the homozygous state, Hb A is absent, thus possibly resulting severe anemia. The interaction of two hemoglobin variants or a combination of a hemoglobin variant and a genetic alteration to thalassemia may still occur (Clark, Thein, 2004). The combined analysis of all methods in the diagnosis of hemoglobinopathies is the most appropriate since each method alone has limitations and can cause a misinterpretation of results (Hughes et al., 2009). Thus, this study aimed to evaluate the prevalence of hemoglobinopathies using conventional and confirmatory laboratory tests such as hemoglobin electrophoresis, quantification of hemoglobin fractions by HPLC, and analysis of the deletion that causes more frequent α-thalassemia (-α3.7) in southeast Brazil (Wenning et al., 2000; Bonini-Domingos, 2004), also occurring in children from public schools in Ribeirão Preto-SP, by molecular biology. Different types of thalassemia are found in the population, the most common being the α and β-thalassemias (Galanello et al., 1998; Galanello, Origa, 2010; Cousens et al., 2010). The α-thalassemia is the most common inherited disease in the world. Recent studies show that α-thalassemia affects 5% of the worldwide population and until recently (Vichinsky, 2010), the frequency of this pathology was underestimated due the use of inadequate diagnostic methods. In Brazil α-thalassemia affects mainly Asian and some African groups. Studies show a frequency of 10-12% in some regions and up to 25% in specific groups. For α-thalassemia in the state of São Paulo in Caucasian descendants, the estimated ratio is 3% (Bonini-Domingos, 2004). Cristiane Fernandes de Freitas Tavares, Jacqueline da Silva Guimarães, Ana Maria de Souza* Onze crianças apresentaram resultado indicando a suspeita de S/β-talassemia; seus pais e/ou irmãos confirmaram a presença de apenas a Hb S. A análise da deleção -α3,7, uma das alterações que estão presentes na α-talassemia, realizada em 207 participantes, identificou 26 crianças (12,6%) com a deleção -α3,7. Dessa forma, 54 (12,6%) das crianças estudadas apresentam hemoglobinopatias. Para a deleção da α-talassemias é necessário utilizar métodos confirmatórios como as análises moleculares e avaliação de membros da família, em casos duvidosos, facilitando o aconselhamento genético nas famílias, sendo a deleção -α3,7 mais frequente no Brasil. Unitermos: Hemoglobinopatias. Eletroforese de hemoglobina. Cromatografia líquida de alta eficiência/ análise quantitativa. Reação em cadeia de polimerase. C. F. F. Tavares, J. S. Guimarães, A. M. Souza 362 MATERIAL AND METHODS The study group included 427 children between six and nine years of age of both genders, recruited from state and municipal schools in the city of Ribeirão Preto- SP, namely: Centro Municipal de Educação Infantil Virgílio Salata, Escola Estadual Antônio Diederichsen, Escola Estadual Dr. Tomas Alberto Whatelly, EMEFEM Dom Luiz Do Amaral Mousinho, and Escola Estadual Dom Alberto José Gonçalves. Participation in this study was at the consent endorsed by their parents. This study was approved by the Ethics Committee of the Faculty of Mutations or deletions may result in a lack of production of globin (β0 and α0) or decreased production of these (β+ and α+) (Cunninghan, 2010). Over 95% of cases of α-thalassemia are due to deletions, being -α5.2 and -α 20.5 generally causing α0-thalassemia and -α3.7, -α4.2 363 Prevalence of hemoglobinopathies in school children: the importance of using confirmatory methods TABLE I - Distribution of hemoglobinopathies in the studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC methodology TABLE I - Distribution of hemoglobinopathies in the studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC methodology Hemoglobinopathies Number of cases and percentage ratios of the total participants Number of cases and percentage ratios in 30 children with hemoglobinopathies β-talassemia 14 / 3,3% 14 / 46,7% Hb S/ Hb A2 ↑ 11 / 2,6% 11 / 36,7% Hb AS 04 / 0,94% 04 / 13,3% Hb AC 01 / 0,24% 01 / 3,3% Dentistry of Ribeirão Preto, University of São Paulo (case no. 2006.1.797.58.5). β0); 11 with increased Hb A2 and the presence of Hb S, consistent with the phenotype of Hb S/β-thalassemia; 4 with the presence of Hb S carriers of sickle cell trait and 1 with the presence of Hb C, indicating heterozygosity for the Hb C disease (Table I, Figure 1). β0); 11 with increased Hb A2 and the presence of Hb S, consistent with the phenotype of Hb S/β-thalassemia; 4 with the presence of Hb S carriers of sickle cell trait and 1 with the presence of Hb C, indicating heterozygosity for the Hb C disease (Table I, Figure 1). MATERIAL AND METHODS Ten milliliters (mL) of blood were collected and placed in a tube containing dipotassium ethylenediaminetetraacetic acid (K2 EDTA) 10%, which was used in the following determinations: hematologic evaluation using automatic counter Micros 45- Horiba ABX®; hemoglobin electrophoresis on cellulose acetate at alkaline pH, according to Naoum (1999); quantification of hemoglobin fractions by HPLC Variant II Bio-Rad® automated system (β-thalassemia Short Program kit); and extraction of genetic material to perform the PCR.i FIGURE 1 - Distribution of hemoglobinopathies in the 427 studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC methodology. For detection of deletion -α3.7 the specific primers (Dodé et al., 1993): 5’ CCA TGC CTG GCA CGT TTG CTG AGG 3’(C9 primer); 5’ GAT GCA CCC ACT GGA CTC CT 3’(C10 primer) were used. The PCR reaction was processed in a thermocycler (Eppendorf Mastercycle), 30 cycles being performed at 94 oC/2 min., 56 oC/1 min., 72 oC/2 min., preceded by an initial heat-denaturation step at 94 oC/5 min. and followed by a final extension step at 72 oC/10 min. The mixture was subjected to horizontal electrophoresis on a 1% agarose gel with ethidium bromide, with a running time of 1 hour and 20 minutes at 90 volts, 220 mA. Then the gel was exposed to UV light for visualization of the presented bands. FIGURE 1 - Distribution of hemoglobinopathies in the 427 studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC methodology. The determination of the levels of Hb A2 of 11 participants with suspected Hb S/β-thalassemia was performed by HPLC and the fractions obtained by elution in electrophoresis on cellulose acetate. The median Hb A2 levels determined by HPLC and elution were 6.3% (± 1.1) and 7.0% (± 1.6), respectively, with no significant statistical difference (p > 0.05). A survey of quantitative and qualitative changes of hemoglobin fractions in 11 parents and/or siblings by HPLC only detected the presence of Hb S, discarding the suspicion S/β-thalassemia in children. Thirteen parents and two brothers participated in the study in order to confirm the results of children who showed the presence of Hb S and increased Hb A2, suggestive of hemoglobinopathy HbS/β-thalassemia. Five mL of blood were collected and placed into a tube containing K2 EDTA, which was used for the identification and quantification of hemoglobin fractions by HPLC. MATERIAL AND METHODS The analysis of deletion -α3.7, featuring α-thalassemia was performed in 207 study participants verifying the presence of 26 children (12.6%) carrying the deletion, being 24 heterozygous and 2 homozygous. RESULTS AND DISCUSSION The results of hemoglobin electrophoresis on cellulose acetate and HPLC of the 427 children studied indicated the presence of 30 children (7%) with hemoglobinopathies, being 14 with increased Hb A2, consistent with the phenotype of β-thalassemia (β+ or Together the analyses showed a total of 54 children, 12.6% of the children studied, showing some qualitative or quantitative change of the globin chains, namely: 15 (3.5%) women with Hb S, 1 (0.24 %) Hb C and 14 (3.3%) with β-thalassemia minor (Table II, Figure 2). C. F. F. Tavares, J. S. Guimarães, A. M. Souza 364 testing and research of family cases of hemoglobinopathies in our population is evident. TABLE II - Distribution of hemoglobinopathies in the studied children utilizing hemoglobin electrophoresis on cellulose acetate, HPLC and PCR methodology Neonatal screening programs follow the standards of the National Program for Newborn Screening (Ministry of Health decree No. 822/01) that recommends the use of two presumptive methods for determining the electrophoretic profile. Currently, screening programs replaced the traditional methods by using isoelectric focusing electrophoresis and HPLC. Despite new presumptive methods having better sensitivity and specificity, they are susceptible to interference and in some cases such as in the diagnosis of α-talassemia the use of molecular methodology is required (Hughes et al., 2009). Our study proved this since the number of patients with hemoglobinopathies rose 5.6% after performing PCR for detection of α-thalassemia. Hemoglobinopathies Number of cases and percentage ratios of the total participants α-talassemia 24 / 11,6% Hb AS 15 / 3,5% β-talassemia 12 / 2,8% α/β-talassemia 02 / 0,96% Hb AC 01 / 0,24% FIGURE 2 – Frequencies of hemoglobinopathies in the 427 studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC, and the 207 studied children utilizing PCR. The use of HPLC has shown to be more sensitive and specific, and with a better reproducibility for the determination of the Hb A2 method. However it is argued that the percentage of Hb A2 measured by HPLC can be affected by the presence of hemoglobin S (HbS) in heterozygous individuals for β-thalassemia. This is due to reduced affinity of the βS to α chains, which leads to the free α to match δ chains, increasing Hb A2, and co- elution of Hb with Hb A2, falsely elevating its percentage (Head et al., 2004; Kalleas et al., 2007). RESULTS AND DISCUSSION No significant differences were observed in quantifying the levels of Hb A2, performed by the methods of electrophoresis on cellulose acetate for elution and HPLC. To confirm suggestive cases of Hb S/β-thalassemia, parents and/or siblings of these children were evaluated, verifying in all cases only the presence of HbS and normal levels of Hb A2, thus inferring that the children had only the sickle cell trait. Probably the increase in Hb A2 was found in these children due to the combination of free α and δ chains and not presenting mutations that compromise the synthesis of β chains. FIGURE 2 – Frequencies of hemoglobinopathies in the 427 studied children utilizing hemoglobin electrophoresis on cellulose acetate and HPLC, and the 207 studied children utilizing PCR. Among patients with hemoglobinopathies, 27 (50%) were anemic according to WHO criteria, i.e., Hb less than 11.5 g/dL for children under 12 years of age of both genders, 8 (14.8%) of them showed only mild hypochromia and the remainder (35.2%) showed no significant changes. In our study the number of carriers of the sickle cell trait (3.5%) was apparently higher than reported in a review by Murao and Ferraz (2007), in which the authors estimate 1.9% of patients with Hb S in the state of Sao Paulo. A prevalence of approximately 3% of patients with β-thalassemia in the state of São Paulo is estimated (Bonini-Domingos, 2004). In our study a prevalence of 3.3%, very close to the above, was observed. There was the presence of 12.6% of α-thalassemia, results close to the literature (Head et al., 2004), which indicates prevalence of between 10 and 20% in our population. These numbers are quite significant and justify the need for more extensive research in the city of Ribeirão Preto to characterize the prevalence in the population. REFERENCES AIGNER, C.P. ; SANDRINI, F.; DUARTE, E.G.; ANDRADE, M.P. ; LARGURA, M. A.; LARGURA, A. Estudo do perfil de hemoglobinas em 9.189 testes realizados no Álvaro Centro de Análises e Pesquisas Clínicas. Rev. Bras. Anal. Clin., v.38, n.2, p.107-109, 2006. HEAD, C. E.; CONROY, M.; JARVIS, M.; PHELAN, L.; BAIN, B. J. Some observations on the measurement of haemoglobin A2 and S percentages by high performance liquid chromatography in the presence and absence of α-thalassemia. J. Clin. Pathol., v.57, n.3, p.276-280, 2004. BERTHOLO, L.C.; MOREIRA, H.W. Focalização isoelétrica na identificação das hemoglobinas. J. Bras. Patol. Med. Lab., v.42, n.3, p.163-168, 2006. GALANELLO, R.; ORIGA, R. Beta-thalassemia. Orphanet. J. Rare Dis., v.5, n.11, p.2-15, 2010. BONINI-DOMINGOS, C.R. Thalassemia screening in Brazil: results for 20 years. Rev. Bras. Hematol. Hemoter., v.26, n.4, p.288-289, 2004. HENDERSON, S.; TIMBS, A.; MCCARTHY, J.; GALLIENNE, A.; MOURIK, M.V. ; MASTERS, G.; MAY, A.; KHALIL, M.S.; SCHUCH, A.; OLD, J. Incidence of haemoglobinopathies in various populations – the impact of immigration. Clin. Biochem., v .42, n.18, p.1745-1756, 2009. BORGES, E.; WENNING, M.R.; KIMURA, E.M.; GERVASIO, S.A.; COSTA, F.F.; SONATI, M.F. High prevalence of alpha-thalassemia among individuals with microcytosis and hypochromia without anemia. Braz. J. Med. Biol. Res., v.34, n.6, p.759-762, 2001. HUGHES, H.Y.; MCKIE, K.; CARMICHAEL, H.; BORA, K.; KUTLAR, A.; KUTLAR, F. Diagnostic complication and molecular characteristics of Hb SC-Chicago disease with α-thal-2 (α3.7 deletion): effects of multiple variant on patient’s phenotype. Ann. Hematol., v.88, n.11, p.1151- 1153, 2009. CLARK, B.E.; THEIN, S.L. Molecular diagnosis of haemoglobin disorders. Clin. Lab. Haematol., v.26, n.3, p.159-176, 2004. CHINELATO-FERNANDES, A.R.; LEONELI, G.G.; CALDERAN, p.O.; OLIVEIRA, R.B.; SILVA, W.A.; HIDALGO, C.A.; BONINI-DOMINGOS, C.R. Avaliação eletroforética, cromatográfica e molecular da Hb D Los Angeles no Brasil. Rev. Bras. Hematol. Hemoter., v.25, n.3, p.161-168, 2003. KALLEAS, C.; TENTES, I.; MARGARITIS, D.; et al. Effect of HbS in the determination of HbA2 with the Biorad Variant II analyzer. Clin. Biochem., v.40, n.9-10, p.744-746, 2007. CHUI, D.H.K.; FUCHAROEN, S.; CHAN, v.Hemoglobin H disease: not necessarily a benign disorder. Blood, v.101, n.3, p.791-799, 2003. MANCA, L.; MASALA, B. Disorders of the synthesis of human fetal hemoglobin. IUBMB Life, v.60, n.2, p.94-111, 2008. MELO, L.M.S.; SIQUEIRA, F.A.M.; CONTE, A.C.F.; BONINI- DOMINGOS, C.R. Rastreamento de hemoglobinas variantes e talassemias com associação de métodos diagnósticos. Rev. Bras. Hematol. Hemoter., v.30, n.1, p.12-17, 2008. COUSENS, N.E.; GAFF, C.L.; METCALFE, S.A.; DELATYCKI, M.B. CONCLUSION We c o n c l u d e t h a t i n o r d e r t o d e t e c t hemoglobinopathies, the use of confirmatory methods such as molecular analysis for the search of α-thalassemia and the investigation of relatives of the doubtful cases are required. Although it has been investigated only for the most frequent deletion for α-thalassemia, α3.7 it is now possible to state the importance of confirmatory methods. Only then we can ensure accurate diagnosis and facilitate genetic counseling in families where there are carriers of these pathologies. On these results the necessity of conducting prenatal 365 Prevalence of hemoglobinopathies in school children: the importance of using confirmatory methods ACKNOWLEDGMENT DODÉ, C.; KRISHNAMOORTHY, R.; LAMB, J.; ROCHETTE, J. Rapid analysis of –α3.7 thalassaemia and ααα anti 3.7 triplication by enzymatic amplification analysis. Br. J. Haematol., v.83, n.1, p.105-111, 1993. The authors wish to thank CNPq for financial support. GALANELLO, R.; SOLLAINO, C.; PAGLIETTI, E.; BARELLA, S.; PERRA, C.; DONEDDU, I.; PIRRONI, M. G.; MACCIONI, L.; CAO, A. α-Thalassemia carrier identification by DNA analysis in the screening for thalassemia. Am. J. Hematol., v.59, n.4, p.273-278, 1998. REFERENCES Carrier screening for beta-thalassaemia: a review of international practice. Eur. J. Hum. Gen., v.18, n.10, p.1077-1083, 2010. CUNNINGHAN, M.J. Update on thalassemia: clinical care and complications. Hematol. Oncol. Clin. North Am., v.24, n.1, p.215-227, 2010. 366 C. F. F. Tavares, J. S. Guimarães, A. M. Souza MELO-REIS, p.R.; NAOUM, p.C.; DINIZ-FILHO, J.A.F.; DIAS-PENNA, K.G.B.; MESQUITA, M.M.; BALESTRA, F.A. Prevalência de talassemias e hemoglobinas variantes no estado de Goiás, Brasil. J. Bras. Patol. Med. Lab., v.42, n.6, p.425-430, 2006. SONATI, M.F.; FARAH, S.B.; RAMALHO, A.S.; COSTA, F.F. High prevalence of alpha-thalassemia in a black population of Brazil. Hemoglobin, v.15, n.4, p.309-311, 1991. VICHINSKY, E. Complexity of alpha thalassemia: growing health problem with new approaches to screening, diagnosis, and therapy. Ann. NY Acad. Sci., v.1202, v.180- 187, 2010. MURAO, M.; FERRAZ, M.H.C. Sickle cell trait: heterozygous for the hemoglobin S. Rev. Bras. Hematol. Hemoter., v.29, n.3, p.223-225, 2007. WAGNER, S.C.; SILVESTRI, M.C.; BITTAR, C.M.; FRIEDRISCH, J. R.; SILLA, L.M.R. Prevalence of thalassemias and variant hemoglobin in patients with non- ferropenic anemia. Rev. Bras. Hematol. Hemoter., v.27, n.1, p.37-42, 2005. NAOUM, p.C. Eletroforese: técnicas e diagnósticos. In: NAOUM, p.C. Hemoglobinopatias. 2.ed. São Paulo: Livraria Editora Santos, 1999. p. 57-100. OLIVEIRA, G.L.V. ; MENDIBURU, C.F.; BONINI- DOMINGOS, C.R. Avaliação do perfil hematológico de portadores de talassemia alfa provenientes das regiões Sudeste e Nordeste do Brasil. Rev. Bras. Hematol. Hemoter., v.28, n.2, p.105-109, 2006. WEATHERALL, D.J. Phenotype-genotype relationships in monogenic disease: lessons from the thalassaemias. Nat. Rev. Genet., v.2, n.4, p.245-255, 2001. WENNING, M.R.S.C.; KIMURA, E.M.; COSTA, F.F.; SAAD, S.T.O.; GERVÁSIO, S.; JORGE, S.B.; BORGES, E.; SILVA, n.M.; SONATI, M.F. α-globin genes: thalassemic and structural alterations in a Brazilian population. Braz. J. Med. Biol. Res., v.33, n.9, p.1041-1045, 2000. WENNING, M.R.S.C.; KIMURA, E.M.; COSTA, F.F.; SAAD, S.T.O.; GERVÁSIO, S.; JORGE, S.B.; BORGES, E.; SILVA, n.M.; SONATI, M.F. α-globin genes: thalassemic and structural alterations in a Brazilian population. Braz. J. Med. Biol. Res., v.33, n.9, p.1041-1045, 2000. Received for publication on 25th March 2014 Accepted for publication on 20th October 2014 ONDEI, L.S.; ZAMARO, p.J.A.; MANGONARO, p.H.; VALÊNCIO, C.R.; BONINI-DOMINGOS, C.R. HPLC determination of hemoglobins to establish reference values with the aid of statistics and informatics. Genet. Mol. Res., v.6, n.2, p.453-460, 2007. Received for publication on 25th March 2014 Accepted for publication on 20th October 2014 Received for publication on 25th March 2014
https://openalex.org/W3096307141	https://link.springer.com/content/pdf/10.1007/s11748-020-01525-w.pdf	English	null	A rare case of pseudoaneurysm at the site of untreated coarctation of the aorta in an adult patient	General thoracic and cardiovascular surgery	2,020	cc-by	2,465	CASE REPORT CASE REPORT Introduction Coarctation of the aorta (CoA) has been reported to account for 4–8% of congenital heart diseases [1–3]. There are few rare cases of detected CoA in adults, which was not detected earlier during childhood, and the natural course of the dis- ease is complicated due to cardiac malformations and car- diovascular comorbidities [2]. Few cases have been reported on the association between pseudoaneurysms and CoA. However, these cases were reported due to post-stenotic dilation. To the best of our knowledge, there have been no reports of pseudoaneurysm at the stenotic site yet. Here, we present a rare case of a patient with CoA who was asymp- tomatic until the age of 50 and had a pseudoaneurysm pro- truding from the site of aortic coarctation. We decided to perform an open surgical repair. The pathological findings of the resected aneurysm wall revealed a saccular pseudoaneu- rysm comprising only an adventitia. Because of the very thin aneurysm wall, the risk of aneurysm rupture was extremely high. Therefore, it is thought that the pseudoaneurysm rup- ture as in this case may be one of the causes of the poor prognosis of adult aortic coarctation. A 50-year-old woman visited a nearby physician for a medical checkup due to complaints of high blood pressure. A 50-year-old woman visited a nearby physician for a medical checkup due to complaints of high blood pressure. A 50-year-old woman visited a nearby physician for a medical checkup due to complaints of high blood pressure. An ankle brachial index (ABI) showed a pressure differ- ence between the upper and lower extremities. A computed tomography scan (CT) showed CoA and a saccular aortic aneurysm protruding from the aortic coarctation. In addition, she had untreated dyslipidemia. The ABIs of the right and left ankles were 0.65 and 0.67, respectively. Transthoracic echocardiographic showed no significant valvular dysfunc- tion or left ventricular remodeling due to heart failure. Nota- bly, no combined vascular malformations were observed. The CT scan showed that CoA originated from the periphery of the left subclavian artery to the aortic isthmus, and clas- sified as a CoA with isthmus hypoplasia [4]. In addition, a saccular aneurysm protruding dorsally from the coarcta- tion was observed (Fig. 1a, b). Calcification in the saccular aneurysm wall was not observed. However, high calcifica- tion in the aortic area except for the aneurysm was observed (Fig. 1c). Takuma Mikami1 · Takeshi Kamada1 · Hiroki Uchiyama1 · Yosuke Kuroda1 · Ryo Harada1 · Syuichi Naraoka1 · Nobuyoshi Kawaharada1 Received: 15 April 2020 / Accepted: 13 October 2020 / Published online: 28 October 2020 © The Author(s) 2020 Abstract Here we report a rare case of pseudoaneurysm at the site of aortic coarctation. Aortic coarctation and a saccular aortic aneu- rysm protruding from the site of this coarctation were detected in a 50-year-old woman. Owing to the shape of the aneurysm and high risk of rupture, an open surgical repair was performed. The pathological findings of the removed aneurysm revealed a pseudoaneurysm consisting of only a thin adventitial wall. Adult uncorrected aortic coarctation has a poor prognosis. One of its causes may be the formation of such a pseudoaneurysm. Keywords Coarctation of the aorta · Pseudoaneurysm · Adult congenital heart disease * Takuma Mikami t.mikami.02.28@gmail.com 1 Department of Cardiovascular Surgery, Sapporo Medical University School of Medicine, 291, Minami 1‑jo Nishi 16‑chome, Chuo‑ku, Sapporo, Hokkaido 060‑8543, Japan General Thoracic and Cardiovascular Surgery (2021) 69:740–743 https://doi.org/10.1007/s11748-020-01525-w General Thoracic and Cardiovascular Surgery (2021) 69:740–743 https://doi.org/10.1007/s11748-020-01525-w Introduction The right common femoral artery and vein were cannulated, and partial cardiopulmonary bypass was established during aortic clamping. The aneurysm mass atrophied easily after clamping (Fig. 2). The saccu- lar aneurysm lacked a defined intima and was formed of a thin single-layered membrane, which was indicative of the adventitia (Fig. 3a). The anastomosis was performed with a 20-mm 1-branch tubular graft (J Graft SHIELD NEO®, Japan Lifeline Co., Ltd. Tokyo, Japan). A drain was placed in the left thoracic cavity. The operative and cardiopulmo- nary bypass time were 194 and 58 min, respectively. In addition, the aortic block time, minimum body temperature and bleeding volume were 54 min, 34.8 °C, and 590 mL, respectively. Improvement was achieved with a conservative treatment (with a low-protein diet), and the drain was removed 7 days after surgery. A postoperatively contrast-enhanced CT scan revealed a complete coarctation resection, and no abnormal findings were observed in the artificial blood vessel anasto- mosis (Fig. 1d). Improvements were observed in postopera- tive ABI (right ABI 0.82, left ABI 0.84). The patient was discharged 23 days postoperatively. Introduction No other abnormalities were observed in the heart or large vessels. Laboratory data showed that erythrocyte sedimentation rate and C-reactive protein levels were within the normal reference ranges. Considering the high risk of aneurysm rupture, we decided to perform an open surgi- cal repair. Written informed consent was obtained from the patient to publish this case report. 1 Department of Cardiovascular Surgery, Sapporo Medical University School of Medicine, 291, Minami 1‑jo Nishi 16‑chome, Chuo‑ku, Sapporo, Hokkaido 060‑8543, Japan Vol:1 .(123456789 3 General Thoracic and Cardiovascular Surgery (2021) 69:740–743 741 Fig. 1 a Preoperative contrast-enhanced CT scan: a saccular aortic aneurysm is observed at the site of aortic coarctation (arrow). Col- lateral circulation appears well developed. b Preoperative contrast- enhanced CT scan, sagittal section: a saccular aortic aneurysm pro- truding dorsally is observed. The aneurysm wall appears very thin (arrow). c Preoperative contrast-enhanced CT scan, axial section: a dorsally protruding saccular aortic aneurysm is observed, the arterial wall other than the aneurysm appears highly calcified. d Postopera- tive contrast-enhanced CT scan: the stenosis was completely resected (arrow). c Preoperative contrast-enhanced CT scan, axial section: a dorsally protruding saccular aortic aneurysm is observed, the arterial wall other than the aneurysm appears highly calcified. d Postopera- tive contrast-enhanced CT scan: the stenosis was completely resected Fig. 1 a Preoperative contrast-enhanced CT scan: a saccular aortic aneurysm is observed at the site of aortic coarctation (arrow). Col- lateral circulation appears well developed. b Preoperative contrast- enhanced CT scan, sagittal section: a saccular aortic aneurysm pro- truding dorsally is observed. The aneurysm wall appears very thin Fig. 2 a Intraoperative image: a saccular aortic aneurysm protruding dorsally at the coarctation site (arrow). b Enlarged image: the aneurysm wall appears thin Fig. 2 a Intraoperative image: a saccular aortic aneurysm protruding dorsally at the coarctation site (arrow). b Enlarged image: the aneurysm wall appears thin Thoracotomy was performed under general anesthesia via the left lower 6th intercostal. No adhesion between the aneurysm and the lung was found. We clamped the aorta between the left common carotid and left subclavian arteries. 1 General Thoracic and Cardiovascular Surgery (2021) 69:740–743 742 Therefore, anastomosis was performed by clamping the cen- tral side of the descending thoracic aorta at the Th8 level without deep hypothermic circulatory arrest. This was per- formed more proximal than the segmental artery connecting to the Adamkiewicz’ artery (AKA) identified preoperatively at the left Th9 level. Fig. 3 a A photo of the resected aneurysm from the intima side: the intima is macroscopi- cally missing at the aneurysm entrance (arrow). b Elastica van Gieson staining of the aneurysm wall: the aneurysm wall has a clearly thinner wall structure than the other parts. c Enlarged findings (within the square): histologically, the intima and media are absent. Several layers of elastic fibers were found (arrow); therefore, a pseudoaneurysm consists of the adventitia only Discussion Many cardiac malformations are associated with CoA [5]. Various factors, such as genetic, environmental, and arterial plaque formation are involved in the pathophysiology of CoA. Cases of adults with CoA that have remained undetected dur- ing childhood are rare, and are often asymptomatic or hyper- tensive. The natural history of patients with CoA is poor. The average survival age of patients with untreated CoA beyond childhood is 34 years, with 75% mortality by age 43 years [2]. The causes of death include heart failure, aortic dissection or rupture, infective endocarditis, and intracranial hemorrhage due to ruptured cerebral aneurysm [6]. Very rare undiagnosed cases (such as the present case) survive without treatment until the age of 50 due to a well-developed collateral circulation from the internal thoracic and intercostal arteries. Peripheral ischemia symptoms were not identified from the coarctation. Histopathologic examination of the excised specimen showed the presence of several layers of elastic fibers, but no intimal or medial structures were observed in the saccu- lar pseudoaneurysm. No acute or chronic inflammation was present. Therefore, these findings, clearly revealed a pseu- doaneurysm consisting of the adventitia only (Fig. 3b, c). The patient’s postoperative course was uneventful. Mild chylous effusion was observed from the left thoracic drain. 1 3 Fig. 3 a A photo of the resected aneurysm from the intima side: the intima is macroscopi- cally missing at the aneurysm entrance (arrow). b Elastica van Gieson staining of the aneurysm wall: the aneurysm wall has a clearly thinner wall structure than the other parts. c Enlarged findings (within the square): histologically, the intima and media are absent. Several layers of elastic fibers were found (arrow); therefore, a pseudoaneurysm consists of the adventitia only 1 3 3 743 General Thoracic and Cardiovascular Surgery (2021) 69:740–743 Funding None. Funding None. There have been some reports on the complications of CoA and aortic aneurysms. The etiological causes of pseudoaneu- rysm formation include: arteriosclerosis due to hypertension; arterial wall weakness due to inflammation (infectious endo- carditis, aortic inflammation); congenital abnormalities in the arterial wall structure and post-stenosis dilation; and incom- pletely occluded arterial ducts [7]. Yet, there are few reports on the association between pseudoaneurysms and CoA. Oi et al. [8] and Prifti et al. [9] reported that an aneurysm was formed at the periphery of the coarctation, and attributed to a jet stream of aortic stenosis. References 1. Doshi AR, Chikkabyrappa S. Coarctation of aorta in children. Cureus. 2018;10:e3690. In CoA, symptomatic or asymptomatic, the difference in blood pressure between the upper and lower limbs (at rest) was 20 mmHg or more. In addition, hypertension or left ventricular hypertrophy were indications for class I surgery according to the European and American guidelines. A class IIa surgery was considered in individuals even when the max- imum pressure difference was 50% or more [10, 11]. In this case, the difference in blood pressure between the upper and lower limbs ranged approximately between 40 and 50 mmHg that was complicated by hypertension. Surgical procedures included end-to-end anastomosis, patch formation, subcla- vian flap aortoplasty, and artificial blood vessel replacement [12]. In the recent years, treatment using thoracic endovas- cular aortic repair (TEVAR) has been reported. However, in this present case, extensive calcification of aortic constriction area and complications were observed in the pseudoaneu- rysm. Therefore, end-to-end anastomosis, patch formation, subclavian flap aortoplasty, and TEVAR were considered as off-label treatments, and artificial blood vessel replace- ment was selected. These findings suggest that a complete resection of the stenosis including calcification and artificial blood vessel replacement are appropriate procedures in cases of CoA with advanced calcification. However, the findings cannot be generalized due to the limited number of cases and further studies are required to validate our findings. 2. Alkashkari W, Albugami S, Hijazi ZM. Management of coarcta- tion of the aorta in adult patients: state of the art. Korean Circ J. 2019;49:298–313. 3. Jurcut R, Daraban AM, Lorber A, Deleanu D, Amzulescu, Zara C, et al. Coarctation of the aorta in adults: what is the best treatment? Case report and literature review. J Med Life. 2011;4:189–95. 4. Amato JJ, Douglas WI, James T, Desai U. Coarctation of aorta. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2000;3:125–41. 5. Ayesha S, Alice C, Ali NZ. The adult with coarctation of the aorta. IntechOpen. 2018. https://doi.org/10.5772/intechopen.79865. 6. Jenkins NP, Ward C. Coarctation of the aorta: natural history and outcome after surgical treatment. QJM. 1999;92:365–71. 7. Yokoyama S, Naito Y, Katayama H, Koh E. A case of multisac- cular thoracic aortic aneurysm associated with coarctation of the aorta. Jpn J Cardiovasc Surg. 2005;34:370–3. p g 8. Oi K, Yoshida T, Takeshita M, Tsuruta G. False aneurysm on distal part of coarctation of the aorta in a parous Turner syndrome patient. Gen Thorac Cardiovasc Surg. 2013;61:531–3. Discussion In addition, the location of aneurysm was more peripheral than that of stenosis. However, in this present case, a pseudoaneurysm was formed in the aortic constriction, and severe calcification was detected around the same area. The pathological findings showed that the intima and media at the site of coarctation were completely unobserved and the very thin aneurysm had formed a few layers of elastic fibers. These findings also suggested that the pseudoaneurysm had a very high risk of rupture, and this is one of the causes of poor progression of coarctation of aorta in adults.f References 9. Prifti E, Kuci S, Krukulli K, Nuellari E. Pseudoaneurysm of the descending aorta complicating an untreated aortic coarctation. Ann Thorac Surg. 2015;99:e3–5. 10. Stout KK, Daniels CJ, Aboulhosn JA, Bozkurt B, Broberg CS, Colman JM, et al. 2018 AHA/ACC guideline for the manage- ment of adults with congenital heart disease. J Am Coll Cardiol. 2019;73:1494–563.f 11. Baumgartner H, Bonhoeffer P, De Groot NM, de Haan F, Dean- field JE, Galie N, et al. ESC guidelines for the management of grown-up congenital heart disease (new version 2010). Eur Heart J. 2010;31:2915–57. 12. Abjigitova D, Mokhles MM, Witsenburg M, van de Woestijne PC, Bekkers JA, Bogers AJJC. Surgical repair of aortic coarctation in adults: half a century of a single centre clinical experience. Eur J Cardiothorac Surg. 2019;56:1175–85. Conflict of interest The authors have no conflicts of interest to declare. Conflict of interest The authors have no conflicts of interest to declare. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Conclusion Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. We studied a rare case of a pseudoaneurysm at the site of aortic coarctation. The natural prognosis of untreated CoA is poor, and pseudoaneurysm at the site of aortic coarctation may be one cause of poor prognosis. 1 3
https://openalex.org/W2021017941	https://www.frontiersin.org/articles/10.3389/fpsyg.2011.00357/pdf	English	null	Lifespan Changes in Global and Selective Stopping and Performance Adjustments	Frontiers in psychology	2,011	cc-by	13,094	INTRODUCTION study is to examine developmental trends in global and selective stopping along the lifespan. The ability to dynamically adjust to the changing environment is critical for survival. One important component of this abil- ity is to refrain from responding when a readied response is no longer required. In the laboratory, the ability to withhold motor responses has been examined frequently by using the stop-signal paradigm (Logan and Cowan,1984;Verbruggen and Logan,2008). In this paradigm, participants are typically presented with a stim- ulus requiring either a speeded left- vs. right-hand response. On some trials, a stop-signal is presented, occasionally and unpre- dictably, just following the onset of the choice stimulus indicating that the response activated by the choice stimulus should be with- held.Assuming that the go process activated by the choice stimulus and the stop process activated by the stop-signal are involved in a race,won by whichever process ﬁnishes ﬁrst,allows for an estimate of the duration of the stop process, or stop-signal reaction time (SSRT, Logan and Cowan, 1984). Most stop studies use a global stop-signal that requires participants to stop whatever response is activated by the choice stimulus. Global SSRT is typically around 200 ms (Logan, 1994; Verbruggen and Logan, 2008). Some stop studies used a selective stop-signal indicating that the response activated by the choice stimulus should be withheld whereas the other (invalid) stop-signal indicates that the readied response is to be executed. Selective SSRT is typically longer than global SSRT, about 30 ms (Bedard et al., 2002; van den Wildenberg and van der Molen, 2004a; van de Laar et al., 2010). The ﬁrst aim of the current Developmental studies showed that global stopping is usu- ally slower in children than adults (Williams et al., 1999; but see Oosterlaan and Sergeant, 1996; Band et al., 2000; Johnstone et al., 2007). With advancing age the speed of global stopping seems to decline during senescence (Kramer et al., 1994; May and Hasher, 1998; Ridderinkhof et al., 1999; Rush et al., 2006; Gamboz et al., 2009) but the age-related decline in global stop- ping speed has been observed to be considerably less than the developmental gain during childhood. Thus, a lifespan analysis of global stopping revealed that older children (11 years) stopped about 50 ms faster than young children (7.5 years). ORIGINAL RESEARCH ARTICLE published: 15 December 2011 doi: 10.3389/fpsyg.2011.00357 Correspondence: e-mail: m.c.vandelaar@uva.nl Keywords: stop-signal paradigm, development, cognitive aging, lifespan, cognitive control Maria C. van de Laar 1,Wery P. M. van den Wildenberg 1, Geert J. M. van Boxtel 2 and Maurits W. van der Molen3 1 Amsterdam Center for the Study of Adaptive Control in Brain and Behavior, Department of Psychology, University of Amsterdam, Amsterdam, Netherlands 2 Department of Psychology, Tilburg University, Tilburg, Netherlands 3 Department of Psychology, University of Amsterdam, Amsterdam, Netherlands This study examined stopping and performance adjustments in four age groups (M ages: 8, 12, 21, and 76 years). All participants performed on three tasks, a standard two-choice task and the same task in which stop-signal trials were inserted requiring either the suppres- sion of the response activated by the choice stimulus (global stop task) or the suppression of the response when one stop-signal was presented but not when the other stop-signal occurred (selective stop task). The results showed that global stopping was faster than selective stopping in all age groups. Global stopping matured more rapidly than selective stopping. The developmental gain in stopping was considerably more pronounced com- pared to the loss observed during senescence. All age groups slowed the response on trials without a stop-signal in the stop task compared to trials in the choice task, the elderly in particular. In addition, all age groups slowed on trials following stop-signal trials, except the elderly who did not slow following successful inhibits. By contrast, the slowing fol- lowing failed inhibits was disproportionally larger in the elderly compared to young adults. Finally, sequential effects did not alter the pattern of performance adjustments.The results were interpreted in terms of developmental change in the balance between proactive and reactive control. Edited by: Shulan Hsieh, National Cheng Kung University, Taiwan Shulan Hsieh, National Cheng Kung University, Taiwan Reviewed by: Cindy Lustig, University of Michigan, USA Antonino Vallesi, Scuola Internazionale Superiore di Studi Avanzati, Italy *Correspondence: Maria C. van de Laar, Department of Psychology, University of Amsterdam, Weesperplein 4, 1018 XA Amsterdam, Netherlands. e-mail: m.c.vandelaar@uva.nl Keywords: stop-signal paradigm, development, cognitive aging, lifespan, cognitive control INTRODUCTION By contrast, the elderly (60–81 years) were only about 20 ms slower in global stopping compared to young adults (Williams et al., 1999). The age-related decline was even absent in a study reported by Kray et al. (2009). Larger declines have been observed by Kramer et al. (1994)andMayandHasher(1998)butinthesestudiestheprimary choice task was more difﬁcult than in the Williams et al. (1999) study and, thus, the cognitive demands imposed by the primary choice might have reduced the resources available for stopping, in particular in the elderly (cf., Williams et al., 1999). Although only a few studies investigated developmental change in selective stopping, the results that emerged from these studies are similar to the pattern observed for global stopping. The lifes- pan study of selective stopping reported by Bedard et al. (2002) December 2011 \| Volume 2 \| Article 357 \| 1 www.frontiersin.org Lifespan stopping and adjustments van de Laar et al. showed that selective stopping was slower; 248 ms in young adults compared to 209 ms for global stopping as observed previously by Williams et al. (1999). Furthermore, it was observed that the speed of selective stopping improved until young adulthood indicating that selective stopping takes longer to mature than global stop- ping. Finally, the difference between young adults and the elderly was about 60 ms whereas it was only 21 ms for global stopping, as observed by Williams et al. (1999), indicating that selective stop- ping may impose a larger burden on the elderly compared to global stopping. We observed a similar developmental pattern in a direct comparison of global vs. selective stopping (van den Wildenberg and van der Molen, 2004a). That is, selective stopping was slower than global stopping, in young adults (i.e., 232 vs. 205 ms), and selective stopping matured slower than global stopping. experiments, participants are typically instructed not to wait for a stop-signal to occur. In spite of these instructions and despite experimental manipulations used to counter waiting strategies (e.g., the use of a tracking algorithm, see below) participants tend to slow their response when stop-signal trials are inserted in the CRT (Logan and Burkell, 1986; Ramautar et al., 2004; van de Laar et al., 2010). INTRODUCTION The slowing of choice reactions when stop-signal trials are inserted is interpreted in terms of a proactive change in response strategy aimed at increasing the success of stopping at the cost of the speed of choice reactions (e.g., Verbruggen and Logan, 2009). Based on the performance data currently available we were led to predict that young adults would slow when stop-signal trials are inserted in the choice task (a manifestation of proac- tive control) and following stop-signal trials (indicating repe- tition priming when the choice stimulus is repeated and reac- tive control for alternating stimuli). Assuming that both chil- dren (e.g., Huizinga and van der Molen, 2010; Somerville and Casey, 2010) and the elderly (e.g., West and Schwarb, 2006; Mayda et al., 2011) are less efﬁcient in implementing control procedures, we assumed that the response slowing associated with proactive and reactive control would be more pronounced in these age groups compared to young adults. Finally, it was predicted that the slowing associated with repetition priming would be more pronounced for children than adults based on recent ﬁndings indicating that the effects of repetition priming decrease when children are growing older (e.g., Crone et al., 2004; Smulders et al., 2005) and less pronounced or even absent for the elderly based on the repetition priming literature showing that repetition priming effects are basically age-invariant dur- ing senescence (e.g., Fleischman, 2007; Bergerbest et al., 2009, for reviews). The current study will provide a direct comparison between age-related change in global vs. selective stopping along the lifes- pan. Such a comparison is still lacking in the stop-signal literature and may reveal potentially interesting differences in the devel- opmental trends associated with global vs. selective stopping. A particular focus of the current study was on the performance on invalid stop-signal trials of the selective stop task. Invalid stop- signal trials are trials on which a stop-signal is presented that should be ignored by the participant. In our previous studies we observed that the speed of responding on invalid stop-signal trials was considerably delayed compared to trials without a stop-signal (vandenWildenbergandvanderMolen,2004a,b;vandeLaaretal., 2010). Thus the stop-signal is not ignored by the participants (or not ignored on a substantial proportion of the invalid trials). INTRODUCTION Pre- viously, we have argued that the delay might be due to the initial inhibition of all responses and the subsequent re-activation of the appropriate response following the complete analysis of the infor- mation provided by the stop-signal (van de Laar et al., 2010). In this regard,the performance on these trials might provide valuable information regarding the strategy employed by the participant. Based on previous ﬁndings (Band et al., 2000), we predicted that the re-initiation strategy would be more prominent in children compared to young adults. The current study would be ﬁrst in providing data regarding invalid stop-signal trial performance in the elderly. December 2011 \| Volume 2 \| Article 357 \| 2 Frontiers in Psychology \| Cognition MATERIALS AND METHODS PARTICIPANTS Four age groups participated in the study. There were two groups of children (a young child group of 8-years-old and an older child group of 12-years-old) and two groups of adults (young adults of 21-years-old and elderly adults of 76-years-old). Participant characteristics are presented in Table 1. Children were recruited from local elementary schools and selected with the help of their teachers. Informed consent was obtained from their primary care- givers. The children received a small present for participation. Young adults were university students who received course credits for participation. Healthy elderly participants were recruited from the metropolitan area. They received 40 euro for their participa- tion. Informed consent was obtained from the adult participants. The second aim of the current study is to examine perfor- mance adjustments related to the insertion of stop-signal trials in the choice reaction task (CRT). Several studies focused on perfor- mance adjustments in the stop-signal paradigm (e.g., Rieger and Gauggel, 1999; Verbruggen and Logan, 2008, 2009; Verbruggen et al., 2008; Aron, 2011; Bissett and Logan, 2011). Rieger and Gauggel(1999)observedthatthespeedof respondingtothechoice stimulus was delayed on trials following a stop-signal trial. The delay was somewhat larger following a failed inhibit compared to following a successful inhibit and larger for trials repeating the choice stimulus compared to trials with alternating stimuli. The delay in responding following successful inhibits has been attrib- uted to repetition priming (Verbruggen et al., 2008) and the added delay after failed inhibits has been interpreted in terms of trial adjustments or reactive control (Bissett and Logan, 2011). Table 1 \| Participant characteristics. Age group Number Age (M, SD) Gender (f/m) Raven percentile Young children 20 7.6 (0.59) 10/10 90 (13.4) Older children 20 11.9 (1.23) 7/13 79 (17.5) Young adults 17 20.8 (3.30) 9/8 88 (3.6) Elderly adults 19 75.6 (5.68) 8/11 92 (4.9) Table 1 \| Participant characteristics. Another type of control that might be exercised in the stop-signal task is referred to as proactive control (Verbruggen and Logan, 2009; see for a review Aron, 2011). In stop-signal Frontiers in Psychology \| Cognition Lifespan stopping and adjustments van de Laar et al. All participants had normal or corrected-to-normal vision. The study was approved by the local Ethics Committee. CRT instruction described above. Trials without a stop-signal are referred to as go trials. On 35% of the trials, the stop-signal occurred instructing the participant to refrain from responding. MATERIALS AND METHODS PARTICIPANTS The stop-signal consisted of either a blue or pink color change of the white arrow (counterbalanced across participants). Partic- ipants were informed about the tracking algorithm and it was explained that a “waiting” strategy would not increase the success of stopping. The GST consisted of four experimental blocks, each containing 100 trials. Chi-square analyses indicated that gender distribution did not differ between groups, χ2(3) = 3.36, p = 0.34. To compare esti- mates of general levels of intelligence between age groups, the standard progressive matrices (SPM; Raven et al., 1985) was administered to children and the elderly, whereas young adults completed the Advanced Progressive Matrices (APM, Raven et al., 1993). Mean percentile scores differed signiﬁcantly across age groups, F(3, 70) = 4.53, p < 0.006. As can be seen in Table 1, the older children had a lower percentile score compared to the other age groups (ps < 0.03). Additional analyses veriﬁed that signif- icant outcomes survived when using Raven percentile scores as covariate. In the selective stop-signal task (SST), participants were instructedtorespondtogotrialsasdescribedabove,butinthistask both stop-signals occurred. Participants were required to discrim- inate between stop-signals and to inhibit their response to one stop-signal (e.g., white-to-pink color change) but to ignore the other stop-signal (e.g., white-to-blue color change). Stop-signal colors were counterbalanced across participants. Stop trials requir- ing participants to stop were coined “valid” stop-signal trials and stop trials on which the stop-signal should be ignored were coined “invalid” stop-signal trials (e.g., van de Laar et al., 2010). Stop- signals occurred on 35% of the trials – half of the trials were valid and the other half of the trials were invalid. Participants performed eight experimental blocks of 100 trials. PROCEDURE The participants performed on all three experimental tasks – CRT, GST, and SST. The tasks were administered during two separate sessions on different days that were scheduled within 5 days. Each session started with the CRT task followed by one of the stop tasks. The order of the two stop tasks was counterbalanced across participants. The duration of an experimental trial block was approximately 5 min and each block was followed by performance feedback. Each task started with one practice block followed by the experimental blocks. A block started with ﬁve warm-up trials that were discarded from analyses. Short breaks were given between blocks and a longer rest was given between tasks. DATA ANALYSES Responses were collected by pressing a left or right zero- displacement force sensor (Honeywell, model FS03) with the left- or right thumb to the direction of the arrow stimulus. Thumbs rested on force sensors that were individually adjusted according to thearmlengthof theparticipant.Acontinuesamplefrequencywas applied (>1000 Hz) and the proportional voltage signal was A/D converted online (Keithley, model KPCI-1802 using DrvrLINX4). The value of 15% of the maximum force served as individual RT threshold which was determined prior to the experiment for each thumb separately (van Boxtel et al., 2001). Data of six young children (CRT data from ﬁve children and GST data from one child), an older child (CRT data), and one elderly participant (SST data) were discarded from analyses, because of unreliable recordings of force pressure. Repeated measures analy- ses of variance (ANOVA) were conducted on latency measures (median RT) and error proportions (errors and omissions). Mul- tiplecomparisonswereusedtoconﬁrmeffects.Degreesof freedom andp-valueswereadjustedusingGreenhouse–Geissercorrections. Since error proportions are not normally distributed, tests were performed on square rooted error values. The integration method based on the horse-race model was used to obtain estimates of stopping latencies (the stop-signal RT or SSRT, Logan and Cowan, 1984). Using this model, the ﬁnish of the stop process can be estimated from the go RT distribution. The left side of the distribution of the RTs on go trials represents fast responses that escape inhibition whereas the right side represents slow responses that will be inhibited. If a given participant actually failed to inhibit on n% of the stop trials, the ﬁnishing time of the stop process will approximately be equal to the nth percentile of the go RT distribution. The mean SSD is then subtracted from the MATERIALS AND STIMULI Participants were seated in a comfortable chair facing a computer screen at a distance of about 1.5 m. Each trial started with a white ﬁxation cross (3 × 3 mm) appearing in the center of the screen for 500 ms followed by the go-signal. The go-signal consisted of a white left- or a right-pointing arrow (2.2 cm × 1.8 cm) presented centrally for 1000 ms against a dark-gray background subtend- ing a visual angle of 1.9˚. The inter-trial interval varied randomly between 1750 and 2250 ms, in steps of 50 ms. During this interval, the ﬁxation cross was presented. Motor responses were collected until the offset of the arrow stimulus. On 35% of the trials the color of the white arrow changed to pink or blue, indicating a stop-signal. The stop-signal delay (SSD; i.e., the interval between the onset of the go-signal and the stop-signal) of the ﬁrst stop trial in the practice block was set at 225 ms and was dynamically adjusted after valid stop trials throughout the experiment as a function of the stopping performance of the participant. Upon successful stopping, SSD on the subsequent stop trial increased with 25 ms, whereas a failure to inhibit decreased SSD on the next stop trial by 25 ms. This tracking algorithm (Levitt, 1971) was set to ensure 50% failed inhibits, which yields accurate estimates of stop-signal RT (Band et al., 2003). www.frontiersin.org PERFORMANCE ON GO TRIALS FIGURE 1 \| Schematic representation of the race model. A distribution FIGURE 1 \| Schematic representation of the race model. A distribution of reaction times (RTs) on go trials (trials without a stop-signal) is shown beneath the curve. These values can be seen as ﬁnishing times of the go process. In stop trials, a stop-signal was shown after the go-signal at a particular stop-signal delay. The ﬁnishing time of the stop process bisects the go-signal RT distribution. The left part consists of go-signal RTs fast enough to escape inhibition (i.e., 50%). The right part (50%) represents slow go-signal RTs that will be inhibited because the stop process ﬁnished before. Stop-signal RT (200 ms) is estimated by subtracting average stop-signal delay (100 ms) from the RT that marks the bisection point (300 ms). FIGURE 1 \| Schematic representation of the race model. A distribution of reaction times (RTs) on go trials (trials without a stop-signal) is shown beneath the curve. These values can be seen as ﬁnishing times of the go process. In stop trials, a stop-signal was shown after the go-signal at a particular stop-signal delay. The ﬁnishing time of the stop process bisects the go-signal RT distribution. The left part consists of go-signal RTs fast enough to escape inhibition (i.e., 50%). The right part (50%) represents slow go-signal RTs that will be inhibited because the stop process ﬁnished before. Stop-signal RT (200 ms) is estimated by subtracting average stop-signal delay (100 ms) from the RT that marks the bisection point (300 ms). Median RTs and error percentages on go trials are presented in Table 2 for each task (CRT, GST, and SST) and age group. The typical lifespan pattern is observed for the RTs of each task – the speed of responding increases into adulthood and slows down in the elderly (see Figure 2). It can be seen that,in spite of the instruc- tions and the dynamic tracking of SSD, all age groups slowed their speed of responding when the task included stop-signals. For all age groups, the slowing seems more pronounced on the GST compared to the SST. These visual impressions were veriﬁed Table 2 \| Median RTs and error percentages associated with go trials of the choice reaction task (CRT), global stop-signal task (GST), and the selective stop-signal task (SST) in each age group (SD between parentheses). EXPERIMENTAL TASKS The choice RT task CRT consisted of go-signals only, stop-signals were not included. Participants were instructed to press the left or right thumb to a white arrow pointing left or right, respectively. The CRT task consisted of one experimental block of 50 trials. All blocks contained equal numbers of right- and left-pointing arrows that were varied randomly within a block. When performing the global stop-signal task (GST), partici- pants were required to respond to go-signals according to the December 2011 \| Volume 2 \| Article 357 \| 3 www.frontiersin.org www.frontiersin.org Lifespan stopping and adjustments van de Laar et al. nth percentile of the go RT distribution, resulting in an estimate of SSRT (see Figure 1). will be addressed refers to age-related changes in the speed of responding when stop-signals are inserted into the task. Based on previous studies, we anticipated that, in spite of the tracking algo- rithm, participants would delay their speed of responding on go trials when they might encounter stop-signals (Verbruggen and Logan, 2009; Aron, 2011). Performance adjustments following a stop trial were examined by comparing trial duplets (go–go vs. successful stop–go vs. failed stop–go vs. invalid stop–go). Within duplets, repetitions vs. alter- nations were distinguished. A repetition duplet consisted of a trial in which the go stimulus indicated the same response as the one on the immediately preceding trial, whereas an alternation duplet consisted of a trial on which the go stimulus indicated the opposite response. In the second section, we focus on age-related changes in the speed of stopping and we will ask whether trends differ between global vs. selective stopping. A second question that will be addressed in this section is on the speed of responding when stopping fails (i.e., failed inhibits). Based on the horse-race model underlying the stop-signal paradigm (Logan and Cowan, 1984), it would be predicted that the failed inhibits are faster than responses on go trials. We anticipated obtaining this pattern for each of the age groups. Finally, we will focus on the speed of respond- ing on invalid stop-signal trials. In previous studies, we observed that the speed of responding on invalid stop-signal trials was con- siderably slower than on go trials (van den Wildenberg and van der Molen, 2004a; van de Laar et al., 2010). RESULTS The results will be presented in three major sections. In the ﬁrst section, we will examine age-related trends in the speed of responding on go trials and will ask whether these trends differ between tasks (CRT, GST, and SST). A particular question that FIGURE 1 \| Schematic representation of the race model. A distribution of reaction times (RTs) on go trials (trials without a stop-signal) is shown beneath the curve. These values can be seen as ﬁnishing times of the go process. In stop trials, a stop-signal was shown after the go-signal at a particular stop-signal delay. The ﬁnishing time of the stop process bisects the go-signal RT distribution. The left part consists of go-signal RTs fast enough to escape inhibition (i.e., 50%). The right part (50%) represents slow go-signal RTs that will be inhibited because the stop process ﬁnished before. Stop-signal RT (200 ms) is estimated by subtracting average stop-signal delay (100 ms) from the RT that marks the bisection point (300 ms). EXPERIMENTAL TASKS Age-related differ- ences in the slowing on invalid stop-signal trials may contribute to our understanding of how stopping is realized in each of the age groups participating in the current study. In the third section, we will examine performance adjustments following stop-signal trials. First, we asked whether age groups differed in the speed of responding following a successful inhibit relative to the speed on a go trial following another go trial. Second, we asked whether the age groups differed in the speed of responding following a failed inhibit. Third, we examined age-related changes in the speed of responding following an invalid stop-signal trial. For all analyses we examined whether performance adjustments differed between repetition vs. alternation trials. Obviously, it would be of considerable interest to assess whether age-related patterns in performance adjustment are similar across trial type suggesting a single underlying mechanism. Signiﬁcant interaction effects including Age Group were ana- lyzed further by transforming RTs to the natural logarithm to reduce the inﬂuence of differences between age groups in baseline performance (e.g., Meiran, 1996; Huizinga and van der Molen, 2010). Thus, interactions resulting from ANOVA on the trans- formed data indicate a disproportional difference in RT between age groups in one condition relative to another. December 2011 \| Volume 2 \| Article 357 \| 4 PERFORMANCE ON STOP-SIGNAL TRIALS Stop-signal reaction time, SSD, and the percentage of successful inhibits are presented in Table 3 for each age group and both stop-signal tasks. It can be seen that the SSRTs are relatively short in both tasks. Typically, young adults’ global SSRT to auditory stop-signals is about 200 ms (Logan, 1994;Verbruggen and Logan, 2008), and SSRT to visual stop-signals is longer compared to audi- tory ones (van der Schoot et al., 2005). The current SSRT to visual stop-signals, however, is only 181 ms in young adults. In a previ- ous study, using a similar GST, adult SSRT was 210 ms (van de Laar et al., 2010). Most likely, the relatively short SSRTs in the present study are due to the speciﬁcs of the response device (see also van Boxtel et al., 2001). Force transducers were used in the current study whereas a computer keyboard was used in the study reported by van de Laar et al. (2010). Typically, computer key- boards involve a considerable delay in transmission time (e.g., Li et al., 2010). A similar analysis was performed on the error percentages. Analysesyieldedamaineffectof Task,F(2,128) = 20.03,p < 0.001, and Age Group, F(3, 64) = 17.51, p < 0.001, but their interaction was not signiﬁcant,F(6,128) = 1.54,p = 0.19. Subsequent analysis revealed higher error rates on the CRT compared to the stopping tasks (ps < 0.001). Young children committed more errors com- pared to older children (p = 0.04), and older children made more errors than the two adult groups (ps < 0.001). Error rates did not differ between the two adult groups (p = 0.91). FIGURE 2 \|The speed of responding (in ms) on go trials (go RT) of the choice reaction task, the global stop-signal task, and the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). The Task ×Age Group ANOVA on SSRT yielded signiﬁcant main effects of Task, F(1,70) = 138.28, p < 0.001, and Age Group, F(3, 70) = 61.6, p < 0.001, which were included in a signiﬁcant Task ×Age Group interaction, F(3,70) = 15.71, p < 0.001, which is plotted in Figure 3. Subsequent analyses for each task, sep- arately, yielded a signiﬁcant effect of Age Group for the GST, F(3, 71) = 40.98, p < 0.001. PERFORMANCE ON GO TRIALS Task CRT GST SST Age group RT (ms) Error (%) Omission (%) RT (ms) Error (%) Omission (%) RT (ms) Error (%) Omission (%) Young children 488 (70.8) 9.6 (5.6) 2.3 (3.1) 613 (59.0) 5.3 (3.0) 4.1 (4.7) 577 (70.7) 6.2 (3.5) 2.7 (2.3) Older children 373 (45.3) 8.0 (5.7) 0.1 (0.5) 462 (56.8) 3.2 (2.5) 1.2 (2.5) 426 (45.6) 4.2 (2.9) 0.5 (0.9) Young adults 296 (22.2) 2.7 (3.0) 0.1 (0.5) 342 (25.9) 1.3 (1.7) 0.2 (0.4) 326 (28.6) 1.5 (1.9) 0.0 (0.1) Elderly adults 435 (61.7) 3.2 (2.8) 0.1 (0.5) 631 (153.7) 1.5 (1.7) 0.6 (1.2) 601 (137.8) 0.9 (1.2) 1.0 (2.3) Frontiers in Psychology \| Cognition December 2011 \| Volume 2 \| Article 357 \| 4 December 2011 \| Volume 2 \| Article 357 \| 4 Lifespan stopping and adjustments van de Laar et al. by submitting median RTs to ANOVA with Task (3), as a within Ss factor, and Age Group as a between Ss factor. The ANOVA yielded main effects of Task, F(2, 128) = 82.66, p < 0.001, and Age Group, F(3, 64) = 49.35, p < 0.001, and their interaction was signiﬁcant also, F(6, 128) = 8.47, p < 0.001. Subsequent analyses indicated that all age groups showed a delay in responding associated with the insertion of stop-signals (ps < 0.001). The delay was more pronounced for the GST compared to the SST in the child groups and young adults (ps < 0.002), but for the elderly adults the delay did not differ between tasks (p = 0.44). Log transformed analyses indicated that the delay was disproportionally larger in the elderly adults compared to the other age groups on both stop-signal tasks (ps < 0.007). The delay was disproportionally larger for the young children compared to young adults on the GST (p < 0.02), and marginally larger for the SST (p = 0.06). Analyses on omission errors (i.e., errors associated with a failure to respond to the go-signal) yielded a signiﬁcant main effect of Task, F(2, 128) = 8.72, p < 0.001, and Age Group, F(3, 64) = 15.64, p < 0.001, but their interaction was not signiﬁ- cant, F(6, 128) = 1.12, p = 0.35. Omission rate was higher on the stopping tasks compared to the CRT (ps < 0.002). Young children committed more omissions than the other age groups (ps < 0.001). December 2011 \| Volume 2 \| Article 357 \| 5 PERFORMANCE ON STOP-SIGNAL TRIALS Analysis on the log transformed data showed that young children stopped slowest and young adults stopped fastest (ps < 0.009), with older children and elderly adults having intermediate stopping times (p = 0.67). The effect of Age Group was signiﬁcant also for the SST, F(3, 71) = 57.18, p < 0.001. Again, analysis on log transformed SSRTs indicated that young children stopped slowest and young adults stopped fastest (ps < 0.03). But now, older children stopped disproportionally faster than elderly adults (p < 0.001). Finally, selective stopping was slower than global stopping in all age groups (ps < 0.001). The analysis on log transformed data revealed that stopping speed on the SST relative to GST was disproportionally slower in the older and young child groups (respectively 1.4 and 1.3 times, FIGURE 2 \|The speed of responding (in ms) on go trials (go RT) of the choice reaction task, the global stop-signal task, and the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). Table 3 \| Mean stop-signal reaction time (SSRT), stop-signal delay (SSD), and percentage inhibits associated with the global stop-signal task (GST) and the selective stop-signal task (SST) in each age group (SD between parentheses). Task GST SST Age group SSRT (ms) SSD (ms) Inhibits (%) SSRT (ms) SSD (ms) Inhibits (%) Young children 276 (36.4) 329 (59.5) 51.1 (1.6) 356 (49.3) 218 (87.8) 49.7 (2.7) Older children 203 (29.6) 256 (61.9) 50.0 (1.3) 277 (47.2) 149 (55.3) 49.0 (2.9) Young adults 181 (17.4) 159 (27.2) 50.6 (0.9) 203 (26.8) 124 (32.9) 49.6 (1.7) Elderly adults 206 (23.5) 416 (154.4) 52.0 (2.6) 224 (23.7) 368 (132.2) 52.1 (2.7) www.frontiersin.org December 2011 \| Volume 2 \| Article 357 \| 5 Lifespan stopping and adjustments van de Laar et al. p = 0.78) compared to the two adult groups (1.1 times, p = 0.72; ps < 0.001). 1984). This observation is consistent with the notion that failed inhibits are fast responses escaping inhibition. For the SST, failed inhibit RTs were shorter than go RTs for the two child groups and elderly adults, Fs > 48.69, ps < 0.001, but not for young adults, F(1, 16) = 2.66, p = 0.12. Failed inhibit RTs did not differ between stop tasks, F < 1, but there was a substantial difference across age groups, F(3, 70) = 45.85, p < 0.001. PERFORMANCE ON STOP-SIGNAL TRIALS Log transformed analy- sis indicated that failed inhibit RT was shortest in young adults compared to all three other age groups (ps < 0.001). Failed inhibit RT was shorter in older children relative to young children and the elderly (ps < 0.001). Failed inhibit RT did not differ between young children and the elderly (p = 0.58). Finally, as can be seen in Table 3, the percentage of successful inhibitswascloseto50%forallgroupsonbothstoppingtasks.This indicates that the tracking algorithm, which was targeted at 50% successful inhibits, worked quite well. The analysis done on the square rooted proportions showed that the proportion of success- ful inhibits was somewhat higher on the GST compared to the SST, F(1, 70) = 5.27, p = 0.03. Stopping success was somewhat higher in the elderly compared to the other age groups (ps < 0.006). Table 4 presents the RTs associated with failed inhibit responses for both stopping tasks and the RTs associated with invalid stop- signal trials of the SST together with the percentage of omissions on this task (i.e., an invalid stop-signal was presented but the par- ticipant failed to respond). The age-related changes in the speed of failed inhibits and responses on invalid stop-signal trials are plotted in Figure 4. Secondly, in Table 4 it can be seen that failed inhibit RTs are basically similar across the two stopping tasks suggesting that the impact of stop-signal processing on the response process must have been minimal or even absent on these trials. In contrast, the responses on invalid stop trials of the SST are very slow; much slower than the speed of responding on (non-signal) go trials. The Task ×Age Group analysis revealed that invalid RTs were longer compared to go RTs in all age groups, F(3, 71) = 4.76, p = 0.004. The analysis on log transformed data indicated that the slowing on invalid stop trials relative to go trials was disproportionally larger in elderly adults compared to the other age groups (ps < 0.009). PERFORMANCE ON STOP-SIGNAL TRIALS First, the Task ×Age Group ANOVA revealed that failed inhibit RTs were shorter than go RTs on the GST for all age groups, Fs > 64.43, ps < 0.001, as predicted by the horse-race model underlying the analysis of stopping speed (e.g., Logan and Cowan, FIGURE 3 \|The speed of stopping (in ms, SSRT) on the global stop-signal task and the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). Finally, performance on invalid stop-signal trials was relatively accurate. The proportion of omissions remained below 5% for each age group. Statistical analysis yielded a signiﬁcant Age Group effect on the proportion of omissions, F(3, 71) = 6.2, p < 0.001. Elderly adults failed to respond somewhat more frequently on invalid stop trials than young children and young children failed to respond somewhat more frequently than older children and young adults (ps < 0.01). PERFORMANCE ADJUSTMENTS Go trials following go trials To assess sequential effects on go RTs following go trials, we per- formed an ANOVA including Task (3), and Sequence (2) as within Ss factors andAge Group (4) as between Ss factor (see Table5). The analysis yielded a signiﬁcant interaction between Age Group and FIGURE 4 \|The speed of responding (in ms) on failed inhibits (FIRT) of the global stop-signal task and the selective stop-signal task and invalid stop-signal trials (IRT) of the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). FIGURE 3 \|The speed of stopping (in ms, SSRT) on the global stop-signal task and the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). Table 4 \| Median failed inhibit reaction time (FIRT) for the global stop task (GST) and the selective stop task (SST). Table 4 \| Median failed inhibit reaction time (FIRT) for the global stop task (GST) and the selective stop task (SST). Task GST SST Age group FIRT (ms) FIRT (ms) IRT (ms) Omissions (%) Young children 519 (53.0) 517 (56.7) 648 (77.4) 3.6 (3.5) Older children 403 (40.1) 400 (41.4) 503 (78.0) 1.1 (1.7) Young adults 316 (21.3) 320 (30.7) 394 (37.0) 0.6 (1.1) Elderly adults 554 (134.4) 532 (110.3) 721 (155.0) 2.6 (3.5) Median invalid reaction time (IRT) and proportion of omissions for the SST in each age group (SD between parentheses). FIGURE 4 \|The speed of responding (in ms) on failed inhibits (FIRT) of the global stop-signal task and the selective stop-signal task and invalid stop-signal trials (IRT) of the selective stop-signal task for each of the four age groups (8-, 12-, 21-, and 76-years-old). December 2011 \| Volume 2 \| Article 357 \| 6 Frontiers in Psychology \| Cognition Lifespan stopping and adjustments van de Laar et al. The delay of responding following failed inhibits compared to go trials was signiﬁcant for all age groups (ps < 0.01). The analysis on the log transformed data failed to reveal disproportional dif- ferences between age groups in this regard (p = 0.50). The delay following a failed inhibit was similar to the delay following a suc- cessful inhibit for young children and young adults (ps > 0.11). Failed inhibit delay was larger than successful inhibit delay for elderly (p = 0.02), whereas the opposite pattern was observed for older children (p = 0.02). Go trials following stop trials The Task (2; GST and SST) × Duplet (3; go–go, successful inhibit– go, and failed inhibit–go) ×Age Group (4) ANOVA yielded a signiﬁcant main effect of Duplet, F(2, 142) = 17.73, p < 0.001, which was included in a signiﬁcant interaction with Age Group, F(6, 140) = 3.49, p = 0.003. A signiﬁcant main effect of Task was found, F(1, 70) = 11.48, p = 0.001; the overall delay was larger on the GST compared to the SST. Neither Age Group nor Duplet interacted with Task, Fs < 2.12. The data are presented in Table 6 and age-related trends are plotted in Figure 5. Performance adjustments following stop-signal trials may depend on trial-by-trial stimulus sequence (e.g.,Verbruggen et al., 2008). The following analysis was conducted to assess the effect of repetition or alternation stimulus sequence on the perfor- mance adjustments following successful and failed inhibited trials. ANOVA was performed on Task (2), Duplet (2; successful inhibit– go, and failed inhibit–go), Sequence (2) ×Age Group (4). The data are presented in Table 7. This analysis revealed a signiﬁcant Sequenceeffect,F(1,70) = 91.93,p < 0.001.Thespeedof respond- ing on repetitions (528 ms) was longer than on alternations (497 ms). Neither Sequence nor Task did interact with Duplet, F < 1. There was a signiﬁcant interaction between Sequence and Age Group, F(3, 70) = 7.89, p < 0.001. Subsequent analysis on log transformed data indicated that the relative alternation–repetition difference was larger in younger children compared to older The delay in responding following successful inhibits, relative to the speed of responding to go trials following non-signal (go) trials, was signiﬁcant for the two child groups and young adults (ps < 0.004), but not for elderly adults (p = 0.97). The analysis on the log transformed data indicated that the delay did not discrimi- nate disproportionally between the child groups and young adults (ps > 0.43),butthedelaywasdisproportionallylongerforthechild groups and young adults compared to elderly adults (ps < 0.02). Table 5 \| Median RT (in milliseconds) for repetition and alternation on go trials following go trials of the choice reaction task (CRT), the global stop-signal task (GST), and the selective stop-signal task (SST) in each age group (SD between parentheses). PERFORMANCE ADJUSTMENTS Go trials following go trials The analysis on log transformed data revealed that elderly adults delayed the go response following a failed inhibit compared to following a successful inhibit dispro- portionally more than the other age groups (ps < 0.02) and young children somewhat more than older children (p = 0.07). Overall,it appearsthatthedelayfollowingstoptrialsisdependentonwhether the response on the immediately preceding trial was successfully stopped or not. Sequence, F(3, 64) = 8.11, p < 0.001. Sequence did not interact with Task, F < 1. Follow-up analysis indicated that the speed of responding on repetitions was similar to alternations in young adults (315 vs. 319 ms, p = 0.09) whereas the three other age groups showed the opposite pattern (564 vs. 526 ms for younger children,p < 0.001; 423 vs. 409 ms for older children,p < 0.01; and 561 vs. 542 ms, p < 0.04, for elderly adults). Subsequent analysis on log transformed data indicated that the relative alternation– repetition difference was larger in younger children and smaller in young adults relative to the two other age groups (ps < 0.03). Older children did not differ from the elderly adults in this regard (p = 0.62). Go trials following stop trials Table 7 \| Median RT (in milliseconds) for repetition and alternation go trials following a successful inhibit (SI–GO), failed inhibit (FI–GO), and invalid stop (I–GO) of the global stop-signal task (GST) and the selective stop-signal task (SST) in each age group (SD between parentheses). Task GST SST FIGURE 5 \|The difference score between the speed of responding (in ms) on go trials following successful inhibit trials (SI–GO) and on go trials following failed inhibit trials (FI–GO) relative to go trials following go trials (GO–GO) associated with the global stop-signal task (A) and speed of responding (in ms) on go trials following successful inhibit trials (SI–GO), on go trials following failed inhibit trials (FI–GO), and on go trials following invalid stop-signal trials (I–GO) relative to go trials following go trials (GO–GO) associated with the selective stop-signal task (B) for each of the four age groups (8-, 12-, 21-, and 76-years-old). successful inhibit trials (SI–GO), on go trials following failed inhibit trials (FI–GO), and on go trials following invalid stop-signal trials (I–GO) relative to go trials following go trials (GO–GO) associated with the selective stop-signal task (B) for each of the four age groups (8-, 12-, 21-, and 76-years-old). Table 7 \| Median RT (in milliseconds) for repetition and alternation go trials following a successful inhibit (SI–GO), failed inhibit (FI–GO), and invalid stop (I–GO) of the global stop-signal task (GST) and the selective stop-signal task (SST) in each age group (SD between parentheses). Task GST SST SI–GO FI–GO SI–GO FI–GO I–GO Age group Repetition Alternation Repetition Alternation Repetition Alternation Repetition Alternation Repetition Alternation Young children 662 (100.9) 600 (76.6) 669 (82.6) 609 (108.6) 606 (76.1) 567 (68.9) 629 (109.2) 574 (81.4) 603 (92.7) 565 (87.6) Older children 502 (83.9) 478 (69.0) 490 (66.0) 461 (77.4) 449 (59.2) 430 (47.5) 449 (65.5) 421 (56.0) 439 (50.1) 421 (50.7) Young adults 358 (33.0) 352 (32.5) 360 (37.5) 358 (34.3) 347 (30.0) 328 (32.9) 344 (31.5) 331 (34.2) 333 (30.1) 329 (27.2) Elderly adults 636 (146.9) 619 (164.5) 681 (180.2) 650 (200.7) 625 (146.2) 578 (126.8) 652 (151.7) 608 (158.6) 635 (163.3) 623 (161.1) ds) for repetition and alternation go trials following a successful inhibit (SI–GO), failed inhibit (FI–GO), and stop-signal task (GST) and the selective stop-signal task (SST) in each age group (SD between parentheses). A ﬁnal analysis was conducted to assess whether the performance adjustments following invalid stop-signal trials differed between repetitions vs. Go trials following stop trials Task CRT GST SST Age group Repetition Alternation Repetition Alternation Repetition Alternation Young children 517 (76.8) 466 (71.0) 619 (53.3) 583 (63.0) 587 (79.8) 559 (61.9) Older children 383 (46.2) 365 (52.7) 458 (59.5) 445 (49.4) 428 (49.4) 417 (43.2) Young adults 297 (25.7) 296 (22.0) 330 (24.7) 337 (25.7) 319 (28.3) 325 (29.6) Elderly adults 445 (64.5) 424 (61.0) 638 (165.4) 621 (152.2) 606 (168.3) 591 (127.8) Table 6 \| Performance adjustments on the global stop signal task (GST) and the selective stop signal task (SST) for four age groups Table 5 \| Median RT (in milliseconds) for repetition and alternation on go trials following go trials of the choice stop-signal task (GST), and the selective stop-signal task (SST) in each age group (SD between parentheses). Table 6 \| Performance adjustments on the global stop-signal task (GST) and the selective stop-signal task (SST) for four age groups. Task GST SST Age group GO–GO SI–GO FI–GO GO–GO SI–GO FI–GO I–GO Young children 601 (20.4) 631 (22.0) 639 (25.6) 573 (18.9) 587 (18.0) 602 (20.8) 584 (21.3) Older children 451 (19.9) 490 (21.4) 475 (25.0) 423 (18.9) 440 (18.0) 435 (20.8) 430 (21.3) Young adults 333 (21.5) 355 (23.3) 359 (27.1) 322 (20.5) 338 (19.5) 337 (22.6) 331 (23.1) Elderly adults 630 (20.4) 628 (22.0) 665 (25.6) 598 (19.9) 601 (18.9) 630 (21.9) 629 (22.4) Median RTs (in milliseconds) are presented associated with go trials following a go trial (GO–GO), go trials following a successful inhibit (SI–GO), go trials following a failed inhibit (FI–GO) and, for the SST, go trials following an invalid stop-signal trial (I–GO; SD between parentheses). adjustments on the global stop-signal task (GST) and the selective stop-signal task (SST) for four age groups. Table 6 \| Performance adjustments on the global stop-signal task (GST) and the selective stop-signal task (S December 2011 \| Volume 2 \| Article 357 \| 7 www.frontiersin.org www.frontiersin.org Lifespan stopping and adjustments van de Laar et al. Go trials following stop trials alternations. The Sequence (2) ×Age Group (4) ANOVA yielded a signiﬁcant Sequence, F(1, 71) = 19.86, p < 0.001, that interacted with Age Group, F(3, 71) = 3.12, p = 0.03. As can be seen in Table 7, the delay fol- lowing an invalid stop-signal trial was longer when the stim- ulus was repeated compared to when it was alternated for the two child groups (ps < 0.005), but not for the two adult groups (ps > 0.16). The log transformed analysis showed that the relative alternation–repetition difference was disproportionally larger in young children compared to the two adults groups (ps < 0.04). children and young adults (ps < 0.006), and tended to be larger compared to elderly adults (p = 0.06). The relative alternation– repetition difference was also larger in elderly adults compared to young adults (p = 0.02), but did not differ from the older children (p = 0.40). Older children did not differ from young adults either (p = 0.11). A separate analysis was done on the SST task to test perfor- mance adjustments following invalid stop trials. The ANOVA Duplet (4; go–go, successful inhibit–go, failed inhibit–go, and invalid stop–go) ×Age Group yielded a signiﬁcant main effect of Duplet, F(3, 213) = 12.34, p < 0.001, that interacted with Age Group, F(9, 213) = 2.57, p = 0.008 (see Table 6). More speciﬁ- cally, the speed of performance adjustment following an invalid stop trial compared to following a go trial was larger in older chil- dren and in the two adult groups (ps < 0.02), but failed to reach signiﬁcance in young children (p = 0.19). The log transformed analysis showed that elderly adults delayed the response following an invalid stop-signal trial relative to following a go trial dispro- portionally more compared to the other age groups (ps < 0.02). The delay following an invalid stop was disproportionally larger compared to following successful inhibits in the elderly adults compared to the other age groups (ps < 0.01). Finally, the delay following an invalid stop-signal trial did not signiﬁcantly differ from those observed following failed inhibit trials in all age groups (ps > 0.07). Go trials following stop trials FIGURE 5 \|The difference score between the speed of responding (in ms) on go trials following successful inhibit trials (SI–GO) and on go trials following failed inhibit trials (FI–GO) relative to go trials following go trials (GO–GO) associated with the global stop-signal task (A) and speed of responding (in ms) on go trials following successful inhibit trials (SI–GO), on go trials following failed inhibit trials (FI–GO), and on go trials following invalid stop-signal trials (I–GO) relative to go trials following go trials (GO–GO) associated with the selective stop-signal task (B) for each of the four age groups (8-, 12-, 21-, and 76-years-old). Table 7 \| Median RT (in milliseconds) for repetition and alternation go trials following a successful inhibit (SI–GO), failed inhibit (FI–GO), and invalid stop (I–GO) of the global stop-signal task (GST) and the selective stop-signal task (SST) in each age group (SD between parentheses). Task GST SST SI–GO FI–GO SI–GO FI–GO I–GO Age group Repetition Alternation Repetition Alternation Repetition Alternation Repetition Alternation Repetition Alternation Young children 662 (100.9) 600 (76.6) 669 (82.6) 609 (108.6) 606 (76.1) 567 (68.9) 629 (109.2) 574 (81.4) 603 (92.7) 565 (87.6) Older children 502 (83.9) 478 (69.0) 490 (66.0) 461 (77.4) 449 (59.2) 430 (47.5) 449 (65.5) 421 (56.0) 439 (50.1) 421 (50.7) Young adults 358 (33.0) 352 (32.5) 360 (37.5) 358 (34.3) 347 (30.0) 328 (32.9) 344 (31.5) 331 (34.2) 333 (30.1) 329 (27.2) Elderly adults 636 (146.9) 619 (164.5) 681 (180.2) 650 (200.7) 625 (146.2) 578 (126.8) 652 (151.7) 608 (158.6) 635 (163.3) 623 (161.1) children and o ng ad lts (ps < 0 006) and tended to be larger A ﬁnal anal sis as cond cted to assess hether the FIGURE 5 \|The difference score between the speed of responding (in ms) on go trials following successful inhibit trials (SI–GO) and on go trials following failed inhibit trials (FI–GO) relative to go trials following go trials (GO–GO) associated with the global stop-signal task (A) and speed of responding (in ms) on go trials following successful inhibit trials (SI–GO), on go trials following failed inhibit trials (FI–GO), and on go trials following invalid stop-signal trials (I–GO) relative to go trials following go trials (GO–GO) associated with the selective stop-signal task (B) for each of the four age groups (8-, 12-, 21-, and 76-years-old). DISCUSSION The current study aimed at examining lifespan changes in global and selective stopping and in post-stopping performance adjust- ments. With regard to the ﬁrst aim, the results showed that global stopping was faster than selective stopping, which is typically interpreted in terms of the time consumed by additional process- ing (e.g., stimulus discrimination) in selective stopping (van den Wildenberg and van der Molen,2004b;Verbruggen et al.,2005;van de Laar et al., 2010). It could be argued that a longer SSRT in the SST is due to the lower probability of valid stop-signals in this task relative to the global task (i.e., 17.5 vs. 35%, respectively). A lower probability of stopping is typically associated with faster responses December 2011 \| Volume 2 \| Article 357 \| 8 Frontiers in Psychology \| Cognition Lifespan stopping and adjustments van de Laar et al. on go trials, as has been observed in both go/no–go paradigms (e.g., Luce, 1986) and in stop-signal paradigms (e.g., van den Wildenberg and van der Molen, 2003; Ramautar et al., 2004; van de Laar et al., 2010). Indeed, in the current study responses on go trials were faster on the selective stopping task compared to the global stopping task (483 vs. 513 ms, respectively). The longer SSRT on the selective stop task is then explained by assuming that it is harder to stop because responses on go trials become more pre-potent when stop-signals are fewer. It should be noted, how- ever,that in studies manipulating stop-signal probability,SSRT did not change (Ramautar et al., 2004; van de Laar et al., 2010) or was even shorter, not longer, when stop-signals were fewer (e.g., van den Wildenberg and van der Molen, 2003). Thus, at this point, the most likely interpretation of longer SSRTs on the SST is in terms of added signal discrimination demands. between young vs. elderly adults was only 203 vs. 224 ms, respec- tively. But the elderly adults differed from young adults in being disproportionally slow on invalid stop-signal trials (721 vs. 394 ms, respectively). One interpretation of the conspicuous absence of a sizable dif- ference in the speed of selective stopping between young vs. elderly adults would be that on a proportion, if not all, of the trials, the elderlyadultsstoppedallresponsesﬁrstandthendecidedtorecruit the appropriate response after determining that inhibition is not required(e.g.,AronandVerbruggen,2008;vandeLaaretal.,2010). DISCUSSION The former has been interpreted in terms of automatic facilitation or priming and the latter has been taken to be a manifestation of subjective expectancy (e.g., Gao et al., 2009). Our data deviate from previous ﬁndings in show- ing an alternation beneﬁt for all age groups with the exception of young adults who failed to show signiﬁcant sequential effects. One interpretation would be that the current discrepancy is due to the insertion of stop-signals into the trial series. This interpretation is unlikely since the alternation beneﬁt associated with invalid stop trials was absent also in the two adult groups. Thus, an alternative interpretation assumes that subjective expectancies are relatively short-lived and dissipate rapidly so that their inﬂuence cannot be detected when inter-trial intervals are relatively long. Along those lines,it would be predicted that with increasing inter-trial intervals alternation beneﬁts should disappear gradually in all age groups. A host of studies indicated that the ability to inhibit pre-potent responses improves rapidly when children are growing older (e.g., Ridderinkhof et al., 1999; Band et al., 2000; van den Wildenberg and van der Molen, 2004a) and starts to decline during senescence (e.g., Williams et al., 1999; Bedard et al., 2002). The current ﬁnd- ings are basically consistent with this literature. More speciﬁcally, the current ﬁndings obtained for the GST showed that the speed of global stopping was slower for young children and faster for young adults compared to older children and elderly adults who did not differ in this regard. This pattern was similar for selective stopping with one exception; the speed of selective stopping did not dis- criminate between the young vs. elderly adults. It should be noted that the developmental gain in global stopping was larger during childhood (young vs. older children) than adolescence (older chil- dren vs. young adults). For selective stopping the developmental gain during childhood was basically similar to the gain during adolescence. This pattern is consistent with the ﬁndings reported previously by van den Wildenberg and van der Molen (2004a) and suggests that selective stopping matures slower than global stopping. Based on a “last-in-ﬁrst-out” hypothesis of lifespan changes in neurocognition (e.g., Davis et al., 2009), one would be led to pre- dict that with advancing age selective stopping would be affected more than global stopping. The current ﬁndings,however,showed the opposite pattern. Global, but not selective, stopping discrim- inated signiﬁcantly between young vs. elderly adults. DISCUSSION The deployment of a“stop-all”strategy in the elderly adults might be another instance of their inclination of preventing errors even when it slows them down doing so. It has been observed that a considerable portion of response slowing in the elderly is due to extra time consumed by response generation at the cortical level, as indicated by brain potentials (e.g., Kolev et al., 2005), the corticospinal level, as indicated by motor-evoked potentials (e.g., Fujiyama et al., 2011), and response activation, as indicated by movement kinematics (e.g., Trewartha et al., 2011). Moreover, it has been observed that the elderly show excessive response activa- tion on no–go trials (e.g., Vallesi and Stuss, 2010; see also Vallesi, 2011; Vallesi et al., 2011). Thus, re-activation following the sup- pression of response-related over-recruited neural circuitry might be particularly time-consuming in the elderly (e.g., Vallesi, 2011; see also Vallesi and Stuss, 2010). Both global and selective SSRTs showed the typical U-shaped relation with advancing age albeit less pronounced than RTs on non-signal (go trials) or signal–respond trials (failed inhibits and invalid stop trials). The U-shaped relation between the processing speed and advancing age is a ubiquitous phenomenon (e.g.,review in Cerella and Hale, 1994) that attracted a score of interpretations, including age-related changes in the setting of response thresholds (Starns and Ratcliff,2010),information loss (Myerson et al.,1990), inhibitory control (Dempster, 1992), frontal lobe function (West, 1996), and neural noise (Kail, 1997). The second aim of the current study was to examine lifespan changes in post-stopping adjustment. Previous studies indicated that post-stopping adjustment is critically dependent on stimulus– response repetitions and alternations across trials. That is, the speed of responding is delayed on post-stopping trials and the delay is largest when the stimulus is repeated (Verbruggen et al., 2008; Bissett and Logan, 2011). Verbruggen et al. (2008) argued that the goal to stop on successful inhibits replaces the go goal on the following go trial if the stimulus is repeated. In previous studies we examined age-related changes in trial-by-trial sequen- tial effects (Melis et al., 2002; Smulders et al., 2005). This research showed that relatively long intervals between trials (i.e., >500 ms) are associated with a repetition beneﬁt for young children (Smul- ders et al., 2005) and an alternation beneﬁt for young and elderly adults (Melis et al., 2002). DISCUSSION Typically, global stopping is slower in the elderly compared to young adults (e.g., Williams et al., 1999; Rush et al., 2006; but see Kramer et al., 1994). The single study of selective stopping including elderlyadults(Bedardetal.,2002)showedapronounceddifference between young vs. elderly adults in the speed of selective stopping; 248 vs. 329 ms, respectively. In the current study, the difference Most important, the current young adult results are consistent with previous reports showing a delay in the speed of responding on trials following successful or failed inhibits in GSTs (e.g., Rieger December 2011 \| Volume 2 \| Article 357 \| 9 www.frontiersin.org Lifespan stopping and adjustments van de Laar et al. stop-signal tasks. Second, on both stop-signal tasks, the slowing followingfailedinhibitswasdisproportionallylargerthaninyoung adults. Third, the slowing following invalid stop-signal trials was disproportionally larger than in young adults. The observation that the elderly adults slowed following failed inhibits but did not following successful inhibits presents a challenge for the orienting interpretation of conﬂict adjustment proposed by Notebaert et al. (2009). According to the orienting account performance adjust- ments following successful and failed inhibits should be the same because the probability of success and failure is identical (around 50%). In this regard, the current ﬁndings are inconsistent with the prediction derived from the orienting account. The current ﬁndings are problematic also with regard to the conﬂict monitor- ing hypothesis (Botvinick et al., 2001). Conﬂict arises on all three stop-signal trials and, thus, performance adjustments should be observed on all trials following a stop-signal trial. The current data of elderly adults do suggest that performance adjustment is triggered by conﬂict but only when this conﬂict is resolved by the execution of an overt response,either correct,as on trials following invalid stop-signal trials, or incorrect, as on trials following failed inhibits. and Gauggel, 1999; Verbruggen et al., 2008). The current results add to this literature by showing that a similar pattern occurs on a SST. The present data revealed that responses following failed inhibits are not signiﬁcantly slower than responses following suc- cessful inhibits. Similarly, responses following invalid stop-signal trials in the selective stop-signal trials were equally slow compared to responses following successful inhibits. In contrast to previous studies, the post-stop trial adjustments in the speed of responding were not inﬂuenced by sequential effects. DISCUSSION Most likely, the occurrence of stop-signals indicated a greater need for caution in the elderly participants relative to the other age groups and, thus, response slowing on go trials is disproportionally larger in the elderly adults. This interpretation is consistent with the typ- ical observation that elderly adults favor accuracy over speed (e.g., Band and Kok, 2000; Starns and Ratcliff, 2010). If elderly adults exercised more proactive control compared to the younger age groups there might be less need for them to resort to reactive con- trol when they are confronted with a stop-signal trial. The current absence of a delay in responding on go trials following success- ful inhibits might suggest that proactive control was sufﬁcient in dealing with the conﬂict between stopping and going. Apparently, proactive control was not sufﬁcient on failed inhibits, which could signal the need for more control on subsequent trials and this is reﬂected in slowing on go trials following failed inhibits. Finally, p g To our surprise, the developmental ﬁndings did not differ from the results obtained in young adults. Response slowing following stop-signal trials on both the global and SSTs was somewhat more pronounced in young adults but not disproportionally so. Stud- ies examining age-performance adjustments following stop-signal trials are scarce and focused on comparisons between clinical groups vs. typically developing children rather than different age groups. Some developmental studies examined post-error slowing but the outcomes of these studies are inconsistent (for a review see Smulders et al., under review). Typically, in studies examining developmental change in post-error slowing, the results are inter- preted in terms of conﬂict monitoring (Botvinick et al., 2001) and, based on event-related brain potentials or bold responses associ- ated with error processing, it is usually suggested that conﬂict monitoring is still immature in children and, in some studies, not fully developed even in adolescents (e.g., Velanova et al., 2008; Braet et al., 2009). Obviously, the current pattern of ﬁndings is difﬁcult to reconcile with the notion of developmental improve- ments in conﬂict monitoring, at least across the age span under study. By contrast, the current pattern suggests that the mech- anisms engaged on stop-signal trials and resulting in response slowing on subsequent trials are already in place and mature in children as young as 7 years of age. DISCUSSION Thus the current results cannot be interpreted easily in terms of repetition priming sug- gesting that the stop-signal on the previous successful inhibit trial automatically primes the stop process on the current trial (e.g., Verbruggen et al., 2008). Instead the current ﬁndings are com- patible with several notions assuming that a stop-signal on the previous trial induces a switch to a more conservative response set resulting in a slowing of response execution. These notions assume that the conﬂict between responding and stopping induced by the stop-signal recruit control processes resulting in performance adjustments on the subsequent trial (e.g., Rieger and Gauggel, 1999; Botvinick et al., 2001) or that the infrequent stop-signal elicits an orienting response reducing the speed of responding on the subsequent trial (e.g., Notebaert et al., 2009). Along these lines it would be predicted that response slowing on trials following a stop-signal does not discriminate between successful vs. failed inhibit vs. invalid stop-signal trials. A uniﬁed account of the data could be provided by assuming age-related changes in the balance between proactive and reac- tive performance adjustments in the stop-signal tasks. Although participants were instructed not to delay their speed of respond- ing, and in spite of the tracking algorithm that should discourage participants from slowing down, the speed of responding on the stop-signal tasks was considerably slower than on the standard choice task in all age groups. The delay in responding could be considered an instance of proactive control exercised by the par- ticipants in an attempt to increase their chances to withhold the response when a stop-signal instructed them to do so. Proactive control in the current study is manifested also by differential slow- ing in the two stop-signal tasks. Larger delays were observed on the GST compared to the SST, presumably because proactive control is guided by the probability of stopping rather than stop-signal occurrence, as stop-signals might turn out to be invalid in the SST. Proactive control in stop-signal performance has been demon- strated in previous studies,both at the behavioral (e.g.,Verbruggen and Logan, 2009) and neurocognitive level (for a review see Aron, 2011). Importantly, the slowing associated with the insertion of stop-signals was most pronounced for the elderly adults. Frontiers in Psychology \| Cognition DISCUSSION A major difference between studies examining developmental change in post-error slowing and the current study focusing on performance adjustments fol- lowing stop-signal trials is that errors are typically less frequent than stop-signals. It would be of considerable interest to examine developmental change in post-error and stop-signal slowing by calibrating the frequencies of error and stop-signal occurrence. The ﬁndings obtained for the elderly adults differed from the results of the other age groups in three respects. First, the elderly adults did not slow following successful inhibits on both December 2011 \| Volume 2 \| Article 357 \| 10 Lifespan stopping and adjustments van de Laar et al. the slowing on go trials following invalid stop-signal trials may suggest that reactive control is elicited on those trials by the added requirement of re-initiating the response that was stopped ﬁrst in reaction to the detection of the stop-signal. The current interpre- tation is similar to the goal-priority hypothesis proposed recently by Bissett and Logan (2011) to account for the dynamic interplay between pro- and reactive control in the stop-signal paradigm. The goal-priority hypothesis assumes that stop-signals indicate the need for more caution and thus participants reduce their speed of responding. This might happen following a stop-signal trial (reactive control) but also during the anticipation of a stop-signal trial (proactive control) when, for example, explicit cues signal the importance of stopping (e.g., Verbruggen and Logan, 2009). It should be of considerable interest to further examine lifespan changes in the dynamic balance between pro- and reactive control in the stop-signal trial by cueing the importance of stopping vs. response speed. young adults in this regard. Consequently, the type of reactive control exercised in the stop-signal paradigm seems to mature very rapidly. This ﬁnding has important implications for devel- opmental theories of cognitive control (e.g., Durston and Casey, 2006), in particular for developmental concepts of conﬂict mon- itoring (e.g., Braet et al., 2009). All age groups delayed the speed of responding when stop-signals were inserted in the trial series and the delay was interpreted in terms of proactive control aimed at enhancing stopping success and in terms of an adjustment of response thresholds (e.g., Bissett and Logan, 2011). The data pat- tern obtained for the elderly adults suggest that they shifted the balance between proactive and reactive cognitive control toward proactive control. ACKNOWLEDGMENTS h d d This study was supported by NWO grants to Maurits W. van der Molen and Wery P. M. van den Wildenberg (MaGW grant 400-03- 261 and VENI grant 451-06-012, respectively). The support of the Netherlands Institute for Advanced Study in the Humanities and Social Sciences is also gratefully acknowledged. We thank Bert van Beek for programming the computer tasks and Marcus Spaan and Bert Molenkamp for technical support. REFERENCES Huizinga, M., and van der Molen, M. W. (2010). Task switching and shift- ing between stopping and going: developmental change in between- trial control adjustments. J. Exp. Child Psychol. 108, 484–503. white matter tracts using diffusion tensor tractography. Neuroimage 46, 530–541. Age-associated reduction of asymmetry in prefrontal function and preservation of conceptual repetition priming. Neuroimage 45, 237–245. Aron, A. R. (2011). From reac- tive to proactive and selective control: developing a richer model for stopping inappropriate responses. Biol. Psychiatry 69, 55–68. Dempster, F. N. (1992). The rise and fall of the inhibitory mechanism: toward a uniﬁed theory of cognitive developmentandaging.Dev.Rev.12, 45–75. trial control adjustments. J. Exp. Child Psychol. 108, 484–503. Bissett, P. G., and Logan, G. D. (2011). Balancing cognitive demands: con- trol adjustments in the stop-signal paradigm. J. Exp. Psychol. Learn. Mem. Cogn. 37, 392–404. Johnstone, S. J., Dimoska, A., Smith, J. L., Barry, R. J., Pleffer, C. B., Chiswick, D., and Clarke, A. R. (2007). The development of stop- signal and go/nogo response inhibi- tion in children aged 7–12 years: per- formance and event-related poten- tial indices. Int. J. Psychophysiol. 63, 25–38. developmentandaging.Dev.Rev.12, 45–75. Aron, A. R., and Verbruggen, F. (2008). Stop the presses: dissociating a selective from a global mecha- nism for stopping. Psychol. Sci. 19, 1146–1153. Durston, S., and Casey, B. J. (2006). What have we learned about cognitive development from neu- roimaging? Neuropsychologia 44, 2149–2157. paradigm. J. Exp. Psychol. Learn. Mem. Cogn. 37, 392–404. Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., and Cohen, J. D. (2001). Conﬂict monitoring and cognitive control. Psychol. Rev. 108, 624–652. Band, G. P. H., and Kok, A. (2000). Age effects on response monitoring in a mental-rotation task. Biol. Psychol. 51, 201–221. Fleischman, D. A. (2007). Repetition priming in aging and Alzheimer’s disease: an integrative review and future directions. Cortex 43, 889–897. Kail, R. (1997). The neural noise hypothesis: evidence from process- ing speed in adults with multiple sclerosis. Aging Neuropsychol. Cogn. 4, 157–165. Braet, W., Johnson, K. ACKNOWLEDGMENTS h d d A., Tobin, C. T., Acheson, R., Bellgrove, M. A., Robertson, I. H., and Garavan, H. (2009). Functional developmen- tal changes underlying response inhibition and error-detection processes. Neuropsychologia 47, 3143–3151. Band, G. P. H., van der Molen, M. W., and Logan, G. D. (2003). Horse-race model simulations of the stop-signal procedure. Acta Psychol. (Amst.) 112, 105–142. Fujiyama, H., Tandonnet, C., and Sum- mers, J. J. (2011). Age-related dif- ferences in corticospinal excitability during a go/nogo task. Psychophysi- ology 48, 1448–1455. Kolev, V., Falkenstein, M., and Yor- danova, J. (2005). Motor-response generation as a source of aging- related behavioural slowing in choice-reaction tasks. Neurobiol. Aging 27, 1719–1730. Band, G. P. H., van der Molen, M. W., Overtoom, C. C., and Verbaten, M. N. (2000). The ability to activate and inhibit speeded responses: separate developmental trends. J. Exp. Child Psychol. 75, 263–290. Gamboz, N., Borella, E., and Brandi- monte, M. A. (2009). The role of switching, inhibition, and working memory in older adults’ perfor- mance in the Wisconsin card sort- ing test. Neuropsychol. Dev. Cogn. B Aging Neuropsychol. Cogn. 16, 260–284. Cerella, J., and Hale, S. (1994). The rise and fall in information-processing rates over the life span. Acta Psychol. (Amst.) 86, 109–197. Kramer, A. F., Humphrey, D. G., Lar- ish, J. F., Logan, G. D., and Strayer, D. L. (1994). Aging and inhibi- tion: beyond a unitary view of inhibitory processing in attention. Psychol. Aging 9, 491–512. Crone, E. A., Jennings, J. R., and van der Molen, M. W. (2004). Developmental change in feedback processing as reﬂected by phasic heart rate changes. Dev. Psychol. 40, 1228–1238. Bedard, A., Nichols, S., Barbosa, J. A., Schachar, R., Logan, G. D., and Tan- nock, R. (2002). The development of selective inhibitory control across the life span. Dev. Neuropsychol. 21, 93–111. Gao, J., Wong-Lin, K., Holmes, P., Simen, P., and Cohen, J. D. (2009). Sequential effects in two-choice reaction time tasks: decomposition andsynthesisof mechanisms.Neural Comput. 21, 2407–2436. Kray, J., Kipp, K. H., and Karbach, J. (2009).Thedevelopmentof selective inhibitory control: the inﬂuence of verbal labeling. Acta Psychol. (Amst.) 130, 48–57. Bergerbest, D., Gabrieli, J. D. E., Whitﬁeld-Gabrieli, S., Kim, H., Stebbins, G. T., Bennett, D. A., and Fleischman, D. A. (2009). Davis, S. W., Dennis, N. A., Buchler, N. G., White, L. E., Madden, D. J., and Cabeza, R. (2009). DISCUSSION Most likely this shift is related to their will- ingness to trade speed for accuracy. Further investigation of the factors inﬂuencing the balance between pro- and reactive control in development and cognitive aging is an important goal for future research. In conclusion, the current ﬁndings indicate that, with regard to global and selective stopping,development,and aging are not sim- plytwosidesof thesamecoin. Developmentalchangeinglobaland selective stopping revealed different trajectories, with global stop- ping maturing more rapidly than selective stopping. The pattern of ﬁndings is inconsistent with a simple“last-in-ﬁrst-out”hypoth- esis of lifespan changes in (neuro)cognition. Importantly,children slowed their speed of responding following stop-signal trials, but the current data revealed little difference between children and ACKNOWLEDGMENTS h d d Assess- ing the effects of age on long December 2011 \| Volume 2 \| Article 357 \| 11 www.frontiersin.org Lifespan stopping and adjustments van de Laar et al. and anxious children: a biologically based model of child psychopathol- ogy. J. Abnorm. Child. Psychol. 24, 19–36. Levitt, H. (1971). Transformed up– down methods in psychoacoustics. J. Acoust. Soc. Am. 49, 467–477. Verbruggen,F.,and Logan,G. D. (2008). Response inhibition in the stop- signal paradigm. Trends Cogn. Sci. 12, 418–424. Vallesi, A., McIntosh, A. R., and Stuss, D. T. (2011). Overrecruitment in the aging brain as a function of task demands: evidence for a compen- satory view. J. Cogn. Neurosci. 23, 801–815. Li, X., Liang, Z., Kleiner, M., and Lu, Z. L. (2010). RTbox: a device for highly accurate response time mea- surements. Behav. Res. Methods 42, 212–225. Verbruggen,F.,and Logan,G. D. (2009). Proactive adjustments of response strategies in the stop-signal para- digm. J. Exp. Psychol. Hum. Percept. Perform. 35, 835–854. Ramautar, J. R., Kok, A., and Rid- derinkhof, K. R. (2004). Effects of stop-signal probability in the stop- signal paradigm: the N2/P3 complex further validated. Brain Cogn. 56, 234–252. Vallesi,A.,and Stuss,D. T. (2010). Exces- sive sub-threshold motor prepara- tionfornon-targetstimuliinnormal aging. Neuroimage 50, 1251–1257. Logan, G. D. (1994). “On the abil- ity to inhibit thought and action: a users’ guide to the stop signal paradigm,” in Inhibitory Processes in Attention, Memory and Language, eds D. Dagenbach and T. H. Carr (San Diego, CA: Academic Press), 189–239. Verbruggen, F., Logan, G. D., Liefooghe, B., and Vandierendonck, A. (2008). Short-term aftereffects of response inhibition: repetition priming or between-trial control adjustments? J. Exp. Psychol. Hum. Percept. Per- form. 34, 413–426. Raven, J., Raven, J. C., and Court, J. H. (1993). Manual for Raven’s Progres- sive Matrices and Mill Hill Vocabu- lary Scales. Oxford: Oxford Psychol- ogists Press. van Boxtel, G. J. M., van der Molen, M. W., Jennings, J. R., and Brunia, C. H. M. (2001). A psychophysiological analysis of inhibitory motor control in the stop- signal paradigm. Biol. Psychol. 58, 229–262. Raven, J. C., Court, J. H., and Raven, J. (1985). Manual for Raven’s Progres- sive Matrices and Vocabulary Scales. Standard Progressive Matrices, Sect. 3. London: H. K. Lewis. Logan, G. D., and Burkell, J. (1986). Dependence and independence in responding to double stimulation: a comparison of stop, change, and dual-task paradigms. J. Exp. Psychol. Hum. Percept. Perform. 12, 549–563. West, R., and Schwarb, H. (2006). ACKNOWLEDGMENTS h d d The inﬂuence of aging and frontal func- tion on the neural correlates of regu- lative and evaluative aspects of cog- nitive control. Neuropsychology 20, 468–481. van de Laar, M. C., van den Wilden- berg, W. P. M., van Boxtel, G. J. M., and van der Molen, M. W. (2010). Processing of global and selective stop signals: application of Donders’ subtraction method to stop-signal task performance. Exp. Psychol. 57, 149–159. Ridderinkhof, K. R., Band, G. P. H., and Logan, G. D. (1999). A study of adaptive behavior: effects of age and irrelevant information on the ability to inhibit one’s actions. Acta Psychol. (Amst.) 101, 315–337. West, R. L. (1996). An application of prefrontal cortex function theory to cognitive aging. Psychol. Bull. 120, 272–292. Logan, G. D., and Cowan, W. B. (1984). On the ability to inhibit thought and action: a theory of an act of control. Psychol. Rev. 91, 295–327. van den Wildenberg, W. P. M., and van der Molen, M. W. (2003). The duration of response inhibition in the stop-signal paradigm varies with response force. Acta Psychol. (Amst.) 114, 115–129. Rieger, M., and Gauggel, S. (1999). Inhibitory after-effects in the stop signal paradigm. Br. J. Psychol. 90, 509–518. Williams, B. R., Ponesse, J. S., Schachar, R. J., Logan, G. D., and Tannock, R. (1999). Development of inhibitory control across the life span. Dev. Psychol. 35, 205–213. Luce, R. D. (1986). Response Times: Their Role in Inferring Elementary Mental Organization. New York, NY: Oxford University Press. Rush, B. K., Barch, D. M., and Braver, T. S. (2006). Accounting for cognitive aging: context processing, inhibition or processing speed? Neuropsychol. Dev. Cogn. B Aging Neuropsychol. Cogn. 13, 588–610. May, C. P., and Hasher, L. (1998). Syn- chrony effects in inhibitory control over thought and action. J. Exp. Psychol. Hum. Percept. Perform. 24, 363–379. van den Wildenberg, W. P. M., and van der Molen, M. W. (2004a). Develop- mental trends in simple and selective inhibition of compatible and incom- patible responses. J. Exp. Child. Psy- chol. 87, 201–220. Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Mayda, A. B. V., Westphal, A., Carter, C. S., and DeCarli, C. (2011). Late life cognitive control deﬁcits are accen- tuated by white matter disease bur- den. Brain 134, 1673–1683. Smulders, S. ACKNOWLEDGMENTS h d d F., Notebaert, W., Meijer, M., Crone, E. A., van der Molen, M. W., and Soetens, E. (2005). Sequential effects on speeded infor- mation processing: a developmen- tal study. J. Exp. Child Psychol. 90, 208–234. van den Wildenberg, W. P. M., and van der Molen, M. W. (2004b). Additive factors analysis of inhibitory pro- cessing in the stop-signal paradigm. Brain Cogn. 56, 253–266. Received: 01 September 2011; paper pending published: 04 October 2011; accepted: 12 November 2011; published online: 15 December 2011. Meiran, N. (1996). Reconﬁguration of processing mode prior to task per- formance. J. Exp. Psychol. Learn. Mem. Cogn. 22, 1423–1442. van der Schoot, M., Licht, R., Horsley, T. M., and Sergeant, J. A. (2005). Effects of stop signal modality, stop signal intensity and tracking method of inhibitory performance as deter- mined by use of the stop sig- nal paradigm. Scand. J. Psychol. 46, 331–341. Somerville, L. H., and Casey, B. J. (2010). Developmental neurobiol- ogy of cognitive control and moti- vational systems. Neurobiology 20, 236–241. Melis, A., Soetens, E., and van der Molen, M. W. (2002). Process- speciﬁc slowing with advancing age: evidence derived from the analysis of sequential effects. Brain Cogn. 49, 420–435. Citation: van de Laar MC, van den Wildenberg WPM, van Boxtel GJM and van der Molen MW (2011) Lifes- pan changes in global and selec- tive stopping and performance adjust- ments. Front. Psychology 2:357. doi: 10.3389/fpsyg.2011.00357 Starns, J., and Ratcliff, R. (2010). The effects of aging on the speed- accuracy compromise: boundary optimality in the diffusion model. Psychol. Aging 25, 377–390. Myerson, J., Hale, S., Wagstaff, D., Poon, L., and Smith, G. (1990). The information-loss model: a mathe- matical theory of age-related cog- nitive slowing. Psychol. Rev. 97, 475–487. Velanova, K., Wheeler, M. E., and Luna, B. (2008). Maturational changes in anterior cingulate and frontopari- etal recruitment support the devel- opment of error processing and inhibitory control. Cereb. Cortex 18, 2505–2522. This article was submitted to Frontiers in Cognition, a specialty of Frontiers in Psychology. Trewartha,K. M.,Penhune,V. B.,and Li, K. Z. (2011). Movement kinematics of prepotentresponsesuppressionin aging during conﬂict adaptation. J. Gerontol. B Psychol. Sci. Soc. Sci. 66, 185–194. Copyright © 2011 van de Laar, van den Wildenberg, van Boxtel and van der Molen. ACKNOWLEDGMENTS h d d This is an open-access article sub- ject to a non-exclusive license between the authors and Frontiers Media SA, which permits use, distribution and reproduc- tioninotherforums,providedtheoriginal authors and source are credited and other Frontiers conditions are complied with. Copyright © 2011 van de Laar, van den Wildenberg, van Boxtel and van der Molen. This is an open-access article sub- ject to a non-exclusive license between the authors and Frontiers Media SA, which permits use, distribution and reproduc- tioninotherforums,providedtheoriginal authors and source are credited and other Frontiers conditions are complied with. Notebaert, W., Houtman, F., Opstal, F. V., Gevers, W., Fias, W., and Verguts, T. (2009). Post-error slowing: an orienting account. Cognition 111, 275–279. Verbruggen, F., Liefooghe, B., and Vandierendonck, A. (2005). On the difference between response inhibi- tion and negative priming: evidence from simple and selective stopping. Psychol. Res. 69, 262–271. Vallesi, A. (2011). Targets and non- targets in the aging brain: a go/nogo event-related potential study. Neu- rosci. Lett. 487, 313–317. Oosterlaan,J.,and Sergeant,J.A. (1996). Inhibition in ADHD, aggressive, December 2011 \| Volume 2 \| Article 357 \| 12 Frontiers in Psychology \| Cognition
https://openalex.org/W4205552178	http://ejournal-balitbang.kkp.go.id/index.php/JSJ/article/download/10719/7650	Indonesian	null	EFEKTIVITAS PENGGUNAAN AUTOMATIC FEEDER PADA BUDIDAYA UDANG VANAME (Litopenaeus vannamei) DI PT. WINDU MARINA ABADI KECAMATAN SAMBELIA, LOMBOK TIMUR	Buletin Jalanidhitah Sarva Jivitam	2,021	cc-by-sa	3,044	Giban Samawi1, Amyda Suryati Panjaitan1, Erni Marlina1, Luchiandini Ika Pamaharyani1, Ofan Bosman2, Dewi Nurmalita Suseno3 1Prodi Teknologi Akuakultur Politeknik Ahli Usaha Perikanan Jl. AUP No. 1 Pasar Minggu-Jakarta Selatan, Jakarta 12520 2Pusat Riset Perikanan Jl. Pasir Putih II, Ancol Timur, Jakarta Utara 14430 3Politeknik Kelautan Perikanan Sidoarjo Jl. Raya Buncitan, Gedangan, Dusun Kp. Baru, Buncitan, Kabupaten Sidoarjo, Jawa Timur 61254 1Prodi Teknologi Akuakultur Politeknik Ahli Usaha Perikanan Jl. AUP No. 1 Pasar Minggu-Jakarta Selatan, Jakarta 12520 2Pusat Riset Perikanan Jl. Pasir Putih II, Ancol Timur, Jakarta Utara 14430 3Politeknik Kelautan Perikanan Sidoarjo Jl. Raya Buncitan, Gedangan, Dusun Kp. Baru, Buncitan, Kabupaten Sidoarjo, Jawa Timur 61254 Pusat Riset Perikanan Jl. Pasir Putih II, Ancol Timur, Jakarta Utara 14430 3Politeknik Kelautan Perikanan Sidoarjo Jl. Raya Buncitan, Gedangan, Dusun Kp. Baru, Buncitan, Kabupaten Sidoarjo, Jawa Timur 61254 j Jl. Raya Buncitan, Gedangan, Dusun Kp. Baru, Buncitan, Kabupaten Sidoarjo, Jawa Timur 61254 Email: luchiika02@gmail.com Email: luchiika02@gmail.com ABSTRAK Budidaya udang vaname menjadi primadona di Indonesia karena memiliki nilai komersil dan memberikan pendapatan bagi negara. Teknologi pemberian pakan salah satu faktor untuk keberhasilan budidaya. Saat ini teknologi berbasis IoT yaitu automatic feeder yang sedang tren pada budidaya udang vaname. Penelitian ini bertujuan untuk mengetahui efektivitas penggunaan automatic feeder pada budidaya udang vaname. Penelitian dilaksanakan pada bulan Maret-Mei 2021 di PT. Windu Marina Abadi, Lombok Timur dengan pengukuran pertumbuhan, kelangsungan hidup, dan FCR serta menganalisis kualitas air. Hasil penelitian menunjukkan pertumbuhan berat terendah terdapat pada petak A5 (7,79 g) dan tertinggi B5 (18,14 g) dengan rata-rata laju pertumbuhan perhari terendah yaitu A5 (0,09 g) dan B5 (0,21 g), kelangsungan hidup pada tambak menggunakan automatic feeder mencapai 97%, FCR tertinggi yaitu pada tambak (manual) A5 (2,45) dan terendah pada tambak B5 (automatic feeder) (1,08). Nilai kualitas air masih kisaran normal untuk budidaya. Tambak dengan penggunaan automatic feeder relatif lebih baik dibandingkan dengan tambak dengan pemberian pakan secara manual. Kata kunci: Udang vaname, automatic feeder; pertumbuhan, teknologi Keywords: Vaname shrimp, automatic feeder, growth, technology Buletin JSJ, 3 (2), 2021, 93-99 Buletin JSJ, 3 (2), 2021, 93-99 Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index PENDAHULUAN Salah satu sektor perikanan budidaya di Indonesia yang sangat potensial dan mempunyai prospek yang besar dalam peningkatan devisa negara (Pantjara et al., 2015). Salah satunya adalah usaha budidaya udang vaname (Litopenaeus vannamei). Permintaan udang vaname di pasar luar negeri yang sangat tinggi dapat meningkatkan devisa negara (Kharisma & Manan, 2012). Selain itu, 77% pembesaran udang vaname diantaranya diproduksi oleh negara-negara Asia termasuk Indonesia (Dahlan et al., 2017). Beberapa keunggulan udang vaname yaitu responsif terhadap pakan/nafsu makan yang tinggi, lebih tahan terhadap serangan penyakit dan kualitas lingkungan yang buruk, pertumbuhan lebih cepat, tingkat kelangsungan hidup tinggi dan waktu pemeliharaan yang relatif singkat yakni sekitar 90-100 hari per siklus (Purnamasari et al., 2017). Menurut Sumeru (2009) udang vaname memiliki tingkat produktivitas tinggi meskipun dengan padat tebar yang tinggi karena mampu memanfaatkan pakan dan ruang secara lebih efisien. Hal ini turut didukung dengan harga yang stabil dan tingginya permintaan pasar domestik maupun ekspor. Nilai ekonomi yang tinggi serta harga yang relatif stabil tersebut membuat peluang pasar budidaya bagi komoditas ini masih terbuka lebar, terutama untuk pasar ekspor (Triyanti & Hikmah, 2015). Menurut Bosman et al. (2021), budidaya udang vaname tergolong memberikan keuntungan secara ekonomi. Perkembangan budidaya udang yang semakin pesat menyebabkan pakan berperan vital dan menjadi variabel terbesar dalam biaya produksi yaitu mencapai 50-60% dari total biaya produksi yang dikeluarkan. Pakan merupakan salah satu unsur penting untuk menunjang pertumbuhan dan kelangsungan hidup dalam budidaya udang (Kurniawan et al., 2016). Pentingnya pengelolaan pakan menyebabkan pemberian pakan harus teratur, terjadwal dan tidak boleh telat karena dapat mempengaruhi perkembangan dan pertumbuhan udang vaname. Salah satu teknologi penting untuk produktivitas yaitu manajemen pemberian pakan (Yi et al., 2018). Automatic feeder bekerja menggunakan tenaga listrik dan dapat diatur sewaktu mengeluarkan pakan (Kordi, 2009). Pelempar pakan otomatis adalah untuk memberi kemudahan kepada petani tambak dalam memberi pakan udang secara efesien, tepat waktu dan terukur. Oleh karena itu, penelitian ini bertujuan untuk mengetahui efektivitas automatic feeder pada budidaya udang vaname. ABSTRACT Vaname shrimp farming is excellent in Indonesia because it has commercial value and provides income for the country. Feeding technology is one of the factors for the success of cultivation. Currently, IoT-based technology, namely automatic feeders, is a trend in vaname shrimp cultivation. This study aims to determine the effectiveness of using an automatic feeder in vaname cultivation. The research was conducted in March-May 2021 at PT. Windu Marina Abadi, East Lombok by measuring growth, survival rate, and FCR as well as analyzing water quality. The results showed that the lowest weight growth was in plots A5 (7.79 g) and the highest B5 (18.14 g) with the lowest average growth rate/day, namely A5 (0.09 g) and B5 (0.21 g), survival rate in ponds using an automatic feeder reaching 97%, the highest FCR was in ponds A5 (manual) (2.45) and the lowest was in ponds B5 (automatic feeder) (1.08). Water quality values were still in the normal range for cultivation. Ponds with automatic feeding were relatively better than ponds with manual feeding. Keywords: Vaname shrimp, automatic feeder, growth, technology Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 93 Buletin JSJ, 3 (2), 2021, 53-62 94 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Penebaran Benur dan Pengelolaan Pakan g Sebelum benur di tebar dilakukan pengukuran suhu, salinitas dan pH pada kantong benur di laboratorium serta melakukan proses penghitungan benur secara manual (pengambilan 2 sampel kantong benur) kemudian dilakukan proses aklimatisasi suhu dan salinitas yang berlangsung selama 15 menit. Melakukan pemberian pakan 4 kali sehari dengan cara blind feeding (saat umur pemeliharaan 1-25 hari) dan anco feeding (saat umur pemeliharaan 26 hari). Persiapan Air Media p Sterilisasi Tandon dengan Kalsium hipoklorit (Ca(ClO)2) sebanyak (150 kg) dengan dosis 37,5 mg/l, pengisian air tambak, sterilisasi tambak dengan menggunakan kupri sulfat (CuSO4) dosis 3,125 mg/l, Trichloroisocyanuric acid (C3Cl3N3O3) dosis 12,5 mg/l, kalsium hipoklorit (Ca(ClO)2) dengan dosis 7,5 mg/l, air netral setelah pemberian kupri sulfat selama 1 hari dan netral setelah pemberian kaporit C3Cl3N3O3 serta Ca(ClO)2 selama 2 hari. Persiapan Tambak p Persiapan tambak meliputi Pengeringan selama 7 hari dan pembersihan dari teritip dan sisa budidaya, persiapan sarana dan perbaikan prasarana, pensucihamaan tambak dengan mengisikan 30% air dan menambahkan Trichloroisocyanuric acid (C3Cl3N3O3) sebanyak (50 kg) dosis 12,5 mg/l dengan cara diaduk secara merata dan ditebar ke tambak. Penyediaan Air Media y Tahapan penyediaan air media meliputi :Pemberian Fermentasi 1 (mengaduk saponin dosis 6,25 mg/l, fermipan dosis 0,025 mg/l, serta aquazyme dosis 0,025 mg/l, setelah 2 hari, fermentasi siap ditebar di depan kincir tambak pemeliharaan; Pemberian Fermentasi 2 (mengaduk pakan dengan dosis 2,5 mg/l, dedak 2,5 mg/l, bungkil kulit kedelai 1,25 mg/l, aquazyme 0,025 mg/l dan fermipan 0,025 mg/l serta air tambak hingga merata, setelah 2 hari, fermentasi siap ditebar dengan menggunakan perahu rakit yang terdapat blong pakan; Pemberian probiotik dengan kandungan bakteri Bacillus sp., Pseudomonas sp., Nitrosomonas sp., Aerobacter sp., dan Nitrobacter sp. (0,625 mg/l); Penebaran kapur kaptan 5 mg/l; Pemberian Fermentasi 3 ( Mengaduk dedak 1,25 mg/l, molase 0,25 mg/l, fermipan 0,05 mg/l, aquazyme 0,025 mg/ dan bungkil kulit kedelai 0,25 mg/l serta air tambak hingga merata, setelah 2 hari, fermentasi siap ditebar di depan kincir tambak pemeliharaan. Monitoring dan Pengelolaan kualitas air Melakukan monitoring kualitas air meliputi suhu, salinitas, kecerahan, pH, DO, alkalinitas, hardnees, nitrit, TAN, TOM, Nitrit, phosphate, bakteri, plankton. Sedangkan untuk kegiatan pengelolaan kualitas air meliputi penambahan air, pengurangan air (penyiponan dan pengetapan), perbaikan kualitas air (pengangkatan klekap, aplikasi probiotik dan aplikasi kapur). Monitoring Pertumbuhan Dan Kesehatan Udang g g Sampling dilakukan pada saat udang berumur 41 hari dengan frekuensi 7 hari sekali menggunakan jala. Mengamati udang di anco, pengamatan hama sekitar tambak dan pengamatan penyakit pada hepatopancreas udang melalui mikroskop. BAHAN DAN METODE Penelitian ini dilaksanakan pada tanggal 02 Maret 2020 sampai dengan 15 Mei 2020, di PT. Windu Marina Abadi, Lombok Timur, Nusa Tenggara Barat. Alat yang digunakan adalah Refraktometer, pH pen, DO meter YSI 20 Pro, Secchi disk, Autofeeder, Mikroskop Olympus CX 23, Haemositometer, Countdown timer, Spektrofotometer MAPADA VD 1100. Bahan yang digunakan yaitu Sativa, probiotik dengan kandungan (Lactobacillus, Bacillus subtilis, Bacillus sp, Super Bacillus sp, Rhodobacter sp, dan Rhodococcus sp), saccharomyces, Kalsium silikat, Kalsium silikat dan Magnesium Klorida, Kapur, molase, Vit C, Vit B1, B2, B6 dan B12. Metode pengumpulan data yang digunakan dalam penelitian ini adalah pertumbuhan, kelangsungan hidup, dan FCR serta analisis kualitas air. Jenis pengumpulan data yang diambil meliputi data primer dan data sekunder. Data yang diperoleh dianalisis dengan tabel dan gambar statistik serta deskripsi. 94 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 94 Buletin JSJ, 3 (2), 2021, 93-99 95 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Tingkat Kelangsungan Hidup g g g p Tingkat kelangsungan hidup budidaya udang vaname disajikan pada Gambar 1, sebagai berikut. Keterangan: Tambak dengan pemberian pakan secara manual (Tambak A), Tambak dengan automatic feeder (Tambak B) 100% 80% 60% 40% 20% 0% A4 A5 A6 A7 A8 B4 B5 B6 B7 Tambak Kelangsungan Hidup (%) Keterangan: Tambak dengan pemberian pakan secara manual (Tambak A), Tambak dengan automatic feeder (Tambak B) Gambar 1. Tingkat kelangsungan hidup budidaya udang vaname. Gambar 1. Tingkat kelangsungan hidup budidaya udang vaname. Berdasarkan hasil tersebut menunjukan bahwa budidaya dengan menggunakan automatic feeder memberikan hasil yang terbaik yaitu mencapai 97% dibandingkan dengan pemberian pakan secara manual. Kelangsungan hidup udang vaname dipengaruhi oleh sistem imun, kualitas, air serta pakan (Erlangga et al., 2021). Automatic feeder memberikan pengaruh pada udang. Menurut Anggraeni et al. (2021) bahwa pakan yang dikonsumsi udang secara fisiologis akan dicerna dan digunakan untuk pertumbuhan somatik. HASIL DAN PEMBAHASAN Tingkat Kelangsungan Hidup Buletin JSJ, 3 (2), 2021, 53-62 Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index HASIL DAN PEMBAHASAN Panen dan Pasca Panen Panen dan Pasca Panen Pemanenan dilakukan dengan metode panen parsial dan panen total. Sortir ukuran dilakukan sebelum masuk pada proses Packing, udang yang telah ditimbang dengan rapi dimemasukkan ke dalam box mobil pembeli. 9 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 95 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 95 Buletin JSJ, 3 (2), 2021, 53-62 96 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Pertambahan Berat Harian Udang (Average Daily Growth/ADG) Pertambahan Berat Harian Udang (Average Daily Growth/ADG) Performa pertumbuhan harian udang yaitu tambak menggunakan automatic feeder memiliki ADG lebih tinggi dari tambak menggunakan manual, hal ini disebabkan pemberian pakan automatic feeder lebih efisien baik secara jumlah pemberian dan juga waktu pemberian. ADG udang yang diperoleh selama kegiatan penelitian disajikan pada Gambar 3. Gambar 3. ADG yang diperoleh perusahaan. Gambar 3. ADG yang diperoleh perusahaan. Adapun tambak unit B memiliki nilai ADG lebih baik dari unit A, hal ini karena unit B memiliki jangka waktu persiapan tambak yang lebih lama dan melakukan sterilisasi tambak yang lebih baik dari tambak unit A (A4, A5, A6 dan A7 tidak melakukan penebaran CuSO4) sehingga menyebabkan tambak pada unit tersebut terserang penyakit yaitu White Feces Disease/WFD. Tanda tambak tersebut terkena WFD yakni dengan munculnya berak putih di pinggir tambak seperti halnya dikemukakan oleh Nur’aini et al. (2019) yaitu gejala klinis udang yang terserang penyakit berak putih yaitu terdapat kotoran putih yang melayang di permukaan air tambak dan diikuti oleh penurunan nafsu makan. Menurut Khasani & Sopian (2013) bahwa salah satu indikator berfungsinya sistem metabolisme dalam tubuh adalah adanya pertumbuhan, baik itu pertambahan bobot maupun panjang. Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index Tidak tercapainya ABW sesuai target disebabkan karena pada tambak tersebut tidak melakukan pengeringan dan sterilisasi tambak dengan baik sehingga pada saat pemeliharaan udang vaname terkena penyakit White Feces Disease (WFD) dan Enterocytozoon hepatopenaei (EHP). Berat Rata-Rata Udang/Ekor (Average Body Weight/ABW) Berat Rata Rata Udang/Ekor (Average Body Weight/ABW) ABW pada tambak dengan pemberian pakan manual memiliki nilai ABW lebih rendah dibandingkan tambak yang menggunakan automatic feeder, hal ini disebabkan pemberian pakan menggunakan automatic feeder lebih baik karena pemberian pakan dilakukan sedikit demi sedikit namun dengan frekuensi yang lebih sering (tiap 5 menit sekali) sehingga pakan yang diberi dapat meminimalisir hilangnya nutrisi penting dalam pakan karena terlalu lama larut dalam air dan dapat dimanfaatkan oleh udang sebagai energi dan juga untuk bertumbuh. ABW udang yang diperoleh selama kegiatan penelitian berlangsung disajikan pada Gambar 2. Gambar 2. ABW yang diperoleh perusahaan. Gambar 2. ABW yang diperoleh perusahaan. 96 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 96 Buletin JSJ, 3 (2), 2021, 93-99 KESIMPULAN Penggunaan automatic feeder sangat efektif untuk diterapkan pada tambak budidaya udang dimana memberikan hasil yang baik dibandingkan dengan pemberian pakan secara manual. Kualitas Air Adapun nilai kualitas air pada budidaya udang vaname, disajikan pada Tabel 1, sebagai berikut. Tabel 1. Nilai kualitas air selama budidaya No Paramater Nilai Kualitas Air 1 Suhu (ºC) 25 - 32 2 Salinitas (ppt) 25 – 30 3 Oksigen terlarut (mg/L) > 4,3 – 7,5 4 pH 7,5 – 8,5 5 Amonia (mg/L) < 0,05 6 Nitrit (mg/L) 0,05 – 0,1 Berdasarkan hasil pengukuran kualitas air dengan nilai parameter suhu menunjukan kisaran normal, hal ini sama dengan penelitian Hukom et al. (2020) di tambak Sidoarjo, dan Karawang (Bosman et al., 2021). Nilai salinitas memiliki kisaran normal untuk budidaya udang. Nilai oksigen terlarut menunjukkan kisaran normal. Kelarutan oksigen di dalam air berpengaruh pada fungsi biologis dan pertumbuhan (Mallya, 2007). Nilai pH masih berada dalam nilai normal, hal ini masih di dalam ambang batas Keputusan Menteri Kelautan dan Perikanan No. 28 tahun 2004. Demikian juga, nilai amonia dan nitrit yang masih dalam ambang batas dengan nilai optimal untuk budidaya. Nitrit memiliki peran yang penting untuk menentukan kualitas air karena bersifat racun (Effendi, 2003). Amri, K., & Kanna, I. (2008). Budi Daya Udang Vaname. Gramedia Pustaka Utama. Anggraeni, F., Anggraeni, F., Malini, D. M., & Imron, I. (2021). Performa reproduksi dan pertumbuhan udang galah betina Macrobrachium rosenbergii setelah pemberian hormon medroxy progesteron acetat melalui pakan. Jurnal Riset Akuakultur, 16(2), 83. https://doi.org/10.15578/jra.16.2.2021.83-91 98 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Feed Convertion Ratio/FCR Nilai FCR tertinggi yaitu pada tambak A5 (manual) sebanyak 2,45 sedangkan FCR terendah yaitu pada tambak B6 (automatic feeder) sebanyak 1,08. Semakin rendah nilai FCR dan semakin tinggi penambahan bobot udang maka pertumbuhan udang semakin bagus (Dahlan y p ( ) y , semakin tinggi penambahan bobot udang maka pertumbuhan udang semakin bagus (Dahlan et al., 2017). Tingginya nilai FCR disebabkan oleh presentase pakan yang diberikan antara pukul 07.00, 11.00, 15.00 dan 20.00 sama rata. Sebaiknya persentase pemberian pakan pada pukul 20.00 dikurangi karena pada malam hari nilai oksigen terlarut ada pada titik terendah sehingga pada saat pemberian pakan, protein pada pakan yang diberikan banyak digunakan untuk menjadi energi dan bukan untuk bertumbuh. Pengecekan anco juga dilakukan terlalu cepat yaitu 1 jam setelah pemberian pakan dan pengecekan hanya dilakukan pada 2 anco, pengeringan pada tambak terlalu cepat yaitu tambak unit A hanya melakukan pengeringan selama 7 hari karena panen pada Bulan Januari dibandingkan tambak unit B yang melakukan panen pada Bulan Desember, kurangnya dosis sterilisasi dan probiotik menyebabkan kualitas air yang belum optimal sehingga penyakit mudah muncul seperti berak putih White Feces Disease (WFD) dan Enterocytozoon heparopenaei (EHP). Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 97 Buletin JSJ, 3 (2), 2021, 53-62 Buletin JSJ, 3 (2), 2021, 53-62 Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index Nilai FCR pada masing- masing tambak disajikan pada Gambar 4. Gambar 4. Nilai FCR tiap tambak. Gambar 4. Nilai FCR tiap tambak. DAFTAR PUSTAKA Anggraeni, F., Anggraeni, F., Malini, D. M., & Imron, I. (2021). Performa reproduksi dan pertumbuhan udang galah betina Macrobrachium rosenbergii setelah pemberian hormon medroxy progesteron acetat melalui pakan. Jurnal Riset Akuakultur, 16(2), 83. https://doi.org/10.15578/jra.16.2.2021.83-91 98 Buletin JSJ, 3 (2), 2021, 93-99 Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index Available online di: http://ejournal-balitbang.kkp.go.id/index.php/JSJ/index Bosman, O., Soesilo, T. E. B., & Rahardjo, S. (2021). Pollution index and economic value of vannamei shrimp (Litopenaeus vannamei) farming in Indonesia. Indonesian Aquaculture Journal, 16(1), 51–60. https://doi.org/10.15578/IAJ.16.1.2021.51-60 Dahlan, J., Sopiyanudin., & Hariyanto. (2017). Pertumbuhan udang vaname (Litopenaeus vannamei) yang dikultur pada sistem bioflok dengan penambahan probiotik. Jurnal Sains dan Inovasi Perikanan. 1 (1):19-27. Effendi, H. (2003). Telaah Kualitas Air Bagi Pengelolaan Sumber Daya dan Lingkungan Perairan. Yogyakarta: Kanisius. Erlangga, E., Nuraini, C., & Salamah, S. (2021). Pengaruh sumber karbon yang berbeda untuk pembentukan flok dan efeknya pada pertumbuhan dan sintasan udang vaname, Litopenaeus vannamei. Jurnal Riset Akuakultur, 16(2),107-115. https://doi.org/10.15578/jra.16.2.2021.107-115 Hukom, V., Nielsen, R., Asmild, M., & Nielsen, M. (2020). Do Aquaculture Farmers Have an Incentive to Maintain Good Water Quality? The Case of Small-Scale Shrimp Farming in Indonesia. Ecological Economics, 176(August 2019), 106717. https://doi.org/10.1016/j.ecolecon.2020.106717 Kharisma, A., & Manan, A. (2012). Kelimpahan bakteri vibrio sp. pada air pembesaran udang vannamei (Litopenaeus vannamei) sebagai deteksi dini serangan penyakit vibriosis. Jurnal Ilmiah Perikanan dan Kelautan, 4(2), 128–134. Khasani, I. & Sopian, A. (2013). Pertumbuhan dan sintasan benih udang galah (Macrobrachium rosenbergii de Man) pada pendederan berbasis sistem heterotrof dengan padat tebar berbeda. Jurnal Riset Akuakultur, 8(3), 373-382. Kordi, M. G. H. (2009). Budi Daya Perairan. Edisi Kedua. Penerbit PT. Citra Aditya Bakti. Kurniawan, L.A., Arief, M., Manan, A., & Nindarwi, D.D. (2016). Pengaruh pemberian probiotik berbeda pada pakan terhadap retensi protein dan retensi lemak udang vaname (Litopenaeus vannamei). Journal of Aquaculture and Fish Health, 6(1). ( p ) q ( ) Mallya, Y. J. (2007). The Effect of Dissolved Oxygen on Fish Growth in Aquaculture. 30. Mallya, Y. J. (2007). The Effect of Dissolved Oxygen on Fish Growth in Aquaculture. 30. Menteri Kelautan dan Perikanan No. 28. (2004). Keputusan Menteri Kelautan dan Perikanan Republik Indonesia Nomor KEP.28/MEN/2004 tentang Pedoman Umum Budidaya Udang di Tambak (p. 26 p.). p. 26 p. Nur’aini, Y.L., Fatmawati, Hanggono, B., & Faries, A. (2019). Penanggulangan penyakit berak putih pada udang vaname (Litopenaeus vannamei). Jurnal Perekayasaan Budidaya Air Payau dan Laut No. 14, Balai Perikanan Budidaya Air Payau Situbondo. Pantjara, B., Syafaat, M. N., & Kristanto, A. H. (2015). Effect of dynamical water quality on shrimp culture in the Integrated Multitropic Aquaculture (IMTA). Indonesian Aquaculture Journal, 10(1), 81. https://doi.org/10.15578/iaj.10.1.2015.81-90. Purnamasari, I., Purnawa, D., dan Utami, M. A. F. 2017. Pertumbuhan udang vaname (Litopenaeus vannamei) di tambak intensif. Buletin JSJ, 3 (2), 2021, 93-99 Jurnal Enggano 2(1). Bengkulu. Sumeru, S. (2009). Pakan Udang. Kanisius. Yogyakarta. Triyanti, R., & Hikmah, H. (2015). Analisis kelayakan usaha budidaya udang dan bandeng: studi kasus di Kecamatan Pasekan Kabupaten Indramayu. Buletin Ilmiah Marina Sosial Ekonomi Kelautan dan Perikanan, 1(1), 1–10. Yi, D., Reardon, T., & Stringer, R. (2018). Shrimp aquaculture technology change in Indonesia: Are small farmers included? Aquaculture, 493, 436–445. https://doi.org/10.1016/j.aquaculture.2016.11.003 Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam Copyright @ 2021, Buletin Jalanidhitah Sarva Jivitam 99
https://openalex.org/W4248534745	https://discovery.ucl.ac.uk/10113592/7/Gilbert_1747021820970156.pdf	English	null	Trait anxiety does not correlate with metacognitive confidence or reminder usage in a delayed intentions task	null	2,020	cc-by	8,660	Original Article https://doi.org/10.1177/1747021820970156 Quarterly Journal of Experimental Psychology 1–11 © Experimental Psychology Society 2020 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/1747021820970156 qjep.sagepub.com Memory; metacognition; anxiety Received: 12 May 2020; revised: 26 August 2020; accepted: 2 October 2020 judgement of their own abilities (metacognitive bias) is related to how optimal these decisions are (see Gilbert et al., 2020). For instance, individuals who are underconfi- dent in their ability will tend to set reminders for delayed intentions which they would have remembered anyway using their own memory. However, metacognitive error cannot explain suboptimal reminder-setting in full. For example, excessive reminder-setting can be observed both in the presence of under- and overconfidence (see Gilbert et al., 2020, Experiment 3). Thus, although confidence may guide one’s decisions to set reminders, there remains a gap in the literature as to what other factors contribute to individuals’ tendency to offload delayed intentions. An 1177_1747021820 1177_1747021820 Peter A Kirk1,2 , Oliver J Robinson1 and Sam J Gilbert1 Peter A Kirk1,2 , Oliver J Robinson1 and Sam J Gilbert1 Abstract Setting external reminders provides a convenient way to reduce cognitive demand and ensure accurate retrieval of information for prospective tasks. Recent experimental evidence has demonstrated that the decision to offload cognitive information to external resources is guided by metacognitive belief, that is, individuals’ confidence in their unaided ability. Other work has also suggested a relationship between metacognitive belief and trait anxiety. In the present study (N = 300), we bridged these two areas by investigating whether trait anxiety correlated with metacognitive belief and—consequently—propensity to offload information in a delayed intentions paradigm. Participants received a financial reward based on their ability to remember targets. However, participants could take a reduced reward per target if they decided to use reminders. We replicated previous findings that participants were biased to use more reminders than would be optimal, and this bias was correlated with metacognitive judgements. However, we show no evidence that trait anxiety held a relationship with metacognitive belief or reminder usage. Indeed, Bayesian analyses strongly favoured the null. Therefore, variation in self-reported trait anxiety does not necessarily influence confidence and strategy when participants remember delayed intentions. 1Institute of Cognitive Neuroscience, University College London, London, UK 2Department of Experimental Psychology, University College London, London, UK Corresponding author: Peter A Kirk, Institute of Cognitive Neuroscience, University College London, London WC1N 3AZ, UK. Email: p.kirk@ucl.ac.uk 1Institute of Cognitive Neuroscience, University College London, London, UK 2Department of Experimental Psychology, University College London, London, UK Introduction As working memory capacity is constrained (Cowan, 2010), it is common for humans to offload relevant infor- mation onto the external environment for prospective tasks (Gilbert, 2015). This is an example of “cognitive offloading” (Finley et al., 2018; Risko & Gilbert, 2016). For example, during a lecture, you may note down ques- tions to ask at the end, or you may create an entry in your calendar to remember your friend’s birthday. However, offloading behaviours vary greatly, and it is intrinsically linked to the way many interact daily with their physical environment (e.g., leaving keys by the door) and techno- logical resources (e.g., smartphone alarms). The use of reminders reduces cognitive demand and improves mem- ory task performance (Gilbert, 2015; Hu et al., 2019; Risko & Dunn, 2015). Recent studies have begun to investigate how individu- als decide whether or not to use reminders. Evidence sug- gests that individual differences in cognitive confidence may guide one’s decision to offload information (Boldt & Gilbert, 2019). Moreover, the accuracy of individuals’ Corresponding author: Peter A Kirk, Institute of Cognitive Neuroscience, University College London, London WC1N 3AZ, UK. Email: p.kirk@ucl.ac.uk Quarterly Journal of Experimental Psychology 00(0) 2 00(0) unexplored avenue is whether affective processes may influence these decisions. metacognition, this does not rule out that anxiety could influence reminder-setting in other ways. In addition to affecting cognitive confidence, anxiety might more directly influence reminder-setting. Some have proposed that anxi- ety may be detrimental to behaviour as worrying thoughts (a primary component of anxiety) use up memory resources, and thus capacity (Eysenck et al., 2007; Seibert & Ellis, 1991), suggesting a potentially greater need for reminders. Recent research has demonstrated however that the increased cognitive load caused by anxiety does not always explain effects on cognition (e.g., time perception, see Sarigiannidis et al., 2020). Attentional control theory (Eysenck et al., 2007) also suggests that anxiety influences behaviour by shifting attentional resources in search of threatening stimuli, reducing attention to the present task (unless it involves threatening stimuli). Irrespective of dimensional underpinnings, no prior research has yet estab- lished whether a correlation exists between anxiety and off- loading behaviour. We therefore conducted a preregistered (https://osf.io/zguhj/) study to explore the relationship between anxiety, metacognition, and reminder-setting. Anxiety is one emotional experience known for its influ- ence on non-affective cognition. Introduction There are a myriad of studies providing evidence that anxiety affects many cogni- tive domains, manifesting in altered behaviours (for a review, see Robinson et al., 2013). In tasks requiring mem- ory for delayed intentions, a mixed picture has emerged, with some studies finding a relationship between anxiety and task performance (Arnold et al., 2015; Kliegel & Jäger, 2006), but not others (Cuttler & Graf, 2008). However, irre- spective of memory performance, no previous study has tested the relationship between anxiety and propensity to set reminders in an experimental task. Prior research has established a link between anxiety and metacognitive belief, such that individuals experiencing higher anxiety often have lower cognitive confidence (Spada et al., 2010; Wells & Cartwright-Hatton, 2004). In the context of delayed intentions, it would then be expected that increased anxiety might increase propensity to use reminders. The correlations observed between metacognition and anxiety largely rely on the Metacognitions Questionnaire 30 (MCQ-30; Wells & Cartwright-Hatton, 2004), which includes asking participants to self-rate how misleading and poor their memory is. Although this may be a clini- cally useful construct (Meyer et al., 1990), this non-spe- cific self-report does not necessarily relate to actual memory ability. An alternative approach is to ask partici- pants to judge their performance level on a specific task, based on a measure such as percent accuracy (e.g., Gilbert et al., 2020). This can then be compared against the actual accuracy level to yield a measure of metacognitive bias, that is, the difference between participants’ subjective rat- ing of accuracy and their actual performance level. To our knowledge, no study has linked global measures of anxiety to task-specific metacognitive bias in a memory task. Consequently, there is no indication as to whether anxious individuals’ lowered confidence is related to actual mem- ory ability and/or whether they engage in suboptimal reminder-setting behaviours. Given the frequency at which we assess our internal memory abilities, and consequently rely on offloading behaviours, significant underconfidence could potentially lead to unhealthy oversetting of remind- ers. Being aware of maladaptive cognitions and behaviour is in fact central to the first-line psychological treatment for anxiety: cognitive-behavioural therapy (Kovacs & Beck, 1978). Thus, delineating whether anxiety results in metacognitive bias and oversetting of reminders presents directly translatable information for clinicians. Introduction For exam- ple, this could provide insight as to whether clinicians should provide additional metacognitive assessments and interventions in the domain of prospective memory for patients with anxiety disorders (Wells, 2009). Our experimental task required participants to remem- ber delayed intentions over a brief time period. This could be considered to involve “prospective memory,” an umbrella term referring to situations where individu- als must remember to perform actions in the future (Brandimonte et al., 1996; Kliegel et al., 2008; Scullin et al., 2015). However, some authors have argued that the term prospective memory is more appropriate for memory tasks involving a longer retention interval (see Graf & Uttl, 2001, for discussion). We note that a task similar to the present one has previously been shown to predict participants’ prospective memory performance over longer retention intervals of up to a week (Gilbert, 2015). Nevertheless, to avoid confusion, we use the more theoretically neutral phrase “delayed intentions task” rather than “prospective memory task” below. Preregistration All hypotheses, experimental methods, and planned analy- ses were preregistered prior to data collection. Materials and source code are uploaded on the Open Science Foundation (https://osf.io/zguhj/). We note the following deviations from our planned analyses: •• We originally planned to conduct frequentist tests for all our analyses. However, we supplemented these with Bayesian equivalents to aid in the inter- pretation of our findings. •• We have provided post hoc, supplementary analy- ses which help address concerns such as counterbal- ancing, screening procedure, and exclusion criteria (Supplementary Materials 5–9). Offloading task. We used a modified version of a previ- ously verified delayed intentions task (Gilbert et al., 2020). Participants chose between remembering intentions using their own memory (in which case they earned full reward for each remembered item) or using external reminders (in which case they earned a smaller reward, which varied from trial to trial). This allowed us to examine not only participants’ propensity to use reminders but also the opti- mality of their reminder-setting strategy. For example, suppose a participant can achieve 55% accuracy using their own memory and 100% accuracy using reminders. Given a choice between 10 points per item (using own memory) or 5 points per item (using reminders), the opti- mal strategy is to use one’s own memory. However, if offered 6 points per item using reminders, it is optimal to use reminders. In this way, we could compare participants’ reminder-setting strategy with the optimal strategy, to investigate the extent to which they were biased towards using external reminders versus their own memory. Hypotheses The main aim of the present study was to test the following six preregistered hypotheses in relation to a delayed inten- tions task: 1. Trait anxiety will negatively correlate with partici- pants’ accuracy. 2. Trait anxiety will negatively correlate with partici- pants’ confidence. 3. Trait anxiety will negatively correlate with partici- pants’ metacognitive bias (trait anxiety results in greater underconfidence or less overconfidence). 4. Trait anxiety will positively correlate with partici- pants’ propensity to use reminders. 5. Trait anxiety will positively correlate with partici- pants’ bias towards the use of reminders (trait Although the literature implies a possible indirect rela- tionship between anxiety and reminder usage mediated via Kirk et al. 3 anxiety results in increased overuse of reminders or reduced underuse of reminders). minimum = 21; maximum = 72); 190 reported their gender as male, 108 as female, and 2 as other. anxiety results in increased overuse of reminders or reduced underuse of reminders). 6. Both anxiety and metacognitive bias will account for unique variance in a model of reminder bias. Design In sum, this study aimed to assess the link between anxiety and metacognitive bias in memory for delayed intentions. Anxiety/worry measures. Trait anxiety was measured using the trait section of the State Trait Anxiety Inventory (STAI; Spielberger, 1983), a 20-item questionnaire which pro- vided us a global measure of individual differences in anxiety. This was the key measure of anxiety. We selected the trait section as this is a temporally stable attribute pre- viously shown to correlate with cognitive confidence (Wells & Cartwright-Hatton, 2004). However, we also report that the trait section appears to already correlate very highly with the state section of the questionnaire, r(1058) = .83, p < .0001, BF10 > 1,000; Supplementary Material 7. As an additional measure for exploratory fol- low-up analyses, we also included the penn state worry questionnaire (PSQW) (Meyer et al., 1990), a 16-item questionnaire which provided us a metric of individual dif- ferences in worrying thoughts. Participants For instance, if a target circle (e.g., 7) initially appeared as orange, par- ticipants needed to remember this instruction while they Quarterly Journal of Experimental Psychology 00(0) 4 Schematic of an example trial: (1) Participants dragged “”circles in sequential order to the bottom edge of a box. Every e was removed from the box, it was replaced with a new one; (2) sometimes, new target circles appeared in a different nifying a delayed intention to later drag the target to a coloured edge; (3) the colour faded after 2 s. (4) participants etimes permitted to set reminders by immediately dragging target circles to the instructed edge when they appeared; (5) s continued to drag the next circles in the sequence to the bottom of the box; and (6) after dragging the appropriate he sequence, they could then execute the delayed intention to drag the target to circle to the correct edge. Figure 1. Schematic of an example trial: (1) Participants dragged “”circles in sequential order to the bottom edge of a box. Every time a circle was removed from the box, it was replaced with a new one; (2) sometimes, new target circles appeared in a different colour, signifying a delayed intention to later drag the target to a coloured edge; (3) the colour faded after 2 s. (4) participants were sometimes permitted to set reminders by immediately dragging target circles to the instructed edge when they appeared; (5) participants continued to drag the next circles in the sequence to the bottom of the box; and (6) after dragging the appropriate circles in the sequence, they could then execute the delayed intention to drag the target to circle to the correct edge. The main experimental paradigm alternated between different trial types. On some trials, participants were forced to either use their own memory or external remind- ers. We refer to these as forced internal and forced exter- nal, respectively. On other trials, participants were given a choice between using their own memory and earning 10 points per remembered item or using reminders and earn- ing a smaller number of points between 2 and 8 (free choice). This allowed us to calculate the optimal strategy (based on accuracy on the forced-internal/external trials) and compare this against actual reminder-setting strategy (based on the free-choice trials). Participants Participants were recruited from the Amazon Mechanical Turk website (http://www.mturk.com), an online market- place in which participants receive payment for comple- tion of web-based tasks (Crump et al., 2013). Ethical approval was received from the UCL Research Ethics Committee (1584/003) and participants provided informed consent before participating in the study. A statistical power analysis was performed with G*Power 3.1 for sample size estimation. No prior research has directly investigated the role of affective systems on reminder-setting. We therefore decided to power our study based on our previous research on metacognitive bias and offloading (Gilbert et al., 2020, Experiment 2, metacogni- tive bias-reminder bias correlation [unadvised group]), which found a correlation of r = –.31. We chose a sample size of 300, which would give us sufficient power to detect a ~50% reduction of this effect (r = .161, α = .05, 1 – β = .8). As in earlier studies (Gilbert, 2015), participation was restricted to volunteers aged at least 18 years, located in the United States. Participants who had already taken part in the present study were blocked to ensure a fresh sample of participants. We also restricted inclusion to participants with a minimum of 90% Mechanical Turk approval rate. Participation took approximately 45 min, for which par- ticipants received a base payment of $2, plus an additional bonus payment of up to $8.67 (see below). Our final sam- ple (N = 300) had a mean age of 37.81 years (SD = 10.97; During each trial, participants used their computer mouse to drag 25 numbered circles in sequential order (1–25) to the bottom edge of a box (Figure 1). Six yellow circles were shown on the screen at once, and each time a circle was removed from the box, it was replaced with a new one (e.g., after dragging “1” to the bottom, a new cir- cle labelled “7” appeared in its place). Sometimes, new circles initially appeared in blue, orange, or purple, which corresponded to the colours of the left, top, and right edges of the box (which was displayed throughout all trials). These circles then faded to yellow after 2 s. This consti- tuted an instruction for a delayed intention to drag these “target” circles to a different edge of the box. Participants Prior to starting the main experiment, participants were instructed and prac- tised the task with and without target circles (forced internal). They were not allowed to continue until they completed these practice trials successfully. Subsequently, they practised again on the forced-internal trial type, after dragged circles 2 to 6 to the bottom of the box (by which time the target circle had faded to yellow). When they fin- ished this, they could then execute the intention to drag 7 to the top. Across an entire trial of 25 circles, 10 target circles were presented. This meant that participants had to encode multiple intentions and were unlikely to remember all of them if they relied on their internal memory ability. Alternatively, if participants set reminders, they could offload the intentions by immediately dragging target cir- cles to the instructed edge when they appeared (e.g., drag- ging an orange 7 next to the top edge of the box; this could be done immediately upon its presentation rather than hav- ing to wait for it to fade to yellow first). Its location then acted as a reminder when the participant reached this num- ber in the sequence, analogous to leaving an object by the front door so that you remember it when leaving the house tomorrow. dragged circles 2 to 6 to the bottom of the box (by which time the target circle had faded to yellow). When they fin- ished this, they could then execute the intention to drag 7 to the top. Across an entire trial of 25 circles, 10 target circles were presented. This meant that participants had to encode multiple intentions and were unlikely to remember all of them if they relied on their internal memory ability. Alternatively, if participants set reminders, they could offload the intentions by immediately dragging target cir- cles to the instructed edge when they appeared (e.g., drag- ging an orange 7 next to the top edge of the box; this could be done immediately upon its presentation rather than hav- ing to wait for it to fade to yellow first). Its location then acted as a reminder when the participant reached this num- ber in the sequence, analogous to leaving an object by the front door so that you remember it when leaving the house tomorrow. Kirk et al. 5 Figure 2. Metacognitive confidence rating screen. Participants After completing a series of practice trials, participants rated what percentage of target circles they believed they could accurately remember. These judgements were collected once at the beginning of the experiment, and separately for the internal and external strategies. For the confidence measure following the forced-external practice, participants were instead prompted with “Now that you have practiced doing the task using reminders, we would like you to tell us how accurately you can perform the task when you use this strategy.” This provided us with our metacognitive confidence measure. Figure 2. Metacognitive confidence rating screen. After completing a series of practice trials, participants rated what percentage of target circles they believed they could accurately remember. These judgements were collected once at the beginning of the experiment, and separately for the internal and external strategies. For the confidence measure following the forced-external practice, participants were instead prompted with “Now that you have practiced doing the task using reminders, we would like you to tell us how accurately you can perform the task when you use this strategy.” This provided us with our metacognitive confidence measure. Figure 3. Example instructions for the free-choice trials. Before beginning a free-choice trial, participants were given the options to use their internal memory or reminders. (a) In the gain condition, if participants selected to use their internal memory, they would score 10 points per target circle remembered. However, if participants chose to use reminders, they would earn a smaller number of points per target circle remembered (2–8). (b) In the loss condition, participants were given points at the beginning of the block and lost either 0 points per target circle remembered if they used internal memory or a greater number of points (2–8) if they chose to use reminders. Every trial (forced internal, forced external, free choice) across both conditions constituted a 25 circle sequence, of which 10 were target circles. In terms of reward, the two conditions are mathematically equivalent. which they provided a measure of how confident they were at their ability to perform the task (Figure 2). Following this, they were made aware of the ability to use reminders in the task. They then practised again, but on the forced-external trial type, after which they pro- vided a measure of how confident they were at their abil- ity to perform the task using reminders. Participants This enabled us to investigate whether participants’ reminder-setting strategy was related to their metacognitive beliefs about their ability to perform the task. The main experiment was split into two counterbal- anced blocks: gain and loss. In the gain condition, partici- pants chose between receiving 10 points for each remembered target circle or a smaller number of points (2–8) to use reminders (Figure 3a), as described above. This matches the version of the task used in previous experiments (Gilbert et al., 2020), and was the focus of analyses for the present study. The loss condition (Figure 3b) was included so that it could be compared with the gain condition as part of a separate project (https://osf. io/8zvf6/), which will be reported in a separate article. During the loss condition, participants received points before the beginning of the block. They were then pre- sented with the choice between (1) using their own mem- ory and keeping all their points (losing 0) each time they correctly remembered target circles, or (2) using remind- ers, and losing points every time they remembered (2–8). In terms of reward, the two conditions are mathematically equivalent. Figure 3. Example instructions for the free-choice trials. Before beginning a free-choice trial, participants were given the options to use their internal memory or reminders. (a) In the gain condition, if participants selected to use their internal memory, they would score 10 points per target circle remembered. However, if participants chose to use reminders, they would earn a smaller number of points per target circle remembered (2–8). (b) In the loss condition, participants were given points at the beginning of the block and lost either 0 points per target circle remembered if they used internal memory or a greater number of points (2–8) if they chose to use reminders. Every trial (forced internal, forced external, free choice) across both conditions constituted a 25 circle sequence, of which 10 were target circles. In terms of reward, the two conditions are mathematically equivalent. For a demonstration, the entire experiment can be accessed here: http://ucl.ac.uk/sam-gilbert/demos/CWPK1/ start.html Apparatus. Participants completed the task via their com- puter’s web browser. Participation was only to be permit- ted if the browser window had dimensions of at least 500 × 500 pixels. Procedure 1. Informed consent. 1. Informed consent. 2. Practice trials and metacognitive judgements. 3. Experimental instructions for gain or loss condi- tion (randomised). •• Actual indifference point (AIP). This is the esti- mated point at which participants were actually indifferent to the two strategy options. As in Gilbert et al. (2020), this was calculated by fitting a sig- moid curve to the strategy choices (0 = own mem- ory; 1 = reminders) across the seven target values (2–8), using the R package “quickpsy” (Linares & López-Moliner, 2016) bounded to the range 2 to 8. 4. Experimental Block 1 (gain or loss). Participants performed a total of 13 trials. On odd-numbered trials, participants were given a free choice between using internal memory (10 points per target circle) or reminders (2–8 points per target circle, pre- sented in random order). On even-numbered trials, participants alternated between the forced-external and forced-internal trials, with the starting trial type (external or internal) randomised between participants, and counterbalanced between gain/ loss conditions. •• Reminder bias. This is defined as OIP – AIP, which will yield a positive value for a participant biased towards using more reminders than would be optimal, and a negative value for a participant biased towards using fewer reminders than would be optimal. 5. Experimental instructions for the other condition (gain or loss). •• Internal metacognitive bias. This is the difference between metacognitive confidence and actual accu- racy on forced-internal trials. A positive number would indicate overconfidence of their own mem- ory abilities. 6. Experimental Block 2 (gain or loss; 13 trials as above). 7. Questionnaires (fixed order: STAI, PSWQ). •• External metacognitive bias. This is the difference between metacognitive confidence and actual accu- racy on forced-external trials. A positive number would indicate overconfidence of their performance when using reminders. Reward. Participants were told that they were scoring points, where 300 points was equivalent to $1. They received 600 points at the beginning of the experiment. Then they were able to earn (or keep) up to 1,300 points (i.e., 100 points per trial) in each half of the experimental trials. Therefore, the earnings could range between 600 points ($2) and 3,200 points ($10.67). The experiment was advertised as having a base payment of $2, which participants received simply for taking part, with the addi- tional earnings sent to participants afterwards as a bonus payment. Reward. Participants The square box containing the circles was sized at 80% of the horizontal or vertical extent of the In addition, the full source code to run the experiment, including all implementation details and instructions, has been uploaded to OSF (https://osf.io/sm3tw/). Quarterly Journal of Experimental Psychology 00(0) 00(0) 6 ACCFE). As in Gilbert et al. (2020), this was calcu- lated as browser window, whichever was smaller. Each circle had a radius of 5.5% of the width/height of the box, and all cir- cles were initially placed so that they fall within a central portion of the box with dimensions sized at 56% of the total width/height, so that no circles were adjacent to any of the edges of the box at the beginning of the trial. OIP 1 ACC ACC FI FE = ( ) / 0× If the OIP was less than 2 or greater than 8, it was set to the relevant lower or upper bound. This was so that the potential values of the OIP would match the potential val- ues of the point at which they were actually indifferent, which was bound by their choices for values 2 to 8. Procedure Procedure Participants were told that they were scoring points, where 300 points was equivalent to $1. They received 600 points at the beginning of the experiment. Each of the previous seven measures was calculated sepa- rately for the gain and loss conditions. •• Internal metacognitive confidence. This is the response made to the metacognitive accuracy pre- diction following practice trials using internal memory (see Figure 1). Analysis Background analyses The following analyses were performed to characterise the basic performance of the experimental task and check whether previous findings were replicated. These analyses do not test any particular hypotheses for the present study but provide further information that may be useful for interpretation of the hypothesis-testing analyses. We have also provided descriptive plots for the distribution of anxi- ety scores (Figure 4a) and metacognitive/reminder meas- ures (Supplementary Material 6). Statistical tests. All Frequentist analyses were run in RStudio (RStudio Team, 2019). These constituted a series of two-tailed paired-sample t tests, one-sample t tests, Pear- son’s correlations, and linear regressions (enter method). As described in the original pre-registration, our analyses were predominantly restricted to the gain condition only, as this condition more closely replicated the procedure used in previous studies. A comprehensive analysis of the loss condition will be reported separately (see https:// osf.io/8zvf6/). We have also supplemented our analyses by providing their Bayesian equivalents in JASP (JASP Team, 2019). Here, we used JASP’s default priors: Bayes- ian paired-sample t test (Cauchy scale = .707); Bayesian one-sample t test (Cauchy scale = .707); Bayesian corre- lation pairs (stretched beta prior width = 1); and Bayesian linear regression (Jeffreys-Zellner-Siow r scale = .354; beta binomial a = 1, b = 1). All Bayes Factors are reported as BF10 and winning models in the linear regressions were defined as those with the highest BF10 relative to the null (intercept only model). The relative predictability of mod- els in the linear regressions was calculated by dividing BF10 between models. To aid in interpretation of Bayes factors, we have used commonly adopted semantic labels for Bayes Factors when describing our findings (anecdo- tal [1–3], substantial [3–10], strong [10–30], very strong [30–100], decisive [>100]; Jeffreys, 1998). We characterised basic performance of the task by com- paring accuracy between the forced-internal and forced- external conditions. Here, we saw decisive evidence for a difference between the two conditions, t(332.03) = –25.3, p < .0001, BF10 > 100, BF01 < .01, wherein accuracy was lower for the forced-internal (M = 64.89%, SD = 21.80) ver- sus forced-external condition (M = 97.60%, SD = 5.13). We then tested for internal metacognitive bias (defined as pre- dicted internal accuracy minus actual accuracy in the forced-internal condition). Analysis •• External metacognitive confidence. This is the response made to the metacognitive accuracy pre- diction following practice trials using reminders. Dependent measures •• Forced-internal accuracy (ACCFI). This is the mean target accuracy (proportion of target circles cor- rectly dragged to the instructed location) on forced- internal trials. Exclusion criteria. Participants were excluded if (1) their accuracy in the forced-internal condition was lower than 10%, averaged across the gain and loss conditions; (2) accuracy in the forced-external condition was lower than 70%, averaged across the gain and loss conditions; (3) accuracy on the forced-internal trials was higher than forced-external trials in either condition; (4) there was a negative point biserial correlation between points offered for correct responses on each trial using reminders (2–8) and choice of strategy (0 = own memory, 1 = reminders; this excludes participants who were more likely to set •• Forced-external accuracy (ACCFE). This is the mean target accuracy (proportion of target circles correctly dragged to the instructed location) on forced-external trials. •• Optimal indifference point (OIP). This is the value for target circles offered with reminders at which an unbiased individual should be indifferent between the two options, based on the accuracy in the forced- internal and forced-external trials (ACCFI and Kirk et al. 7 Figure 4. (a) Distribution of trait anxiety scores. Dashed line refers to US “Working Adult” average (Spielberger, 1983). Plotted using RainCloudPlots (Allen et al., 2019). (b) Correlation coefficient slope with 95% confidence interval between metacognitive bias and reminder bias. Positive reminder bias scores indicate overuse of reminders and negative metacognitive bias scores indicate underconfidence of memory. Figure 4. (a) Distribution of trait anxiety scores. Dashed line refers to US “Working Adult” average (Spielberger, 1983). Plotted using RainCloudPlots (Allen et al., 2019). (b) Correlation coefficient slope with 95% confidence interval between metacognitive bias and reminder bias. Positive reminder bias scores indicate overuse of reminders and negative metacognitive bias scores indicate underconfidence of memory. reminders when it earned them fewer points, suggest- ing random strategy selection); (5) reminder bias score (averaged across the gain and loss conditions) exceeded 3 median absolute deviation units (MAD; Leys et al., 2013); (6) difference in reminder bias scores between the two conditions exceeded 3 MAD units; and (7) internal meta- cognitive bias score exceeded 3 MAD units. Data collec- tion continued until the study had the appropriate power (N = 300) following exclusion (64 excluded). Analysis reminders when it earned them fewer points, suggest- ing random strategy selection); (5) reminder bias score (averaged across the gain and loss conditions) exceeded 3 median absolute deviation units (MAD; Leys et al., 2013); (6) difference in reminder bias scores between the two conditions exceeded 3 MAD units; and (7) internal meta- cognitive bias score exceeded 3 MAD units. Data collec- tion continued until the study had the appropriate power (N = 300) following exclusion (64 excluded). Background analyses We saw very strong evidence that internal metacognitive bias was below 0, M = –6.91, t(299) = –3.62, p < .001, BF10 = 36.78, BF01 = 0.03, indicat- ing that participants were underconfident in their own memory abilities. We also conducted an analogous analysis on the external metacognitive bias score (i.e., predicted accuracy with reminders minus actual accuracy in the forced-external condition). There was decisive evidence that external metacognitive bias was below 0, M = –11.35, t(299) = –12.92, p < .0001, BF10 > 100, BF01 < .01, indicat- ing participants were underconfident in their performance on the task with reminders. We investigated reminder bias scores (defined as OIP minus AIP). There was decisive evi- dence that reminder bias was greater than 0, M = 1.28, Quarterly Journal of Experimental Psychology 00(0) 8 Figure 5. Correlation coefficient slopes with 95% confidence intervals for anxiety and (a) internal confidence, (b) metacognitive bias, (c) reminder usage, and (d) reminder bias. Lower confidence (a) and metacognitive bias (b) scores indicate lower or underconfidence, respectively. Higher reminder use (c) and bias (d) indicate higher or overuse of reminders, respectively. Figure 5. Correlation coefficient slopes with 95% confidence intervals for anxiety and (a) internal confidence, (b) metacognitive bias, (c) reminder usage, and (d) reminder bias. Lower confidence (a) and metacognitive bias (b) scores indicate lower or underconfidence, respectively. Higher reminder use (c) and bias (d) indicate higher or overuse of reminders, respectively. correlation between trait anxiety and internal metacogni- tive bias, r(298) = –.00, p = 1.00, BF10 = 0.07, BF01 = 13.83 (Figure 5b). Fourth, we predicted that trait anxiety would positively correlate with participants’ propensity to use reminders on our delayed intentions task (i.e., AIP). We saw strong evidence to suggest there was no correlation between trait anxiety and participants’ AIP, r(298) = –.02, p = .67, BF10 = 0.08, BF01 = 12.65 (Figure 5c). Fifth, we predicted that trait anxiety would positively correlate with participants’ bias towards the use of reminders in the delayed intentions task. We saw strong evidence to suggest there was no correlation between trait anxiety and partici- pants’ reminder bias, r(298) = .00, p = .94, BF10 = 0.07, BF01 = 13.80 (Figure 5d). t(299) = 10.32, p < .0001, BF10 > 100, BF01 < .01, indicat- ing that participants typically used more reminders than was optimal. Background analyses Finally, we investigated whether internal metacognitive bias was associated with reminder bias, as we have found previously (Gilbert et al., 2020). Replicating previous results, there was decisive evidence that internal metacognitive bias was inversely correlated with reminder bias, r(298) = –.34, p < .0001, BF10 > 100, BF01 < .01 (Figure 4b), that is, underconfidence in memory abilities was associated with increased reminder usage. Additional exploratory analyses outlined a relationship between trait anxiety and metacog- nitive belief (Spada et al., 2010; Wells & Cartwright- Hatton, 2004). Our previous work has established a link between metacognitive belief and use of reminders for future intentions (Boldt & Gilbert, 2019; Gilbert et al., 2020). The present article set to bridge these areas by car- rying out correlational analyses between individual differ- ences in trait anxiety, metacognitive confidence, and offloading behaviour on a delayed intentions paradigm. Specifically, we correlated trait anxiety with the following measures on our task: unaided accuracy, metacognitive confidence in one’s ability to perform the task, internal metacognitive bias (the discrepancy between confidence and performance), reminder usage, and reminder bias (optimality of reminder-setting). Despite an implied rela- tionship from the previous literature and a sample size (N = 300) powered to detect effects bigger than r = .161, our preregistered experiment provided no evidence for any correlation between trait anxiety and the behavioural data on our delayed intentions task. We performed additional exploratory analyses to provide further characterisation of our dataset. Given the large num- ber of tests provided, these were considered as exploratory tests rather than key hypothesis-testing analyses, the full extent of which can be found in the supplemental materials. First, we repeated the above analyses, using the worry measure instead of trait anxiety (Supplementary Material 1). We found no evidence for a relationship in this domain, nor did we find including both trait anxiety and worry as predic- tors in a regression model to alter the inference. We also found that there was a positive correlation between PSWQ and STAI (r = .77, p < .0001, BF10 > 100, Supplementary Material 2), consistent with previous results (r = .64, Meyer et al., 1990). This shows that even though our measure of trait anxiety did not correlate with performance of the exper- imental task, it did relate as expected to our measure of worry. We found no evidence to suggest anxiety or worry correlated with participants’ confidence in their ability to perform the task with reminders, and external metacognitive bias (Supplementary Material 3). Finally, we performed additional analyses using data from the loss rather than the gain condition. Additional exploratory analyses Results were similar: anxiety did not corre- late with internal accuracy, r(298) = –.05, p = .39, BF01 = 0.11, BF10 = 9.52; internal metacognitive bias, r(298) = –.01, p = .90, BF10 = 0.07, BF01 = 13.73; reminder use, r(298) = –.05, p = .43, BF10 = 0.10, BF01 = 10.15; or reminder bias, r(298) = .03, p = .57, BF10 = 0.08, BF01 = 11.80, in the loss condition (Supplementary Material 4), providing further evidence for the null. Our first hypothesis tested whether higher trait anxiety resulted in worse unaided performance on our delayed intentions task. This was motivated by the notion that anxi- ety can result in impaired memory processes (Shackman et al., 2006). On one hand, no relationship may exist between trait anxiety and memory in the context of a delayed intentions paradigm. Alternatively, it has been sug- gested that anxiety only impairs performance at low cogni- tive loads (see Vytal et al., 2012). This may be because anxiety results in disruptive worrying thoughts while at low load, but this could dissipate at higher loads, with atten- tional resources being re-focussed onto the task at hand. The delayed intentions paradigm we utilised is relatively demanding, and performance was well below ceiling (mean accuracy = 65%). Therefore, detrimental effects of anxiety on performance could have been mitigated by difficulty- driven attention. However, our previous work has demon- strated that cognitive load is not always a sufficient explanation of the impact of anxiety on cognition (Sarigiannidis et al., 2020). Nevertheless, future work might seek to retest whether a correlation exists for our task at lower loads (e.g., fewer target circles). We have also provided additional, post hoc analyses which may address outstanding questions. As the order in which conditions were presented (gain first vs. loss first) had a significant impact on internal accuracy for the gain condition, and consequently our calculation of internal metacognitive bias, we re-analysed our data separately for each counterbalancing group (Supplementary Material 8). As STAI scores were not normally distributed, and we did not screen for anxiety disorders, we provide group com- parison analyses between participants on the upper and lower quartiles of STAI scores (Supplementary Material 5), as well as analyses restricted to less anxious partici- pants (Supplementary Material 10). Finally, we have included a re-analysis of the data on all participants with- out any exclusions (N = 364; Supplementary Material 9). Key hypotheses Our key hypotheses were tested using a series of Pearson’s correlations (Hypotheses 1–5) and a linear regression (enter method; Hypothesis 6). First, we tested whether trait anxiety negatively correlated with participants’ unaided accuracy on our delayed intentions task. We saw substan- tial evidence that trait anxiety and accuracy in the forced internal were not correlated, r(298) = –.06, p = .29, BF10 = 0.13, BF01 = 7.97. Second, we tested whether trait anxiety would negatively correlate with participants’ con- fidence in their unaided ability on the task. We saw sub- stantial evidence that there was no correlation between trait anxiety and participants’ confidence in their ability to perform the task, r(298) = –.05, p = .39, BF10 = 0.10, BF01 = 9.54 (Figure 5a). Third, we tested whether trait anx- iety would negatively correlate with participants’ meta- cognitive bias for their ability to perform the delayed intentions task. We saw strong evidence that there was no 01 Finally, we predicted that both anxiety and metacogni- tive bias would account for unique variance in a model of reminder bias. We performed a linear multiple regression on the reminder bias score, with factors trait anxiety and internal metacognitive bias as predictor variables (enter method). Congruent with the previous analyses, internal metacognitive bias retained predictive power (β = –.02, p < .0001) while trait anxiety did not account for unique variance (β = .00, p = .94). For the Bayesian linear regres- sion, the winning model was that which only included internal metacognitive bias (BF10 > 100), which was sub- stantially (7 times) better than the model which included metacognitive bias and trait anxiety (BF10 > 100), deci- sively (>1,000 times) better than the null model (BF10 = 1), and decisively (>1,000 times) better than the trait anxiety only model (BF10 = 0.13). 9 Kirk et al. Additional exploratory analyses None of the above analyses provided any evidence towards the alternative hypotheses, and thus did not change our overall inference. The second and third hypotheses tested whether trait anxiety correlated with confidence to perform the task, and the accuracy of this confidence (metacognitive bias). These hypotheses were driven by work relating scores on trait anxiety to the MCQ-30 (Spada et al., 2010; Wells & Cartwright-Hatton, 2004). This latter measure includes asking participants to rate general confidence in their mem- ory abilities. However, confidence measures for our task did not appear to hold a relationship with trait anxiety. Despite a prior experiment demonstrating correlations between confidence in our paradigm and the MCQ-30 (Boldt & Gilbert, 2019), it may be that the explanatory vari- ance shared between these two measures does not overlap Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. ORCID iDs Peter A Kirk https://orcid.org/0000-0003-0786-3039 Sam J Gilbert https://orcid.org/0000-0002-3839-7045 Peter A Kirk https://orcid.org/0000-0003-0786-3039 Sam J Gilbert https://orcid.org/0000-0002-3839-7045 Peter A Kirk https://orcid.org/0000-0003-0786-3039 Sam J Gilbert https://orcid.org/0000-0002-3839-7045 Conclusion Our present study bridged the gap between two different areas of research, namely, affective processes and meta- cognition. Previous findings suggested a relationship may exist between trait anxiety, metacognitive bias, and reminder bias. Specifically, prior work implied higher trait anxiety results in a general underconfidence in cognitive abilities, and possibly excessive reminder usage. Within the context of a delayed intentions task, we found evidence against a relationship between trait anxiety and memory abilities, confidence, or reminder usage. Highly anxious individuals were similar in their optimality of reminder- setting as low-anxiety individuals. Future work may seek to expand our findings by repeating the task across differ- ent cognitive loads, using clinical comparisons, and manipulating anxiety in-lab. An overarching goal of metacognition research is to develop ways to optimise peoples’ behaviour in line with their cognitive abilities (Gilbert et al., 2020). It would therefore be constructive to examine anxiety in such a con- text. We previously found that providing metacognitive advice can reduce bias in our delayed intentions task (Gilbert et al., 2020). It would be informative to investi- gate whether anxiety acts as a mediating factor between metacognitive advice and updating of confidence. More generally, does anxiety reduce individuals’ propensity to update cognitive confidence through external advice? Our fourth, fifth, and sixth hypotheses investigated the extent to which trait anxiety was related to participants’ use of reminders. We posited that a correlation between reminder usage and anxiety could be due to lowered confi- dence and/or through unique influences external to higher- order cognition. Yet, we did not observe any relationship between anxiety and reminder usage. This suggests there may be no difference between high- and low-anxiety indi- viduals in the frequency and (sub)optimality of reminder- setting. Again, it would be useful for future work to test whether anxiety acts as a barrier (or even facilitator) to the alteration of reminder usage following external advice. The supplementary material is available at qjep.sagepub.com. The supplementary material is available at qjep.sagepub.com. Discussion There is a myriad of experimental evidence demonstrating that anxiety leads to altered behaviour and cognitions (Robinson et al., 2013). In particular, prior studies have Quarterly Journal of Experimental Psychology 00(0) 10 00(0) with trait anxiety. In other words, anxiety may correlate with general feelings of confidence as reported in a ques- tionnaire such as the MCQ-30, but not performance predic- tions in a specific task. Further experiments attempting to replicate our findings should also include the MCQ-30 measures to grasp both a domain-general measure of meta- cognitive belief and task-specific confidence measures. generally high-anxiety traits. This is not something that could have easily been addressed with the state section of the STAI (Supplementary Material 7). Future work could implement within-subjects state anxiety manipulations such as threat-of-shock, anxiety-relevant stimuli, or time pressure paradigms. This would help elucidate whether increases in state anxiety lead to increased metacognitive and reminder bias, despite the lack of a relationship between the latter measures and trait anxiety. Although we took measures of confidence to perform the task, these do not capture the accuracy of beliefs. As such, we calculated metacognitive bias scores, a metric of the discrepancy between self-reported confidence and accuracy. While participants were generally underconfi- dent in their ability to perform the task, we saw no correla- tion between trait anxiety and metacognitive bias. We can infer that high-anxiety individuals show similar levels of metacognitive bias as low-anxiety individuals. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Leverhulme Trust as part of the Doctoral Training Programme for the Ecological Study of the Brain (DS-2017-026). Finally, two caveats should be noted with the present study. First, this work did not explicitly screen for anxiety disorders; rather, we looked across a spectrum of individu- als at the population level. Thus, we do not make a distinc- tion between maladaptive versus adaptive anxiety. As such, we have provided additional analyses (Supplementary Material 5/10) comparing participants on the extremities of the STAI scale, although this did not change our infer- ence. Yet, without explicit screening and population com- parisons, we cannot distinguish to what extent our findings may be driven by the presence (or lack) of maladaptive anxiety disorders. Second, our study was inherently cor- relational in nature and did not explicitly induce state anxi- ety at a within-subjects level. Understanding that our experiment was completed remotely, probably at home, it is feasible that participants were in relatively low anxiety states during the experiment, even if they otherwise held References Allen, M., Poggiali, D., Whitaker, K., Marshall, T. R., & Kievit, R. A. (2019). Raincloud plots: A multi-platform tool for robust data visualization. Wellcome Open Research, 4, Article 63. https://doi.org/10.12688/wellcomeopenres.15191.1 Arnold, N. R., Bayen, U. J., & Böhm, M. F. (2015). Is prospec- tive memory related to depression and anxiety? A hierarchi- cal MPT modelling approach. Memory, 23(8), 1215–1228. https://doi.org/10.1080/09658211.2014.969276 11 Kirk et al. of Experimental Social Psychology, 49(4), 764–766. https:// doi.org/10.1016/j.jesp.2013.03.013 Boldt, A., & Gilbert, S. J. (2019). Confidence guides spontane- ous cognitive offloading. Cognitive Research: Principles and Implications, 4, Article 45. https://doi.org/10.1186/ s41235-019-0195-y Linares, D., & López-Moliner, J. (2016). quickpsy: An R pack- age to fit psychometric functions for multiple groups. http:// diposit.ub.edu/dspace/handle/2445/116040 Brandimonte, M., Einstein, G. O., & McDaniel, M. A. (Eds.). (1996). Prospective memory: Theory and applications. Lawrence Erlbaum. Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28(6), 487–495. https://doi.org/10.1016/0005- 7967(90)90135-6 Cowan, N. (2010). The magical mystery four: How is working memory capacity limited, and why? Current Directions in Psychological Science, 19(1), 51–57. https://doi. org/10.1177/0963721409359277 Risko, E. F., & Dunn, T. L. (2015). Storing information in-the- world: Metacognition and cognitive offloading in a short- term memory task. Consciousness and Cognition, 36, 61–74. https://doi.org/10.1016/j.concog.2015.05.014 Crump, M. J. C., McDonnell, J. C., & Gureckis, T. M. (2013). Evaluating Amazon’s Mechanical Turk as a tool for experi- mental behavioral research. PLOS ONE, 8(3), Article e57410. https://doi.org/10.1371/journal.pone.0057410 Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi. org/10.1016/j.tics.2016.07.002 Cuttler, C., & Graf, P. (2008). Sub-clinical checking compulsions are related to impaired prospective memory independently of depression, anxiety and distractibility. Journal of Anxiety Disorders, 22(4), 642–654. https://doi.org/10.1016/j.janxdis. 2007.06.003 Robinson, O. J., Vytal, K., Cornwell, B. R., & Grillon, C. (2013). The impact of anxiety upon cognition: Perspectives from human threat of shock studies. Frontiers in Human Neuroscience, 7, Article 203. https://doi.org/10.3389/ fnhum.2013.00203 Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional con- trol theory. Emotion, 7(2), 336–353. https://doi.org/10.1037/ 1528-3542.7.2.336 RStudio Team. (2019). RStudio (Version 1.2.5019): Integrated development for R [Computer software]. http://www.rstu- dio.com/ Finley, J. R., Naaz, F., & Goh, F. W. (2018). Memory and technology. Springer International Publishing. https://doi. org/10.1007/978-3-319-99169-6 Sarigiannidis, I., Kirk, P. A., Roiser, J. P., & Robinson, O. J. (2020). References Does overloading cognitive resources mimic the impact of anxiety on temporal cognition? Journal of Experimental Psychology. Learning, Memory, and Cognition, 46(10), 1828–1835. https://doi.org/10.1037/xlm0000845 Gilbert, S. J. (2015). Strategic offloading of delayed intentions into the external environment. The Quarterly Journal of Experimental Psychology, 68(5), 971–992. https://doi.org/ 10.1080/17470218.2014.972963 Scullin, M., Mullet, H., Einstein, G., & Mcdaniel, M. (2015). Prospective memory. In J. D. Wright (Ed.), International encyclopedia of the social & behavioral sciences (pp. 270– 278). Elsevier. https://doi.org/10.1016/B978-0-08-097086- 8.51055-1 Gilbert, S. J., Bird, A., Carpenter, J. M., Fleming, S. M., Sachdeva, C., & Tsai, P.-C. (2020). Optimal use of remind- ers: Metacognition, effort, and cognitive offloading. Journal of Experimental Psychology: General, 149(3), 501–517. https://doi.org/10.1037/xge0000652 Seibert, P. S., & Ellis, H. C. (1991). Irrelevant thoughts, emotional mood states, and cognitive task performance. Memory & Cognition, 19, 507–513. https://doi.org/10.3758/BF03199574 Graf, P., & Uttl, B. (2001). Prospective memory: A new focus for research. Consciousness and Cognition, 10(4), 437–450. https://doi.org/10.1006/ccog.2001.0504 Shackman, A. J., Sarinopoulos, I., Maxwell, J. S., Pizzagalli, D. A., Lavric, A., & Davidson, R. J. (2006). Anxiety selec- tively disrupts visuospatial working memory. Emotion, 6(1), 40–61. https://doi.org/10.1037/1528-3542.6.1.40 Hu, X., Luo, L., & Fleming, S. M. (2019). A role for metamem- ory in cognitive offloading. Cognition, 193, Article 104012. https://doi.org/10.1016/j.cognition.2019.104012 JASP Team. (2019). JASP (Version 0.11.1) [Computer software]. https://jasp-stats.org Spada, M. M., Georgiou, G. A., & Wells, A. (2010). The rela- tionship among metacognitions, attentional control, and state anxiety. Cognitive Behaviour Therapy, 39(1), 64–71. https://doi.org/10.1080/16506070902991791 Jeffreys, S. H. (1998). The theory of probability (3rd ed.). Oxford University Press. Spielberger, C. D. (1983). State-Trait Anxiety Inventory for Adults. https://doi.org/10.1037/t06496-000 Kliegel, M., & Jäger, T. (2006). The influence of negative emo- tions on prospective memory: A review and new data. International Journal of Computational Cognition, 4(1), 1–17. https://doi.org/10.1080/09658211.2014.969276 Vytal, K., Cornwell, B., Arkin, N., & Grillon, C. (2012). Describing the interplay between anxiety and cognition: From impaired performance under low cognitive load to reduced anxiety under high load: Anxiety and cognition. Psychophysiology, 49(6), 842–852. https://doi.org/10.1111/ j.1469-8986.2012.01358.x Kliegel, M., McDaniel, M. A., & Einstein, G. O. (Eds.) (2008). Prospective memory: Cognitive, neuroscience, develop- mental, and applied perspectives. Taylor & Francis Group/ Lawrence Erlbaum Associates. Wells, A. (2009). Metacognitive therapy for anxiety and depres- sion. Guilford Press. Kovacs, M., & Beck, A. T. (1978). Maladaptive cognitive struc- tures in depression. The American Journal of Psychiatry, 135(5), 525–533. https://doi.org/10.1176/ajp.135.5.525 Wells, A., & Cartwright-Hatton, S. (2004). References A short form of the Metacognitions Questionnaire: Properties of the MCQ-30. Behaviour Research and Therapy, 42(4), 385–396. https:// doi.org/10.1016/S0005-7967(03)00147-5 Leys, C., Ley, C., Klein, O., Bernard, P., & Licata, L. (2013). Detecting outliers: Do not use standard deviation around the mean, use absolute deviation around the median. Journal
https://openalex.org/W2985984824	https://europepmc.org/articles/pmc6920764?pdf=render	English	null	Bacteriophages as Potential Tools for Detection and Control of Salmonella spp. in Food Systems	Microorganisms	2,019	cc-by	13,386	Bacteriophages as Potential Tools for Detection and Control of Salmonella spp. in Food Systems 1 Department of Medical Biomaterials Engineering, College of Biomedical Science, Kangwon National University, Chuncheon, Gangwon 24341, Korea; weishuaiws@126.com (S.W.); jalalmbe@kangwon.ac.kr (M.J.U.) 1 Department of Medical Biomaterials Engineering, College of Biomedical Science, Kangwon National University, Chuncheon, Gangwon 24341, Korea; weishuaiws@126.com (S.W.); jalalmbe@kangwon.ac.kr (M.J.U.) j g ( J ) 2 Guangdong Provincial Key Laboratory of Aquatic Product Processing and Safety, College of Food Science and Technology, Guangdong Ocean University, Zhanjiang 524088, China g g y y q g y g and Technology, Guangdong Ocean University, Zhanjiang 524088, China 3 Department of Food Science and Biotechnology, College of Agriculture and Life Science, Kangwon National University, Chuncheon, Gangwon 24341, Korea; ramachandran865@gmail.com (R.C.); rubab.momna@gmail.com (M.R.); deoghwa@kangwon.ac.kr (D.-H.O.) * Correspondence: juheeahn@kangwon.ac.kr; Tel.: +82-33-250-6564 Received: 28 October 2019; Accepted: 15 November 2019; Published: 17 November 2019 and Technology, Guangdong Ocean University, Zhanjiang 524088, China 3 Department of Food Science and Biotechnology, College of Agriculture and Life Science, Kangwon National University, Chuncheon, Gangwon 24341, Korea; ramachandran865@gmail.com (R.C.); rubab.momna@gmail.com (M.R.); deoghwa@kangwon.ac.kr (D.-H.O.) * Correspondence: juheeahn@kangwon ac kr; Tel : +82-33-250-6564 Received: 28 October 2019; Accepted: 15 November 2019; Published: 17 November 2019 Abstract: The global problem of antibiotic resistance in bacteria is quickly developing in most antibiotics used in hospitals and livestock. Recently, the infections with multi-drug resistant (MDR) bacteria become a major cause of death worldwide. Current antibiotics are not very eﬀective in treating MDR Salmonella infections, which have become a public health threat. Therefore, novel approaches are needed to rapidly detect and eﬀectively control antibiotic-resistant pathogens. Bacteriophages (phages) have seen renewed attention for satisfying those requirements due to their host-speciﬁc properties. Therefore, this review aims to discuss the possibility of using phages as a detection tool for recognizing bacterial cell surface receptors and an alternative approach for controlling antibiotic-resistant pathogens in food systems. Keywords: bacteriophage; phage detection; phage control; Salmonella; food system Microorganisms 2019, 7, 570; doi:10.3390/microorganisms7110570 microorganisms www.mdpi.com/journal/microorganisms 1. Introduction A phage as a biorecognition agent provides many advantages for rapid bacterial sensing [27], including target speciﬁcity [28], release of high titer phages, tolerance to environmental stresses [29], safe handling, and eﬀectiveness against live bacteria [30]. The phage-based biosensors include the combination of whole phages or phage-constituents, which can be converted to electrical, colorimetric, ﬂuorescent, or luminescent signals. Phages are thus shown as cheap, fast, sensitive, selective, and speciﬁc tools for detecting bacteria [31]. From a therapeutic viewpoint, phage therapy provides many beneﬁts over chemotherapy, since phages are active against antibiotic-resistant bacteria and no side eﬀect occurs during phage treatment [32]. Microorganisms 2019, 7, x FOR PEER REVIEW 2 of 18 Bacteriophages (phages) are predominant in nature, defined as viruses that can infect bacteria alone [15,16]. In the 1920s, Frederick Twort first observed the glassy and transparent colonies in micrococci cultures, which could cause an acute infectious disease in 1915, but the term “phage” was neither correctly defined nor clearly understood at this time [17]. Two years later, Félix d’Herelle isolated a microbe that showed an antagonistic effect against Shiga bacillus and first described this microbe as an obligate bacteriophage with a high specificity to the host and no pathogenic effects [18]. Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19] Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19]. Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19] Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19]. Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19] Figure 1. Advantages and disadvantages of using bacteriophages for the treatment of Salmonella [19]. [19]. Table 1. Lytic spectrum of bacteriophages against Salmonella and other foodborne pathogens. Microorganism Bacteriophage 1 7 16 18 36 43 Salmonella Typhi ATCC 6539 + 2 + + + + Salmonella Typhimurium ATCC 14028 + + + + + Salmonella Enteritidis ATCC 13076 + + + + + Salmonella Gallinarum ATCC 9184 + + + + + With more scientiﬁcally rigorous approaches in the recent years, more researchers have paid attention toward the bacteriophages as a promising tool to treat the pathogens [33]. 1. Introduction Salmonella, a Gram-negative rod-shaped bacilli bacterium, is the most common foodborne pathogen and is known as one of the four key global causes of diarrheal diseases, according to the World Health Organization (WHO). As estimated by the Centers for Disease Control and Prevention, 23,000 hospitalizations and 450 deaths have been caused by Salmonella every year in the United States. Among the 1.2 million illnesses caused by Salmonella, contaminated food caused 1 million illnesses [1,2]. Contaminated meat, poultry, eggs, and milk are the main reservoirs for Salmonella infections [3]. Some direct contact with infected animals, blood, urine, and feces may also cause the problem to human health [4]. Antibiotic use has been increased for controlling the bacteria in animals or increasing food production, which has accelerated the emergence of antimicrobial-resistant bacteria [5–7]. The antibiotic-resistant bacteria are likely to contaminate in food products throughout the handling or other production stages [8–10]. Fruit and vegetable products are susceptible to cross-contamination during harvest and post-harvest periods [11–13]. Salmonellosis caused by Salmonella is one of the serious infections that is diﬃcult to treat due to the reduced antibiotic activities that become less eﬀective [14]. Thus, novel antibiotics or alternative methods are required to control the antibiotic-resistant bacteria. Bacteriophages (phages) are predominant in nature, deﬁned as viruses that can infect bacteria alone [15,16]. In the 1920s, Frederick Twort ﬁrst observed the glassy and transparent colonies in Microorganisms 2019, 7, 570; doi:10.3390/microorganisms7110570 2 of 22 Microorganisms 2019, 7, 570 micrococci cultures, which could cause an acute infectious disease in 1915, but the term “phage” was neither correctly deﬁned nor clearly understood at this time [17]. Two years later, Félix d’Herelle isolated a microbe that showed an antagonistic eﬀect against Shiga bacillus and ﬁrst described this microbe as an obligate bacteriophage with a high speciﬁcity to the host and no pathogenic eﬀects [18]. However, the interest in phages decreased because of the lack of proper quality controls and reproducible results in the western countries [21–23]. Another important reason was the discovery of antibiotics, which were used as the most powerful reagents for controlling bacterial infections. Ironically, the frequent use of antibiotics has resulted in the development of the multidrug-resistance or superbugs [24–26]. Figure 1 shows the advantages and disadvantages of bacteriophage applications. 1. Introduction Bacteriophages also show additional advantages, including a high speciﬁcity to the host, an ability to diﬀerentiate alive and dead cells, and the most abundant biological entity in nature, which draws renewed attention to the detection and control of antibiotic-resistant pathogens [34–36]. The eﬀectiveness of phage applications in treating pathogenic bacteria is inﬂuenced by several factors, such as the multiplicity of infection (MOI), treatment mode, environmental conditions, phage neutralization, and target bacteria. The bacteriophage survival an adverse environment is also a desired characteristic for therapeutic use. Currently, the commercial products based on bacteriophages targeting Escherichia coli O157:H7 [37–39], 3 of 22 Microorganisms 2019, 7, 570 Listeria monocytogenes [40–42], Salmonella spp. [43,44], and Shigella spp. [45–47] are available in the market. Furthermore, application trials in foods are also performed, which would help enhance the food safety. The lytic activities of speciﬁc bacteriophages against Salmonella and other pathogenic bacteria are compared in Tables 1 and 2. In this review, the detection methods and biocontrol applications based on bacteriophages targeting Salmonella are summarized and discussed in detail. Table 1. Lytic spectrum of bacteriophages against Salmonella and other foodborne pathogens. Microorganism Bacteriophage 1 7 16 18 36 43 Salmonella Typhi ATCC 6539 + 2 + + + + Salmonella Typhimurium ATCC 14028 + + + + + Salmonella Enteritidis ATCC 13076 + + + + + Salmonella Gallinarum ATCC 9184 + + + + + Salmonella Pullorum ATCC 9120 + + + + + Salmonella Abony NCTC 6017 + + + + + Salmonella Choleraesuis ATCC 10708 − + + Salmonella Arizonae ATCC 13314 − − − − − Escherichia coli ATCC 11229 − + + − − Klebsiella pneumoniae ATCC 10031 − − − − − Enterobacter aerogenes ATCC 13048 − − − − − Campylobacter jejuni NCTC 12662 − − − − − Listeria monocytogenes ATCC 7644 − − − − − Enterococcus faecalis ATCC 19433 − − − − − Staphylococcus aureus ATCC 6538 − − − − − 1 Five phages, phiSE 7, 16, 18, 36, and 43, were isolated from chicken feces, and they belong to the Podoviridae family. 2 The phage ability to plaque on diﬀerent pathogens was evaluated. (−) and (+) indicate the absence of phage plaques and the presence of phage plaques, respectively. (Copyright obtained from [20].). Table 1. Lytic spectrum of bacteriophages against Salmonella and other foodborne pathogens. 1 A, B, C, D, and E describe the clearing throughout with a faint hazy background, substantial turbidity throughout the cleared zone, a few individual plaques, and no clearing, respectively. (Copyright obtained from [48].) 2. Phage Plaque Assay as A Traditional Method for Salmonella Detection 2. Phage Plaque Assay as A Traditional Method for Salmonella Detection Th di i l l h d i f l h i f The traditional agar overlay method is a very useful technique for studying bacteriophage, including identiﬁcation, enumeration, puriﬁcation, and isolation of phage mutants. This method is based on lysis plaques, which are formed on lawns of the host bacteria, immobilized in the top soft agar. After incubation, a series of reaction events, such as phage infection, multiplication, and liberation, occur in the host [49,50]. The soft agar overlay method is a common titer assay, which was introduced by Mark Adams [51]. The plaque assay has been improved based on several modiﬁcations, such as the addition of eﬀective supplements to modify growth media and agar composition and the plate spreading method, which can enhance the visibility of the plaques [49,52,53]. However, a long time (more than 24 h) is required for obtaining detection results and it is also labor intensive, which may not meet the demands of rapid pathogen detection. Thus, many researchers devote their eﬀorts to developing reliable and rapid detection methods. Figure 2 shows two diﬀerent life cycles of bacteriophages and lytic and lysogenic cycles. The active virulence phages can produce progeny phages that burst out of the host cell through the lytic cycle, while the lysogenic cycle involves the integration of the temperate phage genome into the host chromosome, which can remain in an inactive state, known as prophage [54]. The distinct life cycles of bacteriophages are used for designing novel diagnostic tools, including reporter bacteriophages [55] and phage display technologies [56]. The traditional agar overlay method is a very useful technique for studying bacteriophage, including identification, enumeration, purification, and isolation of phage mutants. This method is based on lysis plaques, which are formed on lawns of the host bacteria, immobilized in the top soft agar. After incubation, a series of reaction events, such as phage infection, multiplication, and liberation, occur in the host [49,50]. The soft agar overlay method is a common titer assay, which was introduced by Mark Adams [51]. The plaque assay has been improved based on several modifications, such as the addition of effective supplements to modify growth media and agar composition and the plate spreading method, which can enhance the visibility of the plaques [49,52,53]. However, a long time (more than 24 h) is required for obtaining detection results and it is also labor intensive, which may not meet the demands of rapid pathogen detection. 2. Phage Plaque Assay as A Traditional Method for Salmonella Detection 2. Phage Plaque Assay as A Traditional Method for Salmonella Detection Th di i l l h d i f l h i f Thus, many researchers devote their efforts to developing reliable and rapid detection methods. Figure 2 shows two different life cycles of bacteriophages and lytic and lysogenic cycles. The active virulence phages can produce progeny phages that burst out of the host cell through the lytic cycle, while the lysogenic cycle involves the integration of the temperate phage genome into the host chromosome, which can remain in an inactive state, known as prophage [54]. The distinct life cycles of bacteriophages are used for designing novel diagnostic tools, including reporter bacteriophages [55] and phage display technologies [56]. Figure 2. Life cycles of bacteriophages. (Copyright obtained from Kakasis et al., 2019; de Jonge et al., 2019, [57,58].) Figure 2. Life cycles of bacteriophages. (Copyright obtained from Kakasis et al., 2019; de Jonge et al., 2019, [57,58].) Figure 2. Life cycles of bacteriophages. (Copyright obtained from Kakasis et al., 2019; de Jonge et al., 2019, [57,58].) Figure 2. Life cycles of bacteriophages. (Copyright obtained from Kakasis et al., 2019; de Jonge et al., 2019, [57,58].) 1. Introduction 1 Five phages, phiSE 7, 16, 18, 36, and 43, were isolated from chicken feces, and they belong to the Podoviridae family. 2 The phage ability to plaque on diﬀerent pathogens was evaluated. (−) and (+) indicate the absence of phage plaques and the presence of phage plaques, respectively. (Copyright obtained from [20].). Table 2. Eﬃcacy of phages against Salmonella strains and other pathogens. Microorganism Bacteriophage LPST18 LPST23 LPST10 Salmonella Typhimurium ATCC 14028 A 1 A B Salmonella Typhimurium ATCC 13311 C A C Salmonella Typhimurium ST-8 A A A Salmonella Paratyphi B CMCC 50094 A A B Salmonella Enteritidis ATCC 13076 A A C Salmonella Enteritidis SJTUF 10978 E C A Salmonella Enteritidis SJTUF 10984 E E A Salmonella Anatum ATCC 9270 D E D Salmonella Choleraesuis ATCC 10708 E E D Escherichia coli DH5α E E E Escherichia coli BL21 E E E Escherichia coli c83715 E E E Listeria monocytogenes ATCC 19114 E E E Vibrio parahaemolyticus ATCC 33846 E E E Staphylococcus aureus ATCC 29213 E E E Staphylococcus aureus ATCC 6538 E E E Lactobacillus acidophilus ATCC SD5221 E E E 1 A, B, C, D, and E describe the clearing throughout with a faint hazy background, substantial turbidity throughout the cleared zone, a few individual plaques, and no clearing, respectively. (Copyright obtained from [48].) Table 2. Eﬃcacy of phages against Salmonella strains and other pathogens. 4 of 22 t Microorganisms 2019, 7, 570 oug ou e ea obtained from [48].) 3. Rapid Detection Methods Associated with Bacteriophage for Salmonella 3. Rapid Detection Methods Associated with Bacteriophage for Salmonella With the growing need of food safety, several detection methods for targeting Salmonella were developed by combining bacteriophages, including molecular based real-time PCR [59,60], immunomagnetic separation based on fluorescence immunoassays [61,62], enzyme-linked With the growing need of food safety, several detection methods for targeting Salmonella were developed by combining bacteriophages, including molecular based real-time PCR [59,60], immunomagnetic separation based on ﬂuorescence immunoassays [61,62], enzyme-linked immunosorbent assays (ELISAs) [59], matrix-assisted laser desorption/ionization–time-of-ﬂight mass spectrometry (MALDI–TOF MS) [63], and genetically engineered reporter phage [64]. 5 of 22 Microorganisms 2019, 7, 570 3.1. qPCR-Based Detection Bacteriophages have been used for the detection of foodborne pathogens based on their speciﬁcity towards bacteria and ampliﬁcation ability, which is like the “enrichment” procedure and shortens the detection time [36]. In this approach, qPCR is used to directly amplify and detect the nucleic acids of progeny bacteriophage after propagation. The bacteriophages coupled with qPCR were used for the detection of Ralstonia solanacearum, Escherichia coli O157:H7, Mycobacterium avium, and Acinetobacter baumannii [60,65–67]. The ampliﬁcation of Salmonella phage B_SenS_PVP-SE2 combined with qPCR was then developed for the detection of viable Salmonella Enteritidis in chicken samples [68]. The proposed method detected a low concentration level of viable S. Enteritidis (8 colony-forming unit (CFU)/g) in chicken samples within 10 h, which saved much time when compared to the culture-based method and also enhanced sensitivity, speciﬁcity, and accuracy. Thus, this protocol can be used in the food industry for self-monitoring, which successfully completed a “same-day” detection within 10 h. 3.2. Immunomagnetic-Based Detection A novel assay composed of immunomagnetic separation (IMS) and ampliﬁcation of Salmonella bacteriophage SJ2 was developed and optimized for the detection of Salmonella enterica serovar Enteritidis [62]. In the IMS procedure, Dynabeads® anti-Salmonella was used for capturing and concentrating Salmonella. Bacteriophage SJ2 was then added, and the mixture was incubated for attachment and ampliﬁcation. The ﬁnal sample was detected using ﬂuorescence or optical density measurements. This assay showed a detection limit of less than 104 CFU/mL with a short time (4.0–4.5 h). However, the pre-enrichment process was required in food samples. When this technique was applied to contaminated food samples, including skimmed milk, chicken, and beef, at an average of 3 CFU/25 g, S. Enteritidis could be detected within 20 h, including a pre-enrichment time of 16 h [62]. 3.3. Enzyme-Linked Immunosorbent Assay The commercial ELISA kits have been applied for detecting Salmonella in poultry, seafood, milk, and meat [69–72]. ELISA detects the protein in a liquid form using antibodies against the target samples. Instead of using antibodies, bacteriophages can be used in ELISA for detecting bacteria [73]. The ELISA procedure works by replacing antibodies with bacteriophages and was applied for the detection of S. enterica and E. coli [59]. The modiﬁed ELISA showed a detection limit up to 106 cells/mL, which is comparable with other ELISA methods. Thus, bacteriophages in ELISA can be an alternative way to detect pathogenic bacteria without speciﬁc antibodies. Since phages are highly abundant in nature, this assay becomes cheap compared with using speciﬁc antibodies. 3.5. Genetically-Engineered Phages Recent advances in genetically-engineered bacteriophages have been created as a powerful tool for the monitoring and detecting of bacterial pathogens. This novel technique is useful for detecting bacteria from contaminated foods with high selectivity and sensitivity [64,79]. Reporter bacteriophages are genetically modiﬁed bacteriophages that have a reporter gene inserted into their genomes, such as lux, gpf, and lacZ, which are activated by the interaction between bacteriophages and target bacteria [64,80]. The expression of the reporter gene upon infection emits a detectable signal, indicating the presence of target bacteria. The main advantage of using a reporter bacteriophage is the higher speciﬁcity for detecting viable host bacteria. To date, most commonly used reporter bacteriophages are associated with the formation of bioluminescence luciferase protein, which emits light in the presence of aldehyde substrates [81]. Several bioluminescent reporter bacteriophage systems have been designed using the lux operon, luxCDABE. The LuxCDE proteins encode a fatty acid reductase complex, including reductase, synthetase, and transferase, responsible for providing the aldehyde as a substrate, and the LuxAB encodes luciferase α- and β-subunits, which feed the bioluminescent reaction [82,83]. Reporter bacteriophages containing the luxAB gene were the ﬁrst bioluminescent for the detection of Salmonella strains [84]. The luxAB (P22 luxAB) reporter gene without the luxCDE was used to detect S. Enteritidis up to 63 CFU/g in whole eggs [84]. The advantage of the luxAB reporter system is that it can avoid the toxicity and noise signals by emitting a physical signal. However, this system needs the speciﬁc substrates for bacterial detections [82]. To avoid this inconvenience, a complete set of luxABCDE operons was constructed for Salmonella Typhimurium detection in diﬀerent food matrices [83]. The reporter bacteriophage system could detect Salmonella up to 37 CFU/g in sliced pork, 22 CFU/g in iceberg lettuce, and 20 CFU/mL in pure culture. These reporter bacteriophages could be useful for diagnostics and rapid detection of Salmonella spp. in diﬀerent food samples with no substrates and with the reporter host required. Apart from the luciferase-based reporter bacteriophage systems, green ﬂuorescent protein (GFP) and β-galactosidase have also been used for the detection of foodborne pathogens. The GFP gene is originated from Aequorea victoria and has many advantages, including high stability and low toxicity [85,86]. The GFP-labeled PP01 bacteriophage (PP01-GFP) system was applied on the surface of E. coli O157:H7, which can emit a ﬂuorescent signal at an MOI of 1000 at 4 ◦C [87]. 3.4. Matrix-Assisted Laser Desorption/Ionization–Time-of-Flight Mass Spectrometry MALDI-TOF MS is now becoming a most common method for bacterial identiﬁcation, diﬀerentiating from the advantages of high throughput and rapidity and excepting a high cost of the initial installment [74,75]. MALDI-TOF MS has been applied for screening, identiﬁcation, and detection of foodborne bacteria, which can enhance food safety. The MALDI-TOF MS-based detection method depends on progeny bacteriophage proteins or peptides. In this method, progeny bacteriophage proteins are applied to the test plate with a UV-absorbing matrix and then ablated by a laser. After 60–120 min of bacteriophage ampliﬁcation, samples were used for analysis by MALDI-TOF MS. The parameters of MALDI-TOF MS, including matrix preparation, sample preparation, acid added to the matrix, growth medium, and setting parameters, were optimized and a standard protocol was set for the identiﬁcation of Salmonella subspecies, and consequently the classiﬁcation results were comparable to DNA sequence-based methods [76]. A whole-cell MALDI-TOF MS for rapid prescreening of S. enterica subspecies Enterica, isolates based on speciﬁc biomarker ions, rather than antigenic determinants that could reduce the sample numbers for subsequent serotyping analysis [77]. The MALDI-TOF MS combined with selective enrichment broth was developed for the identiﬁcation 6 of 22 Microorganisms 2019, 7, 570 of Salmonella sp. in clinical stool samples. The discrimination of bacteria species was mainly based on the comparison of peaks of peptides and small proteins with the reference database [78]. Strains of Salmonella and E. coli were simultaneously detected based on the characteristics of proteins by using two phages, MS2 and MPSS-1, respectively, [63]. The simultaneous detection of two bacteria, using MALDI-TOF MS coupled with bacteriophage ampliﬁcation, provides the possibility of three or more target detections, which may require the speciﬁc bacteriophage biomarkers. 4. Bacteriophage-Based Biosensors for Detecting Salmonella oorganisms 2019, 7, x FOR PEER REVIEW Over the last few decades, biosensors have been developed as a novel analytical platform for pathogen detection [89,90]. A classical biosensor can be deﬁned as an analytical device that measures biological responses by incorporating bioreceptors (antibodies, enzymes, cells, aptamers, bacteriophage, and organelle) with physical transducers and electrochemical (amperometric, impedimetric, and potentiometric), optical (surface plasmon resonance, surface-enhanced resonance spectroscopy, and ﬂuorescence), and mass-based receptors (magnetoelastic and piezoelectric). The continuous eﬀorts have been successfully developed on bacteriophage-based biosensors for the detection of Salmonella in food samples. The immobilization of bacteriophage receptors on the sensor surface is crucial to develop bacteriophage-based biosensors. The immobilization steps include physical adsorption, covalent attachment, and genetic modiﬁcation of receptors [91,92]. Over the last few decades, biosensors have been developed as a novel analytical platform for hogen detection [89,90]. A classical biosensor can be defined as an analytical device that measures ogical responses by incorporating bioreceptors (antibodies, enzymes, cells, aptamers, teriophage, and organelle) with physical transducers and electrochemical (amperometric, edimetric, and potentiometric), optical (surface plasmon resonance, surface-enhanced resonance ctroscopy, and fluorescence), and mass-based receptors (magnetoelastic and piezoelectric). The tinuous efforts have been successfully developed on bacteriophage-based biosensors for the ection of Salmonella in food samples. The immobilization of bacteriophage receptors on the sensor ace is crucial to develop bacteriophage-based biosensors. The immobilization steps include sical adsorption covalent attachment and genetic modification of receptors [91 92] Appendix A summarizes diﬀerent bacteriophage-based biosensors for the detection of Salmonella. The ﬁrst transducers of bacteriophage biosensors for Salmonella detection are mass-based transducers and magnetoelastic assays (ME), and Figure 3 shows a schematic illustration for the principle of ME biosensors for detection of target analytes [93]. sical adsorption, covalent attachment, and genetic modification of receptors [91,92]. Table S1. summarizes different bacteriophage-based biosensors for the detection of Salmonella. first transducers of bacteriophage biosensors for Salmonella detection are mass-based transducers magnetoelastic assays (ME), and Figure 3 shows a schematic illustration for the principle of ME sensors for detection of target analytes [93] Figure 3. Schematic figure illustrating the working principle of ME biosensors. [93]. Figure 3. Schematic ﬁgure illustrating the working principle of ME biosensors. [93]. hematic figure illustrating the working principle of ME biosensors. [93]. Figure 3. Schematic ﬁgure illustrating the working principle of ME biosensors. [93]. 3.5. Genetically-Engineered Phages Although the speciﬁcity and host range are well deﬁned, this system is not widely used in food because of the interference of food components. The lacZ gene encoding β-galactosidase can catalyze the hydrolysis of β-galactosides. The lacZ-based reporter bacteriophage needs various substrates, such as colorimetric, ﬂuorescent, or luminescent substrates, which emit signals to detect bacteria [36,88]. The detection limits of this system were up to 103 CFU/100 cm2 for the colorimetric method and 10 CFU/100 cm2 for luminescence in beef slice samples [88]. The use of additional substrates can be a drawback of this system as it only allows single time point measurements. Although this reporter bacteriophage system can detect viable Salmonella at high speciﬁcity and a low detection limit, the construction of new reporter bacteriophages is still diﬃcult. Therefore, further studies are needed for the development of reporter bacteriophage systems, which can detect Salmonella spp. in various food matrices. 7 of 22 Microorganisms 2019, 7, 570 7 of 22 4. Bacteriophage-Based Biosensors for Detecting Salmonella oorganisms 2019, 7, x FOR PEER REVIEW ME-based detection methods are the most prominent type of biosensors due to their easy and ap fabrication, composing of amorphous ferromagnetic ribbon that contracts and expands when osed to the external magnetic field and generates magnetic fluxes by binding targets to the sensor ace. The ME biosensor was employed as a transduction platform in bacteriophage biosensors for detection of Salmonella (Table S1). ME biosensors were developed by using filamentous teriophage specific for S. Typhimurium. The bacteriophages were immobilized by a physical orption method for S. Typhimurium detection by the changes in the resonance frequency of the sor [93]. The numbers of S. Typhimurium in fresh tomato surfaces were quantified using the ME sensor detection method by the immobilization of E2 bacteriophages on the sensor surface [94]. tomato surfaces contaminated with S. Typhimurium were measured using a resonance uency with a detection limit of 103 CFU/mL. The results show that E2-bacteriophage-based sensors could detect Salmonella directly on the surface of tomatoes. The same principle for ection of S. Typhimurium in spinach leaves showed a similar detection limit as low as 102 CFU/mL ME-based detection methods are the most prominent type of biosensors due to their easy and cheap fabrication, composing of amorphous ferromagnetic ribbon that contracts and expands when exposed to the external magnetic ﬁeld and generates magnetic ﬂuxes by binding targets to the sensor surface. The ME biosensor was employed as a transduction platform in bacteriophage biosensors for the detection of Salmonella (Table A1). ME biosensors were developed by using ﬁlamentous bacteriophage speciﬁc for S. Typhimurium. The bacteriophages were immobilized by a physical adsorption method for S. Typhimurium detection by the changes in the resonance frequency of the sensor [93]. The numbers of S. Typhimurium in fresh tomato surfaces were quantiﬁed using the ME biosensor detection method by the immobilization of E2 bacteriophages on the sensor surface [94]. The tomato surfaces contaminated with S. Typhimurium were measured using a resonance frequency with a detection limit of 103 CFU/mL. The results show that E2-bacteriophage-based biosensors could detect Salmonella directly on the surface of tomatoes. The same principle for detection of S. Typhimurium in spinach leaves showed a similar detection limit as low as 102 CFU/mL [94]. ]. 5. Bacteriophage-Based Tool for Salmonella Control Bacteriophages have been used for controlling bacterial infections based on their speciﬁcity to the host bacteria [35,97–103]. Bacteriophages kept stable in thermal conditions from 30 to 60 ◦C and pH ranges from 3 to 13 can suggest the possibility of using bacteriophages in variable conditions. Recently, bacteriophages as a biocontrol tool have gained great attention and are recognized as an alternative for antibiotics [104–106]. Listeria bacteriophages were approved by the Food and Drug Administration (FDA) and the United States Department of Agriculture (USDA) in all food products, which were granted as generally recognized as safe (GRAS) [107]. Appendix B summarizes the applications of bacteriophage or bacteriophage-based treatments as biocontrol tools for Salmonella. Bacteriophage control technique has been applied for Salmonella in vivo and food samples. When lytic bacteriophages was applied to the chicken skin contaminated with S. enterica serovar Enteritidis, less than one log reduction was obtained at the MOI of 1 and no viable bacteria were observed at the MOI of 105 [108]. A new virulent bacteriophage, F01-E2, was isolated for controlling S. Typhimurium [109]. F01 belongs to Myoviridae with a double-stranded deoxyribonucleic acid dsDNA genome of 86.2 kb and a broad host range [110,111]. Five log reductions were obtained for turkey deli meat and chocolate milk at 15 ◦C, and three log reductions were observed for hot dogs and seafood, implying that bacteriophage immobilized on the food surfaces were aﬀected by the structure and chemical composition of the foods [112]. Salmonella Enteritidis bacteriophage SE07 showed a potential against S. Enteritidis in both solid and liquid food [113]. The isolated SE07 belongs to Podoviridae and is stable from 28 ◦C to 65 ◦C and pH 4 to 11. As shown in Table 3, two log reductions were obtained for the diﬀerent food matrices after 48 h incubation at 4 ◦C. Additionally, the bacteriophage ΦCJ07 was applied for controlling S. Enteritidis in chicken [114]. Because Salmonella in contaminated, chickens can survive under the acidic conditions in the digestion system. Bacteriophage ΦCJ07 was added as a feed additive, which was eﬀective against Salmonella by protection from other ingested feed constituents [115]. The bacteriophage ΦCJ07 isolated from the sewage eﬄuent showed a lytic activity against most Salmonella spp., including S. Enteritidis, S. Typhimurium, Salmonella Gallinarum, Salmonella Pullorum, Salmonella Choleraesuis, and Salmonella Derby. Further evaluations in vivo also demonstrated the good performance of phage ΦCJ07 in reducing both S. 4. Bacteriophage-Based Biosensors for Detecting Salmonella oorganisms 2019, 7, x FOR PEER REVIEW Other bacteriophage-based biosensor detection systems have been developed using an acoustic ve piezoelectric biosensor combined with filamentous bacteriophages [95] The detection limit was Other bacteriophage-based biosensor detection systems have been developed using an acoustic wave piezoelectric biosensor combined with ﬁlamentous bacteriophages [95]. The detection limit was 8 of 22 Microorganisms 2019, 7, 570 102 CFU/mL for S. Typhimurium by measuring the changes in resonance frequency as a consequence of binding bacteria to the bacteriophage. In addition, the recombinant prophage coupled with a ﬂow cytometer and speciﬁc ﬂuorescence ﬁlter was used for sensitive and speciﬁc detection of Salmonella with a detection limit of 10 CFU/mL [86]. A surface enhanced Raman scattering (SERS) by conjugating bacteriophage tail spike proteins to silica-encapsulated Raman reporter-embedded nanoprobes could detect single Salmonella cells [96]. Therefore, the use of bacteriophage as a bioreceptor in biosensors can contribute to the development of desirable detection tools for Salmonella in food samples. The stability, low cost, environment-friendly production, and genetic modiﬁcation provide beneﬁts for biosensor development. For the successful development of biosensors, the immobilization of phage onto the biosensor surface plays an important role. The genetically modiﬁed phages provide eﬀective immobilization by introducing the functional ligands on their heads. In addition, the ability to manipulate the genetic material provides the possibility of creating novel recognition systems for biosensor applications, such as expanding the host range of phages by manipulating the receptor-binding protein [27]. However, further work should focus on detecting Salmonella in complex food matrices for extending the range of application of bacteriophage-based diagnostic tools from the laboratory to clinical diagnosis, environmental monitoring, and further food analysis in the near future. 1 The bacterial inoculums were 3.2 and 4.2 log colony-forming unit (CFU)/g, respectively, for 18 ◦C and 4 ◦C.2 The phage titers were 7 and 8 log10 plaque-forming unit (PFU)/g, respectively, for 18 ◦C and 4 ◦C. 3 (–) indicates the control samples without phage and (+) denotes the samples treated with the phage cocktail. [116]. 5.1. Phage Cocktails Despite the advantages of bacteriophages, bacteria can become resistance to bacteriophages through surface modiﬁcation, superinfection exclusion, restriction modiﬁcation, abortive infection, and clustered regularly interspaced short palindromic repeatsCRISPR-associated 9 (Cas9) systems [117,118]. Therefore, phage cocktails with diﬀerent host speciﬁcities have been of great interest and are more practical for expanding the bacteriophage application, since the combined bacteriophage cocktails can reduce the development of bacteriophage-resistant mutants. p g p p g A mixture of two bacteriophages was used for controlling Salmonella in sprout seeds [119]. Bacteriophage A belongs to the Myoviridae family, while bacteriophage B is a member of the Siphoviridae family. The reductions of S. Typhimurium, S. Enteritidis, and Salmonella Montevideo were noticeable at the bacteriophage mixture of A and B in broccoli seeds compared to single bacteriophage treatment. The isolated bacteriophages eﬀectively reduced the numbers of S. Typhimurium and S. Enteritidis in chickens [120]. The mixture of three phages, UAB_Phi20, Phi78, and Phi87, showed higher lytic activity than that obtained by any of the three phages alone, while the phage cocktail lysed Salmonella Virchow, Salmonella Hadar, Salmonella Infantis, S. Typhimurium, and S. Enteritidis, showing a broad spectrum lytic capability. The bacteriophage cocktail was applied in diﬀerent food systems [121]. Signiﬁcant reductions of S. Typhimurium and S. Enteritidis were observed for diﬀerent food matrices, including pig skin, chicken breasts, and lettuce. Recently, a phage cocktail (BSPM4, BSP101, and BSP22A) based on targeting diﬀerent cell surface receptors, including ﬂagella, O-antigen, and BtuB, has been developed for the inhibition of Salmonella Typhimurium from fresh produce foods [122]. The multiple receptor-targeting bacteriophage cocktail can reduce Salmonella by up to 4.7–5.5 log CFU/cm2 in iceberg lettuce and 4.8–5.8 log CFU/cm2 in cucumber after 12 h incubation at 25 ◦C, without the development of bacteriophage resistance [122]. At present, the commercial bacteriophage cocktail has been applied for controlling Salmonella in poultry products [123]. 5. Bacteriophage-Based Tool for Salmonella Control Enteritidis colonization and environment contamination levels. This provides a promising alternative of bacteriophage for preventing and controlling S. Enteritidis infections and reducing the incidence of Salmonella food poisoning. 9 of 22 Microorganisms 2019, 7, 570 Table 3. Eﬃcacy of the bacteriophage cocktail in the reduction of Salmonella Enteritidis in raw salmon ﬁllets and smoked salmon slices. Incubation (day) Phage Cocktail Addition 3 Food Sample S. Enteritidis (log CFU/g) 1 Reduction (log CFU/g) Phage cocktail (log PFU/g) 2 18 ◦C 4 ◦C 18 ◦C 4 ◦C 18 ◦C 4 ◦C 3 − Raw salmon ﬁllet 7.51 ± 0.16 4.76 ± 0.20 + 6.76 ± 1.20 1.64 ± 0.36 0.75 3.12 6.57± 0.24 9.32 ± 0.23 6 − 6.70 ± 0.60 5.07 ± 0.17 + 4.13 ± 0.95 2.24 ± 0.45 2.57 2.83 7.32 ± 0.27 9.04 ± 1.82 10 − 5.90 ± 0.49 3.12 ± 0.45 + 2.71 ± 0.98 0.30 ± 0.43 3.19 2.82 7.80 ± 0.40 9.68 ± 0.39 3 − Smoked salmon slice 8.23 ± 0.13 3.84 ± 0.08 + 6.54 ± 0.28 3.34 ± 0.18 1.69 0.5 7.30 ± 0.37 8.32 ± 0.23 6 − 8.34 ± 0.15 3.73 ± 0.26 + 7.32 ± 0.37 3.38 ± 0.19 1.02 0.35 6.61 ± 0.36 8.80 ± 0.07 10 − 6.96 ± 0.42 2.28 ± 0.24 + 5.0 ± 0.48 1.12 ± 0.32 1.96 1.16 6.27 ± 0.19 8.66 ± 0.33 1 The bacterial inoculums were 3.2 and 4.2 log colony-forming unit (CFU)/g, respectively, for 18 ◦C and 4 ◦C.2 The phage titers were 7 and 8 log10 plaque-forming unit (PFU)/g, respectively, for 18 ◦C and 4 ◦C. 3 (–) indicates the control samples without phage and (+) denotes the samples treated with the phage cocktail. [116]. ophage cocktail in the reduction of Salmonella Enteritidis in raw salmon ﬁllets and smoked salmon slices. 1 The bacterial inoculums were 3.2 and 4.2 log colony-forming unit (CFU)/g, respectively, for 18 ◦C and 4 ◦C.2 The phage titers were 7 and 8 log10 plaque-forming unit (PFU)/g, respectively, for 18 ◦C and 4 ◦C. 3 (–) indicates the control samples without phage and (+) denotes the samples treated with the phage cocktail. [116]. 10 of 22 10 of 22 Microorganisms 2019, 7, 570 5.3. Phage Control Combined with Other Preservatives 5.3. Phage Control Combined with Other Preservatives 5.3. Phage Control Combined with Other Preservatives The hurdle concept (or barrier technology) is applied to foods to enhance the microbiological safety and quality. Many preservative methods are employed, together with other barriers to eﬀectively control microbial contamination in food [130]. Many studies have demonstrated bacteriophages as alternative antimicrobials to control bacteria. Several Salmonella bacteriophages, such as Salmonelex™, SalmoFreshTM, and SalmoProTM, have been approved as GRAS by the United States Food and Drug Administration (US FDA) and the US Department of Agriculture’s Food Safety and Inspection Service (USDA-FSIS) [131,132]. The combinations of bacteriophages and antimicrobials or sequential applications showed an eﬀective biocontrol ability against the target bacteria [38,133]. SalmoFreshTM bacteriophages combined with cetylpyridinium chloride (CPC) or lauric arginate (LAE) showed more than 5 log reductions against Salmonella spp. in chicken products [133–136]. However, an in vivo test on chicken breast ﬁllets showed that a lower number of Salmonella (0.5 to 1.3 log CFU/g) was reduced by the combinations of bacteriophages with CPC or LAE, which may be attributed to the complex matrix of the meat components [137]. Sequential treatment of chlorine, CPC, LAE, or peracetic acid (PAA) with concentrations of 50 and 400 ppm, respectively, followed by phage spray, were carried out to evaluate the hurdle eﬀect of Salmonella on chicken skin. The high reductions of 1.7 to 2.2 and 2.2 to 2.5 log CFU/cm2 were obtained with an immersion in 50 and 400 ppm of PAA, followed by phage spray, which may be used in industries for the reduction of Salmonella contamination in cut meat. With growing interest of the combinations of bacteriophages and antimicrobials, further studies are needed to evaluate the inhibitory eﬀect of antimicrobials combined with bacteriophages, the potential synergistic eﬀect of the combination, and the mode of phage application, such as immersion and spraying [38,138–140]. 5.2. Phage Endolysins Bacteriophage endolysins have been used as a novel biocontrol agent and natural food preservatives over the past decades. The endolysins are peptidoglycan hydrolases that can lyse host cells after phage replication and propagation. The endolysins are mainly active against Gram-positive bacteria, which do not contain an outer membrane [36,124]. The outer membrane of Gram-negative bacteria can prevent contact between free endolysins and peptidoglycan. However, some Salmonella bacteriophage endolysins can bypass the outer membrane barriers when combined with diﬀerent outer membrane permeabilizers, such as ethylene diamine tetra-acetic acid (EDTA), citrate, and malate [114,125, 126]. The bactericidal activity of a Salmonella phage endolysin (Lys68) combined with organic acids was increased against Gram-negative bacteria [127]. A Salmonella bacteriophage endolysin, Gp110, has currently proved to show enzymatic activity [128]. In addition, bacteriophage endolysins have also been engineered to increase the bactericidal eﬀect against Gram-negative bacteria. The modiﬁed endolysin combined with lipopolysaccharide (LPS)-destabilizing peptides showed promising results against Pseudomonas aeruginosa, showing more than 5 log CFU/mL reduction. However, there are still some limited eﬀects against Salmonella Typhimurium (<1 log CFU/mL reduction) [129]. Although many bacteriophage endolysins have been introduced and characterized, further optimization is still needed to increase the host speciﬁcity and lytic activity. The genetic engineering endolysin can be one of the useful approaches for satisfying these requirements. 11 of 22 11 of 22 Microorganisms 2019, 7, 570 Conﬂicts of Interest: The authors declare no conﬂict of interest. 6. Conclusions Notably, research related to bacteriophages and their promising applications has increased in recent decades due to frequent outbreaks and the emergence of antibiotic-resistant bacteria. The eﬀective detection and biocontrol of Salmonella, based on the potential bacteriophages, are of importance to reduce the incidence of Salmonella and ensure the food safety. Since many studies have been performed in the laboratory with well-controlled conditions, bacteriophages showed a signiﬁcant eﬀect on the inhibition of bacteria both in vivo and in vitro. Novel hurdle technology-coupled phages with antimicrobials, UV, or antagonistic bacteria are of interest to ﬁnd synergistic eﬀects against pathogens, which provide potential eﬀective ways to be used in industries for control pathogens and alleviate the risk of pathogen contaminations in foods. Author Contributions: M.J.U. conducted all experiments and also contributed to the writing and preparation of the manuscript. J.A. contributed to the experimental design, data interpretation, and manuscript writing. All authors read and approved the ﬁnal manuscript. S.W. worked on the bacteriophage detection. R.C., M.R., and D.H.O. collected and interpreted data associated with phage control. M.J.U. updated recent information and J.A. drafted and revised the manuscript. Funding: This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (NRF-2016R1D1A3B01008304). Conﬂicts of Interest: The authors declare no conﬂict of interest. 12 of 22 Microorganisms 2019, 7, 570 Appendix A Appendix A Table A1. Bacteriophage based sensors for detection of Salmonella. Table A1. Bacteriophage based sensors for detection of Salmonella. Table A1. Bacteriophage based sensors for detection of Salmonella. Transducer Phage Type Phage Immobilization Analyte Sample Detection Limit (CFU/mL) Linear Range (CFU/mL) Reference Magnetoelastic E2 phage Physical adsorption S. typhimurium Tomato surface 5 × 102 5 × 101–5 × 108 [141] Magnetoelastic E2 phage Physical adsorption S. typhimurium Culture 5 × 103 5 × 103–5 × 107 [142] Magnetoelastic C4-22 Physical adsorption and cysteine S. typhimurium Chicken 7.9 × 103 - [143] Magnetoelastic E2 phage - - Tomato surface 1.5 × 103 1.5 × 100–1.5 × 106 [144] Magnetoelastic E2 phage - - Soil 102 104–107 [145] Magnetoelastic E2 phage - - Romaine lettuce 5 × 102 101–108 [146] Capacitive M13 phage clone Phage / Pty/Au electrode using glutaraldehyde linker Salmonella spp. Chicken 2 × 102 2 × 102–1 × 107 [147] Magnetoelastic E2 phage Physical adsorption S. 6. Conclusions typhimurium Fat free milk 5 × 103 - [147] Magnetoelastic E2 phage Physical adsorption S. typhimurium Tomato surface - - [148] SPR M13 Phage derived peptide Phage /Au surface using 1-ethyl−3-(3-dimethyl- aminopropyl) carbodiimid linker S. typhimurium Culture 103 - [149] SPR M13 phages Phage /Au surface using EDC/NHS linker Salmonella spp. Culture 1.3 × 107 - [150] Maxtek acoustic wave device Filamentous phage Physical adsorption S. typhimurium Culture 101 101–107 [95] Microcantilevers M13 phage-derived peptides Phage / Au surface using succinimidyl propionate linker Salmonella spp. Culture 1 × 106 1 × 106−1 × 108 [151] Magnetoelastic E2 phage Physical adsorption S. typhimurium Tomato surface - 102–104 [152] Magnetoelastic Filamentous phage Physical adsorption S. typhimurium Culture 103 5 × 103–5 × 106 [93] SPR P22 Phage TSP Phage /Au surface using EDC/NHS linker Salmonella Culture 103 - [153] Bioluminescence Felix phage or Newport phage - S. newport Culture 103 - [154] oluminescence phage SJ2 - S. enteritidis Culture 103 - [155] Magnetoelastic E2 phage - S. typhimurium Tomato surface 103 103 × 107 [94] Fluorescent Recombinant prophage - S. typhimurium Sea water 10 - [86] Magnetoelastic E2 phage - S. typhimurium Spinach 102 - [156] Magnetoelastic E2 phage Physical adsorption S. typhimurium Eggshells 1.6 × 102 1.6–1.6 × 107 [157] Magnetoelastic Filamentous E2 phage Physical adsorption S. typhimurium Spinach 102 - [39] 13 of 22 Microorganisms 2019, 7, 570 Appendix B Table A2. Applications of bacteriophage or bacteriophage-based treatments for biocontrol of Salmonella. Phage Type Phage Characteristic Target Related Samples Concentration of Phage Treatment Mode Eﬃcacy References One phage Virulent phage F01-E2 Myoviridae family, 86.2 kb dsDNA genome S. Typhimurium RTE foods including Hot dogs, cooked and sliced turkey breast, mixed seafood, chocolate milk, and egg yolk 3 × 108 pfu/g Directly adding in the samples. At 8 ◦C, more than 3 log reduction resulted in no viable cells in all samples; while at 15 ◦C, 5 log reduction on turkey deli meat and in chocolate milk, and by 3 logs on hot dogs and in seafood. Reduction eﬀect only obtained after 2 days in egg yolk. [109] Phage phSE-1 All three belong to order Caudovirales and Siphoviridae family S. Typhimurium In vivo test 107 pfu/mL with a MOI of 100 Directly mixing. 6. Conclusions Signiﬁcant reductions of 1.8, 1.7 and 1.9 log CFU/mL were observed with phSE-1, phSE-2, and phSE-5 respectively [158] Phage phSE-2 Phage phSE-5 A phage cocktail A phage cocktail of UAB_Phi 20, 78, and 87) UAB_Phi 20 and 78 belong to Siphoviridae family and UAB_Phi 87 is a member of Myoviridae family S. Typhimurium and S. Enteritidis Pig skin, chicken breasts, fresh eggs, and packaged lettuce 1010 pfu/mL for pig skin and fresh eggs, 109 pfu/mL for chicken breasts and lettuce Spraying for pig skin and fresh eggs, while agitation 5 min and 60 min for chicken breasts, and lettuce, respectively In pig skin, >4 and 2 log/cm2 for S. Typhimurium and S. Enteritidis were reduced for 6 h, respectively; in chicken breasts, 2.2 and 0.9 log cfu/g for S. Typhimurium and S. Enteritidis were reduced for 7 days, respectively; in lettuce, 9 and 2.2 log cfu/g, respectively; in fresh eggs, a reduction of 0.9 log cfu/cm2 for 2h [120,121] SalmoFreshTM Commercial product S. Newport Whole and fresh-cut cucumbers 1010 pfu/mL Spraying S. Newport was signiﬁcantly lower when treated by phage at 10 °C on day 1 and 4. [159] A phage cocktail of S16 and FO1a Both belong to the order Caudovirales and Myoviridae family S. enterica, S. Heidelberg, S. Newport, and S. Enteritidis C, Se 13 Ground meat including beef and pork trim, and poultry including chicken and turkey thighs 107 or 108 pfu/mL for samples, and 109 pfu/mL for vitro study Tumbling for 2 min at 4 rpm In vitro study, 99% were reduced for all strains; in vivo test, bacteria reductions of 1, 0.8, 1.1 and 0.9 log cfu/g were obtained in beef, pork, chicken, and turkey, respectively. [160] A phage cocktail of vB_SnwM_CGG4-1, 4-2, 3-1, and 3-2 vB_SnwM_CGG4-1, and 4-2 belong to Myoviridae family and vB_SnwM_CGG3-1, and 3-2 belong to Siphoviridae family S. Newport Cherry tomato 106 and 108 pfu/mL In vitro study, 3 log reduction was obtained after up to 7 h incubation; in vivo test, 2 log reduction with a MOI of 103 and about 4.4 log reduction was observed after 2, 3, and 4 days with a MOI of 105 [161] SalmoFreshTM Myoviridae family S. Newport, S. Braenderup, S. Typhimurium, S. Kentucky, and S. References 1. Scallan, E.; Hoekstra, R.M.; Angulo, F.J.; Tauxe, R.V.; Widdowson, M.A.; Roy, S.L.; Jones, J.L.; Griﬃn, P.M. Foodborne illness acquired in the United States-Major pathogens. Emerg. Infect. Dis. 2011, 17, 7–15. [CrossRef] [PubMed] 1. Scallan, E.; Hoekstra, R.M.; Angulo, F.J.; Tauxe, R.V.; Widdowson, M.A.; Roy, S.L.; Jones, J.L.; Griﬃn, P.M. Foodborne illness acquired in the United States-Major pathogens. Emerg. Infect. Dis. 2011, 17, 7–15. [CrossRef] [PubMed] 2. CDC. Salmonella. 2019. Available online: https://www.cdc.gov/salmonella/general/index.html#two (accessed on 1 October 2019). 2. CDC. Salmonella. 2019. Available online: https://www.cdc.gov/salmonella/general/index.html#two (accessed on 1 October 2019). 3. WHO. Foodborne Diseases, Salmonella. Available online: https://www.who.int/foodsafety/areas_work/ foodborne-diseases/salmonella/en/ (accessed on 1 October 2019). 3. WHO. Foodborne Diseases, Salmonella. Available online: https://www.who.int/foodsafety/areas_work/ foodborne-diseases/salmonella/en/ (accessed on 1 October 2019). 4. Chang, Q.; Wang, W.; Regev-Yochay, G.; Lipsitch, M.; Hanage, W.P. Antibiotics in agriculture and the risk to human health: How worried should we be? Evol. Appl. 2015, 8, 240–247. [CrossRef] [PubMed] 5. Agyare, C.; Boamah, V.E.; Zumbi, C.N.; Osei, F.B. Antibiotic use in poultry production and its eﬀects on bacterial resistance. In Antibiotic Resistance-A Global Threat; Kumar, Y., Ed.; IntechOpen: London, UK, 2019. 6. Medeiros, M.A.N.; de Oliveira, D.C.N.; dos Prazeres Rodrigues, D.; de Freitas, D.R.C. Prevalence and antimicrobial resistance of Salmonella in chicken carcasses at retail in 15 Brazillian cities. Rev. Panam. Salud Publica 2011, 30, 555–560. [CrossRef] [PubMed] 7. Khachatourians, G.G. Agricultural use of antibiotics and the evolution and transfer of antibiotic-resistant bacteria. Can. Med. Assoc. J. 1998, 159, 1129–1136. 8. Liu, Y.-Y.; Wang, Y.; Walsh, T.R.; Yi, L.-X.; Zhang, R.; Spencer, J.; Doi, Y.; Tian, G.; Dong, B.; Huang, X.; et al. Emergence of plasmid-mediated colistin resistance mechanism MCR-1 in animals and human beings in China: A microbiological and molecular biological study. Lancet Infect. Dis. 2016, 16, 161–168. [CrossRef] . Marshall, B.M.; Levy, S.B. Food animals and antimicrobials: Impacts on human health. Clin. Microbiol. 2011, 24, 718–733. [CrossRef] [PubMed] Founou, L.L.; Founou, R.C.; Essack, S.Y. Antibiotic resistance in the food chain: A developing country-Perspective. Front. Microbiol. 2016, 7, 1881. [CrossRef] 11. Campos, J.; Mourão, J.; Pestana, N.; Peixe, L.; Novais, C.; Antunes, P. Microbiological quality of ready-to-eat salads: An underestimated vehicle of bacteria and clinically relevant antibiotic resistance genes. Int. J. Food Microbiol. 2013, 166, 464–470. [CrossRef] 12. Schwaiger, K.; Helmke, K.; Hölzel, C.S.; Bauer, J. Antibiotic resistance in bacteria isolated from vegetables with regards to the marketing stage (farm vs. supermarket). Int. J. Food Microbiol. 6. Conclusions Heidelberg Lettuce, mung bean sprouts and seeds 108 pfu/mL Spraying, immersion, Reductions of 0.76 and 0.83 log10 CFU/g were obtained on lettuce and sprouts by spraying, respectively, while 2.43 and 2.16 log10 CFU/g by immersion. [162] Table A2. Applications of bacteriophage or bacteriophage-based treatments for biocontrol of Salmonella. References Pig skin, chicken breasts, fresh eggs, and packaged lettuce Whole and fresh-cut cucumbers Ground meat including beef and pork trim, and poultry including chicken and turkey thighs S. Newport, S. Braenderup, S. Typhimurium, S. Kentucky, and S. Heidelberg Spraying, immersion, 14 of 22 14 of 22 Microorganisms 2019, 7, 570 Table A2. Cont. Phage Type Phage Characteristic Target Related Samples Concentration of Phage Treatment Mode Eﬃcacy References Phage based hurdle treatment A cocktail of 6 phages including F01, P01, P102, P700, P800, and FL 41, combined with Enterobacter asburiae JX1 - S. Agona, S. Berta, S. Enteritidis, S. Hadar, S. Heidelberg, S. Javiana, S. Montevideo, S. Muenchen, S. Newport, S. Saint Paul, and S. Typhimurium DT104 Sprouting mung bean and alfalfa seeds 106 pfu/mL Soaking for 20 min In vivo, reduction of 5.7 to 6.4 log CFU/mL were obtained. In sprouting mung bean sprouts, an additive eﬀect was observed with the combination resulted in a detectable Salmonella only after enrichment. For sprouting alfalfa seeds, no Salmonella was recovered even with enrichment [163] A phage cocktail of S16 and FO1a combined with UV Both belong to the order Caudovirales and Myoviridae family S. Infantis, S. Heidelberg, S. Newport, and S. Enteritidis C, Se 13 Ground beef 109 pfu/mL Tumbling for 2 min at 4 rpm Approximately 1 log CFU/g reduction for bacteriophage and UV, separately, while 2 log CFU/g for combination [160,164] oFreshTM combined with chlorinated water Myoviridae family S. Newport, S. Braenderup, S. Typhimurium, S. Kentucky, and S. Heidelberg Lettuce, mung bean sprouts and seeds 108 pfu/mL Immersion 15 min for lettuce and sprouts, 1 h for mung seeds Reductions of 3.8, and 2.7, 1.28 log CFU/g were obtained by hurdle treatment on lettuce, sprouts, and mung seeds, respectively. [162] 15 of 22 15 of 22 Microorganisms 2019, 7, 570 References 2011, 148, 191–196. [CrossRef] 13. Berger, C.N.; Sodha, S.V.; Shaw, R.K.; Griﬃn, P.M.; Pink, D.; Hand, P.; Frankel, G. Fresh fruit and vegetables as vehicles for the transmission of human pathogens. Environ. Microbiol. 2010, 12, 2385–2397. [CrossRef] 14. WHO. Antibiotic Resistance. Available online: https://www.who.int/news-room/fact-sheets/detail/antibiotic- ( d S b 1 ) 13. Berger, C.N.; Sodha, S.V.; Shaw, R.K.; Griﬃn, P.M.; Pink, D.; Hand, P.; Frankel, G. Fresh fruit and vegetables as vehicles for the transmission of human pathogens. Environ. Microbiol. 2010, 12, 2385–2397. [CrossRef] g , ; , ; , ; , ; , ; , ; , g as vehicles for the transmission of human pathogens. Environ. Microbiol. 2010, 12, 2385–2397. [CrossRef] 14. WHO. Antibiotic Resistance. Available online: https://www.who.int/news-room/fact-sheets/detail/antibiotic- resistance (accessed on 22 September 2019). 14. WHO. Antibiotic Resistance. Available online: https://www.who.int/news-room/fact-sheets/detail/antibiotic- resistance (accessed on 22 September 2019). 15. Doss, J.; Culbertson, K.; Hahn, D.; Camacho, J.; Barekzi, N. A review of phage therapy against bacterial pathogens of aquatic and terrestrial organisms. Viruses 2017, 9, 50. [CrossRef] [PubMed] 16. Salmond, G.P.; Fineran, P.C. A century of the phage: Past, present and future. Nat. Rev. Microbiol. 2015, 13, 777–786. [CrossRef] [PubMed] 7. Twort, F.W. An investigation on the nature of ultramicroscopic viruses. Lancet 1915, 186, 1241–1243. [Cross 17. Twort, F.W. An investigation on the nature of ultramicroscopic viruses. Lancet 1915, 186, 1241–1243. [CrossRef] 18. D’Herelle, F. On an invisible microbe antagonistic toward dysenteric bacilli. Res. Microbiol. 2007, 158, 553–554. [PubMed] 18. D’Herelle, F. On an invisible microbe antagonistic toward dysenteric bacilli. Res. Microbiol. 2007, 158, 553–554. [PubMed] 19. Torres-Barceló, C. Phage therapy faces evolutionary challenges. Viruses 2018, 10, 323. [CrossRef] [PubMed] 19. Torres-Barceló, C. Phage therapy faces evolutionary challenges. Viruses 2018, 10, 323. [CrossRef] [PubMed] 20. Hungaro, H.M.; Mendonça, R.C.S.; Gouvêa, D.M.; Vanetti, M.C.D.; Pinto, C.L.D.O. Use of bacteriophages to reduce Salmonella in chicken skin in comparison with chemical agents Food Res Int 2013 52 75–81 19. Torres-Barceló, C. Phage therapy faces evolutionary challenges. Viruses 2018, 10, 323. [CrossRef] [PubMed] 20. Hungaro, H.M.; Mendonça, R.C.S.; Gouvêa, D.M.; Vanetti, M.C.D.; Pinto, C.L.D.O. Use of bacteriophages to reduce Salmonella in chicken skin in comparison with chemical agents. Food Res. Int. 2013, 52, 75–81. [CrossRef] 21. Sulakvelidze, A.; Alavidze, Z.; Morris, J.G. Bacteriophage therapy. Antimicrob. Agent Chemother. 2001, 45, 649–659. [CrossRef] [PubMed] 22. Summers, W.C. Bateriphage therapy. Ann. Rev. Microbiol. 2001, 55, 437–451. [CrossRef] 23. References Haq, I.; Chaudhry, W.; Akhtar, M.; Andleeb, S.; Qadri, I. Bacteriophages and their implications on future biotechnology: A review. Virol. J. 2012, 9, 9. [CrossRef] 24. Fischbach, M.A.; Walsh, C.T. Antibiotics for emerging pathogens. Science 2009, 325, 1089–1093. [CrossRef] 16 of 22 16 of 22 Microorganisms 2019, 7, 570 25. Alanis, A.J. Resistance to antibiotics: Are we in the post-antibiotic era? Arch. Med. Res. 2005, 36, 697–705. [CrossRef] [PubMed] 26. Clark, J.R.; March, J.B. Bacteriophages and biotechnology: Vaccines, gene therapy and antibacterials. Trends Biotechnol. 2006, 24, 212–218. [CrossRef] [PubMed] 27. Alcaine, S.D.; Law, K.; Ho, S.; Kinchla, A.J.; Sela, D.A.; Nugen, S.R. Bioengineering bacteriophages to enhance the sensitivity of phage ampliﬁcation-based paper ﬂuidic detection of bacteria. Biosens. Bioelectron. 2016, 82, 14–19. [CrossRef] [PubMed] 28. Balasubramanian, S.; Sorokulova, I.B.; Vodyanoy, V.J.; Simonian, A.L. Lytic phage as a speciﬁc and selective probe for detection of Staphylococcus aureus—A surface plasmon resonance spectroscopic study. Biosens. Bioelectron. 2007, 22, 948–955. [CrossRef] 29. Bárdy, P.; Pant˚uˇcek, R.; Benešík, M.; Doškaˇr, J. Genetically modiﬁed bacteriophages in applied microbiology. J. Appl. Microbiol. 2016, 121, 618–633. [CrossRef] 30. Altintas, Z.; Pocock, J.; Thompson, K.-A.; Tothill, I.E. Comparative investigations for adenovirus recognition and quantiﬁcation: Plastic or natural antibodies? Biosens. Bioelectron. 2015, 74, 996–1004. [CrossRef] 31. Wang, F.; Horikawa, S.; Hu, J.; Wikle, H.C.; Chen, I.H.; Du, S.; Liu, Y.; Chin, B.A. Detection of Salmonella Typhimurium on spinach using phage-based magnetoelastic biosensors. Sensors 2017, 17, 386. [CrossRef] 32. Sommer, J.; Trautner, C.; Witte, A.K.; Fister, S.; Schoder, D.; Rossmanith, P.; Mester, P.-J. Don’t shut the stable door after the phage has bolted-The importance of bacteriophage inactivation in food environments. Viruses 2019, 11, 468. [CrossRef] 33. Wittebole, X.; De Roock, S.; Opal, S.M. A historical overview of bacteriophage therapy as an alternative to antibiotics for the treatment of bacterial pathogens. Virulence 2014, 5, 226–235. [CrossRef] 34. Stewart, G.S.; Jassim, S.A.; Denyer, S.P.; Newby, P.; Linley, K.; Dhir, V.K. The speciﬁc and sensitive detection of bacterial pathogens within 4 h using bacteriophage ampliﬁcation. J. Appl. Microbiol. 1998, 84, 777–783. [CrossRef] 35. Hagens, S.; Loessner, M.J. Application of bacteriophages for detection and control of foodborne pathogens. Appl. Microbiol. Biotechnol. 2007, 76, 513–519. [CrossRef] [PubMed] 36. Schmelcher, M.; Loessner, M.J. Application of bacteriophages for detection of foodborne pathogens. Bacteriophage 2014, 4, e28137. [CrossRef] [PubMed] 37. Boyacioglu, O.; Sharma, M.; Sulakvelidze, A.; Goktepe, I. Biocontrol of Escherichia coli O157: H7 on fresh-cut leafy greens. References Bacteriophage 2013, 3, e24620. [CrossRef] [PubMed] 38. Ferguson, S.; Roberts, C.; Handy, E.; Sharma, M. Lytic bacteriophages reduce Escherichia coli O157: H7 on fresh cut lettuce introduced through cross-contamination. Bacteriophage 2013, 3, e24323. [CrossRef] [PubMed] 39. Wang, L.; Qu, K.; Li, X.; Cao, Z.; Wang, X.; Li, Z.; Song, Y.; Xu, Y. Use of bacteriophages to control Escherichia coli O157:H7 in domestic ruminants, meat products, and fruits and vegetables. Foodborne Pathog. Dis. 2017, 14, 483–493. [CrossRef] [PubMed] 40. Perera, M.N.; Abuladze, T.; Li, M.; Woolston, J.; Sulakvelidze, A. Bacteriophage cocktail signiﬁcantly reduces or eliminates Listeria monocytogenes contamination on lettuce, apples, cheese, smoked salmon and frozen foods. Food Microbiol. 2015, 52, 42–48. [CrossRef] 41. Baños, A.; García-López, J.D.; Núñez, C.; Martínez-Bueno, M.; Maqueda, M.; Valdivia, E. Biocontrol of Listeria monocytogenes in ﬁsh by enterocin AS-48 and Listeria lytic bacteriophage P100. LWT-Food Sci. Technol. 2016, 66, 672–677. [CrossRef] 42. Gutiérrez, D.; Rodríguez-Rubio, L.; Fernández, L.; Martínez, B.; Rodríguez, A.; García, P. Applicability of commercial phage-based products against Listeria monocytogenes for improvement of food safety in Spanish dry-cured ham and food contact surfaces. Food Control 2017, 73, 1474–1482. [CrossRef] 43. Grant, A.Q.; Parveen, S.; Schwarz, J.; Hashem, F.; Vimini, B. Reduction of Salmonella in ground chicken using a bacteriophage. Poult. Sci. 2017, 96, 2845–2852. [CrossRef] 44. Moye, Z.D.; Woolston, J.; Sulakvelidze, A. Bacteriophage applications for food production and processing. Viruses 2018, 10, 205. [CrossRef] 45. Soﬀer, N.; Woolston, J.; Li, M.; Das, C.; Sulakvelidze, A. Bacteriophage preparation lytic for Shigella signiﬁcantly reduces Shigella sonnei contamination in various foods. PLoS ONE 2017, 12, e0175256. [CrossRef] [PubMed] 46. Shahin, K.; Bouzari, M. Bacteriophage application for biocontrolling Shigella ﬂexneri in contaminated foods. J. Food Sci. Technol. 2018, 55, 550–559. [CrossRef] [PubMed] 17 of 22 17 of 22 Microorganisms 2019, 7, 570 47. Bernasconi, O.J.; Dona, V.; Tinguely, R.; Endimiani, A. In vitro activity of 3 commercial bacteriophage cocktails against Salmonella and Shigella spp. isolates of human origin. Pathog. Immun. 2018, 3, 72–81. [CrossRef] [PubMed] 48. Huang, C.; Shi, J.; Ma, W.; Li, Z.; Wang, J.; Li, J.; Wang, X. Isolation, characterization, and application of a novel speciﬁc Salmonella bacteriophage in diﬀerent food matrices. Food Res. Int. 2018, 111, 631–641. [CrossRef] [PubMed] 49. Lillehaug, D. An improved plaque assay for poor plaque-producing temperate lactococcal bacteriophages. J. Appl. Microbiol. 1997, 83, 85–90. [CrossRef] [PubMed] 50. Kropinski, A. Measurement of the rate of attachment of bacteriophage to cells. References In Bacteriophages; Clokie, M.J., Kropinski, A., Eds.; Humana Press: Totowa, NJ, USA, 2009; Volume 501, pp. 151–155. 51. Adams, M.H. Bacteriophages; Interscience Publishers Inc.: New York, NY, USA, 1959; pp. 14–15. 52. Cormier, J.; Janes, M. A double layer plaque assay using spread plate technique for enumeration of bacteriophage MS2. J. Virol. Methods 2014, 196, 86–92. [CrossRef] [PubMed] 53. Cornax, R.; Moriñigo, M.A.; Paez, I.G.; Muñoz, M.A.; Borrego, J.J. Application of direct plaque assay for detection and enumeration of bacteriophages of Bacteroides fragilis from contaminated-water samples. Appl. Environ. Microbiol. 1990, 56, 3170–3173. pp 54. Brown-Jaque, M.; Muniesa, M.; Navarro, F. Bacteriophages in clinical samples can interfere with microbiological diagnostic tools. Sci. Rep. 2016, 6, 33000. [CrossRef] 55. Burnham, S.; Hu, J.; Anany, H.; Brovko, L.; Deiss, F.; Derda, R.; Griﬃths, M.W. Towards rapid on-site phage-mediated detection of generic Escherichia coli in water using luminescent and visual readout. Anal. Bioanal. Chem. 2014, 406, 5685–5693. [CrossRef] 56. Petrenko, V.A.; Jayanna, P.K. Phage protein-targeted cancer nanomedicines. FEBS Lett. 2014, 588, 341–349. [CrossRef] 57. Kakasis, A.; Panitsa, G. Bacteriophage therapy as an alternative treatment for human infections. A comprehensive review. Int. J. Antimicrob. Agents 2019, 53, 16–21. [CrossRef] [PubMed] 58. de Jonge, P.A.; Nobrega, F.L.; Brouns, S.J.J.; Dutilh, B.E. Molecular and evolutionary determinants of bacteriophage host range. Trends Microbiol. 2019, 27, 51–63. [CrossRef] [PubMed] 59. Galikowska, E.; Kunikowska, D.; Tokarska-Pietrzak, E.; Dziadziuszko, H.; Los, J.M.; Golec, P.; Wegrzyn, G.; Los, M. Speciﬁc detection of Salmonella enterica and Escherichia coli strains by using ELISA with bacteriophages as recognition agents. Eur. J. Clin. Microbiol. Infect. Dis. 2011, 30, 1067–1073. [CrossRef] [PubMed] 60. Botsaris, G.; Liapi, M.; Kakogiannis, C.; Dodd, C.E.R.; Rees, C.E.D. Detection of Mycobacterium avium subsp. paratuberculosis in bulk tank milk by combined phage-PCR assay: Evidence that plaque number is a good predictor of MAP. Int. J. Food Microbiol. 2013, 164, 76–80. [CrossRef] 61. Bennett, A.R.; Davids, F.G.C.; Vlahodimou, S.; Banks, J.G.; Betts, R.P. The use of bacteriophage-based systems for the separation and concentration of Salmonella. J. Appl. Microbiol. 1997, 83, 259–265. [CrossRef] 62. Favrin, S.J.; Jassim, S.A.; Griﬃths, M.W. Development and optimization of a novel immunomagnetic separation-bacteriophage assay for detection of Salmonella enterica serovar enteritidis in broth. Appl. Environ. Microbiol. 2001, 67, 217–224. [CrossRef] 63. Rees, J.C.; Voorhees, K.J. Simultaneous detection of two bacterial pathogens using bacteriophage ampliﬁcation coupled with matrix-assisted laser desorption/ionization time-of-ﬂight mass spectrometry. Rapid Commun. Mass Spectrom. 2005, 19, 2757–2761. References [CrossRef] p 64. Smartt, A.E.; Xu, T.; Jegier, P.; Carswell, J.J.; Blount, S.A.; Sayler, G.S.; Ripp, S. Pathogen detection using engineered bacteriophages. Anal. Bioanal. Chem. 2012, 402, 3127–3146. [CrossRef] 64. Smartt, A.E.; Xu, T.; Jegier, P.; Carswell, J.J.; Blount, S.A.; Sayler, G.S.; Ripp, S. Pathogen detection using engineered bacteriophages. Anal. Bioanal. Chem. 2012, 402, 3127–3146. [CrossRef] 65. Kutin, R.K.; Alvarez, A.; Jenkins, D.M. Detection of Ralstonia solanacearum in natural substrates using phage ampliﬁcation integrated with real-time PCR assay. J. Microbiol. Methods 2009, 76, 241–246. [CrossRef] 65. Kutin, R.K.; Alvarez, A.; Jenkins, D.M. Detection of Ralstonia solanacearum in natural substrates using phage ampliﬁcation integrated with real-time PCR assay. J. Microbiol. Methods 2009, 76, 241–246. [CrossRef] 66. Luo, J.; Jiang, M.; Xiong, J.; Li, J.; Zhang, X.; Wei, H.; Yu, J. Exploring a phage-based real-time PCR assay for diagnosing Acinetobacter baumannii bloodstream infections with high sensitivity. Anal. Chim. Acta 2018, 1044, 147–153. [CrossRef] 66. Luo, J.; Jiang, M.; Xiong, J.; Li, J.; Zhang, X.; Wei, H.; Yu, J. Exploring a phage-based real-time PCR assay for diagnosing Acinetobacter baumannii bloodstream infections with high sensitivity. Anal. Chim. Acta 2018, 1044, 147–153. [CrossRef] 67. Wang, Z.; Wang, D.; Kinchla, A.J.; Sela, D.A.; Nugen, S.R. Rapid screening of waterborne pathogens using phage-mediated separation coupled with real-time PCR detection. Anal. Bioanal. Chem. 2016, 408, 4169–4178. [CrossRef] 67. Wang, Z.; Wang, D.; Kinchla, A.J.; Sela, D.A.; Nugen, S.R. Rapid screening of waterborne pathogens using phage-mediated separation coupled with real-time PCR detection. Anal. Bioanal. Chem. 2016, 408, 4169–4178. [CrossRef] 18 of 22 18 of 22 Microorganisms 2019, 7, 570 68. Garrido-Maestu, A.; Fuciños, P.; Azinheiro, S.; Carvalho, C.; Carvalho, J.; Prado, M. Speciﬁc detection of viable Salmonella Enteritidis by phage ampliﬁcation combined with qPCR (PAA-qPCR) in spiked chicken meat samples. Food Control 2019, 99, 79–83. [CrossRef] 69. Kumar, R.; Surendran, P.K.; Thampuran, N. Evaluation of culture, ELISA and PCR assays for the dete Salmonella in seafood. Lett. Appl. Microbiol. 2008, 46, 221–226. [CrossRef] [PubMed] 70. Schneid, A.D.S.; Rodrigues, K.L.; Chemello, D.; Tondo, E.C.; Ayub, M.A.Z.; Aleixo, J.A.G. Evaluation of an indirect ELISA for the detection of Salmonella in chicken meat. Braz. J. Microbiol. 2006, 37, 350–355. [CrossRef] 71. Perelle, S.; Dilasser, F.; Malorny, B.; Grout, J.; Hoorfar, J.; Fach, P. Comparison of PCR-ELISA and LightCycler real-time PCR assays for detecting Salmonella spp. in milk and meat samples. Mol. Cell. Probes 2004, 18, 409–420. [CrossRef] 72. References Croci, L.; Delibato, E.; Volpe, G.; Palleschi, G. A rapid electrochemical ELISA for the detection of Salmonella in meat samples. Anal. Lett. 2001, 34, 2597–2607. [CrossRef] 73. Hu, Y.-F.; Zhao, D.; Yu, X.-L.; Hu, Y.-L.; Li, R.-C.; Ge, M.; Xu, T.-Q.; Liu, X.-B.; Liao, H.-Y. Identiﬁcation of bacterial surface antigens by screening peptide phage libraries using whole bacteria cell-puriﬁed antisera. Front. Microbiol. 2017, 8, 82. [CrossRef] 74. Singhal, N.; Kumar, M.; Kanaujia, P.K.; Virdi, J.S. MALDI-TOF mass spectrometry: An emerging technology for microbial identiﬁcation and diagnosis. Front. Microbiol. 2015, 6, 791. [CrossRef] [PubMed] 75. Bell, R.L.; Jarvis, K.G.; Ottesen, A.R.; McFarland, M.A.; Brown, E.W. Recent and emerging innovations in Salmonella detection: A food and environmental perspective. Microb. Biotechnol. 2016, 9, 279–292. [CrossRef] 76. Dieckmann, R.; Helmuth, R.; Erhard, M.; Malorny, B. Rapid classiﬁcation and identiﬁcation of Salmonella at the species and subspecies levels by whole-cell matrix-assisted laser desorption ionization-time of ﬂight mass spectrometry. Appl. Environ. Microbiol. 2008, 74, 7767–7778. [CrossRef] 77. Dieckmann, R.; Malorny, B. Rapid screening of epidemiologically important Salmonella enterica subsp. enterica serovars by whole-cell matrix-assisted laser desorption ionization–time of ﬂight mass spectrometry. Appl. Environ. Microbiol. 2011, 77, 4136–4146. [CrossRef] [PubMed] 78. Sparbier, K.; Weller, U.; Boogen, C.; Kostrzewa, M. Rapid detection of Salmonella sp. by means of a combination of selective enrichment broth and MALDI-TOF MS. Eur. J. Clin. Microbiol. Infect. Dis. 2012, 31, 767–773. [CrossRef] [PubMed] 79. Pires, D.P.; Cleto, S.; Sillankorva, S.; Azeredo, J.; Lu, T.K. Genetically engineered phages: A Review of advances over the last decade. Microbiol. Mol. Biol. Rev. 2016, 80, 523–543. [CrossRef] [PubMed] 80. Chen, J.; Alcaine, S.D.; Jiang, Z.; Rotello, V.M.; Nugen, S.R. Detection of Escherichia coli in drinking water using T7 bacteriophage-conjugated magnetic probe. Anal. Chem. 2015, 87, 8977–8984. [CrossRef] 81. Kim, J.; Kim, M.; Kim, S.; Ryu, S. Sensitive detection of viable Escherichia coli O157:H7 from foods using a luciferase-reporter phage phiV10lux. Int. J. Food Microbiol. 2017, 254, 11–17. [CrossRef] 82. Franche, N.; Vinay, M.; Ansaldi, M. Substrate-independent luminescent phage-based biosensor to speciﬁcally detect enteric bacteria such as E. coli. Environ. Sci. Pollut. Res. 2017, 24, 42–51. [CrossRef] 83. Kim, S.; Kim, M.; Ryu, S. Development of an engineered bioluminescent reporter phage for the sensitive detection of viable Salmonella Typhimurium. Anal. Chem. 2014, 86, 5858–5864. [CrossRef] 84. Chen, J.; Griﬃths, M.W. Salmonella detection in eggs using LuX+ bacteriophages. J. Food Prot. 1996, 59, 908–914. References [CrossRef] , J ; , gg g p g J , , [CrossRef] 85. Tanji, Y.; Furukawa, C.; Na, S.-H.; Hijikata, T.; Miyanaga, K.; Unno, H. Escherichia coli detection by GFP-labeled [CrossRef] 85. Tanji, Y.; Furukawa, C.; Na, S.-H.; Hijikata, T.; Miyanaga, K.; Unno, H. Escherichia coli detection by GFP-labeled lysozyme-inactivated T4 bacteriophage J Biotechnol 2004 114 11–20 [CrossRef] 85. Tanji, Y.; Furukawa, C.; Na, S.-H.; Hijikata, T.; Miyanaga, K.; Unno, H. Escherichia coli detection by GFP lysozyme-inactivated T4 bacteriophage. J. Biotechnol. 2004, 114, 11–20. [CrossRef] 86. Vinay, M.; Franche, N.; Grégori, G.; Fantino, J.-R.; Pouillot, F.; Ansaldi, M. Phage-based ﬂuorescent biosensor prototypes to speciﬁcally detect enteric bacteria such as E. coli and Salmonella enterica Typhimurium. PLoS ONE 2015, 10, e0131466. [CrossRef] 87. Oda, M.; Morita, M.; Unno, H.; Tanji, Y. Rapid detection of Escherichia coli O157:H7 by using green ﬂuorescent protein-labeled PP01 bacteriophage. Appl. Environ. Microbiol. 2004, 70, 527–534. [CrossRef] [PubMed] 88. Goodridge, L.D.; Willford, J. An integrated assay for rapid detection of Escherichia coli O157:H7 on beef samples. Food Prot. Trends 2008, 28, 468–472. 89. Rubab, M.; Shahbaz, H.M.; Olaimat, A.N.; Oh, D.-H. Biosensors for rapid and sensitive detection of Staphylococcus aureus in food. Biosens. Bioelectron. 2018, 105, 49–57. [CrossRef] [PubMed] 90. Hameed, S.; Xie, L.; Ying, Y. Conventional and emerging detection techniques for pathogenic bacteria in food science: A review. Trends Food Sci. Technol. 2018, 81, 61–73. [CrossRef] 19 of 22 Microorganisms 2019, 7, 570 91. Singh, A.; Arutyunov, D.; Szymanski, C.M.; Evoy, S. Bacteriophage based probes for pathogen detection. Analyst 2012, 137, 3405–3421. [CrossRef] 92. Tawil, N.; Sacher, E.; Mandeville, R.; Meunier, M. Bacteriophages: Biosensing tools for multi-drug resistant pathogens. Analyst 2014, 139, 1224–1236. [CrossRef] 93. Lakshmanan, R.S.; Guntupalli, R.; Hu, J.; Petrenko, V.A.; Barbaree, J.M.; Chin, B.A. Detection of Salmonella typhimurium in fat free milk using a phage immobilized magnetoelastic sensor. Sens. Actuators B Chem. 2007, 126, 544–550. [CrossRef] 94. Park, M.-K.; Li, S.; Chin, B.A. Detection of Salmonella typhimurium grown directly on tomato surface using phage-based magnetoelastic biosensors. Food Bioprocess Technol. 2013, 6, 682–689. [CrossRef] 95. Olsen, E.V.; Sorokulova, I.B.; Petrenko, V.A.; Chen, I.H.; Barbaree, J.M.; Vodyanoy, V.J. Aﬃnity-selected ﬁlamentous bacteriophage as a probe for acoustic wave biodetectors of Salmonella typhimurium. Biosens. Bioelectron. 2006, 21, 1434–1442. [CrossRef] 96. Tay, L.-L.; Huang, P.-J.; Tanha, J.; Ryan, S.; Wu, X.; Hulse, J.; Chau, L.-K. References Silica encapsulated SERS nanoprobe conjugated to the bacteriophage tailspike protein for targeted detection of Salmonella. Chem. Commun. 2012, 48, 1024–1026. [CrossRef] 97. Park, S.C.; Nakai, T. Bacteriophage control of Pseudomonas plecoglossicida infection in ayu Plecoglossus altivelis. Dis. Aquat. Org. 2003, 53, 33–39. [CrossRef] [PubMed] 98. Karunasagar, I.; Shivu, M.M.; Girisha, S.K.; Krohne, G.; Karunasagar, I. Biocontrol of pathogens in shrimp hatcheries using bacteriophages. Aquaculture 2007, 268, 288–292. [CrossRef] 99. Hooton, S.P.T.; Atterbury, R.J.; Connerton, I.F. Application of a bacteriophage cocktail to reduce Salmonella Typhimurium U288 contamination on pig skin. Int. J. Food Microbiol. 2011, 151, 157–163. [CrossRef] [PubMed] 100. Silva, Y.J.; Moreirinha, C.; Pereira, C.; Costa, L.; Rocha, R.J.M.; Cunha, Â.; Gomes, N.C.M.; Calado, R.; Almeida, A. Biological control of Aeromonas salmonicida infection in juvenile Senegalese sole (Solea senegalensis) with phage AS-A. Aquaculture 2016, 450, 225–233. [CrossRef] 101. Pereira, C.; Silva, Y.J.; Santos, A.L.; Cunha, A.; Gomes, N.C.M.; Almeida, A. Bacteriophages with potential for inactivation of ﬁsh pathogenic bacteria: Survival, host speciﬁcity and eﬀect on bacterial community structure. Mar. Drugs 2011, 9, 2236–2255. [CrossRef] 102. Costa, P.; Pereira, C.; Gomes, A.T.P.C.; Almeida, A. Eﬃciency of single phage suspensions and phage cocktail in the inactivation of Escherichia coli and Salmonella Typhimurium: An in vitro preliminary study. Microorganisms 2019, 7, 94. [CrossRef] 103. Chen, L.; Yuan, S.; Liu, Q.; Mai, G.; Yang, J.; Deng, D.; Zhang, B.; Liu, C.; Ma, Y. In vitro design and evaluation of phage cocktails against Aeromonas salmonicida. Front. Microbiol. 2018, 9, 1476. [CrossRef] 104. Bao, H.; Zhang, P.; Zhang, H.; Zhou, Y.; Zhang, L.; Wang, R. Bio-control of Salmonella Enteritidis in foods using bacteriophages. Viruses 2015, 7, 2847. [CrossRef] 105. Kim, J.; Jo, A.; Ding, T.; Lee, H.-Y.; Ahn, J. Assessment of altered binding speciﬁcity of bacteriophage for ciproﬂoxacin-induced antibiotic-resistant Salmonella Typhimurium. Arch. Microbiol. 2016, 198, 521–529. [CrossRef] 106. Phongtang, W.; Choi, G.-P.; Chukeatirote, E.; Ahn, J. Bacteriophage control of Salmonella Typhimurium in milk. Food Sci. Biotechnol. 2019, 28, 297–301. [CrossRef] 07. Hagens, S.; Loessner, M.J. Phages of Listeria oﬀer novel tools for diagnostics and biocontrol. Front. Micro 2014, 5, 159. [CrossRef] [PubMed] 108. Goode, D.; Allen, V.M.; Barrow, P.A. Reduction of experimental Salmonella and Campylobacter contamination of chicken skin by application of lytic bacteriophages. Appl. Environ. Microbiol. 2003, 69, 5032–5036. [CrossRef] [PubMed] 109. Guenther, S.; Herzig, O.; Fieseler, L.; Klumpp, J.; Loessner, M.J. Biocontrol of Salmonella Typhimurium in RTE foods with the virulent bacteriophage FO1-E2. Int. References J. Food Microbiol. 2012, 154, 66–72. [CrossRef] [PubMed] 110. Whichard, J.M.; Weigt, L.A.; Borris, D.J.; Li, L.L.; Zhang, Q.; Kapur, V.; Pierson, F.W.; Lingohr, E.J.; She, Y.-M.; Kropinski, A.M.; et al. Complete genomic sequence of bacteriophage felix O1. Viruses 2010, 2, 710–730. [CrossRef] [PubMed] 111. Cherry, W.B.; Davis, B.R.; Edwards, P.R.; Hogan, R.B. A simple procedure for the identiﬁcation of the genus Salmonella by means of a speciﬁc bacteriophage. J. Lab. Clin. Med. 1954, 44, 51–55. [PubMed] 20 of 22 20 of 22 Microorganisms 2019, 7, 570 112. Guenther, S.; Huwyler, D.; Richard, S.; Loessner, M.J. Virulent bacteriophage for eﬃcient biocontrol of Listeria monocytogenes in ready-to-eat foods. Appl. Environ. Microbiol. 2009, 75, 93–100. [CrossRef] 113. Thung, T.Y.; Jayarukshi, K.; Premarathne, K.; Mudiyanselage, J.; San Chang, W.; Loo, Y.Y.; Chin, Y.Z.; Kuan, C.H.; Tan, C.W.; Basri, D.F.; et al. Use of a lytic bacteriophage to control Salmonella Enteritidis in retail food. LWT-Food Sci. Technol. 2017, 78, 222–225. [CrossRef] 114. Lim, J.A.; Shin, H.; Kang, D.H.; Ryu, S. Characterization of endolysin from a Salmonella Typhimurium-infecting bacteriophage SPN1S. Res. Microbiol. 2012, 163, 233–241. [CrossRef] 115. Ma, Y.; Pacan, J.C.; Wang, Q.; Xu, Y.; Huang, X.; Korenevsky, A.; Sabour, P.M. Microencapsulation of bacteriophage felix O1 into chitosan-alginate microspheres for oral delivery. Appl. Environ. Microbiol. 2008, 74, 4799–4805. [CrossRef] 116. Galarce, N.E.; Bravo, J.L.; Robeson, J.P.; Borie, C.F. Bacteriophage cocktail reduces Salmonella enterica serovar Enteritidis counts in raw and smoked salmon tissues. Rev. Argent. Microbiol. 2014, 46, 333–337. [CrossRef] 117. Gill, J.J.; Hyman, P. Phage choice, isolation, and preparation for phage therapy. Curr. Pharm. Biotechnol. 2010, 11, 2–14. [CrossRef] [PubMed] 18. Van Houte, S.; Buckling, A.; Westra, E.R. Evolutionary ecology of prokaryotic immune mechanisms. Micro Mol. Biol. Rev. 2016, 80, 745–763. [CrossRef] [PubMed] 119. Pao, S.; Rolph, S.; Westbrook, E.; Shen, H. Use of bacteriophages to control Salmonella in contaminated sprout seeds. J. Food Sci. 2004, 69, M127–M130. [CrossRef] Pao, S.; Rolph, S.; Westbrook, E.; Shen, H. Use of bacteriophages to control Salmonella in experimentally contaminated sprout seeds J Food Sci 2004 69 M127–M130 [CrossRef] 120. Bardina, C.; Spricigo, D.A.; Cortes, P.; Llagostera, M. Signiﬁcance of the bacteriophage treatment schedule in reducing Salmonella colonization of poultry. Appl. Environ. Microbiol. 2012, 78, 6600–6607. [CrossRef] [PubMed] 121. Spricigo, D.A.; Bardina, C.; Cortés, P.; Llagostera, M. Use of a bacteriophage cocktail to control Salmonella in food and the food industry. Int. J. Food Microbiol. 2013, 165, 169–174. References [CrossRef] [PubMed] 122. Bai, J.; Jeon, B.; Ryu, S. Eﬀective inhibition of Salmonella Typhimurium in fresh produce by a phage cocktail targeting multiple host receptors. Food Microbiol. 2019, 77, 52–60. [CrossRef] [PubMed] 123. Clavijo, V.; Baquero, D.; Hernandez, S.; Farfan, J.C.; Arias, J.; Arevalo, A.; Donado-Godoy, P.; Vives-Flores, M. Phage cocktail SalmoFREE(R) reduces Salmonella on a commercial broiler farm. Poult. Sci. 2019, 1, 1–10. [CrossRef] 124. Bai, J.; Kim, Y.T.; Ryu, S.; Lee, J.H. Biocontrol and rapid detection of food-borne pathogens using bacteriophages and endolysins. Front. Microbiol. 2016, 7, 474. [CrossRef] 125. Legotsky, S.; Vlasova, K.; Priyma, A.; Shneider, M.; Pugachev, V.; Totmenina, O.; Kabanov, A.; Miroshnikov, K.; Klyachko, N. Peptidoglycan degrading activity of the broad-range Salmonella bacteriophage S-394 recombinant endolysin. Biochimie 2014, 107, 293–299. [CrossRef] 126. Walmagh, M.; Boczkowska, B.; Grymonprez, B.; Briers, Y.; Drulis-Kawa, Z.; Lavigne, R. Characterization of ﬁve novel endolysins from Gram-negative infecting bacteriophages. Appl. Microbiol. Biotechnol. 2013, 97, 4369–4375. [CrossRef] 127. Oliveira, H.; Thiagarajan, V.; Walmagh, M.; Sillankorva, S.; Lavigne, R.; Neves-Petersen, M.T.; Kluskens, L.D.; Azeredo, J. A thermostable Salmonella phage endolysin, Lys68, with broad bactericidal properties against Gram-negative pathogens in presence of weak acids. PLoS ONE 2014, 9, e108376. [CrossRef] 128. Rodríguez-Rubio, L.; Gerstmans, H.; Thorpe, S.; Mesnage, S.; Lavigne, R.; Briers, Y. DUF3380 domain from a Salmonella phage endolysin shows potent N-acetylmuramidase activity. Appl. Environ. Microbiol. 2016, 82, 4975–4981. [CrossRef] [PubMed] 129. Briers, Y.; Walmagh, M.; Van Puyenbroeck, V.; Cornelissen, A.; Cenens, W.; Aertsen, A.; Oliveira, H.; Azeredo, J.; Verween, G.; Pirnay, J.-P.; et al. Engineered endolysin-based “Artilysins” to combat multidrug-resistant Gram-negative pathogens. mBio 2014, 5, e01379-14. [CrossRef] [PubMed] 130. Leistner, L. Basic aspects of food preservation by hurdle technology. Int. J. Food Microbiol. 2000, 55, 181–186. [CrossRef] 131. FDA. GARS Notice (GRN) No.603. 2015. Available online: https://www.fda.gov/media/95017/download (accessed on 7 August 2019). 132. USDA-FSIS. FSIS Directive 7120.1-Safe and Suitable Ingredients Used in the Production of Meat, Poultry, and Egg Products. 2013. Available online: https://www.fsis.usda.gov/wps/wcm/connect/7f981741-94f1-468c- b60d-b428c971152d/7120_68.pdf?MOD=AJPERES (accessed on 21 August 2019). 21 of 22 Microorganisms 2019, 7, 570 133. Sukumaran, A.T.; Nannapaneni, R.; Kiess, A.; Sharma, C.S. Reduction of Salmonella on chicken meat and chicken skin by combined or sequential application of lytic bacteriophage with chemical antimicrobials. Int. J. Food Microbiol. 2015, 207, 8–15. [CrossRef] 34. Kim, J.-W.; Slavik, M.F. Cetylpyridinium chloride (CPC) treatment on poultry skin to reduce attac Salmonella. J. Food Prot. 1996, 59, 322–326. [CrossRef] 135. References Rodríguez, E.; Seguer, J.; Rocabayera, X.; Manresa, A. Cellular eﬀects of monohydrochloride of L-arginine, Nα-lauroyl ethylester (LAE) on exposure to Salmonella typhimurium and Staphylococcus aureus. J. Appl. Microbiol. 2004, 96, 903–912. [CrossRef] 136. Sharma, C.S.; Ates, A.; Joseph, P.; Nannapaneni, R.; Kiess, A. Reduction of Salmonella in skinless chicken breast ﬁllets by lauric arginate surface application. Poult. Sci. 2013, 92, 1419–1424. [CrossRef] 137. Burt, S. Essential oils: Their antibacterial properties and potential applications in foods-a review. Int. J. Food Microbiol. 2004, 94, 223–253. [CrossRef] 138. Pimchan, T.; Cooper, C.J.; Eumkeb, G.; Nilsson, A.S. In vitro activity of a combination of bacteriophages and antimicrobial plant extracts. Lett. Appl. Microbiol. 2018, 66, 182–187. [CrossRef] 139. Chaudhry, W.N.; Concepción-Acevedo, J.; Park, T.; Andleeb, S.; Bull, J.J.; Levin, B.R. Synergy and order eﬀects of antibiotics and phages in killing Pseudomonas aeruginosa bioﬁlms. PLoS ONE 2017, 12, e0168615. [CrossRef] [PubMed] 140. Oechslin, F.; Piccardi, P.; Mancini, S.; Gabard, J.; Moreillon, P.; Entenza, J.M.; Resch, G.; Que, Y.-A. Synergistic interaction between phage therapy and antibiotics clears Pseudomonas aeruginosa infection in endocarditis and reduces virulence. J. Infect. Dis. 2017, 215, 703–712. [CrossRef] [PubMed] 141. Li, S.; Li, Y.; Chen, H.; Horikawa, S.; Shen, W.; Simonian, A.; Chin, B.A. Direct detection of Salmonella typhimurium on fresh produce using phage-based magnetoelastic biosensors. Biosens. Bioelectron. 2010, 26, 1313–1319. [CrossRef] [PubMed] 142. Huang, S.; Yang, H.; Lakshmanan, R.S.; Johnson, M.L.; Wan, J.; Chen, I.H.; Wikle, H.C.; Petrenko, V.A.; Barbaree, J.M.; Chin, B.A. Sequential detection of Salmonella typhimurium and Bacillus anthracis spores using magnetoelastic biosensors. Biosens. Bioelectron. 2009, 24, 1730–1736. [CrossRef] 143. Chen, I.H.; Horikawa, S.; Bryant, K.; Riggs, R.; Chin, B.A.; Barbaree, J.M. Bacterial assessment of phage magnetoelastic sensors for Salmonella enterica Typhimurium detection in chicken meat. Food Control 2017, 71, 273–278. [CrossRef] Chai, Y.; Horikawa, S.; Li, S.; Wikle, H.C.; Chin, B.A. A surface-scanning coil detector for real-time, in-situ detection of bacteria on fresh food surfaces. Biosens. Bioelectron. 2013, 50, 311–317. [CrossRef] 145. Partk, M.-K.; Chin, B.A. Novel approach of a phage-based magnetoelastic biosensor for the detection of Salmonella enterica serovar Typhimurium in soil. J. Microbiol. Biotechnol. 2016, 26, 2051–2059. [CrossRef] 146. Mack, J.D.; Yehualaeshet, T.; Park, M.-K.; Tameru, B.; Samuel, T.; Chin, B.A. Phage-based biosensor and optimization of surface blocking agents to detect Salmonella Typhimurium on romaine lettuce. J. Food Saf. 2017, 37, e12299. [CrossRef] 147. References Lakshmanan, R.S.; Guntupalli, R.; Hu, J.; Kim, D.-J.; Petrenko, V.A.; Barbaree, J.M.; Chin, B.A. Phage immobilized magnetoelastic sensor for the detection of Salmonella typhimurium. J. Microbiol. Methods 2007, 71, 55–60. [CrossRef] 148. Park, M.-K.; Wikle, H.C.; Chai, Y.; Horikawa, S.; Shen, W.; Chin, B.A. The eﬀect of incubation time for Salmonella Typhimurium binding to phage-based magnetoelastic biosensors. Food Control 2012, 26, 539–545. [CrossRef] 149. Agrawal, S.; Kulabhusan, P.K.; Joshi, M.; Bodas, D.; Paknikar, K.M. A high aﬃnity phage-displayed peptide as a recognition probe for the detection of Salmonella Typhimurium. J. Bacteriol. 2016, 231, 40–45. [CrossRef] [PubMed] 150. Karoonuthaisiri, N.; Charlermroj, R.; Morton, M.J.; Oplatowska-Stachowiak, M.; Grant, I.R.; Elliott, C.T. Development of a M13 bacteriophage-based SPR detection using Salmonella as a case study. Sens. Actuators B Chem. 2014, 190, 214–220. [CrossRef] 151. Wang, J.; Morton, M.J.; Elliott, C.T.; Karoonuthaisiri, N.; Segatori, L.; Biswal, S.L. Rapid detection of pathogenic bacteria and screening of phage-derived peptides using microcantilevers. Anal. Chem. 2014, 86, 1671–1678. [CrossRef] [PubMed] 152. Horikawa, S.; Chai, Y.; Wikle, H.C.; Suh, S.-J.; Barbaree, J.M.; Chin, B.A. Direct detection of Salmonella on fresh produce. ECS Trans. 2015, 69, 25–31. [CrossRef] 22 of 22 Microorganisms 2019, 7, 570 153. Singh, A.; Arya, S.K.; Glass, N.; Haniﬁ-Moghaddam, P.; Naidoo, R.; Szymanski, C.M.; Tanha, J.; Evoy, S. Bacteriophage tailspike proteins as molecular probes for sensitive and selective bacterial detection. Biosens. Bioelectron. 2010, 26, 131–138. [CrossRef] Blasco, R.; Murphy, M.J.; Sanders, M.F.; Squirrell, D.J. Speciﬁc assays for bacteria using phage mediated release of adenylate kinase. J. Appl. Microbiol. 1998, 84, 661–666. [CrossRef] 155. Wu, Y.; Brovko, L.; Griﬃths, M.W. Inﬂuence of phage population on the phage-mediated bioluminescent adenylate kinase (AK) assay for detection of bacteria. Lett. Appl. Microbiol. 2001, 33, 311–315. [CrossRef] 156. Park, M.-K.; Park, J.W.; Wikle, H.C.; Chin, B.A. Evaluation of phage-based magnetoelastic biosensors for direct detection of Salmonella Typhimurium on spinach leaves. Sens. Actuators B Chem. 2013, 176, 1134–1140. [CrossRef] 157. Chai, Y.; Li, S.; Horikawa, S.; Park, M.-K.; Vodyanoy, V.; Chin, B.A. Rapid and sensitive detection of Salmonella Typhimurium on eggshells by using wireless biosensors. J. Food Prot. 2012, 75, 631–636. [CrossRef] 158. Pereira, C.; Moreirinha, C.; Lewicka, M.; Almeida, P.; Clemente, C.; Cunha, A.; Delgadillo, I.; Romalde, J.L.; Nunes, M.L.; Almeida, A. Bacteriophages with potential to inactivate Salmonella Typhimurium: Use of single phage suspensions and phage cocktails. Virus Res. 2016, 220, 179–192. [CrossRef] 159. References Sharma, M.; Dashiell, G.; Handy, E.T.; East, C.; Reynnells, R.; White, C.; Nyarko, E.; Micallef, S.; Hashem, F.; Millner, P.D. Survival of Salmonella Newport on whole and fresh-cut cucumbers treated with lytic bacteriophages. J. Food Prot. 2017, 80, 668–673. [CrossRef] [PubMed] 160. Yeh, Y.; Purushothaman, P.; Gupta, N.; Ragnone, M.; Verma, S.C.; de Mello, A.S. Bacteriophage application on red meats and poultry: Eﬀects on Salmonella population in ﬁnal ground products. Meat Sci. 2017, 127, 30–34. [CrossRef] [PubMed] 161. El-Dougdoug, N.K.; Cucic, S.; Abdelhamid, A.G.; Brovko, L.; Kropinski, A.M.; Griﬃths, M.W.; Anany, H. Control of Salmonella Newport on cherry tomato using a cocktail of lytic bacteriophages. Int. J. Food Microbiol. 2019, 293, 60–71. [CrossRef] [PubMed] 162. Zhang, X.; Niu, Y.D.; Nan, Y.; Stanford, K.; Holley, R.; McAllister, T.; Narvaez-Bravo, C. SalmoFresh eﬀectiveness in controlling Salmonella on romaine lettuce, mung bean sprouts and seeds. Int. J. Food Microbiol. 2019, 305, 108250. [CrossRef] 163. Ye, J.; Kostrzynska, M.; Dunﬁeld, K.; Warriner, K. Control of Salmonella on sprouting mung bean and alfalfa seeds by using a biocontrol preparation based on antagonistic bacteria and lytic bacteriophages. J. Food Prot. 2010, 73, 9–17. [CrossRef] 164. Yeh, Y.; de Moura, F.H.; Van Den Broek, K.; de Mello, A.S. Eﬀect of ultraviolet light, organic acids, and bacteriophage on Salmonella populations in ground beef. Meat Sci. 2018, 139, 44–48. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2800788635	https://europepmc.org/articles/pmc5937055?pdf=render	English	null	2017 Dutch Report Card+: Results From the First Physical Activity Report Card Plus for Dutch Youth With a Chronic Disease or Disability	Frontiers in pediatrics	2,018	cc-by	6,803	2017 Dutch Report Card+: Results From the First Physical Activity Report Card Plus for Dutch Youth With a Chronic Disease or Disability Marcella Burghard, Nynke B. de Jong, Selina Vlieger and Tim Takken* Shared Utrecht Pediatric Exercise Research Lab, Child Development & Exercise Center, Wilhelmina Children’s Hospital, University Medical Centre Utrecht, Utrecht, Netherlands Background: The Dutch Active Healthy Kids (AHK) Report Card+ (RC+) consolidates and translates research and assesses how the Netherlands is being responsible in providing physical activity (PA) opportunities for youth (<18 years) with a chronic disease or disability. The aim of this article is to summarize the results of the Dutch RC+. Keywords: children, youth, chronic disease, disability, health, exercise Correspondence: Tim Takken Correspondence: Tim Takken t.takken@umcutrecht.nl Specialty section: This article was submitted to Child Health and Human Development, a section of the journal Frontiers in Pediatrics INTRODUCTION Received: 14 February 2018 Accepted: 13 April 2018 Published: 30 April 2018 According to the World Health Organization (WHO) physical inactivity is the fourth leading risk factor for mortality. Regular physical activity (PA) reduces the risk of many diseases including cardiovascular disease, diabetes, breast and colon cancer, and depression (1). Noting that the more physically active the child the greater the health beneﬁt, speciﬁc research showed that PA has positive eﬀects on musculoskeletal health, cardiovascular health, and mental health (2). It has been indicated as well that the earlier in life one starts engaging in sports and exercise, the longer one beneﬁts from it (3). Therefore, PA is important. However, according to the Global Matrix 2.0, in which Report Cards from 38 countries, including the ﬁrst Dutch Physical Activity Report Card, ORIGINAL RESEARCH published: 30 April 2018 doi: 10.3389/fped.2018.00122 Edited by: Saralee Glasser, Women and Children’s Health Research Unit, Gertner Institute for Epidemiology and Health Policy Research, Israel Edited by: Saralee Glasser, Women and Children’s Health Research Unit, Gertner Institute for Epidemiology and Health Policy Research, Israel Methods: Nine indicators were graded using the AHK Global Alliance RC development process, which includes a synthesis of best available research, surveillance, policy and practice ﬁndings, and expert consensus. Two additional indicators were included: weight status and sleep. Reviewed by: Melanie M. Adams, Keene State College, United States Shahanawaz Syed, School of Physiotherapy, Mahatma Gandhi Mission Institute of Health Sciences, India Results: Grades assigned were: Overall Physical Activity, D; Organized Sports Participation, B–; Active Play, C–; Active Transportation, A–; Sedentary Behavior, C; Sleep C; For Weight Status, Family and Peers, School, Community and Built Environment, Government Strategies, and Investments all INC. Conclusions: The youth with disabilities spend a large part of the day sedentary, since only 26% of them met the PA norm for healthy physical activity. Potential avenues to improve overall physical activity are changing behaviors regarding sitting, screen time, and active play. The Netherlands is on track regarding PA opportunities for youth with disabilities, however they are currently not able to participate unlimited in sports and exercise. 1For a more detailed description about the Report Card, see Colley et al. (5) Abbreviations: BMI, Body mass index; CBS, Statistics Netherlands; INC, Incomplete; KSC, Knowledge Centre for Sports Netherlands; MET, Metabolic equivalent; NHS, National Health Survey; NIVEL, Netherlands Institute for health service research; NNGB, Dutch Physical Activity Guideline; NOC∗NSF, Dutch Olympic Committee∗Dutch Sports Federation; PA, Physical activity; PE, Physical Education; RIVM, National Institute for Public Health and Environment; RWG, Research work group; SCP, the Netherlands Institute for Social Research; WHO, World Health Organization. Citation: Burghard M, de Jong NB, Vlieger S and Takken T (2018) 2017 Dutch Report Card+: Results From the First Physical Activity Report Card Plus for Dutch Youth With a Chronic Disease or Disability. Front. Pediatr. 6:122. doi: 10.3389/fped.2018.00122 April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 1 2017 Dutch Report Card+ Burghard et al. were compared regarding PA behavior, norms were often not met by typically developing youth (4). The Report Card is an annual update or “state of the nation” that assesses how a country is doing as a nation at promoting and facilitating PA opportunities for children and youth and grades outcomes using an academic letter grade approach (i.e., A, B, C, D, F). Data to grade the outcomes are drawn from several sources, including the research literature, governmental agencies, and non-governmental organizations1 (5). FIGURE 1 \| Front cover of the 2017 Dutch Physical Activity Report Card+. Next to typically developing children, also many children with disabilities are not physically active (6). Even though it might be especially important for this group of children to engage in sports and exercise, because of the positive health eﬀects in the physical, mental, and social domain (4, 7–10). Because of multiple barriers, this group should perhaps be more stimulated and encouraged to engage in an active lifestyle in a broad sense: from PA during sports and play activities and reducing sedentary behavior, to behavior related to sleep, and weight/nutrition (11). Over the past few years, changes have occurred to facilitate the sports and exercise behaviors of people with disabilities in the Netherlands. Many organizations, foundations, and governmental bodies developed or funded projects that focus on improving PA and sports participation among people with disabilities. However, it is not yet clear what the overall eﬀects of these projects were and where the gaps are. Do people with disabilities feel less restricted in the opportunities they have to participate in sports? In the Netherlands, there was no overview yet of the actual status of PA behavior, sleeping behavior and weight status for youth with disabilities. Citation: Regarding the proven and potential positive eﬀects of exercise for a good health, it was considered useful to fulﬁll this gap, by Active Healthy Kids the Netherlands, which consists of a group of researchers in the ﬁeld of PA in children and youth, with a mission to inspire the nation to engage all children and youth in PA by providing expertise and direction to policy makers and public on how to increase, and eﬀectively allocate resources and attention toward PA for Dutch children and youth. This is also the ﬁrst Report Card in the world that was speciﬁcally developed for this group of children. FIGURE 1 \| Front cover of the 2017 Dutch Physical Activity Report Card+. METHODS For the developmental process, guidelines of the Active Healthy Kids Canada framework were followed (5). Eleven indicators were graded in this Report Card. Nine of the indicators were part of this standard international framework. It was decided to add sleep behavior and weight status as additional indicators. The indicators were divided over three categories, except weight status, which did not ﬁt in any of the categories (Figure 2). The grades were based on the percentage that met the single or multiple benchmarks. With this Report Card+, we want to gain more insight in the PA levels and patterns of the Dutch youth with a chronic disease or disability and answer the question “how (un)limited are the possibilities for the Dutch youth with disabilities to be physical active?” (Figure 1). In line with this, another aim was to compare the results of the Report Card for Dutch typically developing youth with the Report Card+ for Dutch youth with disabilities. The principal investigator and project manager formed a research work group (RWG) together with seven researchers of the University Medical Centre Utrecht, Utrecht University, Utrecht University of Applied Sciences and Center of Excellence in Rehabilitation Medicine Utrecht. An expert group was formed with the involvement of National Institute for Public Health and Environment (RIVM), Mulier Institute, Dutch Olympic Committee∗Dutch Sports Federation (NOC∗NSF), Windesheim University of Applied Sciences, Knowledge Centre for Sports Netherlands (KCS), Hanze University of Applied Sciences Groningen, Amsterdam University of Applied Sciences, Institute for Health and Care Research, Netherlands Institute for health service research April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 2 2017 Dutch Report Card+ Burghard et al. FIGURE 2 \| Overview categories and related indicators. FIGURE 2 \| Overview categories and related indicators. (NIVEL), and an advisory role for the Primary Education Board [PO-Raad]. gathered by Statistics Netherlands (CBS) and the RIVM as the primary source. These organizations annually collect data about several lifestyle themes, the Lifestyle Monitor, part of which is the National Health Survey (NHS) (12). Most of the grading was based on this survey. The Lifestyle Monitor divides youth in two age groups: 4–11 years and 12–17 years. As a consequence of this, both age groups were assessed for each indicator. Unfortunately, the sample sizes for the years 2011 up to 2014 were too small for a subgroup analysis. Frontiers in Pediatrics \| www.frontiersin.org 2Cluster I: Schools for visual impaired children or children with multiple disabilities who are visually impaired or blind. Cluster II: Schools for deaf children and hearing impaired children, children with speech or language diﬃculties and children with communicative problems, as with some forms of autism. Cluster III: Schools for children with motor and/or mental disabilities, chronically ill children and children with epilepsy. Cluster IV: Schools for children with psychiatric disorders or severe behavioral problems and schools that are related to pedagogical institutes. Benchmarks The ﬁrst six indicators are Overall Physical Activity, which is also the ﬁrst category, and the behaviors that contribute to that: Organized Sports Participation, Active Play, Active Transportation, Sedentary Behavior, and Sleep. For all of these indicators the grading was based on data from the NHS for children with disabilities in general and, when available, data from the Mulier Institute were used to describe the situation of children attending special education (13). The next category, Settings and Sources of Inﬂuence, consists of the indicators Family and Peers, School and Community and Environment. No data of the NHS regarding these indicators were present. Thus, no general information was present. Other sources were used for assessment of the indicators in this category. The criteria of Family were: “percentage of parents who facilitate PA and sports opportunities for their children (e.g., volunteering, coaching, driving, paying for membership fees, and equipment),” “percentage of parents who meet the PA guidelines for adults” and “percentage of parents who are physically active with their kids.” For Peers the criteria were: “percentage of children and youth with disabilities with friends and peers who encourage and support them to be physically active” and “percentage of children and youth who encourage and support their friends and peers to be physically active.” However, as there was no consistent data for children with disabilities in general nor for children attending special education (not all clusters2), the RWG and experts decided that this indicator could not be graded. The available numbers of some of the clusters in special education from the Mulier Institute and other sources were used to get some insight in this indicator. For the ﬁrst indicator, Overall Physical Activity, the grading was based on the percentage of children who met the Dutch Physical Activity Guideline [NNGB: Dutch Guidelines Healthy Physical Activity; to be at least moderate active (at least 5 MET) for at least 60 min every day]. For this indicator data on children attending special education were available. The RWG and expert group reached consensus to use data regarding engaging sports on a weekly basis, thus the grading of Organized Sports Participation was based on the percentage of children and youth who participated in organized sports and/or PA programs weekly. METHODS For 2015, 142 children were included for the Both the RWG and the expert group were responsible for the interpretation and evaluation of the data sources and evidence and had to decide about deﬁnitions and benchmarks of the indicators for the grading and were responsible for the ﬁnal grading. For the evaluation of the indicators, data of the period 2011 up to 2015 have been included. When available, we used data April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 3 2017 Dutch Report Card+ Burghard et al. age range 4–11 years and 232 children were included for the age range 12–17 years. For the youngest age category (4–11 years), the answers were parent reported. For the older age group (12–17 years) the NHS was a self-report questionnaire. If the required data to grade an indicator could not be provided by the primary sources, other governmental and non-governmental sources were used. sedentary time. The number of children who watch television or sit in front of the computer less than 2 h a day outside school hours determined the grade for this indicator. No numbers where available on sedentary time of scholars attending special education. The indicator Sleep was assessed by the amount of children meeting the sleep duration recommendation for their age group. The sleep duration recommendations used are described in the study of Hirshkowitz et al. (14). These recommendations are for healthy individuals with normal sleep. The appropriate sleep duration for school-aged children is considered between 9–11 h each night and for adolescents this is 8–10 h (14). age range 4–11 years and 232 children were included for the age range 12–17 years. For the youngest age category (4–11 years), the answers were parent reported. For the older age group (12–17 years) the NHS was a self-report questionnaire. If the required data to grade an indicator could not be provided by the primary sources, other governmental and non-governmental sources were used. Children with disabilities in the Netherlands could attend regular education or special education at special schools. The situation of scholars attending special education was described when reports were available (13). The grades of the indicators were based on the data about youth with disabilities in general from the NHS. Frontiers in Pediatrics \| www.frontiersin.org METHODS For the additional indicator Weight Status, the grade was based on the percentage of children with a normal body weight (BMI between 18.5 and 25 kg/m2 was classiﬁed as normal weight) (15). Data for this indicator were taken from the NHS for the children with disabilities in general. Data of scholars attending special education were used from reports from the Mulier Institute (13). As this Report Card+ was developed following a standard framework that was also used for the ﬁrst Dutch Physical Activity Report Card for typically developing children, the results of both Report Cards could be compared. 3In 2014, the regulation “Appropriate Education” [Wet Passend Onderwijs] was introduced, which aims that every student should attend a school that provides education suited to their talents and capabilities. Schools should adapt their teaching to the individual child’s development and oﬀer extra assistance. This applies to the school where the child is currently registered, another mainstream school or a school providing special education (16). Benchmarks For children attending special education, it was known how many children were a member of a sports club and how many played non-school based sports at least once a week. Regarding Active Play, the grading was determined by the percentage of children who played outside for at least 60 min after school, for 7 days a week. The NHS does not include questions about active play behavior in 12–17 year old youth, therefore the grade was based only on 4–11 year old children. For the scholars attending special education the percentages of children who played outside 5–7 times a week were reported. For School the following criteria were set: “the percentage of schools with an active school policy (e.g., oﬀering sports- and exercise activities next to physical education (PE) or activities during recess, collaboration with communities and/or sports clubs, presence of annual planning),” “percentage of schools with a PE specialist,” “the percentage of schools were the students have at least 90 min of PE per week,” and lastly “the percentage of students who have at least 45 min of outside play time Active Transportation was assessed by the percentage of children who use active transportation (walking and cycling) to get to and from places (school and/or work) for at least 3 days a week. Of children attending special education, only the amount of children who used active transport was known and not the weekly frequencies. For Sedentary Behavior only the amount of time spent in front of a screen (screen time) was surveyed, so the grading was based only on this criteria even though this does not cover all April 2018 \| Volume 6 \| Article 122 4 2017 Dutch Report Card+ Burghard et al. during school for 5 days per week.” Again, however, it was decided to grade this indicator as Incomplete. Data was present about regular education and special education. However, as a consequence of the regulation “Appropriate Education” [Wet Passend Onderwijs]3, some children with disabilities attended regular schools and participate in regular PE. The speciﬁc situation for these children was unknown. were based on the percentages of youth meeting the deﬁned benchmark. Some indicators are stand-alone, while others are comprised of several components. A was 81 to 100%, B was 61 to 80%, C was 41 to 60%, D was 21 to 40%, F was 0 to 20%. Benchmarks INC was incomplete data or not enough available evidence to assign a grade to the indicator or absence of clear well-established criteria. This grading system is in accordance to the Canadian Report Card framework (5). The last indicator of this category, Community and Environment, also had several criteria: “the percentage of children and parents who perceive their community/municipality is doing a good job at promoting PA (e.g., variety, location, cost quality),” “the percentage of communities/municipalities that report they have policies promoting PA,” “the percentage of communities/municipalities that report they have infrastructure (e.g. sidewalks, trails, paths, bike lanes) speciﬁcally geared toward promoting PA,” “the percentage of children or parents who report having facilities, programs, parks, and playgrounds available to them in their community,” “the percentage of children or parents who report living in a safe neighborhood where they can be physically active,” “the percentage of children who report having well-maintained facilities, parks and playgrounds in their community that are safe to use” and ﬁnally, “the percentage of children and parents who report that in organizations like sports clubs, they (their child) are socially accepted and that social accessibility is present.” Also for this indicator it was decided to mark it as an Incomplete. In the Report Card for typically developing children, data of the Leisure time Omnibus [Vrijetijdsomnibus] of the CBS and the Netherlands Institute for Social Research (SCP) was used to grade this indicator. Unfortunately, the sample size of children with disabilities was too low for both 2012 and 2014 to use the results. When the data about scholars attending special education showed that the situation for that particular indicator was considerably better or worse for these children, the grade was given a plus or minus respectively. Overall Physical Activity Levels: D Overall Physical Activity Levels: D The grade for Overall Physical Activity levels was a D. In 2015, 26% of both children and youth (4–17 year olds) met the Dutch PA guideline of Healthy Physical Activity (NNGB). Scholars of cluster II schools were the most physically active compared to the other clusters. Of the scholars attending cluster II schools 35% exercised 8 or more hours per week (excluding sports) (13). For the cluster I and III scholars this was 21% and in cluster IV 27% (13). Organized Sports Participation: B– Of the 4–11 year olds 69% and of the 12–17 year olds 73% was considered a weekly athlete (12) Among scholars attending special schools, the sports participation was lower. Cluster IV scholars had the highest sports participation, namely 45 vs. 25, 37, 26% for cluster I, II, III, respectively (13). The last category, Strategies and Investments was divided in the indicators Government and Non-Government. The criteria that were set were: “evidence of leadership and commitment in providing PA opportunities for all children and youth,” “allocation of funds and resources for the implementation of PA promotion strategies and initiatives for all children and youth” and “demonstrated progress through the key stages of public policy making (i.e., policy agenda, policy formation, policy implementation, policy evaluation, and decisions about the future).” No clear numbers were available to state that policy is eﬃcient or how much ﬁnancing is acceptable. Therefore, the decision was made to grade this indicator with an Incomplete. Multiple governmental documents were studied and reports of the Mulier Institute on diﬀerent policies and programs were evaluated. For Non-Government, annual reports and websites of several national and regional foundations and organizations were considered. RESULTS The 2017 Dutch Report Card+ is the ﬁrst ever assessment of PA behaviors, settings, and sources of inﬂuence and government strategies and investments for children with a chronic disease or disability. The grades are summarized in Table 1. April 2018 \| Volume 6 \| Article 122 Active Transportation: A– Of the children in the age of 4–11 years 39% cycled 3 or more days to or from school or work and this was 38% for walking 3 or more days per week. Of the 12–17 year olds 71.8% cycled 3 or more days to or from school or work and this was 15.8% for walking 3 or more days per week (12). Concerning active school policies, 71% of the special schools oﬀered their students other sports and exercise activities, next to PE (10). All cluster I and II schools, had a PE specialist, and 84.2 and 94% of the cluster III and IV schools had a PE specialist respectively. All cluster schools oﬀered twice a week PE (7, 8, 13) The number of average minutes PE per week varied between 63 min per week in cluster III to 103 min in cluster IV (7, 8). Only 4% of the children in cluster I schools used active transportation to get to their school (13). This was 18% in cluster II, 13% in cluster III, and 30% in cluster IV schools (7, 13). Sleep: C Of the 4–11 year old children with disabilities 26% met the sleep recommendations. This was 63% in the 12–17 year old age group (12). No data was present about sleep behavior of scholars attending special schools. Active Play: C– Of the 4–11 year old children with disabilities 53% played outside for at least 60 min after school, on all days of the week (12). TABLE 1 \| Overview of indicators and corresponding grades. Indicator Grades Overall physical activity D Organized sports participation B– Active play C– Active transportation A– Sedentary behavior C Sleep C Weight status INC Family and peers INC School INC Community and the built environment INC Government strategies and investments INC April 2018 \| Volume 6 \| Article 122 TABLE 1 \| Overview of indicators and corresponding grades. Indicator Grades Overall physical activity D Organized sports participation B– Active play C– Active transportation A– Sedentary behavior C Sleep C Weight status INC Family and peers INC School INC Community and the built environment INC Government strategies and investments INC April 2018 \| Volume 6 \| Article 122 TABLE 1 \| Overview of indicators and corresponding grades. Active play Active transportation Sedentary behavior Sleep Weight status Family and peers School Community and the built environment Government strategies and investments The RWG and experts evaluated the evidence for each of the indicators and discussed the proposed grading. The grades Frontiers in Pediatrics \| www.frontiersin.org 5 Burghard et al. 2017 Dutch Report Card+ children signiﬁcantly more (p = 0.05) to sports and exercise, than parents whose child is not a sports club member (9). Scholars of cluster II schools, most often played 5–7 times per week outside (45%), compared to cluster I (31%), III (30%), and IV (33%) scholars. The average amount of minutes of active playtime outside school hours was 529 min per week for the 4–11 year old children with disabilities (13). Government Strategies and Investments: INC The sample size of the NHS was unfortunately too small, to grade this indicator. These data showed however, that the mean BMI of the 4–11 year olds was 16.5 and 20.8 kg/m2 in the 12–17 year old age group (12). This indicator about the current policy of the government could not be judged. There have been several initiatives that have to resulted in a more physically active youth. Unfortunately, no clear criteria and monitors were present to evaluate the eﬀectiveness of these initiatives and policies. When evaluating the scholars who attended special schools (all clusters together), 68% of the children had a normal weight, 11% was underweight, 17% was overweight, and 4% obese. When comparing the diﬀerent clusters, the highest percentage of overweight and obese children (combined) was found in cluster III schools (25%) (13). With regard to foundations, we saw that proportionally more foundations were founded to help or facilitate children with disabilities in their possibilities to play sports or exercise compared to foundations for typically developing children. School: INC Data was present about regular education and special education. However, as a consequence of the regulation “Appropriate Education” [Wet Passend Onderwijs]3, some children with disabilities attend regular schools and participate in regular PE. The speciﬁc situation for these children was unknown and consequently an Incomplete was graded. Key ﬁndings about the situation in special schools will be given. Community and the Built Environment: INC Community and the Built Environment: INC As mentioned in the methods, the sample size of children with disabilities in the Leisure time Omnibus [Vrijetijdsomnibus] of the CBS and SCP was unfortunately too low to grade this indicator. A smaller study showed that 12% of the parents of children with disabilities reported that play sets/grounds are not nearby enough. Only 2% of these parents reported that the play sets/equipment are not safe and/or badly maintained and only 1% considered them not safe (for younger children). Only 9% of these parents reported that it was not safe for their children to play in the neighborhood, due to traﬃc safety (16). The 4–11 year olds sat/lay on average 7.9 h per day on a school day, compared to 11.1 h for the 12–17 year olds. On a day oﬀfrom school, the younger age group sat/lay on average 6.5 h, compared to 9.2 h in the older age group (12). Frontiers in Pediatrics \| www.frontiersin.org Sedentary Behavior: C Regarding playtime during school recess, 50% of the 4–11 year old students played at least 45 min outside during school time for 5 days per week and the average active play time at school was 284 min per week for this age group (12). Of the 4–11 year old children 45.5% sat in front of the computer or watched TV, less than 2 h a day (average day of the week), outside school. This was only 23.2% for 12–17 year old children (12). No data concerning sedentary behavior was available for scholars in special schools. DISCUSSION No data of the NHS regarding “Family & Peers” were present. Thus, no general information was present to grade this indicator. Data was only available on parents of children in cluster III or IV schools. No information about the parental behavior in the other two clusters was present, consequently an Incomplete was graded. Of the parents of cluster IV scholars 59% considered it important that their child engages in sports or exercise frequently. Of the parents 72% encouraged their child to play sports or exercise frequently (8). A smaller study showed that parents of whom the child joins a sports club, stimulate their The primary aim of this Report Card+ was to provide an overview of the methods and results of the ﬁrst Dutch Report Card+ for youth with disabilities. The results showed that about a quarter of the Dutch youth with disabilities met the PA norm. In 2016 the results of the ﬁrst Dutch Physical Activity Report Card were published (17). These results were compared with the Report Card+ results (Figure 3). A notable ﬁnding was that the percentages of children that met the Dutch April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 6 2017 Dutch Report Card+ Burghard et al. FIGURE 3 \| Comparison of the results of the Report Card and the Report Card+. FIGURE 3 \| Comparison of the results of the Report Card and the Report Card+. Physical Activity Guidelines was the same for children with and without disabilities (26%). Assessing the diﬀerent indicators that contribute to Overall Physical Activity (Organized Sports, Active Play, Active Transport, and Sedentary Behavior), it was clear that the youth with disabilities used active transport less often than their typically developing peers. Regarding youth attending special education, norms were less often met than in youth attending normal education. The diﬀerences between healthy children and children attending special education may be caused by the (social) accessibility and by the diversity of disorders/disabilities. Noteworthy, was that in the Report Card+, only six of the 11 indicators could be graded and ﬁve were graded an Incomplete, thus we stated that the national monitoring in youth with disabilities is unfortunately lacking. Therefore, it was diﬃcult to make powerful statements about possible causes (17). Fortunate, a large part of the youth with disabilities engaged in sports weekly and chose an active mode of transportation for their way to school. DISCUSSION It is important that the conditions for these indicators will remain this high in the future. Solutions should be developed to make it possible for more scholars in special schools to travel to school (partly) using active transportation. Furthermore, sports clubs need to educate their staﬀand volunteers more properly so children and their parents experience less barriers to join a sports club. Physical Activity Guidelines was the same for children with and without disabilities (26%). Assessing the diﬀerent indicators that contribute to Overall Physical Activity (Organized Sports, Active Play, Active Transport, and Sedentary Behavior), it was clear that the youth with disabilities used active transport less often than their typically developing peers. Regarding youth attending special education, norms were less often met than in youth attending normal education. The diﬀerences between healthy children and children attending special education may be caused by the (social) accessibility and by the diversity of disorders/disabilities. Noteworthy, was that in the Report Card+, only six of the 11 indicators could be graded and ﬁve were graded an Incomplete, thus we stated that the national monitoring in youth with disabilities is unfortunately lacking. Therefore, it was diﬃcult to make powerful statements about possible causes (17). The role of the parents and family is of high importance as well in this group of children. Even though no grade could be assigned to this indicator, results demonstrated that parents should be more informed about their large inﬂuence as a role model for all behaviors and that their home rules are of high relevance as well. Stimulating parents to engage in sports and exercise activities with their whole family should be more promoted. In addition, strategies that promote sports opportunities for children with disabilities, such as sports and play activities in the neighborhood and foundations who can help families with less ﬁnancial back up, should be improved. Currently, too many children and parents are not familiar with these possibilities and sports opportunities. Other indicators for which improvement is warranted are sedentary behavior and active play. The Dutch youth with disabilities spent a large part of the day sitting or lying and/or behind a screen, especially during school times. Though, around half of the children with disabilities engaged in daily active play for at least 60 min, the other half did not. Frontiers in Pediatrics \| www.frontiersin.org ACKNOWLEDGMENTS The authors also thank Knowledge Centre for Sports Netherlands, S. W. van de Berg (RIVM) and M. Wissing for their contributions to the 2017 Dutch Report Card+. B. van Leeuwen MSc is acknowledged for the layout of the Dutch Report Card+. This work was supported by a seed grant from the Utrecht University focus area Sport & Society and Knowledge Centre for Sports Netherlands. ADDITIONAL INFORMATION The long form Report Card+, with more background information about the developmental process, methods, indicators, and recommendations, is available online: http:// www.super-lab.nl/reportcarddownloads/. DISCUSSION Thus, changing the behaviors regarding, sitting (at school), screen time, and active play, seems most likely to improve overall activity levels. As the indicator sedentary behavior showed, the youth with disabilities sat the most during school hours. Strategies to April 2018 \| Volume 6 \| Article 122 7 2017 Dutch Report Card+ Burghard et al. CONCLUSION interrupt the long sitting duration should be developed and implemented, for example physically active academic lessons. As school is the place where all children can be reached, strategies, and ﬁnancial resources are needed to enlarge the duration of PE lessons and to realize higher intensities during these lessons. Based on the results of this Physical Activity Report Card+, only 26% of the Dutch youth with a chronic disease or disability met the current national PA guidelines. The most important behaviors to change that will most likely result in improvement of overall PA levels seem to be sitting (at school), screen time, and active play. In the past few years, many initiatives, possibilities, and policies were developed and the Netherlands is on track, but currently, the Dutch youth with disabilities is not yet able to participate completely unlimited in sports and exercise. Further, collaborations between all sectors should be stimulated. Problems in the accommodation and oﬀer of sports and other active activities will beneﬁt from this. Furthermore, it is important to involve parents, PE specialists and teachers in realizing and improving the sports opportunities for children with disabilities. Both parents and teachers know the child and his/her possibilities and disabilities the best and can search together with the sports clubs for the most appropriate sports activity. Strengths and Limitations This is the ﬁrst ever developed Physical Activity Report Card+ for children and youth with a chronic disease or disability. This Report Card provides a comprehensive overview about how the Netherlands is doing, regarding PA opportunities, overall PA levels and the role of sources of inﬂuence for children with disabilities. AUTHOR CONTRIBUTIONS TT was the principal investigator and MB was the project manager according to the international Report Card framework. NdJ and SV supported TT and MB in their work in the Report Card developmental process (for example, literature search, analyzing the results, writing of ﬁnal Report Card, and the manuscript). Strength of this Report Card+ is the participation of many experts and organizations in this area, which made that many important data sources were identiﬁed and included. Unfortunately, not all indicators were integrated in national surveys yet (e.g., family and peers) and in the national surveys no clear demarcation was present for children with disabilities. No subcategories could be made and the size of the researched population is small. Furthermore, only the data of 2015 from the NHS could be used because the sample sizes in the years 2011-2014 were too small to analyze. With this in mind, one can question whether these results actually represented the current situation for people with disabilities and youth in particular. Making appropriate policies based on the results of this monitoring should therefore be questioned. REFERENCES 6. Bloemen MA, Verschuren O, van Mechelen C, Borst HE, de Leeuw AJ, van der Hoef M, et al. Personal and environmental factors to consider when aiming to improve participation in physical activity in children with Spina Biﬁda: a qualitative study. BMC Neurol. (2015) 15:11. doi: 10.1186/s12883-015-0 265-9 1. McGillivray J, McVilly K, Skouteris H, Boganin C. Parental factors associated with obesity in children with disability: a systematic review. Obes Rev. (2013) 14:541–4. doi: 10.1111/obr.12031 1. McGillivray J, McVilly K, Skouteris H, Boganin C. Parental factors associated with obesity in children with disability: a systematic review. Obes Rev. (2013) 14:541–4. doi: 10.1111/obr.12031 2. Papas MA, Trabulsi JC, Axe M, Rimmer JH. Predictors of obesity US sample of high school adolescents with and without disabilities. J Sch Health (2016) 86:803–12. doi: 10.1111/josh.12436 7. van den Dool R, van Lindert C, Smits F, Breedveld K. Monitor Special Heroes in Cluster 3. End Situation of Participating Cluster 3 Schools and Their Students. [Monitor Special Heroes in Cluster 3. Eindsituatie van Deelnemende Cluster 3 Scholen en Hun Leerlingen]. Utrecht: Mulier Institute (2013). 3. Brown CL, Irby MB, Houle TT, Skelton JA. Family-based obesity treatment in children with disabilities. Acad Pediatr. (2015) 15:197–203. doi: 10.1016/j.acap.2014.11.004 3. Brown CL, Irby MB, Houle TT, Skelton JA. Family-based obesity treatment in children with disabilities. Acad Pediatr. (2015) 15:197–203. doi: 10.1016/j.acap.2014.11.004 8. van Lindert C, Bruining J-W, Smits F, Nootebos W. Monitor Special Heroes in Cluster 4. Ending Situation of Participating Cluster 4 Schools and Their Students. [Monitor Special Heroes in Cluster 4. Eindsituatie van Deelnemende Cluster 4 Scholen en Hun Leerlingen]. Utrecht: Mulier Institute (2013). 4. Tremblay MS, Barnes JD, Gonzalez SA, Katzmarzyk PT, Onywera VO, Reilly JJ, et al. Global matrix 2.0: report card grades on the physical activity of children and youth comparing 38 countries. J Phys Act Health (2016) 13(11Suppl. 2):S343–66. doi: 10.1123/jpah.2016-0594 4. Tremblay MS, Barnes JD, Gonzalez SA, Katzmarzyk PT, Onywera VO, Reilly JJ, et al. Global matrix 2.0: report card grades on the physical activity of children and youth comparing 38 countries. J Phys Act Health (2016) 13(11Suppl. 2):S343–66. doi: 10.1123/jpah.2016-0594 Cluster 4 Scholen en Hun Leerlingen]. Utrecht: Mulier Institute (20 9. van Bakel E, Moresi S, Borghouts L, van Lindert C. Parent Involvement. [Ouderbetrokkenheid]. Utrecht: Fontys Sports College/Mulier Institute (2013). 5. Colley RC, Brownrigg M, Tremblay MS. 7. van den Dool R, van Lindert C, Smits F, Breedveld K. Monitor Special Heroes in Cluster 3. End Situation of Participating Cluster 3 Schools and Their Students. [Monitor Special Heroes in Cluster 3. Eindsituatie van Deelnemende Cluster 3 Scholen en Hun Leerlingen]. Utrecht: Mulier Institute (2013). 6. Bloemen MA, Verschuren O, van Mechelen C, Borst HE, de Leeuw AJ, van der Hoef M, et al. Personal and environmental factors to consider when aiming to improve participation in physical activity in children with Spina Biﬁda: a qualitative study. BMC Neurol. (2015) 15:11. doi: 10.1186/s12883-015-0 265-9 REFERENCES A model of knowledge translation in health: the active healthy kids Canada report card on physical activity for children and youth. Health Promot Pract. (2012) 13:320–30. doi: 10.1177/1524839911432929 10. Lucassen J, Cevaal A, Scholten V, van der WerﬀH. Physical Education in Special and Vocational Education, Baseline Measurement 2015. [Bewegingsonderwijs in Het Speciaal Onderwijs en Praktijkonderwijs, Nulmeting 2015]. Utrecht: Mulier Institute (2016). 5. Colley RC, Brownrigg M, Tremblay MS. A model of knowledge translation in health: the active healthy kids Canada report card on physical activity for children and youth. Health Promot Pract. (2012) 13:320–30. doi: 10.1177/1524839911432929 April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 8 Burghard et al. 2017 Dutch Report Card+ 11. Bloemen M, Van Wely L, Mollema J, Dallmeijer A, de Groot J. Evidence for increasing physical activity in children with physical disabilities: a systematic review. Dev Med Child Neurol. (2017). 59:1004–10. doi: 10.1111/dmcn. 13422 16. Milder IEJ, Cloostermans L, van den Dool R, Preller L, Wendel-Vos GCW. Environmental and Financial Determinants of Sports, Movement and Sedentairy Behavior. [Ruimtelijke en Financiële Determinanten van Sporten, Bewegen en Sedentair Gedrag]. Bilthoven: RIVM (2012). 17. Burghard M, Knitel K, van Oost I, Tremblay MS, Takken T. Is our youth cycling to health? Results from the Netherlands’ 2016 report card on physical activity for children and youth. J Phys Act Health (2016) 13(11Suppl. 2):S218– 24. doi: 10.1123/jpah.2016-0299 12. Van den Berg S, Vos W. Health Survey /Life style Monitor [Gezondheidsenquete/Leefstijl monitor] 2015. Hague; Utrecht: Statistics Netherlands in collaboration with the National Institute for Public Health and the Environment (2015). 13. von Heijden A, van den Dool R, van Lindert C, Breedveld K. (Un)limited Sportive. Monitor Physical Activity and Sport Participation of People with a Disability.[(On)beperkt sportief. Monitor Sport- en Beweegdeelname van Mensen met een Beperking]. Utrecht: Mulier Institute (2013). Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 14. Hirshkowitz M, Whiton K, Albert SM, Alessi C, Bruni O, DonCarlos L, et al. National Sleep Foundation’s sleep time duration recommendations: methodology and results summary. Sleep Health (2015) 1:40–3. doi: 10.1016/j.sleh.2014.12.010 Copyright © 2018 Burghard, de Jong, Vlieger, and Takken. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Frontiers in Pediatrics \| www.frontiersin.org REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 15. Guideline Dignostics and Treatment of Obesity in Adults and Children. [Richtlijn Diagnostiek en Behandeling van Obesitas bij Volwassenen en Kinderen]. Utrecht: Quality institute for healthcare. [Kwaliteitsintituut voor de gezondheidszorg (CBO)] (2008). April 2018 \| Volume 6 \| Article 122 Frontiers in Pediatrics \| www.frontiersin.org 9
https://openalex.org/W4361273618	https://documentserver.uhasselt.be//bitstream/1942/40440/1/Dynamics%20of%20non-household%20contacts%20during%20the%20COVID-19%20pandemic%20in%202020%20and%202021%20in%20the%20Netherlands.pdf	English	null	Dynamics of non-household contacts during the COVID-19 pandemic in 2020 and 2021 in the Netherlands	Scientific reports	2,023	cc-by	8,400	Dynamics of non‑household contacts during the COVID‑19 pandemic in 2020 and 2021 in the Netherlands OPEN Jantien A. Backer 1, Laurens Bogaardt 1, Philippe Beutels 2, Pietro Coletti 3, W. John Edmunds 4, Amy Gimma 4, Cheyenne C. E. van Hagen 1, Niel Hens 2,3, Christopher I. Jarvis 4, Eric R. A. Vos 1, James Wambua 3, Denise Wong 1, Kevin van Zandvoort 4 & Jacco Wallinga 1,5 Jantien A. Backer 1, Laurens Bogaardt 1, Philippe Beutels 2, Pietro Coletti 3, W. John Edmunds 4, Amy Gimma 4, Cheyenne C. E. van Hagen 1, Niel Hens 2,3, Christopher I. Jarvis 4, Eric R. A. Vos 1, James Wambua 3, Denise Wong 1, Kevin van Zandvoort 4 & Jacco Wallinga 1,5 The COVID-19 pandemic was in 2020 and 2021 for a large part mitigated by reducing contacts in the general population. To monitor how these contacts changed over the course of the pandemic in the Netherlands, a longitudinal survey was conducted where participants reported on their at-risk contacts every two weeks, as part of the European CoMix survey. The survey included 1659 participants from April to August 2020 and 2514 participants from December 2020 to September 2021. We categorized the number of unique contacted persons excluding household members, reported per participant per day into six activity levels, defined as 0, 1, 2, 3–4, 5–9 and 10 or more reported contacts. After correcting for age, vaccination status, risk status for severe outcome of infection, and frequency of participation, activity levels increased over time, coinciding with relaxation of COVID-19 control measures. From early 2020 onwards, the COVID-19 outbreak grew into a global pandemic, taking only a few months to affect nearly all countries worldwide. SARS-CoV-2 virus that causes COVID-19 is transmitted between persons during contact events with physical proximity, i.e. at-risk contacts. Reducing transmission with non-pharmaceutical measures can be achieved by lowering the transmission probability per contact, e.g. by maintaining a safe distance from each other or using protective equipment such as a face mask, and by reducing the number of contacts per person, e.g. by closing schools, theaters and restaurants, teleworking or inviting fewer people at home. p p In the Netherlands, the first nation-wide lockdown was imposed on March 15th, 2020 involving all of the above measures. It was lifted stepwise on May 11th (partly opening of schools), June 2nd (further opening of schools, museums and terraces open a.o.) and July 6th (bars and restaurants open a.o.). www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| (2023) 13:5166 Methods i CoMix participant recruitment and survey design in the Netherlands. This contact survey was part of the CoMix study that was conducted in several EU countries and the UK16. Participants were recruited by the market research company Ipsos-MORI, and filled out the contact survey online. The study was carried out in two survey periods: 8 survey rounds from 16 April to 5 August 2020, referred to as the 2020 series, and 20 survey rounds from 23 December 2020 to 22 September 2021, referred to as the 2021 series. At the start of the 2020 series, 1500 participants of 18 years and older were recruited, reflecting the age distribution of the Dutch population. An overall drop-out rate of 10% per round was allowed, otherwise additional participants were recruited for the following rounds. However, the drop-out rates differed per age group leading to very few participants in the youngest age group of 18–24 in the last rounds of the 2020 series (Fig. 1A). For this reason, a preferential sampling scheme was adopted at the start of the 2021 series, oversampling the younger age groups. In total 1200 participants of 18 years and older were included and additionally 300 children between 0 and 17 years of age were included by asking adult participants to fill out the contact survey for one of their children. After 10 rounds, the age cohorts were supplemented to include 1500 participants in total again in May 2021.h t g pp p p g y The contact survey was repeated every two weeks and a survey round lasted at most 7 days, except for the survey rounds where many new participants were recruited that were allowed more time to fill out the questionnaire. In each survey round participants were asked about characteristics that can change over time, such as risk perception, risk status and—in the 2021 series—vaccination status against COVID-19. High risk participants were people with self-reported chronic respiratory disease, chronic heart disorder, chronic kidney disease, diabetes, reduced resistance to infection (due to illness or medication), morbid obesity (BMI over 40), and pregnant women, i.e. reflecting the medical indication for influenza vaccination. Missing data in the risk status of a participant were completed with the most common risk status reported over the survey rounds by the participant. The risk status of participants could change by survey round. Dynamics of non‑household contacts during the COVID‑19 pandemic in 2020 and 2021 in the Netherlands OPEN The rise of the Alpha variant in winter 2020/2021 called for an even stricter lockdown including an evening curfew. The measures were relaxed from April 28th, 2021 onwards as vaccination coverages were increasing. It is not known how the change in control measures affected the number of at-risk contacts in the Dutch general population.f f g p p To learn about the effect of control measures on contact behaviour, many countries set up contact surveys1–9. The POLYMOD study10 served as a blueprint for these as it did for many pre-COVID-19 studies11. But while these pre-COVID-19 studies aimed to measure ‘normal’ contact behaviour, the unprecedented measures that were taken during the pandemic gave a unique opportunity to determine contact behaviour under restricted conditions. Understanding contact patterns between age groups is especially important when restrictions affect age groups differently. Physical distancing measures were often aimed at reducing contacts with old or frail persons, as they have a higher chance of hospitalisation and death after infection12–14. During the first wave of the COVID-19 pandemic, the in-person contacts of elderly persons decreased15, and also surveys in the general population showed a substantial reduction of contacts1–9. p p As part of a European project, the CoMix contact survey was set up in March/April 2020 in Belgium, the UK, and the Netherlands, later followed by 19 other European countries16. In this contact study, participants 1National Institute for Public Health and the Environment, Bilthoven, The Netherlands. 2University of Antwerp, Antwerp, Belgium. 3UHasselt, Data Science Institute and I-BioStat, Hasselt, Belgium. 4London School of Hygiene and Tropical Medicine, London, UK. 5Leiden University Medical Center, Leiden, The Netherlands. *email: jantien.backer@rivm.nl \| https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 www.nature.com/scientificreports/ www.nature.com/scientificreports/ were asked to regularly report their contact behaviour, i.e. the number of unique persons they had contact with in certain locations on the day preceding the survey day, as well as some characteristics of the contacted persons such as age and sex. Participants were asked to report only at-risk contacts, specifically face-to-face conversational contacts (within 1.5 meters) and/or physical contacts. What sets the CoMix survey apart from other contact surveys is that it uses a standardised questionnaire in multiple countries and is repeated at regular intervals in the same participants. p p At the same time the CoMix survey started, a large-scale population-based SARS-CoV-2 serosurveillance study was initiated in the Netherlands17,18. This Pienter Corona (PiCo) study was repeated every 4 to 6 months in a representative sample of the Dutch population involving a few thousand participants of all ages per survey round, and is still ongoing. It also included a contact question where the participants were asked to report the number of unique persons they contacted the day before in specific age groups. An early analysis of the first two survey rounds in 2020 showed that the first lockdown decreased the number of community contacts per participant on average by 76% compared to pre-COVID behaviour4. The PiCo survey rounds are too far apart to closely monitor the contact behaviour over time, but will be used to compare to the CoMix survey results at certain points in time.f Our aim is to describe how control measures affected the contact behaviour of the Dutch general population, using the CoMix data that was collected in the Netherlands during the survey periods in 2020 and 2021. Although contacts with household members are thought to be more intensive than contacts with non-household members, they are not subject to any of the control measures. For this reason, we will focus on the number of contacts per participant outside the household because these can be affected by the control measures. The results are compared to four survey rounds of the PiCo survey that coincided with the CoMix study periods. Methods i 2020 2021 May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0 100 200 300 400 Number of participants Age group 0−11 12−17 18−24 25−34 35−44 45−54 55−64 65+ a 2020 2021 18−24 25−34 35−44 45−54 55−64 65+ 0−11 12−17 18−24 25−34 35−44 45−54 55−64 65+ 0.00 0.25 0.50 0.75 Age group High risk fraction Study population General population b 2020 2021 May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0 100 200 300 400 Number of participants a 2020 2021 18−24 25−34 35−44 45−54 55−64 65+ 0−11 12−17 18−24 25−34 35−44 45−54 55−64 65+ 0.00 0.25 0.50 0.75 Age group High risk fraction Study population General population b b b Age group 55−64 65+ 25−34 35−44 45−54 0−11 12−17 18−24 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Jan Feb Mar Apr May Jun Jul Aug Sep Oct Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 2021 Vaccination coverage Mean vaccination start date for non−risk groups Mean vaccination coverage for general population Mean vaccination coverage (and 95% CI) for study population c 55−64 25−34 0−11 Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Vaccination coverage c c Figure 1. Characteristics of study population, compared to general population. (a) Number of participants included per survey round in the 2020 and 2021 series, by age group. After 10 survey rounds in May 2021, the study population was supplemented to meet the target numbers of the first survey round again. (b) The fraction of high risk participants of the 2020 and 2021 series (points with 95% confidence interval), and of the general population19 (bars). Only participants with unambiguous risk status are included. (c) Vaccination coverage in 2021 per age group of the study population (points with 95% confidence interval) and of the general population20 (line). The dashed line denotes the average date when non-risk groups were invited for vaccination. The dashed line for the 55–64 age group is in fact the average for 55–59 year olds, as 60–64 year olds were invited by their general practitioner. Analysis of number of contacts. Methods i https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \| www.nature.com/scientificreports/ 2020 2021 May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0 100 200 300 400 Number of participants Age group 0−11 12−17 18−24 25−34 35−44 45−54 55−64 65+ a 2020 2021 18−24 25−34 35−44 45−54 55−64 65+ 0−11 12−17 18−24 25−34 35−44 45−54 55−64 65+ 0.00 0.25 0.50 0.75 Age group High risk fraction Study population General population b 55−64 65+ 25−34 35−44 45−54 0−11 12−17 18−24 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Jan Feb Mar Apr May Jun Jul Aug Sep Oct Jan Feb Mar Apr May Jun Jul Aug Sep Oct 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 2021 Vaccination coverage Mean vaccination start date for non−risk groups Mean vaccination coverage for general population Mean vaccination coverage (and 95% CI) for study population c Figure 1. Characteristics of study population, compared to general population. (a) Number of participants included per survey round in the 2020 and 2021 series, by age group. After 10 survey rounds in May 2021, the study population was supplemented to meet the target numbers of the first survey round again. (b) The fraction of high risk participants of the 2020 and 2021 series (points with 95% confidence interval), and of the general population19 (bars). Only participants with unambiguous risk status are included. (c) Vaccination coverage in 2021 per age group of the study population (points with 95% confidence interval) and of the general population20 (line). The dashed line denotes the average date when non-risk groups were invited for vaccination. The dashed line for the 55–64 age group is in fact the average for 55–59 year olds, as 60–64 year olds were invited by their general practitioner. Methods i The fraction of high risk participants is calculated per age group using only the participants with unambiguous risk status, and compared to the population fraction with a medical indication for influenza vaccination19.h p pl The 2021 series coincided with the roll-out of the COVID-19 vaccination campaign. Health care workers, persons with a high medical risk and care home residents were vaccinated with priority, while the non-risk groups were vaccinated from elderly to young. Most persons were vaccinated with either BionTech/Pfizer, Moderna or to a smaller extent Janssen vaccine by the Municipal Health Services, except for 60–64 year olds who were vaccinated at an early stage with AstraZeneca by their general practitioner. Participants are considered to be vaccinated after they have reported to have had at least one vaccination dose. The fraction of vaccinated participants is compared to the vaccination coverage of one vaccine dose in the general population over time20. p p p g g p p In the main part of the survey, participants reported their contact behaviour of the previous day. For each unique person they had conversational or physical contact with between 5 am on the previous day and 5 am of the survey day, participants filled out the characteristics of the contacted person, such as age group and gender, as well as duration and location of the contact event. Multiple contact events with the same person were to be considered as one contact, and only contact events with a close physical proximity were to be reported, i.e. excluding online or phone contacts. Besides these individual contacts, it was also possible to report group contacts from the third survey round onwards in late May 2020. Group contacts were reported as the total number of contacted persons in broad age groups (0–17, 18–64, 65 +) at work, school or another location. More details on the CoMix study and questionnaire can be found in previous publications16,21,22.hi q p p The Medical Research Ethics Committee (MREC) NedMec confirmed that the Medical Research Involving Human Subjects Act (WMO) does not apply to the CoMix study in the Netherlands (research protocol number 22/917). Therefore an official approval of this study by the MREC NedMec is not required under the WMO. The study was in accordance with relevant GDPR guidelines and regulations. Participants gave informed consent to participate in the study before taking part. Results St d Study population. In total, 1659 participants were included in the 2020 series, and 2514 participants in the 2021 series. Participants were equally distributed over men and women in 2020, while female participants were slightly overrepresented in 2021 (Table 1). Household sizes of 2 persons were most common (Table 1). Each series started with a target population of 1500 participants, and the number of participants declined in each survey round (Supplementary Table S1, Supplementary Fig. S1). In 2020 the drop-out rate was on average 4% per round in the oldest age group and 14% in the youngest age group; in 2021 drop out rates were more comparable for all age groups (Fig. 1A).h g g p g The fraction of self-reported high risk adult participants was higher than in the general population, especially in the 2020 series (Fig. 1B). The vaccination coverage of the study population increased over time largely in agreement with the general population (Fig. 1C). The participants in the age groups of 18–34 years achieved a higher vaccination coverage at an earlier stage than the general population. The participants in the age group of 65 years and older lagged behind the general population. Contact behaviour. The distribution of the number of contacts reveals that the adult and elderly age groups in both series exhibit heavy tails (Fig. 2). The fitted lines on log-log scale (inset in Fig. 2) for these four groups have slopes ranging from − 1.1 to − 1.5. These values indicate that the estimates of the sample mean are stable but the estimates of the sample variance are not23, which makes the choice to model the number of contacts with a parametric distribution less obvious. The number of contacts in the child age group lacks a heavy tail, because of the abundance of children reporting contact numbers in the range of a class size. The numbers of contacts per participant per round are categorized in activity levels of 0, 1, 2, 3–4, 5–9, ≥ 10 contacts per participant per round.fif Separate analyses are carried out for the different series and age groups. Of the fixed effects only the weekends significantly reduces the activity levels in all analyses (Supplementary Table S2). For the youngest age group, also holidays significantly reduce the activity levels as they also include school holidays. www.nature.com/scientificreports/ where is the logistic function, θl are the thresholds for the activity levels l, βid are the random intercepts and the s() function indicates that splines are used. The fits are assessed by explained deviance and by comparing the fitted activity levels to the observations.i i y Next, the fitted results are used in a synthetic population that has the same size as the study population, but reflects the general population, with respect to age24, risk status19 and vaccination status20 at each day of the study periods. Participant id’s are sampled with replacement from the study population, and participant round is set to 1 for each day during the study period. 200 parameter sets are sampled from the estimated model parameters, and combined with 200 synthetic populations, to capture both the uncertainty of the model parameters and the uncertainty caused by the sample size of the study population. The predicted activity levels of the synthetic population are compared to the activity levels measured in PiCo survey rounds 1 (Apr 2020), 2 (Jun 2020), 4 (Feb 2021) and 5 (Jun 2021). All analyses are performed in R version 4.1.325 using the mgcv package26 with family ocat for ordered categorical variables. Code is available on Github27, and data is published on Zenodo28 in the standardised format of socialcontactdata.org. Results St d Activity levels increase over time during the two series, indicating that the underlying number of non-household contacts per participant increases (Supplementary Fig. S2), corresponding to the lifting of COVID-19 control measures. In all analyses the activity levels decrease with participant round. This means that participants tend to report fewer contacts the more often they participate. This is especially the case for the first few rounds, after which the effect stabilises.h thitf The deviance explained by this model ranges from 14 to 19% (Supplementary Table S2). Using these model results the fitted activity levels in the study population are compared to the data (Fig. 3). With the exception of some rounds, notably the first round in 2020 and the fourth round in 2021, the fitted activity levels agree well with the observed activity levels. y Next, the model results are extrapolated to the general population over the full course of the study periods (Fig. 4). The saw-tooth pattern is caused by the weekend effect, which is largest for the 0–17 age group. Over time activity levels generally increase in all age groups in both series (Fig. 4). As a comparison, the results of the PiCo study4 are plotted, of which rounds 1, 2, 4 and 5 fell within the CoMix study periods. Participants of the PiCo study reported higher activity levels than CoMix participants. The discrepancy is largest for the 0–17 age group, whereas the 65 + age group all but agrees with the PiCo study results. Methods i The participants are stratified into three age groups (0–17, 18–64, 65 +) and two series (2020, 2021). Only contacts with non-household members are included, as these are most likely to change over the course of the study period. Twelve occasions were excluded where participants reported more than 1000 contacts in a single survey round, and 58 participants were excluded that participated in four or more rounds but did not report any contact. These excluded participants did not have a significantly different age than the included participants, but they did have a significantly smaller household size.h p p y gi y The distribution of the number of contacts per participant per round had a heavy tail, which can result in unstable estimates of the variance23. Instead of using a parametric distribution, we define activity levels of 0, 1, 2, 3–4, 5–9, ≥ 10 contacts per participant per round and analyse these using ordinal regression. The activity levels are used as variables of a generalised additive model with fixed effects for participant vaccination status, participant risk status, weekends and (school) holidays, cubic splines for calendar time, participant age, participant round (i.e. the nth time a participant participated), and a random intercept for each participant. The statistical model can be described as: Pr(activity ≤l) = [θl −(βid + β0 + β1 ∗s(date) + β2 ∗s(part_round) + β3 ∗s(part_age)+ β4 ∗part_vacc + β5 ∗part_risk + β6 ∗weekend + β7 ∗holiday)], Pr(activity ≤l) = [θl −(βid + β0 + β1 ∗s(date) + β2 ∗s(part_round) + β3 ∗s(part_age)+ β4 ∗part_vacc + β5 ∗part_risk + β6 ∗weekend + β7 ∗holiday)], Pr(activity ≤l) = [θl −(βid + β0 + β1 ∗s(date) + β2 ∗s(part_round) + β3 ∗s(part_age)+ β4 ∗part_vacc + β5 ∗part_risk + β6 ∗weekend + β7 ∗holiday)], Scientific Reports \| (2023) 13:5166 \| https://doi.org/10.1038/s41598-023-32031-7 www.nature.com/scientificreports/ Discussion 2020 series n % Referencea (%) Total 1659 Age group 18–24 84 5 11 25–34 236 14 16 35–44 257 15 15 45–54 337 20 18 55–64 312 19 17 65 + 433 26 24 Gender Female 814 49 51 Male 845 51 49 High risk Mixed 359 b No 744 58 78 Yes 540 42 22 Missing 16 b Household size 1 409 25 17c 2 716 43 31 3 241 15 17 4 203 12 23 5 + 90 5 12 2021 series n % Reference (%) Total 2514 Age group 0–11 240 10 12 12–17 290 12 7 18–24 511 20 9 25–34 378 15 13 35–44 254 10 12 45–54 271 11 14 55–64 230 9 14 65 + 340 14 20 Gender Female 1383 55 50 Male 1131 45 50 High risk Mixed 385 b No 1720 81 81 Yes 400 19 19 Missing 9 b Household size 1 573 23 17c 2 766 30 31 3 444 18 17 4 496 20 23 5 + 235 9 12 Table 1. Participants in the CoMix surv general population based on 18+ popula population only calculated for participan 2020 series n % Referencea (%) Total 1659 Age group 18–24 84 5 11 25–34 236 14 16 35–44 257 15 15 45–54 337 20 18 55–64 312 19 17 65 + 433 26 24 Gender Female 814 49 51 Male 845 51 49 High risk Mixed 359 b No 744 58 78 Yes 540 42 22 Missing 16 b Household size 1 409 25 17c 2 716 43 31 3 241 15 17 4 203 12 23 5 + 90 5 12 2021 series n % Reference (%) Total 2514 Age group 0–11 240 10 12 12–17 290 12 7 18–24 511 20 9 25–34 378 15 13 35–44 254 10 12 45–54 271 11 14 55–64 230 9 14 65 + 340 14 20 Gender Female 1383 55 50 Male 1131 45 50 High risk Mixed 385 b No 1720 81 81 Yes 400 19 19 Missing 9 b Household size 1 573 23 17c 2 766 30 31 3 444 18 17 4 496 20 23 5 + 235 9 12 Table 1. Participants in the CoMix survey in the Netherlands in 2020 and 2021. a Characteristics for 2020 general population based on 18+ population to match study population. Discussion In this paper we showed how the number of at-risk contacts per person outside the household increased as the COVID-19 control measures were lifted in the Netherlands, based on the Dutch data of the European CoMix survey.hi These findings are consistent with those found in the PiCo studies, even though the activity levels reported in the PiCo data are higher than in the CoMix data. The two studies differed in how the participants report their contacts. In the CoMix study contacts were reported individually with a lot of detail on location, duration, distance from contact, protection measures, etc. Also for the group contacts at work, school, or other places additional details were requested. In the PiCo study a participant reported the total number of contacts in specific age classes, which made reporting many contacts easier. This may partly explain the higher activity levels found in the PiCo study.i y We defined six activity levels and used ordinal regression to describe the contact behaviour of the study population. Other studies have used parametric distributions instead, with a maximum number of contacts22. Scientific Reports \| (2023) 13:5166 \| https://doi.org/10.1038/s41598-023-32031-7 www.nature.com/scientificreports/ Table 1. Participants in the CoMix survey in the Netherlands in 2020 and 2021. a Characteristics for 202 general population based on 18+ population to match study population. b Percentage of risk status in stu population only calculated for participants with unambiguous risk status to allow comparison with gener population. c Expected household size distribution per Dutch citizen, based on Dutch households consist one (38%), two (33%), three (12%), four (12%) and five or more (5%) persons4. Discussion Distribution by series (line type) and age group (color), with complementary cumulative distribution function on y-axis. The inset shows the same plot on log–log scale. 2020 2021 0−17 18−64 65+ May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 Fraction of participants Number of contacts > 0 > 1 > 2 > 4 > 9 Observed Fitted Figure 3. Fitted and observed activity levels over time by series (columns) and age group (rows). Activity levels are shown as the fraction of participants that report more than a certain number of non-household contacts per day. With five limits (> 0, > 1, > 2, > 4 and > 9) six activity levels are defined, e.g. the fraction between the limits of > 2 and > 4 is the activity level that represents 3 or 4 contacts per participant. The model fits per round are shown by the median (lines) and 95% interval (shaded), from the first to last participation date of that survey round. The observed activity levels per round (open circles) are placed at the mean participation date of that survey round. Holidays and school holiday periods are shaded in grey. 2020 2021 0−17 18−64 65+ May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 Fraction of participants Number of contacts > 0 > 1 > 2 > 4 > 9 Observed Fitted Figure 3. Fitted and observed activity levels over time by series (columns) and age group (rows). Activity levels are shown as the fraction of participants that report more than a certain number of non-household contacts per day. With five limits (> 0, > 1, > 2, > 4 and > 9) six activity levels are defined, e.g. the fraction between the limits of > 2 and > 4 is the activity level that represents 3 or 4 contacts per participant. The model fits per round are shown by the median (lines) and 95% interval (shaded), from the first to last participation date of that survey round. The observed activity levels per round (open circles) are placed at the mean participation date of that survey round. Holidays and school holiday periods are shaded in grey. Discussion b Percentage of risk status in study population only calculated for participants with unambiguous risk status to allow comparison with general population. c Expected household size distribution per Dutch citizen, based on Dutch households consisting of one (38%), two (33%), three (12%), four (12%) and five or more (5%) persons4. Table 1. Participants in the CoMix survey in the Netherlands in 2020 and 2021. a Characteristics for 2020 general population based on 18+ population to match study population. b Percentage of risk status in study population only calculated for participants with unambiguous risk status to allow comparison with general population. c Expected household size distribution per Dutch citizen, based on Dutch households consisting of one (38%), two (33%), three (12%), four (12%) and five or more (5%) persons4. Table 1. Participants in the CoMix survey in the Netherlands in 2020 and 2021. a Characteristics for 2020 general population based on 18+ population to match study population. b Percentage of risk status in study population only calculated for participants with unambiguous risk status to allow comparison with general population. c Expected household size distribution per Dutch citizen, based on Dutch households consisting of one (38%), two (33%), three (12%), four (12%) and five or more (5%) persons4. This maximum is necessary to yield the same trends in contact behaviour as our results (Supplementary Fig. S3) but also disregards the heavy tails of the distribution of number of contacts. Like the ordinal regression model, machine learning models such as tree models and neural network models29 are agnostic methods. For this reason https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \| www.nature.com/scientificreports/ www.nature.com/scientificreports/ 0.00 0.25 0.50 0.75 1.00 0 250 500 750 Number of non−household contacts Fraction of participants reporting more contacts Series 2021 2020 Age group 0−17 18−64 65+ 1e−04 1e−03 1e−02 1e−01 1e+00 1 10 100 1000 Figure 2. Distribution of number of contacts per participant per round excluding household members. Distribution by series (line type) and age group (color), with complementary cumulative distribution function on y-axis. The inset shows the same plot on log–log scale. 0.00 0.25 0.50 0.75 1.00 0 250 500 750 Number of non−household contacts Fraction of participants reporting more contacts Series 2021 2020 Age group 0−17 18−64 65+ 1e−04 1e−03 1e−02 1e−01 1e+00 1 10 100 1000 Figure 2. Distribution of number of contacts per participant per round excluding household members. Discussion 2020 2021 0−17 18−64 65+ May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 Fraction of population Figure 4. Predicted activity levels over time by series (columns) and age group (rows) for the general population. Activity levels are shown as the fraction of the population that has more than a certain number of non-household contacts per day. With five limits (> 0, > 1, > 2, > 4 and > 9) six activity levels are defined, e.g. the fraction between the limits of > 2 and > 4 is the activity level that represents 3 or 4 contacts per person. The model predictions are shown by the median (lines) and 95% interval (shaded). The activity levels observed in the contact survey of the independent PiCo study are shown for comparison per round at the mean participation date (points), with bootstrapped 95% confidence intervals. Holidays and school holiday periods are shaded in grey. of the participant rounds over the survey rounds22. Also, a lower frequency of the survey rounds may lead to a smaller fatigue effect, but this would conflict with the aim of monitoring contact behaviour close to real time.if gf ,l g To extrapolate our findings to the general population, the fatigue effect needed to be corrected for, and differences between the study population and the general population had to be taken into account. Particularly the age distribution of the study participants varied over the survey rounds, because drop-out rates differed between age groups. Also vaccination and risk status of the participants did not always reflect the vaccination and risk status in the general population. The participants in the age groups of 18–34 years achieved a higher vaccination coverage at an earlier stage than the general population. This could indicate that these groups contained relatively many health care workers, who were vaccinated with priority and may have a higher willingness to get vaccinated. The participants in the age group of 65 years and older lagged behind the general population, because the study population probably did not contain many care home residents who were vaccinated with priority. The fraction of high risk adult participants was higher than in the general population. Data availability The datasets analysed in the current study are available in the Zenodo repository (https://doi.org/10.5281/zenodo. 4790347)28 under CC BY 4.0 license. Discussion we would not expect qualitatively different results as all information is contained in the data. For robustness we also reanalysed the data without distinction between age groups, but using participant age as a covariable over the full age range. Although not markedly different from the main results (Supplementary Fig. S4), this analysis fails to capture any interactions between age group and calendar time.hf Participants tend to report fewer contacts when they participate in more rounds. This fatigue effect is also seen in the CoMix data of other countries, but the fatigue effect in the Netherlands is found to be extreme compared to these30. A complicating factor in interpreting the fatigue effect is that the participant rounds are not equally distributed over the study period. Most participants have their first participant round in survey round 1 in 2020, and survey rounds 1 and 11 in 2021, due to the study design. In an alternative design used in the UK, the study population is supplemented to reach the target size in every survey round, which leads to a better distribution https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \| www.nature.com/scientificreports/ 2020 2021 0−17 18−64 65+ May Jun Jul Aug Jan Feb Mar Apr May Jun Jul Aug Sep 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75 Fraction of population Number of contacts > 0 > 1 > 2 > 4 > 9 PiCo survey Figure 4. Predicted activity levels over time by series (columns) and age group (rows) for the general population. Activity levels are shown as the fraction of the population that has more than a certain number of non-household contacts per day. With five limits (> 0, > 1, > 2, > 4 and > 9) six activity levels are defined, e.g. the fraction between the limits of > 2 and > 4 is the activity level that represents 3 or 4 contacts per person. The model predictions are shown by the median (lines) and 95% interval (shaded). The activity levels observed in the contact survey of the independent PiCo study are shown for comparison per round at the mean participation date (points), with bootstrapped 95% confidence intervals. Holidays and school holiday periods are shaded in grey. Discussion The risk status was reported by participants themselves and compared to the fraction of the population that is invited for the annual influenza vaccination. Although the high risk definition is the same for both populations, participants may have assessed their risk status differently than their GP would. This is supported by the finding that the high risk participants are mainly overrepresented in the highest age groups in 2020 (Fig. 1B, Table 1) when risks may have been perceived to be higher than in 2021 when mass-vaccination against COVID-19 was implemented. This would explain the discrepancy, but it cannot be excluded that they were really high risk participants.i p y y y g p p Our findings suggest that the contact data collected in the CoMix survey describe trends in contact behaviour in the general population. As such it can be useful in near real-time monitoring of the transmission potential22,31, determining the effect of control measures that are aimed at contact reduction32,33, describing determinants for contact behaviour34,35, or serving as input for infectious disease models36,37. These results can help guide policy in future waves of COVID-19 or other emerging respiratory diseases by monitoring the interaction between contact behaviour, control measures and compliance, and emphasize that contact surveys such as the CoMix study are indispensable for providing a quantitative basis to public health policy. Data availability The datasets analysed in the current study are available in the Zenodo repository (https://doi.org/10.5281/zenodo. 4790347)28 under CC BY 4.0 license. Received: 18 October 2022; Accepted: 21 March 2023 Received: 18 October 2022; Accepted: 21 March 2023 Received: 18 October 2022; Accepted: 21 March 2023 https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \| www.nature.com/scientificreports/ References h References 1. Zhang, J. et al. Changes in contact patterns shape the dynamics of the COVID-19 outbreak in China. Science 368, 1481–1486. https://doi.org/10.1126/science.abb8001 (2020). 1. Zhang, J. et al. Changes in contact patterns shape the dynamics of the COVID-19 outbreak in China. Science 368, 1481–1486. https://doi.org/10.1126/science.abb8001 (2020). p g 2. Latsuzbaia, A., Herold, M., Bertemes, J. P. & Mossong, J. Evolving social contact patterns during the COVID-19 crisis in Luxembourg. PLoS ONE 15, e0237128. https://doi.org/10.1371/journal.pone.0237128 (2020).h g 2. Latsuzbaia, A., Herold, M., Bertemes, J. P. & Mossong, J. Evolving social contact patterns during the COVID-19 crisis in Luxembourg. PLoS ONE 15, e0237128. https://doi.org/10.1371/journal.pone.0237128 (2020).h g p g j p 3. Quaife, M. et al. The impact of COVID-19 control measures on social contacts and transmission in Kenyan informal settlements BMC Med. 18, 316. https://doi.org/10.1186/s12916-020-01779-4 (2020). p g ( ) 4. Backer, J. A. et al. Impact of physical distancing measures against COVID-19 on contacts and mixing patterns: repeated cross- sectional surveys, the Netherlands, 2016–2017, April 2020 and June 2020. Euro Surveill. 26, 1–10. https://doi.org/10.2807/1560- 7917.ES.2021.26.8.2000994 (2021).i p g 4. Backer, J. A. et al. Impact of physical distancing measures against COVID-19 on contacts and mixing patterns: repeated cross- sectional surveys, the Netherlands, 2016–2017, April 2020 and June 2020. Euro Surveill. 26, 1–10. https://doi.org/10.2807/1560- 7917.ES.2021.26.8.2000994 (2021).i ( ) 5. Bosetti, P. et al. Lockdown impact on age-specific contact patterns and behaviours, France, April 2020. Euro Surveill.https://doi. org/10.2807/1560-7917.ES.2021.26.48.2001636 (2021). ( ) 5. Bosetti, P. et al. Lockdown impact on age-specific contact patterns and behaviours, France, April 2020. Euro Surveill.https://doi org/10.2807/1560-7917.ES.2021.26.48.2001636 (2021). g ( ) 6. Feehan, D. M. & Mahmud, A. S. Quantifying population contact patterns in the United States during the COVID-19 pandemic Nat. Commun. 12, 893. https://doi.org/10.1038/s41467-021-20990-2 (2021). g 6. Feehan, D. M. & Mahmud, A. S. Quantifying population contact patterns in the United States during the COVID-19 pandemic. Nat. Commun. 12, 893. https://doi.org/10.1038/s41467-021-20990-2 (2021). p g 7. Tomori, D. V. et al. Individual social contact data and population mobility data as early markers of SARS-CoV-2 transmission dynamics during the first wave in Germany-an analysis based on the COVIMOD study. BMC Med. 19, 271. https://doi.org/10 1186/s12916-021-02139-6 (2021). p g 7. Tomori, D. V. et al. Individual social contact data and population mobility data as early markers of SARS-CoV-2 transmission dynamics during the first wave in Germany-an analysis based on the COVIMOD study. BMC Med. 19, 271. https://doi.org/10. 1186/s12916-021-02139-6 (2021). 8. References h Epidemic Models: Their Spread, Analysis and Invasions in Scale-Free Networks 1–25 (Springer, 2015).t h 24. Statistics Netherlands. Bevolking; geslacht, leeftijd en burgerlijke staat, 1 januari 2021. (2021). https://opendata.cbs.nl/statline/#/ CBS/nl/dataset/7461bev/table. Accessed 1 July 2021. h 24. Statistics Netherlands. Bevolking; geslacht, leeftijd en burgerlijke staat, 1 januari 2021. (20 CBS/nl/dataset/7461bev/table. Accessed 1 July 2021. 5. R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2022). 6. Wood, S. Generalized Additive Models: An Introduction with R 2nd edn. (Chapman and Hall/CRC, 2017). 25. R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2022). 26. Wood, S. Generalized Additive Models: An Introduction with R 2nd edn. (Chapman and Hall/CRC, 2017). g g f p g ( p 26. Wood, S. Generalized Additive Models: An Introduction with R 2nd edn. (Chapman and Hall/CRC, 2017). ( p ) 7. RIVM. CoMixNL. https://github.com/rivm-syso/CoMixNL (2022). 27. RIVM. CoMixNL. https://github.com/rivm-syso/CoMixNL (2022). 27. RIVM. CoMixNL. https://github.com/rivm-syso/CoMixNL (2022). p g y 8. Backer, J. A., Bogaardt, L. & Wallinga, J. CoMix social contact data (Netherlands). (2022). https://zenodo.org/record/7276465#.Y Sk28nMKUk, https://doi.org/10.5281/zenodo.7276465. Accessed 9 Feb 2023.h 28. Backer, J. A., Bogaardt, L. & Wallinga, J. CoMix social contact data (Netherlands). (2022). https://zenodo.org/record/7276465#.Y- Sk28nMKUk, https://doi.org/10.5281/zenodo.7276465. Accessed 9 Feb 2023.h p g 29. Hastie, T., Tibshirani, R. & Friedman, J. The Elements of Statistical Learning 2nd edn. (Springer, 2009). h 0. Wong, K. L. M. et al. Social contact patterns during the COVID-19 pandemic in 21 European countries: Evidence from a two-yea study. MedRxivhttps://doi.org/10.1101/2022.07.25.22277998 (2022).h h 30. Wong, K. L. M. et al. Social contact patterns during the COVID-19 pan study. MedRxivhttps://doi.org/10.1101/2022.07.25.22277998 (2022).h 30. Wong, K. L. M. et al. Social contact patterns during the COVID-19 pandemic in 21 European countries: E study. MedRxivhttps://doi.org/10.1101/2022.07.25.22277998 (2022).h 30. Wong, K. L. M. et al. Social contact patterns during the COVID 19 pandemic in 21 European countries: Evidence from a two year study. MedRxivhttps://doi.org/10.1101/2022.07.25.22277998 (2022). 31. Jarvis, C. I. et al. The impact of local and national restrictions in response to COVID-19 on social contacts in England: A study. MedRxivhttps://doi.org/10.1101/2022.07.25.22277998 (2022). 31. Jarvis, C. I. et al. The impact of local and national restrictions in response to COVID-19 on social contacts in England: A l d l l d h d ( ) y p g 1. Jarvis, C. I. et al. References h McCreesh, N. et al. Impact of the Covid-19 epidemic and related social distancing regulations on social contact and SARS-CoV-2 transmission potential in rural South Africa: analysis of repeated cross-sectional surveys. BMC Infect. Dis. 21, 928. https://doi org/10.1186/s12879-021-06604-8 (2021).h g ( ) 9. Drolet, M. et al. Time trends in social contacts before and during the COVID-19 pandemic: The CONNECT study. BMC Public Health 22, 1032. https://doi.org/10.1186/s12889-022-13402-7 (2022). g ( ) 9. Drolet, M. et al. Time trends in social contacts before and during the COVID-19 pandemic: The CONNECT study. BMC Public Health 22, 1032. https://doi.org/10.1186/s12889-022-13402-7 (2022). p g ( ) 10. Mossong, J. et al. Social contacts and mixing patterns relevant to the spread of infectious diseases. PLoS Med. 5, e74. https://doi. org/10.1371/journal.pmed.0050074 (2008). p g 0. Mossong, J. et al. Social contacts and mixing patterns relevant to the spread of infectious diseases. PLoS Med. 5, e74. https://doi org/10.1371/journal.pmed.0050074 (2008). g j p 1. Hoang, T. et al. A systematic review of social contact surveys to inform transmission models of close-contact infections Epidemiology 30, 723–736. https://doi.org/10.1097/EDE.0000000000001047 (2019).i p gy , p g ( ) 12. Levin, A. T. et al. Assessing the age specificity of infection fatality rates for COVID-19: Systematic review, meta-analysis, and public policy implications. Eur. J. Epidemiol. 35, 1123–1138. https://doi.org/10.1007/s10654-020-00698-1 (2020).hf p gy p g 12. Levin, A. T. et al. Assessing the age specificity of infection fatality rates for COVID-19: Systematic review, meta-analysis, and public policy implications. Eur. J. Epidemiol. 35, 1123–1138. https://doi.org/10.1007/s10654-020-00698-1 (2020).hf p y p p p g 3. Hewitt, J. et al. The effect of frailty on survival in patients with COVID-19 (COPE): A multicentre, European, observational cohort study. Lancet Public Health 5, e444–e451. https://doi.org/10.1016/s2468-2667(20)30146-8 (2020). y p g 14. Blomaard, L. C. et al. Frailty is associated with in-hospital mortality in older hospitalised COVID-19 patients in the Netherla The COVID-OLD study. Age Ageing 50, 631–640. https://doi.org/10.1093/ageing/afab018 (2021). g t al. Frailty is associated with in-hospital mortality in older hospita h y g g g p g g g 5. Freedman, V. A., Hu, M. & Kasper, J. D. Changes in older adults’ social contact during the COVID-19 pandemic. J. Gerontol. B 77, e160–e166. https://doi.org/10.1093/geronb/gbab166 (2022). p g g g 16. Verelst, F. et al. SOCRATES-CoMix: A platform for timely and open-source contact mixing data during and in between COVID-19 surges and interventions in over 20 European countries. BMC Med. 19, 254. References h https://doi.org/10.1186/s12916-021-02133-y (2021). 17 V E R A t l N ti id l f SARS C V 2 d id tifi ti f i k f t i th l l ti f th 6. Verelst, F. et al. SOCRATES-CoMix: A platform for timely and open-source contact mixing data during and in between COVID-19 surges and interventions in over 20 European countries. BMC Med. 19, 254. https://doi.org/10.1186/s12916-021-02133-y (2021) 16. Verelst, F. et al. SOCRATES-CoMix: A platform for timely and open-source contact mixing data during and in between COVID-19 surges and interventions in over 20 European countries. BMC Med. 19, 254. https://doi.org/10.1186/s12916-021-02133-y (2021). 17. Vos, E. R. A. et al. Nationwide seroprevalence of SARS-CoV-2 and identification of risk factors in the general population of the i surges and interventions in over 20 European countries. BMC Med. 19, 254. https://doi.org/10.1186/s12916-021-02133-y (2021) 7. Vos, E. R. A. et al. Nationwide seroprevalence of SARS-CoV-2 and identification of risk factors in the general population of the Netherlands during the first epidemic wave. J. Epidemiol. Community Healthhttps://doi.org/10.1136/jech-2020-215678 (2020). 7. Vos, E. R. A. et al. Nationwide seroprevalence of SARS-CoV-2 and identification of risk factors in the general population of the Netherlands during the first epidemic wave. J. Epidemiol. Community Healthhttps://doi.org/10.1136/jech-2020-215678 (2020). i 18. Vos, E. R. A. et al. Associations between measures of social distancing and severe acute respiratory syndrome coronavirus 2 seropositivity: A nationwide population-based study in the Netherlands. Clin. Infect. Dis. 73, 2318–2321. https://doi.org/10.1093/ cid/ciab264 (2021). 19. Hooiveld, M., Heins, M., Hendriksen, J. & Korevaar, J. Aantal mensen met medische indicatie voor vaccinatie tegen COVID-19 (Number of people with medical indication for vaccination against COVID-19). (2021). https://www.nivel.nl/sites/default/files/ bestanden/1003980.pdf. Accessed 8 Aug 2022. p g 20. RIVM. Cumulatieve opkomst tenminste één COVID-19 vaccinatie naar geboortejaar en week (Cumulative coverage of at least one COVID-19 vaccination per birth year and calendar week). (2022). https://www.rivm.nl/covid-19-vaccinatie/cijfers-vaccinatie programma. Accessed 14 April 2022. p g p 1. Jarvis, C. I. et al. Quantifying the impact of physical distance measures on the transmission of COVID-19 in the UK. BMC Med 18, 124. https://doi.org/10.1186/s12916-020-01597-8 (2020). p g 22. Gimma, A. et al. Changes in social contacts in England during the COVID-19 pandemic between March 2020 and March 2021 as measured by the CoMix survey: A repeated cross-sectional study. PLoS Med. 19, e1003907. https://doi.org/10.1371/journal.pmed. 1003907 (2022).h ( ) 23. Tagore, S. Acknowledgementsh g The authors would like to thank all participants of the CoMix survey. The CoMix study in the Netherlands was made possible through funding from the European Union’s Horizon 2020 research and innovation programme— project EpiPose (Grant Agreement Number 101003688). This work reflects only the authors’ view. The European Commission is not responsible for any use that may be made of the information it contains. We are grateful to the EpiPose management team for coordinating the CoMix survey so data collection was made possible in the Netherlands. References h The impact of local and national restrictions in response to COVID-19 on social contacts in England: A longitudinal natural experiment. BMC Med. 19, 52. https://doi.org/10.1186/s12916-021-01924-7 (2021).t . Jarvis, C. I. et al. The impact of local and national restriction g p p g 2. Coletti, P. et al. CoMix: Comparing mixing patterns in the Belgian population during and after lockdown. Sci. Rep. 10, 21885 https://doi.org/10.1038/s41598-020-78540-7 (2020). p g ( ) 3. Munday, J. D. et al. Estimating the impact of reopening schools on the reproduction number of SARS-CoV-2 in England, using weekly contact survey data. BMC Med. 19, 233. https://doi.org/10.1186/s12916-021-02107-0 (2021).hl y y g 4. Wambua, J. et al. The influence of risk perceptions on close contact frequency during the SARS-CoV-2 pandemic. Sci. Rep. 12 5192. https://doi.org/10.1038/s41598-022-09037-8 (2022). p g 35. Wong, K. L. M. et al. Pregnancy during COVID-19: Social contact patterns and vaccine coverage of pregnant women from Co in 19 European countries. BMC Pregnancy Childbirth 22, 757. https://doi.org/10.1186/s12884-022-05076-1 (2022). p g 35. Wong, K. L. M. et al. Pregnancy during COVID-19: Social con 35. Wong, K. L. M. et al. Pregnancy during COVID-19: Social contact patterns and vaccine coverage of pregnant women from C in 19 European countries. BMC Pregnancy Childbirth 22, 757. https://doi.org/10.1186/s12884-022-05076-1 (2022). 36 Coletti P et al A data driven metapopulation model for the Belgian COVID 19 epidemic: Assessing the impact of lockdow p g y p g 36. Coletti, P. et al. A data-driven metapopulation model for the Belgian COVID-19 epidemic: Assessing the impact of lockdow exit strategies. BMC Infect. Dis. 21, 503. https://doi.org/10.1186/s12879-021-06092-w (2021).if g f p g 37. Franco, N. et al. Inferring age-specific differences in susceptibility to and infectiousness upon SARS-CoV-2 infection based on Belgian social contact data. PLoS Comput. Biol. 18, e1009965. https://doi.org/10.1371/journal.pcbi.1009965 (2022). https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \| www.nature.com/scientificreports/ Author contributions P.B., J.E., N.H. and J.W. initiated the study; N.H., P.B., C.J., K.v.Z., J.E. designed the survey questionnaire; A.G., C.J., K.v.Z, P.C., J.W., L.B. collected and cleaned the data; J.B. and L.B. analysed the data; J.B. wrote the manuscript; E.V., C.v.H. and D.W. collected and cleaned the PiCo data. All authors reviewed the manuscript and approved the final version for publication. Competing interests h The authors declare no competing interests. Additional informationh Supplementary Information The online version contains supplementary material available at https://doi.org/ 10.1038/s41598-023-32031-7. Correspondence and requests for materials should be addressed to J.A.B. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2023 https://doi.org/10.1038/s41598-023-32031-7 Scientific Reports \| (2023) 13:5166 \|
https://openalex.org/W3030175780	https://www.rasi.vr.uff.br/index.php/rasi/article/download/333/92	Portuguese	null	A interação com atores da hélice tríplice e as perspectivas de desenvolvimento de capacidades empreendedoras	Revista de Administração, Sociedade e Inovação	2,019	cc-by	9,567	RESUMO Esta pesquisa tem o propósito de incentivar o fortalecimento da identidade dos Institutos Federais com a pesquisa aplicada, a partir da interação dos atores da hélice tríplice. Com base no método de estudo de caso descritivo sob uma abordagem qualitativa, esta pesquisa objetivou analisar o caso de cooperação do Campus Pinheiral com a empresa Rica Alimentos. Os resultados indicam que a experiência em análise tem potencial de desenvolvimento e que o Marco Legal da Inovação exerce importante papel nesse contexto. Espera-se, assim, que essa análise auxilie decisões administrativas e incentive a promoção de desenvolvimento socioeconômico através da pesquisa aplicada. A originalidade encontra-se na análise de configurações únicas de um estudo de caso específico que traz contribuições para a rede dos Institutos Federais. p q ç p ALAVRAS-CHAVES: Hélice Tríplice; Cooperação Academia-Empresa; Institutos Federais; Marco Legal da novação. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 A Interação com Atores da Hélice Tríplice e as Perspectivas de Desenvolvimento da Cooperação Academia-Empresa: Reflexões sobre a experiência do IFRJ Campus Pinheiral Greicianne Sousa de Oliveira (IFRJ) - greicisousa@gmail.com Thiago Borges Renault (UFRRJ) – thiagorenault@gmail.com http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 A Interação com Atores da Hélice Tríplice e as Perspectivas de Desenvolvimento da Cooperação Academia-Empresa: Reflexões sobre a experiência do IFRJ Campus Pinheiral Greicianne Sousa de Oliveira (IFRJ) - greicisousa@gmail.com Thiago Borges Renault (UFRRJ) – thiagorenault@gmail.com http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 A Interação com Atores da Hélice Tríplice e as Perspectivas de Desenvolvimento da Cooperação Academia-Empresa: Reflexões sobre a experiência do IFRJ Campus Pinheiral Greicianne Sousa de Oliveira (IFRJ) - greicisousa@gmail.com Thiago Borges Renault (UFRRJ) – thiagorenault@gmail.com A Interação com Atores da Hélice Tríplice e as Perspectivas de Desenvolvimento da Cooperação Academia-Empresa: Reflexões sobre a experiência do IFRJ Campus Pinheiral Greicianne Sousa de Oliveira (IFRJ) - greicisousa@gmail.com Thiago Borges Renault (UFRRJ) – thiagorenault@gmail.com A Interação com Atores da Hélice Tríplice e as Perspectivas de Desenvolvimento da Cooperação Academia-Empresa: Reflexões sobre a experiência do IFRJ Campus Pinheiral Greicianne Sousa de Oliveira (IFRJ) - greicisousa@gmail.com Thiago Borges Renault (UFRRJ) – thiagorenault@gmail.com KEYWORDS: Triple Helix; University-Industry Cooperation; Federal Institutes; Legal Framework of Innovation 1 Introdução Apesar do potencial científico do Brasil, o país assiste a uma baixa incorporação do conhecimento em inovação. No ranking que mede a relevância de países na ciência, considerando as publicações em revistas de prestígio, o Brasil foi líder na América Latina e obteve o 23º lugar na classificação geral em 2018. No entanto, no mesmo ano, o Brasil ocupou a 64ª posição o ranking de avaliação do desempenho global em inovação, distante do líder da América Latina, o Chile, que obteve o 37º lugar1. Diante desse contexto e ao considerar um cenário de mais de 13 milhões de pessoas à busca de emprego no país2, o Brasil tem o desafio de transformar seu potencial científico em oportunidades de crescimento social e econômico. Para mais, o país enfrenta, ainda, o desafio de fortalecer suas principais fontes produtoras e disseminadoras de ciência e tecnologia, as instituições públicas de ensino e pesquisa. As universidades federais obtiveram, em 2017, o menor repasse de verbas em sete anos (Moreno, 2018). A Rede Federal de Educação Profissional, Científica e Tecnológica mais que duplicou a quantidade de alunos desde 2013, no entanto, com as reduções de orçamento, a verba de 2018 foi a mesma de 20133. Além disso, a situação agrava-se pela Emenda Constitucional nº 95 (EC 95), de 15 de dezembro de 2016, que limitou os gastos públicos por 20 anos. p p Ao considerar tais problemáticas apresenta-se que a integração entre academia, empresa e governo, pela abordagem da hélice tríplice, é fonte de transformação do potencial científico da academia em inovação e fortalecimento da academia. Com base nesses pressupostos, justifica-se a necessidade de se discutir e incentivar a cooperação entre academia-empresa em interação com o governo no cenário brasileiro. Tal importância se intensifica ao ponderar a necessidade de minimizar o hiato entre possibilidades legais, como Marco Legal da Inovação (Lei nº 13.243/2016), e práticas acadêmicas para desenvolvimento e transferência de tecnologia. Posto tudo isso, esta pesquisa tem como foco analisar a experiência de uma instituição de ensino técnico e superior em um caso de cooperação academia-empresa. O Colégio Agrícola Nilo Peçanha, em 2008, após uma longa história de formação de recursos humanos de nível técnico, incorpora-se ao Instituto Federal de Educação, Ciência e Tecnologia do Rio de Janeiro (IFRJ). 1 Dados do Nature Index Global [NIG], 2018 e do Índice Global da Inovação [IGI], 2018. 2 Dados do Instituto Brasileiro de Geografia e Estatística (2018) 3 Informação fornecida pelo presidente do Conselho Nacional das Instituições da Rede Federal de Educação Profissional, Científica e Tecnológica (Conif) e reitor do IFTM, Roberto Gil Rodrigues Almeida, em evento de celebração dos 10 anos dos Institutos Federais na Assembleia Legislativa de Minas Gerais (ALMG), realizado no 12 de junho de 2018. ABSTRACT This research aims to encourage the strengthening of the identity of the Federal Institutes with applied research, based on the interaction of the triple helix actors. Based on the descriptive case study method under a qualitative approach, this research aimed to analyze the case of cooperation of Campus Pinheiral with the company Rica Alimentos. The results indicate that the experience in analysis has development potential and that the Legal Framework of Innovation plays an important role in this context. Thus, this analysis is expected to aid administrative decisions and encourage the promotion of socioeconomic development through applied research. Originality lies in the analysis of unique configurations of a specific case study that brings contributions to the network of Federal Institutes. KEYWORDS: Triple Helix; University-Industry Cooperation; Federal Institutes; Legal Framework of Innovation Copyright © 2020 RASI. Todos os direitos, até mesmo de tradução, são reservados. É permitido citar parte de artigos sem autorização prévia, desde que seja identificada a fonte. Recebido em 31/01/2019 Aceito em 31/08/2019 Oliveira & Renault \| 25 10.20401/rasi.6.1.333 Oliveira & Renault \| 25 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 2.1 Hélice Tríplice A abordagem da hélice tríplice se estrutura no papel da academia de promover desenvolvimento socioeconômico através da aplicação da ciência (Etzkowitz & Zhou, 2017). De acordo com Jia, Zhou e Etzkowitz (2016), em contraste com as teorias que enfatizam o papel do governo ou das empresas em inovação, a hélice tríplice centra-se na universidade como propulsora de empreendedorismo e de tecnologia. Leydesdorff (2017) defende, ainda, que a tese da hélice tríplice se estrutura no argumento de que as relações entre as instituições tornaram-se baseadas em conhecimento e no crescimento em rede. Já Leydesdorff e Ivanova (2016) argumentam que o princípio dessa cooperação tríade surge da necessidade de mediação administrativa ou intervenção política das relações entre universidades e empresas. Na dinâmica da hélice tríplice, cada ator mantém a sua função principal, assim como, conserva uma identidade distinta. Para Jia, Zhou & Etzkowitz, 2016, na transcendência de papéis entre os atores, a universidade tem o papel do mercado ao estimular o desenvolvimento de novas empresas de pesquisa e ao introduzir o objetivo acadêmico de capitalização do conhecimento. Já as empresas desenvolvem treinamento para níveis cada vez mais elevados e compartilham conhecimentos, assumindo, dessa forma, atitudes tradicionais da universidade. E os governos agem como capitalistas de risco, além de suas atividades regulatórias. Para Etzkowitz e Leydesdorff (2000), a singularidade do papel do conhecimento para o desenvolvimento socioeconômico na sociedade pós-industrial está na intensidade e na aproximação entre a ciência e a sua aplicação. À vista disso, a universidade — enquanto um ator institucional de produção e difusão de conhecimentos — conquista destaque no cenário de desenvolvimento socioeconômico e se aproxima de um modelo empreendedor (Etzkowitz & Zhou, 2017). 1 Introdução Essa nova institucionalidade, vivenciada, em 2019, em conjunto com 14 campi de diferentes realidades, tornou o antigo colégio agrícola uma instituição verticalizada que integra cursos técnicos, de graduação e de pós-graduação. Essa nova identidade, como então Campus Pinheiral, incorpora, também, as missões de pesquisa e extensão à unidade. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br Oliveira & Renault \| 26 10.20401/rasi.6.1.333 Oliveira & Renault \| 26 10.20401/rasi.6.1.333 Oliveira & Renault \| 26 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral O IFRJ, apesar de possuir dez anos de existência, ainda não possui regras institucionalizadas para realização de parcerias com empresas para desenvolvimento e transferência de tecnologia. No entanto, apesar da ausência de regulamentação específica, o Campus Pinheiral possui uma parceria com uma empresa do setor de carnes de aves, a Rica Alimentos. As principais questões de pesquisa a serem examinadas são: quais os principais conflitos para a cooperação academia-empresa, identificados a partir da análise da experiência do Campus Pinheiral? A pesquisa questiona, ainda, quais são motivações e incentivos para o desenvolvimento do empreendedorismo acadêmico no IFRJ e a satisfação resultante desse processo de interação. 2.2 Universidade empreendedora A sociedade pós-industrial preconiza a inovação como fonte de diferencial competitivo. Assim sendo, o cenário contemporâneo não se compatibiliza — como outrora na sociedade industrial — com a longa lacuna de tempo entre a descoberta e a utilização dos resultados de pesquisa. Nesse contexto, a economia do conhecimento necessita de velocidade no processo de transferência de conhecimento para sobreviver em um mercado instável, competitivo e de inovações contínuas (Leydesdorff, 2017). Ao intensificar os processos de transferência de tecnologia e se aproximar de empresas, a universidade segue para a construção de uma identidade empreendedora. Para http://www.rasi.vr.uff.br http://www.rasi.vr.uff.br http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 27 10.20401/rasi.6.1.333 Oliveira & Renault \| 27 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Etzkowitz (2015b), o desenvolvimento de uma identidade empreendedora ocorre por três estágios. Em um primeiro momento, a universidade define suas prioridades e estabelece fontes diversificadas de recursos. Em seguida, a instituição desenvolve mecanismos de comercialização da propriedade intelectual, sendo eles produtos de atividades de pesquisa. Por fim, a universidade assume um papel ativo no desenvolvimento socioeconômico regional. p p g Para Corsi e Prencipe (2016), a pesquisa acadêmica, em um contexto empreendedor, alcança a missão de promover ativamente o desenvolvimento socioeconômico. Dessa forma, ao assumir a missão de gerar progresso econômico através da aplicação da ciência, a universidade adquire o potencial de oferecer retornos dos investimentos públicos à sociedade. Tal retorno se apresenta na forma de produtos, empregos e alternativas de crescimento de um país. De acordo com Bessant e Tidd (2009), além de ganhos econômicos, as inovações têm potencial para melhorar a qualidade de vida da sociedade. À vista disso, a geração de novos conhecimentos e a inovação estimulam o desenvolvimento econômico e social (Kruss; Adeoti & Nabudere, 2015). De acordo com Segatto-Mendes e Sbragia (2002), o processo de cooperação academia-empresa amplia a participação da academia no desenvolvimento tecnológico do país e aprimora sua atuação no tripé ensino, pesquisa extensão. Bonaccorsi e Piccaluga (1994) apresentam um modelo teórico de análise do processo de cooperação academia-empresa, ilustrado pela Figura 1. A esquematização propõe a compreensão de fatores que podem definir o sucesso da cooperação. À vista disso, aspectos como motivações e resultados, caso não considerados e alinhados, podem dificultar a continuidade do trabalho. Da mesma forma, a identificação de barreiras permite trabalhos a fim de minimizar e administrar conflitos. 2.2 Universidade empreendedora Assim como, o reconhecimento de facilitadores tem potencial para direcionar a ampliação de resultados. E, por fim, o modelo propõe a identificação da satisfação resultante dos envolvidos com a pesquisa. Figura 1 - Modelo Teórico do Processo de Cooperação entre Academia-empresa Fonte: Adaptado de Bonaccorsi & Piccaluga (1994) Motivações Processo de Cooperação Satisfação Resultante Barreiras e/ou Facilitadores Figura 1 - Modelo Teórico do Processo de Cooperação entre Academia-empresa Fonte: Adaptado de Bonaccorsi & Piccaluga (1994) Oliveira & Renault \| 28 10.20401/rasi.6.1.333 Oliveira & Renault \| 28 10.20401/rasi.6.1.333 Oliveira & Renault \| 28 10.20401/rasi.6.1.333 Oliveira & Renault \| 28 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral como, na permissão que as academias compartilhem seus laboratórios e equipamentos e demais instalações (Brasil, 2016). como, na permissão que as academias compartilhem seus laboratórios e equipamentos e demais instalações (Brasil, 2016). Para Etzkowitz (2009), são fontes de inovação a circulação de profissionais de vida dupla com atuação na academia e na empresa. Ao encontro de tal pressuposto, o Marco Legal da Inovação ampliou o número de horas disponíveis ao pesquisador público para exercício de atividades fora da universidade. A carga horária de atividades externas se estendeu de 120 para 416 horas anuais. Além disso, a legislação permitiu que pesquisadores públicos, em regime de dedicação exclusiva, exerçam atividades externas ligadas à ciência, à tecnologia e inovação em empresas com remuneração (BRASIL, 2016, 2018). Faucher e Ribeiro (1995) analisam que o Estado tem forte demanda de conteúdo tecnológico, representando uma porcentagem importante da demanda total de tecnologia de um país. Os autores argumentam, ainda, que a propriedade de tecnologias, a exemplo de energia, representa independência política externa de um país. Nesse contexto, Marco Legal da Inovação traz, ainda, a prerrogativa de utilizar o poder de compra do Estado para fomento à inovação e o decreto regulamentador da legislação traz diretrizes a respeito do assunto. No entanto, mesmo diante dos avanços novo marco legal e seu decreto regulamentador (Decreto Federal nº 9.283 / 2018), indefinições persistem sobre as práticas e o modo de operação da gestão da inovação por academias públicas em parcerias com empresas privadas (RAUEN; POMPO, 2016). Pedro (2016) afirma que os próximos passos são definir as políticas institucionais; testar e corrigir os mecanismos; equalizar a legislação dos Estados e prosseguir a agenda de reformas legais. 2.3 Marco Legal da Inovação A Lei de Inovação (Lei nº 10.973/2004) não foi capaz de promover a cooperação academia-empresa e alterar a dinâmica de pesquisa no Brasil. Dessa forma, a fim de reduzir obstáculos legais e burocráticos desse processo, o Marco Legal da Ciência, Tecnologia e Inovação (CT&I) nº 13.243/2016 alterou a Lei de Inovação e outras leis relacionadas ao tema, como a Lei de Licitações (Rauen, 2016). Uma importante possibilidade de cooperação academia-empresa prevista na nova legislação consiste na autorização de cessão de imóveis de instituições públicas a empresas para a instalação de “ambientes promotores da inovação” como parques tecnológicos. Assim RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br 3. Metodologia Este estudo optou pela pesquisa qualitativa ao considerar que, segundo Creswell (2014), tal abordagem busca compreender o contexto no qual estão inseridos os participantes de um estudo. A metodologia de estudo de caso norteou o caráter da investigação. A escolha do método ocorreu devido ao objetivo de investigar um fenômeno contemporâneo in loco, a partir de múltiplas fontes de evidências (Yin, 2005; Eisenhart,1989). p p ( ) A metodologia descritiva norteou o estudo de caso com o objetivo de descrever os principais conflitos para a cooperação academia-empresa e identificar as perspectivas do empreendedorismo acadêmico para o IFRJ. Gil (p. 28, 1994) afirma que tal abordagem tem como “objetivo primordial a descrição das características de determinada população ou fenômeno”. Já a seleção dos sujeitos de pesquisa foi intencional e considerou a importância da alta administração para a compreensão do caso, assim como preconizou os sujeitos envolvidos na experiência de cooperação academia-empresa em análise (Creswell, 2014). Quanto aos entrevistados, os quadros 1 e 2 apresentam a caracterização desses sujeitos. Destaca-se que os códigos E07 e EB fazem referência a perspectivas diferentes de um mesmo entrevistado. Sendo elas as seguintes: a de diretor-geral do Campus Pinheiral gestão 2018-2022 e a de responsável do laboratório em análise, desde o princípio da cooperação com a Rica até a presente data. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 3. Metodologia 2020 http://www.rasi.vr.uff.br Quadro 1 – Caracterização dos entrevistados: gestores do IFRJ e do Campus Pinheiral Código Formação Acadêmica Função: IFRJ E01 Doutor em Informática na Educação Pró-Reitor de Pesquisa, Inovação e Pós-graduação Mestre em Ensino de ciências da saúde e do ambiente Especialista em Marketing Graduado em Tecnologia em Processamento de Dados E02 Doutora em Ciências Diretora da Agência de Inovação Mestre em Propriedade Intelectual e Inovação Mestre em Ciências Biológicas E03 Mestre em Ciências Ambientais e Florestais Pró-Reitoria de Desenvolvimento Institucional, Valorização de Pessoas e Sustentabilidade Especialista em Sociedade e Agricultura no Brasil e em Ciências Ambientais Graduado em Licenciatura em Ciências Agrícolas Campus Pinheiral E04 Doutor em Nutrição Animal Diretor-Geral do Campus Pinheiral Mestre em Produção Animal Graduado em Médico Veterinário Mestre em Zootecnia Graduação em Medicina Veterinária E05 Tecnólogo em Gestão Pública Diretor de Administração Fonte: Próprios autores Quadro 2 - Caracterização dos entrevistados: responsáveis pela atividade de cooperação Rica- Campus Pinheiral Código Identificação Caracterização EA Representante da empresa responsável pelas atividades de parceria com o Campus Pinheiral Administrador EB Pesquisador do Campus Pinheiral responsável pelas atividades de cooperação com a empresa em análise Professor, pesquisador, médico- veterinário, ex-funcionário da Rica Alimentos. Fonte: Próprios autores Para análise, esta pesquisa utilizou a estrutura teórica de estudos da cooperação Quadro 2 - Caracterização dos entrevistados: responsáveis pela atividade de cooperação Rica- Campus Pinheiral Código Identificação Caracterização EA Representante da empresa responsável pelas atividades de parceria com o Campus Pinheiral Administrador EB Pesquisador do Campus Pinheiral responsável pelas atividades de cooperação com a empresa em análise Professor, pesquisador, médico- veterinário, ex-funcionário da Rica Alimentos. Fonte: Próprios autores Para análise, esta pesquisa utilizou a estrutura teórica de estudos da cooperação academia-empresa, proposto por Bonaccorsi e Piccaluga (1994). A partir dessa esquematização, o presente estudo identificou fatores de motivações, processo de cooperação, barreiras e/ou facilitadores e a satisfação resultante do caso em estudo. Esta pesquisa norteou suas análises pelas estratégias propostas pelo estudo de caso (Yin, 2005). Tal escolha ocorreu com o propósito de explorar o máximo do método ao utilizá-lo para desenho, desenvolvimento e análise do trabalho. A análise utilizou de citações com comparativo com a teoria e reuniu diferentes eventos com base na frequência de falas com sentidos equivalentes e seleções por relevância. O exame seguiu com a comparação dos resultados com a literatura e citações (Eisenhardt, 1989; Yin, 2005). RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 30 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral empresa e do professor do campus responsável pela atividade. Os pontos de vistas desses dois atores também apresentaram a descrição dos conflitos da atividade de cooperação academia- empresa. Já a análise das perspectivas do relacionamento academia-empresa e das motivações para essa cooperação academia-empresa contaram, também, com os pontos dos gestores do Campus Pinheiral e do IFRJ. O Quadro 3 apresenta alguns destaques do percurso metodológico. Quadro 3 - Destaques do percurso metodológico Objetivos Específicos - Identificar os principais conflitos para a cooperação entre academia-empresa, a partir da experiência do Campus Pinheiral com a empresa Rica Alimentos; - Discutir sobre as perspectivas da cooperação academia-empresa no Campus Pinheiral; - Apresentar as motivações para a cooperação academia-empresa no Campus Pinheiral. Unidade de análise A cooperação academia-empresa. Unidades de observação Conflitos, motivações e incentivos da cooperação academia-empresa. Levantamento de dados Entrevista com roteiro semiestruturado, pesquisa documental e bibliográfica (Duarte, 2004; Yin, 2005; Gil, 2008; Vergara, 2012). Análise dos Dados Estudo de caso (Eisenhardt, 1989; Yin, 2005). Fonte: Próprios autores Quadro 3 - Destaques do percurso metodológico 4.1. Caracterização da unidade de pesquisa: o Campus Pinheiral O Campus Pinheiral é uma instituição centenária de vocação agrícola que adquiriu uma nova institucionalidade em 2008 ao integrar-se ao Instituto Federal do Rio de Janeiro. Após 40 anos de trabalho na formação de mão de obra técnica — como então Colégio Agrícola Nilo Peçanha vinculado à Universidade Federal Fluminense — a instituição ao integra-se ao IFRJ se depara com desafios como o de promover ensino, pesquisa e extensão (Brasil, 2008). Essa nova institucionalidade, vivenciada em conjunto com 14 campi de diferentes realidades em 2019, tornou o antigo colégio agrícola em uma instituição verticalizada que integra cursos técnicos, de graduação e de pós-graduação. Além disso, essa identidade junto ao IFRJ incorpora a instituição ao projeto institucional dos Institutos Federais de promover o desenvolvimento socioeconômico regional a partir da pesquisa aplicada. Para Perucchi e Mueller (2016), enquanto as universidades primam pela pesquisa básica, os Institutos Federais têm o compromisso de se dedicarem, prioritariamente, à pesquisa aplicada em articulação com governo e empresas. Em síntese, o projeto de integração ao IFRJ apresenta ao antigo colégio agrícola a necessidade de construção de uma nova identidade. Dessa forma, ao encontro da missão de promover o desenvolvimento regional e atuar na pesquisa aplicada, o caso emblemático da instituição encontra-se em sua experiência com a empresa Rica Alimentos. 3. Metodologia ( ) As informações para apresentar a trajetória da experiência do Campus Pinheiral com a Rica foram organizadas em ordem cronológica sob as perspectivas do representante da http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 4.2. O caso de cooperação academia-empresa http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 31 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral O trabalho do Campus Pinheiral com a Rica Alimentos iniciou com atividades de prestação de serviços da unidade à empresa em 1994. A parceria se ampliou, a partir de 2010, impulsionada pelo interesse da empresa em obter ganhos competitivos através da pesquisa. Nesse cenário, destaca-se, ainda, a importância da experiência do pesquisador responsável pelo laboratório de avicultura no campus, como ex-funcionário da instituição, e as novas percepções desse ator a partir de uma visita técnica nos EUA. Como eu havia feito um curso na Geórgia em 2010 e vi que lá nos Estados Unidos essa parceria entre instituições de pesquisas e instituições federais é muito forte, na verdade lá é nacional ou estadual, é... Eu trouxe a ideia então pro Brasil, um centro de pesquisas avícolas que ia atender tanto a demanda da indústria, da iniciativa privada como a necessidade da instituição de fazer pesquisa e extensão, além do ensino (Entrevistado B, 2017). Porque a avicultura como outras atividades também é muito dinâmica, então, você precisa estar buscando melhorias, à busca de resultados econômicos. Se você não fica atualizado, faz com que você não fique competitivo no mercado. Então, por isso, nos motivamos a continuar e fortalecer a parceria com o Campus Pinheiral (Entrevistado A, 2017). A partir da proposta de parceria para desenvolvimento e transferência de tecnologia, o laboratório de avicultura da unidade se transformou em um galpão experimental. Como então Centro de Pesquisas Avícolas (CPA), o setor redirecionou suas atividades para além das ações de ensino. O laboratório superou o caráter de produção animal — ligado à sua origem de colégio técnico — e alcançou ações de pesquisa e de extensão. O laboratório de avicultura da unidade passou a desenvolver pesquisas a partir de problemas apresentados pela empresa e, em contrapartida, o setor recebe investimentos. A empresa já investiu, desde 2010, o valor estimado de R$ 45 mil em equipamentos e obras de infraestrutura no setor. A empresa fornece, ainda, regularmente, aves filhotes ao setor para atividades de ensino, pesquisa, extensão e demandas da empresa. O laboratório de avicultura da unidade passou a desenvolver pesquisas a partir de problemas apresentados pela empresa e, em contrapartida, o setor recebe investimentos. 4.3 Principais conflitos para a cooperação entre academia-empresa: observação do caso Rica Da experiência entre o Campus Pinheiral e a empresa em análise emergem conflitos como a dualidade entre: a falta de segurança jurídica e as perspectivas de garantias legais, a partir do Marco Legal da Inovação. O Entrevistado 1 relata suas expectativas quanto à nova legislação: “O marco traz essa segurança jurídica para que os setores públicos possam fazer, porque ninguém quer fazer uma coisa que possa lá na frente ter algum tipo de transtorno”. Devido a questões controversas do relacionamento em análise, observa-se ausência de diálogo sobre a temática e restrições em trabalhos de divulgação do laboratório. O caso em que o jornal institucional da unidade relatou a evolução do espaço sem menção à empresa representa a dificuldade de se debater o tema (Informativo do Campus Pinheiral, 2017). Da mesma forma, essa problemática é simbolizada pela condução da campanha do pesquisador líder do laboratório, em 2017, à direção geral da unidade com pouca referência a esse trabalho com a empresa e sua posição sobre o empreendedorismo acadêmico. p p ç p Leydesdorff & Ivanova (2016) argumentam que o princípio das relações entre universidades, empresas e governo necessitam de mediação administrativa ou intervenção política. Sob a ótica do pesquisador do Campus Pinheiral responsável pelas atividades de cooperação com a Rica, a principal dificuldade da parceria se concentra na ausência, depois de nove anos, de um contrato oficial. Até 2019, o IFRJ não conseguiu oferecer esse suporte à parceria, mesmo diante de tentativas do pesquisador. Dessa forma, sem as definições de um contrato, o pesquisador relata que o principal conflito enfrentado ao longo deste período da parceria é a mudança da condução da parceria por parte da direção da instituição. Ele relata que a cada gestão no Campus Pinheiral há a interferência nos acordos firmados anteriormente com a empresa. Na avaliação do pesquisador, isso ocorre devido à ausência de contrato garantidor das condições acordadas. Envolve, também, o relacionamento em análise, o dilema entre a continuidade das ações (devido aos benefícios mútuos percebidos) e a interrupção (devido à fragilidade vinda da ausência de um contrato estruturante). O Entrevistado A (2017) releva que a empresa tem interesse em fortalecer a parceria. “A empresa está disposta a cada vez mais crescer essa relação entre iniciativa privada e setor público”. Por outro lado, há a fragilidade devido à falta de um contrato: “Sem contrato fica difícil até de continuar a parceria. 4.2. O caso de cooperação academia-empresa A empresa já investiu, desde 2010, o valor estimado de R$ 45 mil em equipamentos e obras de infraestrutura no setor. A empresa fornece, ainda, regularmente, aves filhotes ao setor para atividades de ensino, pesquisa, extensão e demandas da empresa. De acordo com o Informativo do Campus Pinheiral (2017) e relatos complementares, os resultados da parceria para o Campus Pinheiral se apresentam nas atividades apresentadas a seguir: g • Estágios internos: 30 vagas de estágios em média por ano • Estágios externos: 8 vagas de estágios por ano em média, • Intercâmbios internacionais: 2 vagas de intercâmbios internacionais por ano. • Dissertações: Até 2018, foram 10 concluídas; • Teses: Até 2018, foram 8 concluídas; • Artigos publicados em revistas nacionais e internacionais: duas publicações por ano em média. • Artigos publicados em revistas nacionais e internacionais: duas publicações por ano em média. O ensino do Campus Pinheiral é, majoritariamente, técnico de nível médio. Diante disso, as produções são impulsionadas por parcerias com instituições de ensino superior e organizações como a Associação Fluminense de Avicultura e Suinocultura e o Comitê Estadual de Sanidade Avícola do Rio de Janeiro. As parcerias ocorrem a partir da oferta de oportunidades de estudos, estruturadas com base em demandas da empresa. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br Oliveira & Renault \| 32 10.20401/rasi.6.1.333 Oliveira & Renault \| 32 10.20401/rasi.6.1.333 Oliveira & Renault \| 32 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral 4.3 Principais conflitos para a cooperação entre academia-empresa: observação do caso Rica 4.3 Principais conflitos para a cooperação entre academia-empresa: observação do caso Rica 4.3 Principais conflitos para a cooperação entre academia-empresa: observação do caso Rica Já foi interrompido, por um tempo devido à falta de contrato”. Espera-se que o amparo legal proporcione mais segurança de prosseguimento da parceria. O Entrevistado B (2018) complementou: “A empresa vai poder também ter uma garantia de que, vamos dizer assim, de continuidade. Ela vai investir, mas ela vai ter também um documento que oficializa que vai ter um retorno para ela também”. Nota-se, ainda, no relacionamento em análise, desacordos sobre os propósitos da cooperação academia-empresa, a contraposição entre o trabalho em prol da sociedade e o trabalho para atender as necessidades do capital. Existe uma, internamente vem também pela formação acadêmica e tudo mais, existe uma linha aqui meio que talvez demonize o capital com todos os problemas do sistema capitalista, as pessoas entendem porque são servidoras públicas, elas não devem fazer parceria com empresa privada, porque a parceria de empresa privada vai fortalecer empresa privada e quando vai fortalecer empresa privada, vai fortalecer a capacidade de lucro daquela empresa privada e nós não estamos aqui a serviço de aumentar o lucro de determinada empresa A, B ou C (Entrevistado 4). RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br Oliveira & Renault \| 33 10.20401/rasi.6.1.333 Oliveira & Renault \| 33 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Os diferentes tempos das organizações também estão entre os conflitos da cooperação academia-empresa no caso do Campus Pinheiral. O pesquisador líder da unidade avalia que a empresa tem uma velocidade maior do que a instituição pública. Ele analisa, ainda, que é frequente que a instituição pública passe por dificuldades internas de viabilizar experimento por falta de matéria-prima e mão de obra e não atenda o tempo que a empresa necessita. p p p q p A gestão do IFRJ, na avaliação do Entrevistado 2, considera que parcerias como a Rica sofrem problemas operacionais no âmbito da instituição. Tudo isso devido à ausência de uma fundação de apoio para atuar na gestão financeira dos projetos e na emissão de notas fiscais de prestação de serviços. No entanto, a expectativa é que, em 2019, a Agência de Inovação do IFRJ solucione esta questão com a contratação de uma fundação de apoio. Até 2018, a contrapartida que o Campus Pinheiral recebeu foi em equipamentos, insumos, rações, aves, entre outros investimentos ao laboratório, no entanto, sem compensações financeiras. 4.4 Perspectivas da cooperação academia-empresa no IFRJ A aproximação do Campus Pinheiral ao setor produtivo para desenvolvimento e transferência de tecnologia vai ao encontro da missão institucional dos IFs de realizar pesquisa aplicada e promover desenvolvimento regional. Para o Entrevistado 2, o Marco Legal da Inovação representa uma oportunidade ao IFRJ para atender o seu compromisso institucional de desenvolver pesquisa aplicada e promover o desenvolvimento regional. “A oportunidade que o Marco Legal está oferecendo é a necessidade de atender a lei de criação dos institutos. A própria lei, ela define que a parceria com empresas é uma das atividades para gente cumprir com a missão institucional”. A partir de uma institucionalidade empreendedora, os gestores do IFRJ projetam expectativas de fortalecimento das atividades de ensino, de pesquisa e de extensão. “Então a diversificação, saindo para além da Matriz Conif, da matriz orçamentária do governo, seria aumentar os nossos recursos próprios, através de, aí uma coisa vai ligada a outra, que tem mais pesquisa, mais extensão, mais aulas práticas”, avaliou o Entrevistado 7. Tais concepções vão ao encontro da abordagem da hélice tríplice que preconiza que cada ator envolvido (academia, governo e empresa) aprimora suas tradicionais missões ao assumir as funções tradicionalmente atribuídas aos demais agentes (Etzkowitz & Leydesdorff, 1998, 2000). A expectativa ainda que o caráter empreendedor traga ao IFRJ reconhecimento social. Hoje, todos os campis reclamam: ‘Ah a gente não tem visibilidade, ninguém conhece a gente’. O empreendedorismo acadêmico é uma forma da gente lançar a nossa marca. Então, isso não só seria importante para formação profissional como seria a forma de fincar a bandeira: cheguei e cheguei mesmo. O Instituto Federal ele só vai, como um todo, fincar o seu nome na história, quando ele mudar a realidade da comunidade local em que está inserido (Entrevistado 2). Dentre as perspectivas para o desenvolvimento de um instituto empreendedor, espera- se, ainda, atender a um cenário de crise social e econômica do país. O Entrevistado 1 analisa o compromisso do IFRJ em oferecer respostas ao cenário de desemprego do país. Ressalta-se, também, a expectativa de que um instituto empreendedor atenda a necessidade de superar a percepção da academia como a formação de mão de obra, a partir de uma visão instrumentista. Reconhece-se que na era da criação, em oposição à reprodução, o valor está no indivíduo e em seu saber, no seu potencial de criar, em oposição às máquinas que replicam. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 4.4 Perspectivas da cooperação academia-empresa no IFRJ 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 34 10.20401/rasi.6.1.333 Oliveira & Renault \| 34 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Para Segatto-Mendes e Sbragia (2002), o processo de cooperação academia-empresa amplia a participação da academia no desenvolvimento tecnológico do país e aprimora sua atuação no tripé ensino, pesquisa extensão. A respeito do caso do Campus Pinheiral com a empresa avícola, o Entrevistado B (2017) avalia o progresso decorrente da parceria, que trouxe ganhos para as atividades de ensino, pesquisa e extensão. Para o referido entrevistado, a experiência é um caso de sucesso do Campus “com baixo custo para a instituição e alto poder científico”. Dentre os ganhos dessa parceria à empresa, o Entrevistado A (2017) indica os resultados econômicos. “A empresa é privada, empresa privada se não tem lucro, não consegue também se manter no mercado”. Os resultados das pesquisas orientam as medidas futuras da empresa e resultam em modificações no processo produtivo. O referido entrevistado destaca que o fortalecimento do ensino também se apresenta benéfico à empresa, a partir da disponibilidade de profissionais com mais experiência no seu setor de avicultura. Para o Entrevistado A, a academia tem o potencial de auxiliar em necessidades de pesquisas que a iniciativa privada não consegue absorver. Isso devido a altos custos de profissionais habilitados e de infraestrutura. Enfim, de acordo com a visão do Entrevistado A (2017), a parceria com o Campus Pinheiral tem potencial inexplorado. O desejo da empresa é desenvolvê-lo, ela possui demanda e interesse de investimento. No entanto, a não oficialização do convênio por contrato impede esse progresso. Para o representante da empresa, a normatização de direitos e deveres de atores envolvidos se faz importante para alcançar esse desenvolvimento. 4.5 Motivações para a cooperação academia-empresa no IFRJ Com o declínio de repasses governamentais, a administração do IFRJ e do Campus Pinheiral percebem as parcerias com agentes externos como uma alternativa de fortalecimento institucional. Por outro lado, os gestores defendem a condição do Estado como agente provedor da instituição. Eu não posso fazer com que a instituição trabalhe para buscar nos financiamentos externos mecanismos para sua sobrevivência. Mas ela pode sim buscar em financiamentos externos mecanismos para melhorar a sua qualidade e a sua excelência (Entrevistado 4) O desenvolvimento do conceito da universidade empreendedora tanto quanto a promoção da abordagem da hélice tríplice considera a possibilidade de outras fontes de financiamento da academia que não o próprio Estado. A abordagem da hélice tríplice preconiza o papel do governo como provedor da infraestrutura básica de desenvolvimento do capital intelectual (Etzkowitz, 2009). Já Clark (2003) defende o papel da diversificação de renda da academia, tanto para a sustentabilidade da pesquisa como para a sustentabilidade da instituição. O desenvolvimento do conceito da universidade empreendedora tanto quanto a promoção da abordagem da hélice tríplice considera a possibilidade de outras fontes de financiamento da academia que não o próprio Estado. A abordagem da hélice tríplice preconiza o papel do governo como provedor da infraestrutura básica de desenvolvimento do capital intelectual (Etzkowitz, 2009). Já Clark (2003) defende o papel da diversificação de renda da academia, tanto para a sustentabilidade da pesquisa como para a sustentabilidade da instituição. p p q p ç A partir das aberturas jurídicas do Marco Legal da Inovação, a instituição observa o fortalecimento de caminhos legais para parcerias entre academia e empresa. O Entrevistado 1 analisa: “O marco traz essa segurança jurídica para que os setores públicos possam plantar essa ideia, porque ninguém quer fazer uma coisa que possa lá na frente ter algum tipo de transtorno”. Já Entrevistado 2 enfatiza a segurança jurídica a partir do Marco Legal da Inovação: “Quase pode-se dizer que o Marco Legal é marco zero desse processo, porque a partir de agora existe, existem condições legais de se fazer”. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 35 10.20401/rasi.6.1.333 Oliveira & Renault \| 35 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral A ausência de uma legislação clara, pela ótica das universidades brasileiras, já foi uma das barreiras para integração com o setor empresarial para atividades de cooperação técnica (Costa & Cunha, 2001). 4.5 Motivações para a cooperação academia-empresa no IFRJ Em 2019, legislações como o Marco Legal da Inovação buscam contornar tal cenário. Dentre tais mudanças, destaca-se a ampliação do número de horas disponíveis ao pesquisador público para se dedicar a atividades fora da universidade. A carga horária de atividades externas se estendeu de 120 para 416 horas anuais, ou 8 horas semanais. Além disso, a legislação permitiu que pesquisadores públicos em regime de dedicação exclusiva exerçam atividades externas ligadas à ciência, à tecnologia e inovação em empresas com remuneração (Brasil, 2016, 2018). ç ( ) Pela ótica dos dirigentes do IFRJ e do Campus Pinheiral, o contexto jurídico anterior ao Marco Legal da Inovação não era estimulante para a promoção de parcerias com agentes externos com vistas à capitalização do conhecimento e ampliação de fontes de renda nas ICTs públicas. O Entrevistado 2 analisa: “Não tinha realmente uma abertura. Se você não tiver um Marco Legal que diga que possa fazer, alterando as leis do serviço público, o trabalho é muito grande, porque você vai esbarrar numa lei que vai impedir que você faça”. À vista disso, o Marco Legal da Inovação alterou nove leis, entre elas a Lei da Inovação e a Lei nº 12.772/ 2012 (Carreira de Magistério). ( g ) Além disso, o contexto institucional não era estimulante, a falta de amparo jurídico e assistência institucional, para assinatura de um contrato, por exemplo, marcaram a história da instituição. O Entrevistado 7 declara: “Tentei uma parceria durante muitos anos e aí a reitoria sempre tinha um problema para poder assinar o processo”. Já o Entrevistado 4 relata uma experiência em que buscou amparo jurídico para a realização de um trabalho em parceria com uma empresa. No entanto, não encontrou respostas dos órgãos jurídicos da instituição e ao final precisou responder legalmente pela ação. Naquele tempo não tinha isso (diz se referindo ao Marco Legal da Inovação), a gente ficava desesperado. A gente procurava ajuda do procurador da UFF para poder saber. Como se faz isso? É um termo de convênio? É termo de parceria? É um termo daqui? E eles não sabiam fazer, não sabiam explicar para gente, não sabia orientar a gente. Então, isso nos deixa muito fragilizado, eu fui fruto de uma investigação, tem outro nome, não é investigação, do Ministério Público Federal, que foi denunciado por causa desse projeto como eu tivesse embolsado o dinheiro disso (Entrevistado 4). RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 4.5 Motivações para a cooperação academia-empresa no IFRJ No cenário de 2019, observa-se uma gestão com interesse em desenvolver o empreendedorismo: “Olha, a principal barreira existia na gestão anterior, que era o fato da alta gestão não comprar essa ideia. Com o Rafael, o reitor, ele compra essa ideia, entendeu? (Entrevistado 2)”. Declarações de dirigentes da unidade de Pinheiral também corroboram com essa percepção: “Então, eu acho que, agora, talvez com essa nova gestão, mais empreendedora, né? (Entrevistado 9)”. Além disso, o cenário para o desenvolvimento de políticas empreendedoras é favorável, visto que o pesquisador líder da experiência com a Rica assumiu a direção do Campus Pinheiral em 2018. Nota-se, ainda, uma conjuntura favorável na Reitoria do IFRJ, ao considerar que o professor responsável pela Pró-Reitoria de Pesquisa, Inovação e Pós- Graduação foi propulsor de uma experiência empreendedora em outro campus do IFRJ. Ao encontro de uma gestão à busca de ações empreendedoras, a supracitada pró- reitora inseriu em sua composição setores ligados à área, sendo eles os seguintes: Diretoria da Agência de Inovação, Diretoria de Pesquisa Básica e Aplicada, Coordenação Geral de http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 36 10.20401/rasi.6.1.333 Oliveira & Renault \| 36 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Prospecção e Empreendedorismo. O Quadro 4 apresenta um comparativo da composição da pró-reitoria na gestão atual de 2019 e a sua antecessora. Quadro 4 - Comparativo da organização da Pró-Reitoria de Pesquisa, Inovação e Pós-Graduação Gestão 2014-2018 Gestão 2018-2022 Pró-Reitoria de Pesquisa Inovação e Pós-graduação Pró-Reitoria de Pesquisa Inovação e Pós-graduação Pró-Reitoria Adjunta de Pesquisa Inovação e Pós- graduação Diretoria da Agência de Inovação Diretoria Geral de Pesquisa e Pós-Graduação Diretoria de Pesquisa Básica e Aplicada Coordenação de Programas e Projetos Coordenação Geral de Programas e Projetos Coordenação de Propriedade Intelectual Coordenação Geral de Prospecção e Empreendedorismo Coordenação de Transferência de Tecnologia Coordenação de Transferência de Tecnologia e Propriedade Intelectual Coordenação de Pós-Graduação Fonte: Próprios autores No entanto, mesmo diante do estímulo do Marco Legal da Inovação e o interesse da alta administração, existem dificuldades de estabelecer políticas institucionais no IFRJ. Tais dificuldades ocorrem, de acordo com os gestores, principalmente devido à falta de conhecimento da comunidade acerca da nova legislação para ciência e tecnologia. De acordo com análise do Entrevistado 2: “O maior entrave é que a nossa comunidade não conhece o Marco Legal. 4.5 Motivações para a cooperação academia-empresa no IFRJ Então, eu posso chegar falando tudo com todo mundo, que as pessoas nunca ouviram falar”. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br Para as empresas: - Formar de mão de obra mais qualificada para recrutamento futuro; - Reduzir riscos e custos para a realização de pesquisas; - Solucionar problemas técnicos; - Possuir acesso a novos conhecimentos; - Ampliar sua lucratividade e competividade. Fonte: Próprios autores Oliveira & Renault \| 37 10.20401/rasi.6.1.333 Fonte: Próprios autores Motivações Para o Campus Pinheiral - Cumprir o projeto institucional dos IFs com a pesquisa aplicada e desenvolvimento regional; - Fortalecer as atividades de ensino, pesquisa e extensão; - Atender à função social da academia de promover o desenvolvimento socioeconômico; - Ampliar o conhecimento dos problemas locais; - Divulgar a imagem da instituição; - Ampliar seu reconhecimento social; - Obter recursos financeiros adicionais; - Obter recursos materiais adicionais. Para as empresas: - Formar de mão de obra mais qualificada para recrutamento futuro; - Reduzir riscos e custos para a realização de pesquisas; - Solucionar problemas técnicos; - Possuir acesso a novos conhecimentos; - Ampliar sua lucratividade e competividade. g - Fortalecer as atividades de ensino, pesquisa e extensão; - Atender à função social da academia de promover o desenvolvimento socioeconômico; - Ampliar o conhecimento dos problemas locais; - Divulgar a imagem da instituição; - Obter recursos materiais adicionais. Oliveira & Renault \| 37 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Figura 2 - Síntese analítica do processo de cooperação academia-empresa no IFRJ - Pinheiral Figura 2 - Síntese analítica do processo de cooperação academia-empresa no IFRJ - Pinheiral Processo de Cooperação Satisfação Resultante o a e e a o s a o s s s o a s s e - Agência de Inovação do IFRJ - Relações pessoais formais e informais Para o campus: - Desejo de continuidade da parceria já existente - Desejo de novas parcerias para outros laboratórios Para a empresa em análise: - Desejo de continuidade e fortalecimento da parceria. Barreiras Facilitadores - Tempos diferentes; - Grau de incerteza dos projetos, devido à falta de um contrato; - Cultura interna contra a entrada de capital externo na instituição pública; - Dificuldades operacionais (ex.: recebimento de contrapartidas financeiras). - Apoio da alta gestão; - Estímulo do Marco Legal da Inovação; - Contexto econômico favorável à busca de novas alternativas de verbas. Fonte: Próprios autores Motivações Processo de Cooperação Para o Campus Pinheiral - Cumprir o projeto institucional dos IFs com a pesquisa aplicada e desenvolvimento regional; - Fortalecer as atividades de ensino, pesquisa e extensão; - Atender à função social da academia de promover o desenvolvimento socioeconômico; - Ampliar o conhecimento dos problemas locais; - Divulgar a imagem da instituição; - Ampliar seu reconhecimento social; - Obter recursos financeiros adicionais; - Obter recursos materiais adicionais. Para as empresas: - Formar de mão de obra mais qualificada para recrutamento futuro; - Reduzir riscos e custos para a realização de pesquisas; - Solucionar problemas técnicos; - Possuir acesso a novos conhecimentos; - Ampliar sua lucratividade e competividade. - Agência de Inovação do IFRJ - Relações pessoais formais e informais Barreiras Facilitadores - Tempos diferentes; - Grau de incerteza dos projetos, devido à falta de um contrato; - Cultura interna contra a entrada de capital externo na instituição pública; - Dificuldades operacionais (ex.: recebimento de contrapartidas financeiras). - Apoio da alta gestão; - Estímulo do Marco Legal da Inovação; - Contexto econômico favorável à busca de novas alternativas de verbas. Fonte: Próprios autores 4.6 Síntese analítica dos resultados A partir da estrutura teórica de estudos proposta por Bonaccorsi e Piccaluga (1994), a análise do caso do Campus Pinheiral com a empresa avícola tem a esquematização conforme apresenta a Figura 2. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 38 10.20401/rasi.6.1.333 Oliveira & Renault \| 38 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral geral, elas não encontram barreiras diretas. O interesse do Campus Pinheiral em cumprir o projeto institucional dos IFs com a pesquisa aplicada e desenvolvimento regional tem potencial de se concretizar com o trabalho de cooperação. Assim como, a empresa tem possibilidade de atingir o objetivo de formar de mão de obra mais qualificada para recrutamento futuro. geral, elas não encontram barreiras diretas. O interesse do Campus Pinheiral em cumprir o projeto institucional dos IFs com a pesquisa aplicada e desenvolvimento regional tem potencial de se concretizar com o trabalho de cooperação. Assim como, a empresa tem possibilidade de atingir o objetivo de formar de mão de obra mais qualificada para recrutamento futuro. Esta pesquisa apresentou dados que intencionam subsidiar a compreensão da realidade de instituição e, dessa forma, amparar ações administrativas rumo ao empreendedorismo acadêmico. Nota-se que o principal conflito da cooperação em análise está na falta de segurança jurídica, situação que gera conflitos quanto à continuidade do trabalho e indefinições das ações frente à troca de gestões do IFRJ. Em suma, diante de uma história recente e de formação heterogênea, os Institutos Federais carecem de estudos que direcionam o seu desenvolvimento. Como desenvolver o empreendedorismo acadêmico em uma instituição verticalizada? Quais são as especificidades de um Instituto Empreendedor? Como integrar os diferentes níveis de ensino em um projeto empreendedor? Em 2019, a exemplo deste estudo, a cooperação academia-empresa nos Institutos Federais é observada pelas concepções da universidade empreendedora. Esta pesquisa se concentrou na visão dos gestores da instituição. Dessa forma, tais limitações apresentam oportunidades de pesquisas, a partir da ótica dos servidores, alunos e empresas. Por fim, os Institutos Federais fazem parte de um projeto inédito de educação, ciência e tecnologia no país, que se distingue pelo compromisso com a aplicada em articulação com o desenvolvimento regional. Diante disso, o Campus Pinheiral — entre a história egressa de uma escola técnica formadora de mão obra e a referência de desenvolvimento a partir da universidade — necessita definir sua identidade institucional ao encontro do empreendedorismo acadêmico e do projeto dos IFs. Para mais, os IFs necessitam definir sua identidade perante a sociedade para, assim, consolidar seu projeto institucional. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br 5. Considerações finais O Brasil tem o desafio de transformar seu potencial científico em inovação e desenvolvimento socioeconômico. Além disso, os Institutos Federais enfrentam a necessidade de construir sua identidade institucional junto à pesquisa aplicada e ao desenvolvimento regional. Nesse cenário, a hélice tríplice apresenta caminhos para o alcance desses objetivos e o Marco Legal da Inovação traz perspectivas de desenvolvimento do processo de cooperação academia-empresa. A análise da experiência do Campus Pinheiral com a empresa Rica Alimentos apresenta que a cooperação em estudo tem potencial para se fortalecer. Para tal afirmação se considera as possiblidades do Marco Legal da Inovação e o interesse da alta administração da unidade, pontos de partida para transpor barreiras como a incerteza jurídica da cooperação e dificuldades operacionais. Ao observar as motivações à cooperação, nota-se ainda que, em http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Tudo isso em um cenário em que se busca, através fontes de renda complementares, fortalecimento orçamentário frente ao declínio de repasses públicos. http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 Oliveira & Renault \| 39 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral REFERÊNCIAS Bessant, J.; & Tidd, J. (2009). Inovação e empreendedorismo. Porto Alegre: Bookman. Bessant, J.; & Tidd, J. (2009). Inovação e empreendedorismo. Porto Alegre: Bookman. Bonaccorsi, A., & Piccaluga, A. (1994). A theoretical framework for the evaluation of university-industry relationships. R&D Management, 24 (3), p. 229-247. Bonaccorsi, A., & Piccaluga, A. (1994). A theoretical framework for the evaluation of university-industry relationships. R&D Management, 24 (3), p. 229-247. Brasil (2018). Decreto n. 9.283, de 7 de fevereiro de 2018. Regulamenta a Lei nº 10.973, de 2 de dezembro de 2004, a Lei nº 13.243, de 11 de janeiro de 2016, o art. 24, § 3º, e o art. 32, § 7º, da Lei nº 8.666, de 21 de junho de 1993, o art. 1º da Lei nº 8.010, de 29 de março de 1990, e o art. 2º, caput, inciso I, alínea "g", da Lei nº 8.032, de 12 de abril de 1990, e altera o Decreto nº 6.759, de 5 de fevereiro de 2009, para estabelecer medidas de incentivo à inovação e à pesquisa científica e tecnológica no ambiente produtivo, com vistas à capacitação tecnológica, ao alcance da autonomia tecnológica e ao desenvolvimento do sistema produtivo nacional e regional. Brasília. Brasil (2016). Lei n. 13.243, de 11 de janeiro de 2016. (2016). Dispõe sobre estímulos ao desenvolvimento científico, à pesquisa, à capacitação científica e tecnológica e à inovação e altera a Lei nº 10.973, de 2 de dezembro de 2004, a Lei nº 6.815, de 19 de agosto de 1980, a Lei nº 8.666, de 21 de junho de 1993, a Lei nº 12.462, de 4 de agosto de 2011, a Lei nº 8.745, de 9 de dezembro de 1993, a Lei nº 8.958, de 20 de dezembro de 1994, a Lei nº 8.010, de 29 de março de 1990, a Lei nº 8.032, de 12 de abril de 1990, e a Lei nº 12.772, de 28 de dezembro de 2012, nos termos da Emenda Constitucional nº 85, de 26 de fevereiro de 2015. Brasília. Brasil (2008). Lei nº 11.892, de 29 de dezembro de 2008. Institui a Rede Federal de Educação Profissional, Científica e Tecnológica, cria os Institutos Federais de Educação, Ciência e Tecnologia, e dá outras providências. Brasília. Brasil (2008). Lei nº 11.892, de 29 de dezembro de 2008. REFERÊNCIAS Institui a Rede Federal de Educação Profissional, Científica e Tecnológica, cria os Institutos Federais de Educação, Ciência e Tecnologia, e dá outras providências. Brasília. Clark, B. (2003). Creating Entrepreneurial Universities. Oxford: IAU Press, Elsevier Science. Costa, V. M. G., & Cunha, J. C. da. (2001). A universidade e a capacitação tecnológica das empresas. Revista de Administração Contemporânea, 5(1), 61-81. https://dx.doi.org/10.1590/S1415-65552001000100005 Corsi, C., & Prencipe, A. (2016). Improving Innovation in University Spin-Offs: The Fostering Role of University and Region. Journal of technology management & innovation, 11(2), 13-21. https://dx.doi.org/10.4067/S0718-27242016000200002 Creswell, J. W. (2014). Investigação qualitativa e projeto de pesquisa: escolhendo entre cinco abordagens. (3ª ed.). Porto Alegre: Penso. Duarte, R. (2004). Entrevistas em pesquisas qualitativas. Educ. rev. [online]. 24, pp.213-225. ISSN 0104-4060. Disponível em: <http://dx.doi.org/10.1590/0104-4060.357>. Acesso em: 25 mai. 2017. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br Oliveira & Renault \| 40 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Etzkowitz, H. (2015a). Making a humanities town: knowledge-infused clusters, civic entrepreneurship and civil society in local innovation systems. Triple Helix, 2(1), p. 1- 22. Etzkowitz, H. (2015b). The Entrepreneurial University as a technopole platform: a global phenomenon. In: MIAO, J. T.; BENNEWORTH, P.; PHELPS, N. A. (Org.). Making 21st Century Knowledge Complexes: technopoles of the world revisited. New York: Routledge, p. 23-40. Etzkowitz, H. (2009) Hélice Tríplice-universidade-indústria-governo: inovação em movimento. (1ª.ed). (C. Hintz, Trad.). Porto Alegre: Edipucrs. (Obra original publicada em 2008). Etzkowitz, H., & Leydesdorff, L. (1998) The triple hélix as a model for innovation studies. Science and Public Policy. Inglaterra, 25(3), p.195-203. Etzkowitz, H., & Leydesdorff, L. (2000). The dynamics of innovation: from national systems and “mode 2 to a triple helix of university–industry–government relations. Research Policy, 29(2), p. 109–123. Etzkowitz, H., & Zhou, C. (2017). Hélice Tríplice: inovação e empreendedorismo universidade-indústria-governo. Estudos Avançados, 31(90), p. 23-48. https://dx.doi.org/10.1590/s0103-40142017.3190003 Eisenhardt, K. M. (1989). Building Theories from Case Study Research. The Academy of Management Review, (14) 4, pp. 532-550. Faucher, P., & Ribeiro, M. (1995). Desenvolvimento tecnológico: novos espaços de convergência entre o público e o privado. Revista do Serviço Público, Brasília, 119, p 29-54, mai./dez. Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, P., & Trow, M. (1994). The new production of knowledge - the dynamics of science and research in contemporary societies. London, Thousand Oaks, California: Sage Publications. REFERÊNCIAS Gil, A. C. (1994). Como elaborar projetos de pesquisa. (3ª ed.). São Paulo: Atlas. Gil, A. C. (2008). Como elaborar projetos de pesquisa. (4º ed.) São Paulo: Atlas. Informativo do Campus Pinheiral. Centro de pesquisas avícolas comemora 7 anos de ensino, pesquisa e extensão. Pinheiral, (jul. 2017) 8 (58). Disponível em https://bit.ly/2NTWtnQ. Acesso em: 30 mai de 2018. Jia, Y., Zhou, C., & Etzkowitz, H. (2016). Working Paper Nine-Quadrants of the " Endless Frontier ”: Triple Helix Techno-, (2014), 1–20. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.ras RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br http://www.rasi.vr.uff.br Oliveira & Renault \| 41 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Kruss, G, Adeoti, J. O., & Nabudere, D. (2015). Bracing for change: making universities and firms partners for innovation in sub-Saharan Africa, In: Developing National Systems of Innovation, chapter 1, pages 31-54 Edward Elgar Publishing. Leydesdorff, L. (2018). Synergy in Knowledge-Based Innovation Systems at National and Regional Levels: The Triple-Helix Model and the Fourth Industrial Revolution, Journal of Open Innovation: Technology, Market, and Complexity 4(2), p. 30. Leydesdorff, L., & Etzkowitz, H. (1998). The Triple Helix as a model for innovation studies. Science and Public Policy, (25), 3, pp. 195-203. Leydesdorff, L. (2017). Triple Helix of University-Industry-Government Relations. In: Encyclopedia of Creativity, Invention, Innovation and Entrepreneurship (3ª ed.). Nova York: Springer. Leydesdorff, L., & Ivanova, I. (2016). “Open innovation” and “triple helix” models of innovation: can synergy in innovation systems be measured?," Journal of Open Innovation: Technology, Market, and Complexity, MDPI, Open Access Journal, vol. 2(3), p. 1-12, Agost. Moreno, A. C. (2018). 90% das universidades federais tiveram perda real no orçamento em cinco anos; verba nacional encolheu 28%. G1, 29 jul.2018. Disponível em: https://g1.globo.com/educacao/noticia/90-das-universidades-federais-tiveram-perda-real- no-orcamento-em-cinco-anos-verba-nacional-encolheu-28.ghtml. Acesso em: 13 agosto. 2018. Perucchi, V., & Mueller, S. P. M. (2016). Produção de conhecimento científico e tecnológico nos Institutos Federais de Educação, Ciência e Tecnologia: uma investigação sobre a sua natureza e aplicação. Perspectivas em Ciência da Informação, 21(1), 134-151. https://dx.doi.org/10.1590/1981-5344/2503 Pombo, R. G. de F. (2016). A Lei 13.243/2016 (Marco Legal da Ciência, Tecnologia e Inovação): Alteração na Legislação sobre Licitação e Contratos Administrativos. São Paulo: Justen, Pereira, Oliveira & Talamini, p. 1-4. 2016. Rauen, C. V. (2016). O Novo Marco Legal da Inovação no Brasil: o que muda na relação ICTempresa? Radar: tecnologia, produção e comércio exterior, 43, 21-35. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 REFERÊNCIAS Segatto-Mendes, A. P., & Sbragia, R. (2002). O processo de cooperação academia-empresa em universidades brasileiras. Revista de Administração, 37(4), 58-71. Stal, E., & Fujino, A. (2016). A evolução das relações das universidades com o setor empresarial no Brasil: o que as publicações nacionais entre 1980 e 2012 revelam. Revista de Administração, 51 (1), 72-86. https://doi.org/10.5700/rausp1224 Vergara, S. C. (2012). Métodos de pesquisa em administração. São Paulo: Atlas. RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020 http://www.rasi.vr.uff.br http://www.rasi.vr.uff.br Oliveira & Renault \| 42 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Yin, R. K. (2005). Estudo de caso: planejamento e métodos (3ª ed.). (D. Grassi, Trad.). Porto Alegre: Bookman. (Obra original publicada em 1984). Oliveira & Renault \| 42 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Yin, R. K. (2005). Estudo de caso: planejamento e métodos (3ª ed.). (D. Grassi, Trad.). Porto Alegre: Bookman. (Obra original publicada em 1984). Oliveira & Renault \| 42 10.20401/rasi.6.1.333 Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Avaliado pelo sistema Double Blind Review Editor: Marcelo G. Amaral Yin, R. K. (2005). Estudo de caso: planejamento e métodos (3ª ed.). (D. Grassi, Trad.). Porto Alegre: Bookman. (Obra original publicada em 1984). http://www.rasi.vr.uff.br http://www.rasi.vr.uff.br RASI, Volta Redonda/RJ, v. 6, n. 1, pp. 24-42, jan./abr. 2020
https://openalex.org/W4220652790	https://www.researchsquare.com/article/rs-1418093/latest.pdf	English	null	Factors influencing men’s decisions whether or not to utilize sexual and reproductive health services in Low-Middle-Income countries: A narrative review	Research Square (Research Square)	2,022	cc-by	11,174	Factors influencing men’s decisions whether or not to utilize sexual and reproductive health services in Low-Middle-Income countries: A narrative review Mpumelelo Nyalela (  nyalelam@ukzn.ac.za ) University of Kwa-Zulu Natal Thembelihle Dlungwane (  dlungwane@ukzn.ac.za ) University of Kwa-Zulu Natal Mpumelelo Nyalela (  nyalelam@ukzn.ac.za ) University of Kwa-Zulu Natal Thembelihle Dlungwane (  dlungwane@ukzn.ac.za ) University of Kwa-Zulu Natal Systematic Review Keywords: Men, Sexual and Reproductive Health services, factors influencing SRH service utilization, factors facilitating the SRH service utilization, factors inhibiting SRH service utilization Posted Date: March 7th, 2022 Results A total of 2219 articles were retrieved, from which 51 articles met the inclusion criteria. We reveal lack of access and availability of SRH services, poor health-seeking behavior among men and, SRH facilities often not being seen as "male-friendly” spaces and not providing men's services under one roof as the underlying factors contributing to the poor uptake of SRH services by men. Further influences to decline in SRH utilization by men included lack of focus on adult men's sexual and reproductive health; the lack of inclusion of men in a meaningful way; and the insufficient evidence about large-scale and implementable approaches to address men’s SRH needs. Our reviews reveal that dealing with negative or inhibiting factors to manage low utilization of SRH services had improved men's SRH utilization where this was carried out. We conclude by proposing a framework to illustrate men's engagement with SRH services. Background Low-Middle-Income countries (LMIC) present with a high aversion to the utilization of Sexual and Reproductive Health (SRH) services by men. Nevertheless, in both Low-Middle-Income and High-Income countries, low SRH utilization is mainly related to men's poor Health Seeking Behaviour (HSB). Identifying SRH inhibitors and facilitators remains an essential approach to improve SRH utilization and avert higher mortality and earlier morbidity associated with poor HSB among men. This literature review identifies factors contributing to men's decisions not to utilize SRH services, particularly to inform a conceptual model illustrating men's SRH utilization in LMIC. Methods In this narrative review, we searched articles published between 2004 and 2021 from international databases, including Google Scholar, Science Direct, EBSCOhost, Scopus, PubMed, Medline, and reference lists of retrieved published articles. DOI: https://doi.org/10.21203/rs.3.rs-1418093/v1 DOI: https://doi.org/10.21203/rs.3.rs-1418093/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. R d F ll Li License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/36 Page 1/36 Plain English Sexual and Reproductive Health service utilization refers to accessibility, affordability, availability, and quality of sexual health services that are provided in a convenient and suitable manner. The utilization of SRH services provides a state of physical, mental, and emotional well-being. SRH services include assistance with contraception, prevention and treatment of HIV/AIDS and other Sexually Transmitted Infections (STIs), Medical Male Circumcision (MMC), condom use, psychosocial interventions such as sexual health counseling and education, Vasectomy, prevention and treatment of male cancers such as Page 2/36 prostate and testicular, infertility services, sexual dysfunction developmental anomalies, malignancy, trauma, developmental and lifecycle issues, sexual identity and orientation, and partnership issues. We have found that there is a lack of literature investigating factors associated with adult men’s utilization of SRH. To our knowledge, this is the only review that comprehensively focuses on adult men’s utilization of SRH services. Empirical evidence indicates the enormous under-utilization of SRH services by men. Very few men are reported to be accessing and utilizing SRH services. Consequently, men die from curable SRH problems that can be prevented, contributing to higher mortality and earlier morbidity among men. The factors influencing men’s decision to utilize SRH services include cultural, religious, psychosocial, economic, and socio-demographic characteristics. We conducted a systematic review of articles between 2004 and 2021 and identified an initial 2219 papers, of which 51 were included in the analysis and then analyzed narratively. We extracted information about the factors influencing SRH service utilization among adult men from published and unpublished peer-reviewed literature. Identifying SRH inhibitors and facilitators remains an essential approach to improve SRH utilization. Eligibility criteria This review included quantitative and qualitative papers published in English between 2004 and 2021. Studies reporting exclusively on the factors influencing SRH service utilization among adolescents or women were excluded. Studies reporting on the interventions and the evaluation of interventions to improve SRH utilization; clinical decision-making algorithms; studies conducted in HICs; and studies not reporting on barriers or facilitators of SRH utilization were also not considered. Methodology The systematic quantitative literature review method was followed to search and identify relevant literature. The search included published peer-reviewed articles, reports, and grey literature. Peer-reviewed articles were searched in the following databases: Google Scholar, Science Direct, Scopus, and EBSCOhost. Additionally, electronic databases: CINHAL, PubMed, Medline, Academic Search Complete, Health Source - Consumer Edition, Health Source: Nursing/Academic Edition and MEDLINE, electronic journals, and reference lists of retrieved published articles were also searched. The following MESH terms were used to retrieve relevant articles for this review: "sexual and reproductive health services", "men", "factors influencing / inhibiting / facilitating SRH service utilization", "Utilization of Sexual and Reproductive Health services by adult men". To cover comprehensive SRH services, studies were also searched by individual SRH services, e.g., utilization of condoms. Studies that investigated Health Care Workers’ (HCW) knowledge and perceptions about men's use of SRH services were also included. Studies conducted in 2004 were sourced because the World Health Organisation (WHO) recognized the immense global health burden (including extensive mortality and morbidity) associated with sexual and reproductive health conditions and poor quality of care for people's sexual and reproductive health in 2004 [12]. The database was constructed to extract important dimensions of factors influencing SRH service utilization. We extracted only relevant results about populations eligible for inclusion in this narrative review. For example, where data was collected for males and females, we extracted only data presented on males. Background Sexual and Reproductive Health (SRH) is a global public health concern [1, 2]. SRH problems account for major health challenges and constitute almost 14% of the disease burden which contributes to higher mortality and earlier morbidity in men [3]. Globally, SRH is one of the fundamental human rights for every individual [1], but only less than one-quarter of men report accessing and utilizing SRH services [4]. The utilization of SRH services refers to having suitable and convenient access to these services [5, 10]. The services provide a state of physical, mental, and emotional well-being related to sexuality and reproduction, and are essential for the socio-economic development of communities and countries [2, 6– 8]. SRH services include providing contraception, prevention and treatment of mental disorders, communicable and non-communicable diseases, medical male circumcision, and psychosocial interventions such as sexual health counseling, [5, 8–12]. Under-utilization of SRH services is evident despite several international conventions and adopted programmes and policies that seek to engage men and boys in the context of reproductive health services [11]. The Lancet Commission (2018) posited that 4.3 billion people of reproductive age would inadequately utilize sexual and reproductive health services throughout their lives. The Commission further stated that two million people were newly infected with HIV while more than 350 million men and women needed treatment for curable STIs [2]. Factors influencing men's under-utilization are multifaceted. These factors include the lack of focus on men's SRH by international programmes such as Sustainable Development Goals (SDGs) and Family Planning 2020 (FP2020), which largely focus on women and youth [14–17]. Furthermore, SRH service provision is often fragmented or poorly structured to address men's health needs, and is mostly needlessly expensive [3, 17, 18]. Page 3/36 Page 3/36 The purpose of the review is to synthesize available evidence on factors contributing towards adult men's engagement with SRH services in LMICs. The review focused on LMICs because of resource constraint challenges that largely lead to relative poorer access to and higher unmet need for SRH services. The literature review was conducted as part of a doctoral research project. These findings contribute to the development of appropriate interventions to guide men's engagement with SRH services. Assessment of Publications The database search generated 2219 records which were grey literature, health services reports, research reports, theses, and dissertations. The Principal Investigator assessed titles and abstracts for inclusion. Page 4/36 On a review of titles, 2042 were excluded because they were duplicates unrelated to this narrative review topic or exclusively focused on adolescents and women. On a review of abstracts, 96 papers were excluded because they reported on the evaluation of SRH services, management of SRH problems, or male involvement in partner family planning. These were policies and documents that did not report on factors facilitating or hindering SRH service utilization. The latter were systematic reviews. Of 73 potentially relevant full-text papers, 22 were excluded because results did not yield factors hindering or facilitating SRH service utilization. Fifty-one papers met the inclusion criteria and were captured in the database. Secondary to the heterogeneity of papers, we present a narrative synthesis describing study characteristics and key findings. We further summarize overarching themes and the consistency of key findings. Discussions between team members were used to build consensus on the narrative synthesis. Constructing the database A database was constructed to summarise the studies identified for the review. The following information was captured in the database (Table 1): author (s), year of publication, the country(s) where the study was conducted, the age and target population for the study, the study design, barriers to and facilitators of the utilization of SRH services, and SRH issues investigated. Figure 1 presents the search algorithm indicating the number of identified studies, included and excluded studies, and reasons for exclusion. Results A total of 2219 articles were retrieved, from which 51 studies met the inclusion criteria for review. Table 1 presents the author, year of publication and study location, participant characteristics (age and gender), and study design. Table 2 presents barriers, facilitators, and sexual and reproductive health issues. All articles (qualitative and quantitative studies) extracted from the literature were descriptive. The results are summarized narratively. 4. Study setting Forty-nine percent of studies were conducted in both urban and rural settings. In some studies, the exact locations where data was collected were not specified [33, 34, 40, 43, 45, 46, 56, 59, 60, 61, 63, 65-68, 70]. Some of the studies were conducted in public and private-sector health facilities (clinics and hospitals) [24, 25, 29, 58, 64]; private rooms or offices in community centres; Non-Governmental Organizations sites [27, 32, 35], and households in towns and rural villages [28, 48, 69]. In studies conducted exclusively in urban settings (27%), data was collected in towns and townships in which households were visited. Some participants were recruited on the streets [21, 23, 30, 47], district health offices [26], hospital and clinical settings [36, 38, 44, 53, 62], and there were two unspecified studies [42, 54]. Studies solely conducted in rural settings (24%) approached households in villages [22, 37, 39, 41, 49, 51, 52, 55, 57, 71], traditional healers’ places, school and community halls, community tuck-shops, faith-based organizations’ premises, hospital and clinic waiting rooms. [31, 50]. 2. Age and gender The studies included all targeted adult men. In studies that mixed male and females, as well as those which included younger ages, we only extracted information pertaining to adult men. The majority (55%) of studies identified exclusively targeted men [22-26, 28, 30, 31, 33, 35, 37, 39, 41, 43, 44, 45, 48, 49, 53, 57, 59, 60, 62, 63-66, 71]. The balance (45%) focused on both men and women [21, 27, 29, 32, 34, 36, 38, 40, 42, 46, 47, 50-52, 54-56, 58, 61, 67-70]. The majority (86%) of studies targeted adults 26 years and above [21-39, 41-44, 46-51, 53, 57-66, 71]. Studies which targeted adolescents were excluded, however those targeting youth or young adults age between 15-25 years (14%) were included [40, 45, 52, 54-56, 67- 70]. 3. Study design The majority (55%) of studies that met the inclusion criteria were conducted qualitatively, and employed either Focus Group Discussions (FGDs) (16%) or In-depth Interviews (IDIs) (18%), and some both FGDS and IDIs (24%) [22, 24, 26 ,28, 31-34, 36, 38, 39-41, 45-47, 49, 50, 52, 56, 59, 60-62, 65, 67, 68, 70, 71]. Twenty-five percent of quantitative studies employed cross-sectional study design [23, 25, 27, 29, 42, 43, 45, 51, 53, 54, 57, 58, 69], and two (4%) employed National Health and Demographic Survey (NHDS) [35, 66]. The later (16%) used mixed [21, 30, 37, 44, 48, 55, 63, 64]. There was only one mixed study that included Randomized Control Trial (RCT). 1. Study location Most of the studies were conducted in South Africa (22%) [21-31], and five (10%) of studies were conducted in Kenya [32-36], Uganda [37-41], and Ghana [42-46]. Four (8%) studies were published in Zimbabwe [47-50]. Three (6%) studies were conducted in Nigeria [51-53] and Nepal [54-56]. In Rwanda [57, 58], there were two (4%) studies identified, whilst only one (2%) study each conducted in Swaziland [59], Papua New Guinea [60], Tanzania [61], Zambia [62], Botswana [63], Lesotho [64], Mexico [65], Philippines [66], Ethiopia [67], Egypt [68], Myanmar [69], Lao [70] and one both in Zimbabwe and South Africa [71]. Page 5/36 Page 5/36 5. SRH issues The SRH issues identified were Vasectomy; Family Planning (FP); Medical Male Circumcision (MMC); condom use; management and prevention of STIs; HIV services; Prostate Cancer (PC) screening; and Erectile Dysfunction (FD). 6. Barriers to men’s decisions to utilize or not utilize Sexual and Reproductive Health Services 6. Barriers to men’s decisions to utilize or not utilize Sexual and Reproductive Health Services Page 6/36 Page 6/36 Several barriers have been associated with the decision to utilize or not utilize S.R.H. services. These can be summarized into individual/personal factors, knowledge, socio-cultural and religious, socio-economic factors, geographical, and health service system factors, as detailed below. Several barriers have been associated with the decision to utilize or not utilize S.R.H. services. These can be summarized into individual/personal factors, knowledge, socio-cultural and religious, socio-economic factors, geographical, and health service system factors, as detailed below. Individual / personal factors In this review, individual/personal factors are described as behavior and characteristics demonstrated by participants which determined whether they utilized SRH services or not. Individual factors presented as barriers to SRH utilization were identified. Fear emerged as a barrier to utilizing SRH services. Among the studies focused on PC screening, fear of post-operative complications such as pain, delayed recovery, infections, morning erections, as well as the post-procedure abstinence period deterred most men from undergoing MMC [21,23, 32, 33, 39, 48, 49, 53, 57, 61-64, 59,71]. The requirement to do HIV testing before getting circumcised deterred most men from undergoing Medical Male Circumcision (MMC). They feared the possibility of testing positive, and hence death related to HIV complications. because of self- knowledge of reckless lifestyles. Men also feared the stigma associated with HIV, and the possibility of being blamed and rejected by significant others such as partners, family members, and friends, as well as the pressure from employees to quit the job after being diagnosed [24, 26, 27,31, 43, 47-52, 64, 69]. Post circumcision myths such as the inability to sexually satisfy partners, less natural lubrication, and decreased penile sensitivity on a circumcised penis also deterred individual men [21, 22, 33, 37, 39, 46, 59]. Factors associated with barriers involving Condom use included poor quality of condoms and embarrassment to buy condoms [28, 44, 66, 70]. Furthermore, the perception of low-risk also hindered men from using condoms and undergoing MMC [32, 33, 50, 57, 69, 70]. In Hassan's (2015) study, some men in Nigeria were reluctant to screen for Prostate Cancers they perceived it as a low-risk since there was no family history [53]. Factors associated with barriers involving Condom use included poor quality of condoms and embarrassment to buy condoms [28, 44, 66, 70]. Furthermore, the perception of low-risk also hindered men from using condoms and undergoing MMC [32, 33, 50, 57, 69, 70]. In Hassan's (2015) study, some men in Nigeria were reluctant to screen for Prostate Cancers they perceived it as a low-risk since there was no family history [53]. Lack of Knowledge Men's underutilization of SRH services was associated with a lack of awareness of disease screening services such as HIV and prostate cancer, and the lack of knowledge of the existence of SRH services. Consequently, some men did not know where to go for SRH services [27, 35, 46, 53-56, 67, 69, 70]. Lack of knowledge and understanding of HIV was evident when men often inferred their status from their female partners' results [27, 31, 52]. As a result of this lack of knowledge and of reliable information, men claimed less or no benefits of MMC if they were already HIV-positive, had good hygiene, were already practising other HIV prevention methods such as the ‘Abstinence’ ‘Be faithful’ ‘Condomise’ (ABC) method, or their partner(s) was sexually satisfied [33, 49]. Furthermore, in non-circumcising communities in South Africa, men did not know that the MMC service was offered for free at the local clinic or hospital [21]. Furthermore, some men associated infertility with circumcision [39, 48]. In addition to lack of knowledge, myths resulted in men opting not to use condoms after getting circumcised, citing reasons such as the circumcised penis would tear a condom; or putting a condom on Page 7/36 Page 7/36 an exposed circumcised penis would cause pain [47]. Furthermore, misconceptions such as the nutritional benefit from sperm, condom porousness, lubricant related infections, and the belief that white men had infected condoms with HIV hindered Condom use [28, 40, 45, 50, 56, 70]. Men often confused Vasectomy with castration and wrongly associated it with loss of libido, decreased sexual activity, and loss of masculinity [46, 66]. Socio-economic factors SRH service utilization was hindered by the perceived high costs of SRH products such as condoms, particularly in rural areas [40, 56, 67]. In addition, the inaccessibility and unavailability of condoms due to fewer shops contributed to low utilization [23, 25, 29, 30, 44, 45, 69]. Fear of losing income (work) due to long-period off work due to pain or post-procedure healing period deterred some men from undergoing MMC [21, 32, 33, 39, 61, 62]. Traveling costs to healthcare establishments also emerged as a major deterrent [27]. Socio-cultural and religious factors Factors associated with culture, such as the threat to masculinity, prevented men from undergoing MMC as non-circumcising communities presumed circumcision to be an alien culture or part of a foreign religion [21, 23, 30, 39, 45, 47, 49, 59, 60]. Some men alluded to undergoing MMC as tampering with God's creation [33, 47, 49, 59]. Most cultural and religious practices consider discussing sexual matters a taboo. Therefore, the sensitivity of discussing SRH issues had hindered most men from accessing psychosocial health therapy and counseling [35, 42, 45, 46, 55, 56, 67]. Geographical factors In most LMICs, health facilities are concentrated in the urban area, whereas most men in these populations reside in rural areas [67, 56, 63]. Consequently, long distances and poor transport (especially in rural areas) to the health facility became barriers to accessing and utilizing SRH services [29, 33, 40, 42, 54, 69]. The distance of health centres from men's workplaces often leads to men failing to utilize SRH services. Furthermore, poor transport infrastructure can prevent access to services in rural areas [30]. 7. Facilitators to men’s engagement with Sexual and Reproductive Health Services There was a limited number of studies that focused on factors facilitating the utilization of SRH services Facilitators to men's decisions to utilize SRH services are summarized into individual/personal issues, knowledge, socio-cultural and religious factors, socio-economic factors, and health service system factors as detailed below. Individual / personal factors Some studies discussed above concluded that men were reluctant to undergo MMC due to poor sexual performance. Conversely, other studies found that sexual appeal and satisfying women motivated men to perform MMC sexually. Men believe that post circumcision, they better satisfy women as they last longer before ejaculating, and also wearing condoms was much easier after the foreskin was removed [21, 22, 33, 38, 47, 60, 61, 64, 59]. Curiosity to feel the difference between sex with an uncircumcised penis and a circumcised one motivated some men [47]. Furthermore, personal gain or prestige from research activities such as free medical care for ailments, financial incentives, and a sense of being responsible by participating and contributing to research, motivated men to undergo MMC [27, 39, 50]. Role modeling positive HIV status disclosure and adherence to Anti-Retroviral Therapy (ART) motivated men to engage in HIV treatment initiatives [24, 27]. Furthermore, an individual desire to limit family size encouraged some men to undergo a Vasectomy since it was perceived as a permanent method with a low risk of complications, thus limiting the side effects of other female-controlled hormonal methods [58]. Health service system (Physical accessibility, availability, accessibility, affordability) Poor quality of the services and lack of materials such as condoms and medicines hindered the utilization of SRH services by men [29, 42, 54, 55, 63, 67, 68, 70]. Inconvenient service hours or limited opening hours at the delivery point were also barriers to SRH services utilization [27, 36, 40, 54, 55, 67, 68]. The inconvenience of the location of the SRH services at the local clinic seems to be a barrier to accessing the SRH care, especially for young men [44, 67]. Long waiting times because of queues influenced men's decisions to avoid coming to a healthcare centre [29, 30, 36, 40, 42, 50, 63, 64, 67]. Some studies revealed that most health care providers lack the necessary training and knowledge to make men feel comfortable discussing SRH issues [36, 56, 67]. Physicians indicated discomfort in Page 8/36 Page 8/36 counseling on sexuality and safer sex for unmarried youth [68]. Lack of privacy, respect, and potential breaches of confidentiality from the health care workers at the health facilities deterred men from utilizing SRH services [25, 36, 40, 41, 44, 55, 67, 70]. Men avoided utilization of SRH services due to hostile and judgmental attitudes from female health service providers [27, 29, 30, 40, 45, 56], especially towards young and unmarried men. There is still much sensitivity needed regarding pre-marital sexuality [26, 29, 36, 56, 67, 69, 70]. Consequently, the non- availability of same-sex health workers was a barrier since some men felt embarrassed to discuss their health issues and be examined by female health workers [30, 33, 45, 55, 64, 68]. Frequently changing HIV prevention programmes led to distrust in the health system, discouraging men from testing and treating HIV [49, 59]. A further deterrent was partial protection of condoms and lack of empirical efficacy of MMC for HIV prevention as people can die even after circumcision [39]. Knowledge Knowing the benefits of undergoing MMC, such as protection against diseases and improved hygiene motivated most men to perform the procedure [21, 33, 34, 37-39, 47, 48, 57, 60-62, 64, 59,]. Some men were motivated by myths such as the increased size of the penis post circumcision (Humphries 2015). Health service system (Physical accessibility, availability, accessibility, affordability) Health workers' welcoming friendly attitudes, and respect for men's privacy and confidentiality motivated men to access and utilize SRH services [55, 63, 68]. Access to the right information about SRH services via advertisements such as pamphlets and radio/television programmes [27, 63] and the support from healthcare providers also played a vital role in encouraging SRH service use by men [24, 41, 50]. Socio-economic factors Page 9/36 Page 9/36 Financial constraints motivated undergoing a Vasectomy [58]. The fact that some SRH services are provided at no cost, also motivated men [38]. Discussion Page 10/36 Under-utilization of SRH services was associated with a range of cultural barriers. Most studies indicated that socio-cultural factors inhibited men's utilization of SRH services. For example, cultural and religious backgrounds that perceived discussing sexual matters as a taboo deterred men's engagement with SRH services [35, 42, 45, 46, 55, 56, 67]. The threat to masculinity was also a major barrier, and was associated with post MMC or Vasectomy procedure misconceptions and myths. These are beliefs about the inability to produce more children, the loss of manhood, and infertility [48, 39, 40]. Despite the availability of services, poor utilization was associated with limited knowledge of the availability of the different SRH services. Lack of knowledge or awareness of the existence of SRH services such as vasectomies, screening for prostate cancer, STI / HIV testing, MMC, centres for counseling, and provision of information, hindered men's utilization [27, 35, 46, 53–56, 67, 69, 70]. Under-utilization of SRH services was associated with a range of cultural barriers. Most studies indicated that socio-cultural factors inhibited men's utilization of SRH services. For example, cultural and religious backgrounds that perceived discussing sexual matters as a taboo deterred men's engagement with SRH services [35, 42, 45, 46, 55, 56, 67]. The threat to masculinity was also a major barrier, and was associated with post MMC or Vasectomy procedure misconceptions and myths. These are beliefs about the inability to produce more children, the loss of manhood, and infertility [48, 39, 40]. Despite the availability of services, poor utilization was associated with limited knowledge of the availability of the different SRH services. Lack of knowledge or awareness of the existence of SRH services such as vasectomies, screening for prostate cancer, STI / HIV testing, MMC, centres for counseling, and provision of information, hindered men's utilization [27, 35, 46, 53–56, 67, 69, 70]. Cost-related issues prevented access to and utilization of SRH services such vasectomies, condoms, HIV testing. Although most SRH services are provided for free in LMICs, some men, especially those from rural areas, alluded to cost as a hindrance to accessing SRH services [40, 56, 67]. Such difficulties in accessing SRH services may be due to unemployment. In rural areas, most health facilities are far from the communities. As a result, people in rural communities have to walk long distances due to a lack of money and transport. Discussion This literature review aimed to establish the factors that influence men's decisions to utilize or not to utilize SRH services. In this review, only studies conducted in LMICs were included for synthesis. Identified studies were conducted mostly in the Africa regions (Southern, Eastern, and Western regions), and ten percent in South-Asia. Although studies were conducted in either rural or urban or in both rural and urban areas, findings in these settings remained comparable. Qualitative methods were the most used methods, either employed Focus Group Discussions (FGDs) or In-depth Interviews (IDIs) and some utilized both FGDS and IDIs. All studies included in this review targeted adult men. In studies that mixed males and females, as well as those that included younger ages, we only extracted information pertaining to adult men. Adult men were targeted because the literature review revealed less focus on studies investigating SRH service utilization among adult men. The most prevalent SRH issues identified incorporated Medical Male Circumcision, management and prevention of STIs, HIV/AIDS services, Family Planning (FP), and condom use. Less noted were Prostate Cancer (PC) screening, and Erectile Dysfunction (ED), and Vasectomy surgeries. This may be because services for these SRH issues are expensive, and most men in LMICs cannot afford them and sometimes they are against their cultures and religious beliefs [46, 58]. The identified barriers and facilitators of SRH service utilization encompassed individual or personal factors, lack of knowledge, socio-cultural and religious factors, socio-economic factors, geographical and health service systems. The most prevalent barrier identified was the personal perceptions (personal factors) in which men largely demonstrated embarrassment, insecurity (which have to do with their self- esteem and reputation), and fear. Fear played a vital role in inhibiting men’s use of SRH services. Consequently, men were discouraged to utilize SRH services such as MMC, Vasectomy services, family planning, STI / HIV services, screening for prostate cancer, and Condom use [23, 26, 27, 33, 42, 43, 46, 47, 49, 50, 59, 63, 64, 66]. Some studies revealed that men were largely deterred by fear of MMC post- procedure complications [32, 33, 39, 61, 62]. Most men find it embarrassing to go to SRH services and consider it a very negative experience when they are seen, ridiculed, and disrespected by people known to them [28, 44, 66, 70]. These negative experiences may motivate men not to go back to the clinic. Discussion In the health service system, health care workers' conduct significantly influenced men's decisions to utilize or not to utilize SRH services. Predominantly, men’s utilization of SRH services was deterred by bad health workers' attitudes and lack of privacy and confidentiality. The review identified only a few studies that focused on factors facilitating the SRH service utilization by men. Among studies that investigated SRH facilitating factors, virility was found to have motivated men to undergo MMC. Having done MMC, some men claimed to be stronger sexually, and that wearing condoms was much easier [21, 22, 33, 38, 47, 60, 61, 64, 59]. Conversely, some studies found that men were reluctant to undergo MMC in case it resulted in poor sexual performance. Barriers to SRH service were prevalent among societies who did not regard MMC as part of life, whereas, facilitators to SRH service utilization were predominant among those who embraced MMC. Individual factors such as curiosity motivated SRH utilization. For example, some men were inquisitive about the differences experienced during the sexual act with an uncircumcised penis as compared to a circumcised one [47]. Moreover, personal gain from research activities cannot be ruled out as most men showed interest in financial incentives post MMC [27, 39, 50]. Knowledge was instrumental in encouraging SRH service utilization. Men who understood the benefits of undergoing MMC were motivated to perform the procedure [21, 33, 34, 37–39, 47, 48, 57, 60–62, 64, 59,]. Bad attitudes from healthcare workers can be a barrier to SRH service utilisation, however, professional and good treatment from healthcare workers was found to facilitate SRH service utilization [55, 63, 68]. This review of the literature revealed a limited number of studies conducted exclusively in the area of adult men's SRH service utilization. More studies that focus on adult men need to be conducted, especially in LMICs. Given the lack of focus on men's SRH by international programmes such as SDGs and Family Planning 2020 (FP2020) [14–17], research focusing on men may bring about evidence-based Page 11/36 Page 11/36 knowledge that can be utilized by policy makers to ensure accessibility and availability of SRH services for adult men. Subsequently, high morbidity and mortality among men adult men may decrease. knowledge that can be utilized by policy makers to ensure accessibility and availability of SRH services for adult men. Subsequently, high morbidity and mortality among men adult men may decrease. Limitations of the review This review restricted the inclusion criteria to studies conducted in LMICs. Therefore, findings may not be generalizable to other settings. However, during analysis, the authors extracted findings of adult men only. The writers of the review also restricted the criteria only to include studies published in English due to a lack of resources for translation. This may limit how findings from this review can be transferable to all low and middle-income countries. Another limitation of this review is that data extraction was not performed in duplicate, which could result in bias results. We did not do additional study quality assessments or remove studies based on the risk of bias, given our goal to describe the relevant studies identified. Strengths of the review This review consolidated knowledge about barriers and facilitators that influence men's decisions to utilize SRH services. While reviews conducted in past decades have identified barriers and facilitators to SRH services, this current systematic review is the first to focus on adult men only and to comprehensively focus on all SRH services. The findings from this review have implications for clinical practice and policy. Discussion The analysis also reveals that men generally ignore the warning signs of sexual health problems, and are reluctant to seek and engage with SRH services for anything other than severe illness [67, 41]. Nevertheless, if factors influencing men’s decisions to seek and utilize SRH services are known and understood comprehensively and in the local context, men's engagement with SRH services could improve. Furthermore, this analysis can be used by policymakers, service administrators, and service providers who want to identify the barriers experienced by men to make services more accessible. Conclusion Despite the need for SRH utilization by men, there is little evidence suggesting the implementation of proposed interventions. Consequently, low SRH utilization has become a major concern regarding men's health. Men's SRH service utilization needs to be investigated comprehensively to comprehend men's problems fully. None of the identified studies comprehensively investigated SRH issues. There is also a lack of studies investigating vasectomies and their influence on infertility. Page 12/36 Furthermore, most papers identified described adolescents and young adults, predominantly focusing on barriers rather than facilitators. There is, therefore, a need to explore the utilization of SRH services among adult men. There is a need to motivate men to participate in SRH services so that the high level of morbidity and mortality among men can be averted. Although SRH includes mental health issues, little is Page 12/36 known about mental health issues related to sexual health. The same applies to the impact of physical disabilities and chronic illnesses on sexual well-being. Further research may be warranted in this regard. Governments need to increase the awareness and education of the public to fight against myths and misconceptions linked to the utilization of SRH services and to improve the capacity of the healthcare providers to engage men better. Men perceive many hindrances in obtaining SRH services from the time they access the services until they leave. These hurdles must be reduced as they could prevent them from even initiating seeking advice or care. Some obstacles (e.g., access, service delivery) are easier to remove than others (e.g., living place issues). There is a need to advertise the services widely, to indicate who can seek these, and to offer convenient opening hours. Abbreviations Pa SRH Sexual and Reproductive Health LMICs Low- and Middle-Income countries HSB Health Seeking Behaviour HIV Human Immunodeficiency Virus AIDS Acquired Immunodeficiency Disease Syndrome STI Sexually transmitted infection UNFPA United Nations Population Fund WHO World Health Organization HCW Healthcare worker MMC Medical Male Circumcision PC Prostate Cancer ED Erectile Dysfunction FP Family Planning IPPF Acknowledgements Not applicable. Ethics approval and consent to participate Ethics approval and consent to participate Not applicable. Not applicable. Authors’ contributions M.N. conceived the study and developed the design of this review with valuable contributions from T.D. M.N. performed the searches, conducted the title and abstract screening and the full-text screening, and performed the data abstraction and thematic analysis. M.N. drafted the first manuscript. T.D. oversaw the project and contributed valuable feedback to the manuscript. All authors read and gave approval of the final manuscript. Funding The author(s) received no financial support for this article's research, authorship, and publication. Consent for publication Not Applicable. Competing interests The authors declare that they have no competing interests. SRH Page 13/36 International Planned Parenthood Federation S.D.G.s Sustainable Development Goals F.G.D.s Focus Group Discussions I.D.I.s In-depth Interviews. Availability of data and materials Not applicable. Author details Mpumelelo Nyalela (MPH), Lecturer. Email: nyalelam@ukzn.ac.za Mpumelelo Nyalela (MPH), Lecturer. Email: nyalelam@ukzn.ac.za Page 14/36 Page 14/36 Thembelihle Dlungwane (PhD), Senior Lecturer. Email: dlungwane@ukzn.ac.za Thembelihle Dlungwane (PhD), Senior Lecturer. Email: dlungwane@ukzn.ac.za Thembelihle Dlungwane (PhD), Senior Lecturer. Email: dlungwane@ukzn.ac.za References 1. Jobson M. Structure of the health system in South Africa. 2015. South Africa Khulumani Support Group. 2. Starrs AM, Ezeh AC, Barker G, Basu A, Bertrand JT, et al. Accelerate progress—sexual and reproductive health and rights for all: report of the Guttmacher–Lancet Commission. The Lancet. 2018 Jun 30;391(10140):2642–92. 2. Starrs AM, Ezeh AC, Barker G, Basu A, Bertrand JT, et al. Accelerate progress—sexual and reproductive health and rights for all: report of the Guttmacher–Lancet Commission. The Lancet. 2018 Jun 30;391(10140):2642–92. 3. De Silva VP. World Book of Family Medicine:Sexual and Reproductive Health in Primary Care: Where Do We Go from Here?. Wonca Europe: Iberoamericana Edition. 2016;26:1–4. 4. Same RV, Bell DL, Rosenthal SL, Marcell AV. Sexual and reproductive health care: Adolescent and adult men’s willingness to talk and preferred approach. American Journal of Preventive Medicine. 2014 Aug 1;47(2):175 – 81. 5. Department of Health (DOH). Sexual and Reproductive Health and Rights: Fulfilling Our Commitments 2011–2021 and Beyond. 2011. http://www.agenda.org.za. Accessed 21 Jun 2020. 6. World Health Organization. Developing sexual health programmes: A framework for action. World Health Organization. 2010. https://www.gov.uk. Accesed 17 May 2021. 7. Forrest KA. Men's reproductive and sexual health. Journal of American College Health. 2001 May 1;49(6):253–66. 8. Ghebreyesus TA, Kanem N. Defining sexual and reproductive health and rights for all. The Lancet. 2018 Jun 30;391(10140):2583–5. 9. Shand T, Zamir J, Marcell AV, Perlson S. Global Sexual and Reproductive Health Service Package for Men and Adolescent Boys. London: IPPF and New York City: UNFPA.2017. https://www.unfpa.org. Accessed 15 Aug 2020. 10. World Health Organization. Sexual health and its linkages to reproductive health: an operational approach. 2017. https://www.who.int. Accessesd 16 Jul 2020. 11. Griffin S. Universal access to sexual and reproductive health services. Eldis Health Key Issues. 2006. https://assets.publishing.service.gov.uks. Accessed 22 Jul 2020 12. Green A. Man to man, sexual health needs are better met at male-only clinics. Sonke Gender Justice: Cape Town, South Africa. 2015. 13. Health Communication Capacity Collaborative (HC3). Guide for Promoting Sexual and Reproductive Health Products and Services for Men. Baltimore, MD: Johns Hopkins Center for Communication Programs (CCP). 2017. https://healthcommcapacity.org. Accessed 12 Sep 2020. 14. EngenderHealth. Men’s Reproductive Health Curriculum: Introduction to Men’s Reproductive Health Services. 2008. http://www.medbox.org. Accessed 13 Jun 2021. 15. Kalmuss D, Tatum C. Patterns of men’s use of sexual and reproductive health services. Perspectives on Sexual and Reproductive Health. 2007 Jun;39(2):74–81. Page 15/36 Page 15/36 16. Mnyulwa B. References One Man Can’s First Clinic in Gugulethu: Our Pride. Sonke Gender Justice: Cape Town, South Africa. 2011. https://genderjustice.org.za. Accessed 23 Oct 2020. 17. Sonke Gender Justice. Men for change, health for all: A policy discussion paper on men, health and gender equity. 2008. Sonke Gender Justice Network: Cape Town, South Africa. https://genderjustice.org.za. Accessed 17 Jun 2020. 18. Zuma T. mHealth for Men: Development of a home-based intervention to test and start (HITS) to support HIV testing and early linkage to care amongst men in rural KwaZulu-Natal, South Africa. Africa Health Research Institute. 2018. 19. Baker P, Shand T. Men’s health: time for a new approach to policy and practice?. Journal of Global Health. 2017 Jun;7(1). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5344012/. Accessed 26 Jul 2020. 20. Hoffman JR, Arendse KD, Larbi C, Johnson N, Vivian LM. Perceptions and knowledge of voluntary medical male circumcision for HIV prevention in traditionally non-circumcising communities in South Africa. Global public health. 2015 Jul 3;10(5–6):692–707. 20. Hoffman JR, Arendse KD, Larbi C, Johnson N, Vivian LM. Perceptions and knowledge of voluntary medical male circumcision for HIV prevention in traditionally non-circumcising communities in South Africa. Global public health. 2015 Jul 3;10(5–6):692–707. 21. Humphries H, van Rooyen H, Knight L, Barnabas R, Celum C. ‘If you are circumcised, you are the best’: understandings and perceptions of voluntary medical male circumcision among men from KwaZulu- Natal, South Africa. Culture, health & sexuality. 2015 Aug 9;17(7):920 – 31. 21. Humphries H, van Rooyen H, Knight L, Barnabas R, Celum C. ‘If you are circumcised, you are the best’: understandings and perceptions of voluntary medical male circumcision among men from KwaZulu- Natal, South Africa. Culture, health & sexuality. 2015 Aug 9;17(7):920 – 31. 22. Lissouba P, Taljaard D, Rech D, Dermaux-Msimang V, Legeai C, Lewis D, Singh B, Puren A, Auvert B. Adult male circumcision as an intervention against HIV: an operational study of uptake in a South African community (ANRS 12126). BMC infectious diseases. 2011 Dec;11(1):1–2. 23. Zissette S, Watt MH, Prose NS, Mntambo N, Moshabela M. “If you don’t take a stand for your life, who will help you?”: Men’s engagement in HIV care in Kwazulu-Natal, South Africa. Psychology of men & Masculinity. 2016 Jul;17(3):265. 24. Khan R, Yassi A, Engelbrecht MC, Nophale L, van Rensburg AJ, Spiegel J. Barriers to HIV counselling and testing uptake by health workers in three public hospitals in Free State Province, South Africa. AIDS care. References 2015 Feb 1;27(2):198–205. 24. Khan R, Yassi A, Engelbrecht MC, Nophale L, van Rensburg AJ, Spiegel J. Barriers to HIV counselling and testing uptake by health workers in three public hospitals in Free State Province, South Africa. AIDS care. 2015 Feb 1;27(2):198–205. 25. Mambanga P, Sirwali RN, Tshitangano T. Factors contributing to men's reluctance to seek HIV counselling and testing at primary health care facilities in Vhembe District of South Africa. African journal of primary health care and family medicine. 2016 Jan 1;8(2):1–7. 26. Mwisongo A, Mohlabane N, Tutshana B, Peltzer K. Barriers and facilitators associated with HIV testing uptake in South African health facilities offering HIV Counselling and Testing. Health SA Gesondheid. 2016 Jan 1;21(1):86–95. 27. Stern E, Cooper D, Rau A. Sexual and reproductive health perceptions and practices as revealed in the sexual history narratives of South African men living in a time of HIV/AIDS. SAHARA: Journal of Social Aspects of HIV/AIDS Research Alliance. 2014 Jan 1;11(1):233 – 44. DOI:10.1080/17190376.2014.985701 28. Morison T, Lynch I. Use and perceptions of public sexual and reproductive health services: A quantitative situational analysis in OR Tambo and Gert Sibande districts. 2016. Page 16/36 Page 16/36 29. Mthembu ME. Factors affecting men's health care seeking behaviour and use of services: a case study of Inanda Township, Durban (Doctoral dissertation). 2015. 30. Chikovore J, Gillespie N, McGrath N, Orne-Gliemann J, Zuma T. Men, masculinity, and engagement with treatment as prevention in KwaZulu-Natal, South Africa. 2016;28(sup3):74–82. 31. Evens E, Lanham M, Hart C, Loolpapit M, Oguma I, Obiero W. Identifying and addressing barriers to uptake of voluntary medical male circumcision in Nyanza, Kenya among men 18–35: a qualitative study. PLoS One. 2014 Jun 5;9(6): e98221. 32. Herman-Roloff A, Otieno N, Agot K, Ndinya-Achola J, Bailey. Acceptability of medical male circumcision among uncircumcised men in Kenya one year after the launch of the national male circumcision program. 2011;6(5): e19814. 32. Herman-Roloff A, Otieno N, Agot K, Ndinya-Achola J, Bailey. Acceptability of medical male circumcision among uncircumcised men in Kenya one year after the launch of the national male circumcision program. 2011;6(5): e19814. 33. Macintyre K, Andrinopoulos K, Moses N, Bornstein M, Ochieng A, et al. Attitudes, perceptions and potential uptake of male circumcision among older men in Turkana County, Kenya using qualitative methods. PLoS One. 2014; 9(5), p.e83998. 34. Withers M, Dworkin SL, Onono M, Oyier B, Cohen CR, Bukusi EA, Newmann SJ. References Men's perspectives on their role in family planning in Nyanza Province, Kenya. Studies in Family Planning. 2015 Jun;46(2):201 – 15. 35. Godia PM, Olenja JM, Lavussa JA, Quinney D, Hofman JJ, Van Den Broek N. Sexual reproductive health service provision to young people in Kenya; health service providers’ experiences. BMC health services research. 2013 Dec;13(1):1–3. 36. Mbonye M, Kuteesa M, Seeley J, Levin J, Weiss H, Kamali A. Voluntary medical male circumcision for HIV prevention in fishing communities in Uganda: the influence of local beliefs and practice. African journal of AIDS research. 2016 Sep 30;15(3):211-8. DOI: 10.2989/16085906.2016.1179652 37. Muhamadi L, Ibrahim M, Wabwire-Mangen F, Peterson S, Reynolds SJ. Perceived medical benefit, peer/partner influence and safety and cost to access the service: client motivators for voluntary seeking of medical male circumcision in Iganga district eastern Uganda, a qualitative study. Pan African Medical Journal. 2013;15(1). 38. Ssekubugu R, Leontsini E, Wawer MJ, Serwadda D, Kigozi G, Kennedy CE, Nalugoda F, Sekamwa R, Wagman J, Gray RH. Contextual barriers and motivators to adult male medical circumcision in Rakai, Uganda. Qualitative health research. 2013 Jun;23(6):795–804. 38. Ssekubugu R, Leontsini E, Wawer MJ, Serwadda D, Kigozi G, Kennedy CE, Nalugoda F, Sekamwa R, Wagman J, Gray RH. Contextual barriers and motivators to adult male medical circumcision in Rakai, Uganda. Qualitative health research. 2013 Jun;23(6):795–804. 39. Nalwadda G, Mirembe F, Byamugisha J, Faxelid E. Persistent high fertility in Uganda: young people recount obstacles and enabling factors to use of contraceptives. BMC public health. 2010 Dec;10(1):1–3. 39. Nalwadda G, Mirembe F, Byamugisha J, Faxelid E. Persistent high fertility in Uganda: young people recount obstacles and enabling factors to use of contraceptives. BMC public health. 2010 Dec;10(1):1–3. 40. Siu GE, Seeley J, Wight D. Dividuality, masculine respectability and reputation: how masculinity affects men's uptake of HIV treatment in rural eastern Uganda. Social science & medicine. 2013 Jul 1; 89:45–52. 40. Siu GE, Seeley J, Wight D. Dividuality, masculine respectability and reputation: how masculinity affects men's uptake of HIV treatment in rural eastern Uganda. Social science & medicine. 2013 Jul 1; 89:45–52. 41. Sunnu E, Adatara P, Opare FY, Kuug A, Nyande F. Factors Influencing the Utilisation of Family Planning contraceptives among Men and Women in the Ho Municipality of Ghana. Int. J. Heal. Sci. Res. 2016; 6:204. Page 17/36 Page 17/36 42. Dako-Gyeke P, Agyapong EB. References Utilization of HIV counselling and testing services in urban and rural communities in Eastern region of Ghana: Implications for achieving universal HCT coverage. Academic Journal of Interdisciplinary Studies. 2014 Jun 29;3(4):239-. 43. Leblanc NM, Andes KL. An exploration of men's knowledge, attitudes, and perceptions of HIV, HIV risk, and willingness to test for HIV in Yendi District, Northern Ghana. Journal of the Association of Nurses in AIDS Care. 2015 May 1;26(3):281 – 95. 43. Leblanc NM, Andes KL. An exploration of men's knowledge, attitudes, and perceptions of HIV, HIV risk, and willingness to test for HIV in Yendi District, Northern Ghana. Journal of the Association of Nurses in AIDS Care. 2015 May 1;26(3):281 – 95. 44. Addo IY, Gyamfuah AI. Utilization of reproductive health services among young people in Sawua, Ghana. Clinical Medicine Reviews in Patient Care. 2015 Aug 9;2015(3):1–6. 44. Addo IY, Gyamfuah AI. Utilization of reproductive health services among young people in Sawua, Ghana. Clinical Medicine Reviews in Patient Care. 2015 Aug 9;2015(3):1–6. 45. Adongo PB, Tapsoba P, Phillips JF, Tabong PT, Stone A, Kuffour E, Esantsi SF, Akweongo P. “If you do vasectomy and come back here weak, I will divorce you”: a qualitative study of community perceptions about vasectomy in Southern Ghana. BMC International Health and Human Rights. 2014 Dec;14(1):1–8. 46. Adongo PB, Tapsoba P, Phillips JF, Tabong PT, Stone A, Kuffour E, Esantsi SF, Akweongo P. “If you do vasectomy and come back here weak, I will divorce you”: a qualitative study of community perceptions about vasectomy in Southern Ghana. BMC International Health and Human Rights. 2014 Dec;14(1):1–8. 47. Hatzold K, Mavhu W, Jasi P, Chatora K, Cowan FM, Taruberekera N, et al. Barriers and motivators to voluntary medical male circumcision uptake among different age groups of men in Zimbabwe: results from a mixed methods study. 2014;9(5):e85051. 47. Hatzold K, Mavhu W, Jasi P, Chatora K, Cowan FM, Taruberekera N, et al. Barriers and motivators to voluntary medical male circumcision uptake among different age groups of men in Zimbabwe: results from a mixed methods study. 2014;9(5):e85051. 48. Moyo S, Mhloyi M, Chevo T, Rusinga O. Men's attitudes: A hindrance to the demand for voluntary medical male circumcision–A qualitative study in rural Mhondoro-Ngezi, Zimbabwe. Global public health. 2015 Jul 3;10(5–6):708–20. DOI:10.1080/17441692.2015.1006241 48. Moyo S, Mhloyi M, Chevo T, Rusinga O. References Men's attitudes: A hindrance to the demand for voluntary medical male circumcision–A qualitative study in rural Mhondoro-Ngezi, Zimbabwe. Global public health. 2015 Jul 3;10(5–6):708–20. DOI:10.1080/17441692.2015.1006241 49. Skovdal M, Campbell C, Madanhire C, Mupambireyi Z, Nyamukapa C, Gregson SJG, et al. Masculinity as a barrier to men's use of HIV services in Zimbabwe. 2011;7(1):13. 49. Skovdal M, Campbell C, Madanhire C, Mupambireyi Z, Nyamukapa C, Gregson SJG, et al. Masculinity as a barrier to men's use of HIV services in Zimbabwe. 2011;7(1):13. 50. Thomas KA. HIV/AIDS Voluntary Counselling and Testing (VCT): Perspectives of Rural Youths in Oyo State, Nigeria. International Journal of Agricultural Economics and Rural Development. 2015 Nov 1;7(1):52 – 8. 51. Jones OO. Factors affecting the uptake of voluntary counselling and testing among youth in rural Nigeria. University of Lethbridge (Canada); 2017. 51. Jones OO. Factors affecting the uptake of voluntary counselling and testing among youth in rural Nigeria. University of Lethbridge (Canada); 2017. 52. Hassan R. Knowledge, Perception, Risk Factors and Utilization of Prostate Cancer Screening Services among male staff of the University College Hospital, Ibadan, Nigeria. 2015. (Doctoral dissertation). 52. Hassan R. Knowledge, Perception, Risk Factors and Utilization of Prostate Cancer Screening Services among male staff of the University College Hospital, Ibadan, Nigeria. 2015. (Doctoral dissertation). 53. Tamang L, Raynes-Greenow C, McGeechan K, Black KI. Knowledge, experience, and utilisation of sexual and reproductive health services amongst Nepalese youth living in the Kathmandu Valley. Sexual & Reproductive Healthcare. 2017 Mar 1;11:25–30. 53. Tamang L, Raynes-Greenow C, McGeechan K, Black KI. Knowledge, experience, and utilisation of sexual and reproductive health services amongst Nepalese youth living in the Kathmandu Valley. Sexual & Reproductive Healthcare. 2017 Mar 1;11:25–30. 54. Gautam P, Soomro MH, Sapkota S, Gautam KR, Kasaju A. Barriers to utilization of sexual health services among young people in district dang Nepal: A qualitative study. Journal of Medicine. 2018 Jul 10;19(2):79–83. 54. Gautam P, Soomro MH, Sapkota S, Gautam KR, Kasaju A. Barriers to utilization of sexual health services among young people in district dang Nepal: A qualitative study. Journal of Medicine. 2018 Jul 10;19(2):79–83. Page 18/36 55. Regmi PR, Van Teijlingen E, Simkhada P, Acharya DR. Barriers to sexual health services for young people in Nepal. Journal of health, population, and nutrition. 2010 Dec;28(6):619. 56. Gasasira RA, Sarker M, Tsague L, Nsanzimana S, Gwiza A, Mbabazi J, Karema C, Asiimwe A, Mugwaneza P. References Determinants of circumcision and willingness to be circumcised by Rwandan men, 2010. BMC public health. 2012 Dec;12(1):1–8. 57. Shattuck D, Wesson J, Nsengiyumva T, Kagabo L, Bristow H, Zan T, Ngabo F. Who chooses vasectomy in Rwanda? Survey data from couples who chose vasectomy, 2010–2012. Contraception. 2014 Jun 1;89(6):564 – 71. 58. Adams A, Moyer E. Sex is never the same: Men's perspectives on refusing circumcision from an in- depth qualitative study in Kwaluseni, Swaziland. Global Public Health. 2015 Jul 3;10(5–6):721 – 38. DOI: 10.1080/17441692.2015.1004356 59. Kelly A, Kupul M, Fitzgerald L, Aeno H, Neo J, Naketrumb R, Siba P, Kaldor JM, Vallely A. " Now we are in a different time; various bad diseases have come." understanding men's acceptability of male circumcision for HIV prevention in a moderate prevalence setting. BMC Public Health. 2012 Dec;12(1):1–3. 60. Plotkin M, Castor D, Mziray H, Küver J, Mpuya E, Luvanda PJ, Hellar A, Curran K, Lukobo-Durell M, Ashengo TA, Mahler H. “Man, what took you so long?” Social and individual factors affecting adult attendance at voluntary medical male circumcision services in Tanzania. Global Health: Science and Practice. 2013 Mar 1;1(1):108 – 16. 61. Price JE, Phiri L, Mulenga D, Hewett PC, Topp SM, Shiliya N, Hatzold K. Behavior change pathways to voluntary medical male circumcision: narrative interviews with circumcision clients in Zambia. PLoS One. 2014 Nov 6;9(11):e111602. 61. Price JE, Phiri L, Mulenga D, Hewett PC, Topp SM, Shiliya N, Hatzold K. Behavior change pathways to voluntary medical male circumcision: narrative interviews with circumcision clients in Zambia. PLoS One. 2014 Nov 6;9(11):e111602. 62. Sabone M, Magowe M, Busang L, Moalosi J, Binagwa B, Mwambona J. Impediments for the uptake of the Botswana government's male circumcision initiative for HIV prevention. The Scientific World Journal. 2013 Jan 1;2013. 62. Sabone M, Magowe M, Busang L, Moalosi J, Binagwa B, Mwambona J. Impediments for the uptake of the Botswana government's male circumcision initiative for HIV prevention. The Scientific World Journal. 2013 Jan 1;2013. 63. Skolnik L, Tsui S, Ashengo TA, Kikaya V, Lukobo-Durrell M. A cross-sectional study describing motivations and barriers to voluntary medical male circumcision in Lesotho. BMC public health. 2014 Dec;14(1):1–0. 63. Skolnik L, Tsui S, Ashengo TA, Kikaya V, Lukobo-Durrell M. A cross-sectional study describing motivations and barriers to voluntary medical male circumcision in Lesotho. BMC public health. 2014 Dec;14(1):1–0. 64. Yabeny TE. Assessing Condom Use among Navajo Men in the Southwest. References 2018. (Doctoral dissertation, Walden University). 64. Yabeny TE. Assessing Condom Use among Navajo Men in the Southwest. 2018. (Doctoral dissertation, Walden University). 65. Parcon CR. MEN, FAMILY PLANNING AND CONTRACEPTIVE USE IN WESTERN VISAYAS. 2010. 65. Parcon CR. MEN, FAMILY PLANNING AND CONTRACEPTIVE USE IN WESTERN VISAYAS. 2010. 66. Muntean N, Kereta W, Mitchell KR. Addressing the sexual and reproductive health needs of young people in Ethiopia: an analysis of the current situation. African journal of reproductive health. 2015 Sep 1;19(3):87–99. 66. Muntean N, Kereta W, Mitchell KR. Addressing the sexual and reproductive health needs of young people in Ethiopia: an analysis of the current situation. African journal of reproductive health. 2015 Sep 1;19(3):87–99. 67. Oraby DM. Sexual and reproductive health among young people in Egypt: the role and contribution of youth-friendly services. Sex Education. 2013 Jul 1;13(4):470-7. DOI:10.1080/14681811.2012.756810 67. Oraby DM. Sexual and reproductive health among young people in Egypt: the role and contribution of youth-friendly services. Sex Education. 2013 Jul 1;13(4):470-7. DOI:10.1080/14681811.2012.756810 67. Oraby DM. Sexual and reproductive health among young people in Egypt: the role and contribution of youth-friendly services. Sex Education. 2013 Jul 1;13(4):470-7. DOI:10.1080/14681811.2012.756810 Page 19/36 68. Thin Zaw PP, Liabsuetrakul T, Htay TT, McNeil E. Equity of access to reproductive health services among youths in resource-limited suburban communities of Mandalay City, Myanmar. BMC health Page 19/36 68. Thin Zaw PP, Liabsuetrakul T, Htay TT, McNeil E. Equity of access to reproductive health services among youths in resource-limited suburban communities of Mandalay City, Myanmar. BMC health Page 19/36 Page 19/36 Page 19/36 services research. 2012 Dec;12(1):1–2. 69. Thongmixay S, Essink DR, Greeuw TD, Vongxay V, Sychareun V, Broerse JE. Perceived barriers in accessing sexual and reproductive health services for youth in Lao People’s Democratic Republic. PLoS One. 2019 Oct 29;14(10):e0218296. 69. Thongmixay S, Essink DR, Greeuw TD, Vongxay V, Sychareun V, Broerse JE. Perceived barriers in accessing sexual and reproductive health services for youth in Lao People’s Democratic Republic. PLoS One. 2019 Oct 29;14(10):e0218296. 70. Khumalo-Sakutukwa G, Lane T, van-Rooyen H, Chingono A, Humphries H, Timbe A, Fritz K, Chirowodza A, Morin SF. Understanding and addressing socio-cultural barriers to medical male circumcision in traditionally non-circumcising rural communities in sub-Saharan Africa. Culture, health & sexuality. 2013 Oct 1;15(9):1085 – 100. DOI: 10.1080/13691058.2013.807519 Tables Table 1: Author, year of publication and study location, participant characteristics (age and gender), study design and study setting Page 20/36 Authors Year Study location Participant characteristics (age and gender) Study design Study setting (Urban or Rural) Hoffman et al. 2015 South Africa Men and women 18 and above Mixed - analytical cross- sectional study design Urban Humphries et al. 2015 South Africa Men 18-54 Qualitative - FGDs Rural Lissouba et al. 2011 South Africa Men 15-49 Quantitative - Community- based cross-sectional study Urban Zissette et al. 2016 South Africa Men 24-80 Qualitative – IDIs Urban Khan et al. 2014 South Africa Men 18-50 Survey questionnaire Urban Mambanga et al. 2016 South Africa Men 18 and above Qualitative - semi-structured interviews Urban Mohlabane et al. 2016 South Africa Men and women 16 and above Quantitative – cross-sectional survey (semi-structured questionnaire) Urban and Rural Stern et al. 2014 South Africa Men 18-55 Qualitative – FGDs Urban and Rural Morison et al. 2016 South Africa Men and women 20-90 Quantitative – Survey - A Healthcare User questionnaire Urban and Rural Mthembu et al. 2015 South Africa Men 18 and above Quantitative - Survey questionnaire Qualitative - IDIs Urban Chikovore et al. 2016 South Africa Men 17-64 Qualitative - IDIs, FGDs Rural Evens et al. 2014 Kenya Men and women Qualitative - FGDs, IDIs Urban and Rural Herman- Roloff et al. 2011 Kenya Men 18-40 Qualitative - FGDs Urban and Rural Macintyre et al. 2014 Kenya Men and women 25-49 Qualitative - FGDs, IDIs Urban and Rural Withers et al. 2015 Kenya Men 15-64 Quantitative - Kenya Demographic and Health Survey (KDHS). Urban and Rural Godia et al. 2013 Kenya Men and women 27-50 Qualitative – FGDs, IDIs Urban and Rural Mbonye et al. 2016 Uganda Men 18-35 Mixed: Quantitative - interim serosurvey; Qualitative - IDIs Rural Muhamadi et al. 2013 Uganda Men and women 18 and above Qualitative - IDIs, FGDs Urban Ssekubugu et al. 2013 Uganda Men 15-49 Quantitative - RCT Qualitative - FGDs Rural Nalwadda et al. 2010 Uganda Men and women 15-25 Qualitative - FGDs Urban and Rural Siu et al. 2013 Uganda Men 27-51 Qualitative - IDIs Rural Sunnu et al. 2016 Ghana Men and women 15-65 Quantitative - cross-sectional survey Urban Dako-Gyeke et al. 2014 Ghana Men 15-49 Quantitative - cross-sectional survey Urban and Rural Leblanc et al. Tables 2015 Ghana Men 18 and above Mixed: Quantitative – self- administered structured questionnaire. Qualitative – IDIs, FGDs Urban 15-25 15-25 Adongo et al. 2014 Ghana Men and women 18 and above Qualitative – IDIs, FGDs Urban and Rural Chikutsa et al. 2015 Zimbabwe Men and women 18-49 Qualitative - FGDs, IDIs Urban Hatzold et al. 2014 Zimbabwe Men 15-49 Mixed: Quantitative - National, population-based survey, Qualitative - FGDs Urban and Rural Moyo et al. 2015 Zimbabwe Men 15-79 Qualitative - FGDs Rural Skovdal et al. 2011 Zimbabwe Men and women 18 and above Qualitative - IDIs, FGDs Rural Thomas et al. 2015 Nigeria Men and women 18-30 Quantitative - interview schedule - pre-tested questionnaire (scale) Rural Jones et al. 2017 Nigeria Men and women 18-24 Qualitative - IDIs Rural Hassan et al. 2015 Nigeria Men 25-60 Quantitative - descriptive cross-sectional Urban Tamang et al. 2017 Nepal Men and women 15-24 Quantitative - cross-sectional household survey Urban Gautam et al. 2018 Nepal Men and women 15-24 Qualitative - IDIs Rural Regmi et al. 2010 Nepal Men and women 18-22 Qualitative - IDIs and FGDs Urban and Rural Gasasira et al. 2012 Rwanda Men Quantitative - cross sectional study Rural Shattuck et al. 2014 Rwanda Men and women 24-45 Quantitative - cross-sectional descriptive Urban and Rural Adams et al. 2015 Swaziland Men 18-49 Qualitative – FGDs, IDIs, participant observation Urban and Rural Kelly et al. 2012 Papua New Guinea Men 16 and above Qualitative – FGDs, IDIs Urban and Rural Plotkin et al. 2013 Tanzania Men and women 18-39 Qualitative - FGDs Urban and Rural Price et al. 2014 Zambia Men 18 and above Qualitative - semi-structured interviews Urban Sabone et al. 2013 Botswana Men 18 and above Mixed: Quantitative - survey was cross-sectional, - Qualitative – IDIs, FGDs Urban and Rural Skolnik et al. 2014 Lesotho Men 18 and above Mixed: Quantitative – cross-sectional Qualitative - FGDs Urban and Rural Yabeny et al. 2018 Mexico Men 20-39 Qualitative - IDIs Urban Parcon et al. 2010 Philippines (Western Visayas) Men 15-54 Quantitative - National Health and Demographic Survey (NDHS) Urban and Rural Muntean et al. 2015 Ethiopia Men and women 15-24 Qualitative - IDIs Urban and Rural Oraby et al. 2013 Egypt Men and women 15-24 Qualitative – IDIs, FGDs Urban and Rural Thongmixay et al. Tables 2019 Lao Men and women Qualitative - IDIs Urban and Rural women 15-24 study and Rural Khumalo- Sakutukwa et al. 2013 Zimbabwe, South Africa Men 18 and above Qualitative – IDIs, FGDs Rural Table 2: Barriers, Facilitators, and SRH issues Table 2: Barriers, Facilitators, and SRH issues Table 2: Barriers, Facilitators, and SRH issues Page 25/36 Authors Barriers Facilitators SRH issues Hoffman et al. Intra and Post and procedure complications Low-risk perception, Lack of social support High costs Fear HIV test Protection against diseases (HIV, STIs, cancers) acquisition Hygiene Virility Good societal standing MMC Humphries et al. Post-procedure complications Low-risk perception Virility Social support MMC Lissouba et al. Intra-procedure complications Non-culture High costs Safety Social support MMC Zissette et al. Threat to masculinity (stigma around HIV) Fear of losing manhood Social support Good standing example HIV testing and management Khan et al. Lack of confidentiality High costs HIV counselling and testing (HCT) Mambanga et al. Fear HIV test (stigma around HIV) Cultural practices preference Low-risk perception None HIV counselling and testing (HCT) Mohlabane et al. Fear of HIV test (stigma around HIV) Staff attitudes Lack of knowledge (testing sites, understanding HIV) Inconvenient opening hours Fear of death Ignorance (lack of condom use) Low-risk perception High costs (traveling) L k f i l t Knowledge Incentives for those who test for HIV Good staff attitudes Role modes Page 26/36 urge) Lack of knowledge of HIV transmission Substance abuse Low-risk perception Morison et al. Long waiting times Unavailability of medicines and equipment Staff attitudes Lak of privacy and confidentiality High costs (traveling) None HCT Family Planning Condoms use HIV and STI treatment and counselling Mthembu et al. Staff attitudes Long waiting times High costs Absence of male health workers HCT Chikovore et al. Fear HIV test Low-risk perception Preference of traditional medicine None HCT and management Evens et al. Myths (circumcised penis would tear a condom) Nonculture / Nonreligion Lack of trust in government witchcrafts beliefs Low-risk perception Fear HIV test Virility MMC Herman- Roloff et al. Fear of losing a job Noncultural, nonreligious Intra and Post procedure complications Lack of knowledge Vulnerability to ignorance Distance to health facilities Female service providers Hygiene Social acceptance Virility Protection against diseases Convenience (easier to use condom) MMC HCT Protection against diseases Social support Withers et al. Lack of knowledge or awareness Religious prohibition Unvirility Hindrance to community development None Family planning Godia et al. Tables Limited knowledge and competency of HSP Staff attitudes Lack of medicines and equipment Lack of confidentiality and privacy Long waiting times High costs. Inconvenient hours None Family Planning STI/HIV services, Condom use Mbonye et al. Post-procedure compilations Fear HIV test Protection against diseases Hygiene MMC Muhamadi et al. None Protection against diseases Hygiene Social support Virility MMC Ssekubugu et al. Post-procedure compilations Fear HIV test Myths (infertility) Partial protection against diseases Long waiting times Fear of losing a job Protection against diseases Hygiene Social support Incentives. MMC Nalwadda et al. Myths (infertility, porous and infectious condoms) Virility k None Family Planning Incontrollable sexual urge Staff attitudes Lack of privacy and confidentiality, Lack medicines and equipment High costs Long distance to health facilities Inconvenient opening hours, Long waiting times Condom use Siu et al. Threat to masculinity Fear HIV test Long waiting times Lack of confidentiality and privacy None HIV services Sunnu et al. Lack social support Nonreligious, noncultural Staff attitudes Distance to health facilities None Family planning Dako-Gyeke et al. Fear HIV test None HCT Leblanc et al. Cultural beliefs (non-condom use) Embarrassment Desire for children Fear of HIV test (stigma) High costs Lack of confidentiality None HIV services Condom use Yeboah et al. Embarrassment High costs Female health Staff attitudes Religious doctrines None Family planning STI services Adongo et Lack of social support None Vasectomy Condom use HIV services Vulnerability to ignorance Intra and Post procedure complications Lack of knowledge Religious doctrines Chikutsa et al. Myths (circumcised penis would tear a condom, foreskin may be used in satanic rituals.) Nonculture, nonreligion Lack of trust in government Low-risk perception Fear HIV test Virility Protection against diseases MMC Hatzold et al. Intra procedure complications Low-risk perception Lack of social support, High costs Fear of an HIV test Myths (infertility) Protection against diseases Hygiene, Virility Set good example MMC Moyo et al. Lack of knowledge Fear of HIV testing Myths (foreskin to be used in witchcraft) Lack of trust in government Threat to masculinity Female health care provider Unvirility None MMC Skovdal et al. Fear of HIV test Stigma Embarrassment Threat to masculinity Low risk perception Social support Role models HIV services Page 30/36 Thomas et al. Fear of HIV test Stigma Nonculture None HIV services Jones et al. Stigma Inaccessibility Lack of knowledge Inaccessibility Low-risk perception None HIV services Hassan et al. Tables Low-risk perception Lack of time Intra procedure complications Fear of PC test outcome Lack of knowledge None Prostate Cancer screening Tamang et al. Embarrassment Poor health services Lack of knowledge Inaccessibility None Condom use Family planning Gautam et al. Embarrassment Lack of knowledge Staff attitudes Lack of privacy and confidentiality Lack of medicines and equipment None All Regmi et al. Embarrassment Poor health services Lack of knowledge Substance abuse Inaccessibility High costs None Condom use None HIV services HIV services Prostate Cancer screening Shattuck et al. None Financial relief (limit family size) Complications of other family plaining methods Permanent method Low-risk of complications Vasectomy services Adams et al. Threat to masculinity Virility Intra and post procedure complications Partially protective against diseases Lack of trust in government Noncultural, nonreligious Lack of knowledge Myths (foreskins used for witchcraft) Virility Protection against diseases Convenience (wearing condoms) MMC Kelly et al. Vulnerable to ignorance Noncultural, nonreligious Protection against diseases Hygiene Culturally appropriate Virility MMC Plotkin et al. Post procedure complications Fear of losing job Protection against diseases Hygiene Social support Virility MMC Price et al. Intra and post procedure complications Lack of time Social support Hygiene Protection against diseases MMC Sabone et al Lack of medicines and equipment Advertisement MMC Page 32/36 Low-risk of complications Adams et al. Threat to masculinity Virility Intra and post procedure complications Partially protective against diseases Lack of trust in government Noncultural, nonreligious Lack of knowledge Myths (foreskins used for witchcraft) Virility Protection against diseases Convenience (wearing condoms) MMC Kelly et al. Vulnerable to ignorance Noncultural, nonreligious Protection against diseases Hygiene Culturally appropriate Virility MMC Plotkin et al. Post procedure complications Fear of losing job Protection against diseases Hygiene Social support Virility MMC Price et al. Intra and post procedure complications Lack of time Social support Hygiene Protection against diseases MMC Sabone et al. Lack of medicines and equipment Long waiting times Poor transport to health facilities Post procedure complications Advertisement Social support Right staff attitudes MMC Skolnik et al. Intra procedure complications Fear of HIV test High costs Female health workers Long waiting times Protection against diseases Hygiene Social support Virility MMC Yabeny et al. Low-risk perception Fear of illness Mistrust in relationship Condom use Parcon et al. Lack of knowledge Myths (loss of libido) Unvirility decreased sexual activity, loss of vitality, Inconvenience Embarrassment Protection against diseases Vasectomy Condom use Muntean et al. High costs Long distance to the health facility Staff attitudes Inconvenient location of facilities Inconvenient hours Lack of privacy and confidentiality Embarrassment Lack of knowledge Noncultural None All Oraby et al. Inaccessibility Absence of male health workers None All Thongmixay et al. Lack of medicines and equipment Low-risk perception Lack of knowledge Lack of privacy and confidentiality Protection against diseases Social acceptance Condom use Family planning All All Embarrassment High costs Zaw et al. None High costs Lack of confidentiality Lack of transport Staff attitudes Fear of HIV test Embarrassment Lack of knowledge None Family planning HIV services Condom use Khumalo- Sakutukwa et al. Noncultural Intra and post procedure complications None MMC Figures Page 34/36 Figure 1 Flow chart mapping out the number of articles identified, screened, excluded together wit exclusion Figure 1 Flow chart mapping out the number of articles identified, screened, excluded together with reasons for exclusion Page 35/36 Figure 2 Conceptual Framework Conceptual Framework Page 36/36
https://openalex.org/W2584488565	https://europepmc.org/articles/pmc5295584?pdf=render	English	null	A Network of Chromatin Factors Is Regulating the Transition to Postembryonic Development in<i>Caenorhabditis elegans</i>	G3	2,017	cc-by	10,394	KEYWORDS genome-wide RNAi screen chromatin development germline P granules Copyright © 2017 Erdelyi et al. doi: 10.1534/g3.116.037747 Manuscript received August 24, 2016; accepted for publication December 11, 2016; published Early Online December 22, 2016. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Supplemental material is available online at www.g3journal.org/lookup/suppl/ doi:10.1534/g3.116.037747/-/DC1. 1Present address: Institute of Biomedical Sciences, Fudan University, Shanghai, China 200433. 2Corresponding author: Department of Biology, University of Fribourg, Ch. du Musée 10, 1700 Fribourg, Switzerland. E-mail: chantal.wicky@unifr.ch MUTANT SCREEN REPORT MUTANT SCREEN REPORT A Network of Chromatin Factors Is Regulating the Transition to Postembryonic Development in Caenorhabditis elegans Peter Erdelyi, Xing Wang,1 Marina Suleski, and Chantal Wicky2 Department of Biology, University of Fribourg, 1700 Fribourg, Switzerland ABSTRACT Mi2 proteins are evolutionarily conserved, ATP-dependent chromatin remodelers of the CHD family that play key roles in stem cell differentiation and reprogramming. In Caenorhabditis elegans, the let-418 gene encodes one of the two Mi2 homologs, which is part of at least two chromatin complexes, namely the Nucleosome Remodeling and histone Deacetylase (NuRD) complex and the MEC complex, and functions in larval development, vulval morphogenesis, lifespan regulation, and cell fate determination. To explore the mechanisms involved in the action of LET-418/Mi2, we performed a genome-wide RNA interference (RNAi) screen for suppressors of early larval arrest associated with let-418 mutations. We identiﬁed 29 suppressor genes, of which 24 encode chromatin regulators, mostly orthologs of proteins present in transcriptional activator complexes. The remaining ﬁve genes vary broadly in their predicted functions. All suppressor genes could suppress multiple aspects of the let-418 phenotype, including developmental arrest and ectopic expression of germline genes in the soma. Analysis of available transcriptomic data and quantitative PCR revealed that LET-418 and the suppressors of early larval arrest are regulating common target genes. These suppressors might represent direct competitors of LET-418 complexes for chromatin regulation of crucial genes involved in the transition to postembryonic development. The concerted action of transcription factor networks and epigenetic regulators is required to ensure proper development of a multicellular organism. Together, these factors tightly control the transcriptional activity of the genome to allow a cell or a group of cells to acquire a speciﬁc fateata giventime of development.The highly conserved, ATP- dependent epigenetic modiﬁerMi2 ispartof an abundant multi-protein complexinmammaliancellscalled NuRD(nucleosomeremodelingand deacetylase). The ﬁrst evidence for a developmental function of NuRD came from studies in mouse embryos, where a lack of the NuRD component Mbd3 compromised the differentiation potential of ES cells (Kaji et al. 2006, 2007). Molecular studies in ESCs revealed that Mi2b, as well as other NuRD components, suppresses the expres- sion of pluripotency genes to allow transcriptional heterogeneity and, ﬁnally, proper lineage commitment (Reynolds et al. 2012, 2013). In the mouse hematopoietic system and during skin devel- opment the central component of NURD, Mi2b, is required for stem cell homeostasis and lineage choice (Kashiwagi et al. 2007; Yoshida et al. 2008). 2Corresponding author: Department of Biology, University of Fribourg, Ch. du Musée 10, 1700 Fribourg, Switzerland. E-mail: chantal.wicky@unifr.ch A Network of Chromatin Factors Is Regulating the Transition to Postembryonic Development in Caenorhabditis elegans The Drosophila dMi2, together with Polycomb group proteins and the Hunchback transcription factor, regulates the tran- scriptional activity of the HOX genes during embryonic patterning (Kehle et al. 1998; McDonel et al. 2009). In addition, in Arabidopsis, the Mi2 homolog Pickle represses embryonic traits in root meristem cells and is required for proper postembryonic development (Ogas et al. 1997, 1999; Eshed et al. 1999). (Kaji et al. 2006, 2007). Molecular studies in ESCs revealed that Mi2b, as well as other NuRD components, suppresses the expres- sion of pluripotency genes to allow transcriptional heterogeneity and, ﬁnally, proper lineage commitment (Reynolds et al. 2012, 2013). In the mouse hematopoietic system and during skin devel- opment the central component of NURD, Mi2b, is required for stem cell homeostasis and lineage choice (Kashiwagi et al. 2007; Yoshida et al. 2008). The Drosophila dMi2, together with Polycomb group proteins and the Hunchback transcription factor, regulates the tran- scriptional activity of the HOX genes during embryonic patterning (Kehle et al. 1998; McDonel et al. 2009). In addition, in Arabidopsis, the Mi2 homolog Pickle represses embryonic traits in root meristem cells and is required for proper postembryonic development (Ogas et al. 1997, 1999; Eshed et al. 1999). The genome of Caenorhabditis elegans encodes two Mi2 homo- logs, LET-418 and CHD-3. LET-418 is required for postembryonic development, repression of the germline expression program in somatic cells, proper patterning of the vulva, and lifespan regulation (Solari and Ahringer 2000; von Zelewsky et al. 2000; Unhavaithaya et al. 2002; De Vaux et al. 2013). Lack of chd-3 activity causes no obvious defects; however, when both let-418 and chd-3 functions are impaired, worm embryos arrest their development at the twofold stage. Genetic and biochemical analysis has revealed that LET-418 Volume 7 \| February 2017 \| 343 343 amount of embryos were transferred to 20 ml of bacterium-free and OP50-containing S Medium. During the test, animals were shaken slowly at 25. After every 24 hr, 100 ml of each sample were trans- ferred to three OP50-seeded NGM plates; the survival rate was ﬁrst determined and then the animals were kept at 15 to recover. Recov- ery rates were checked after 5–14 d. and CHD-3 act as multi-protein complexes. As part of the NuRD complex, they regulate early development. Microscopy Microscopy analyses were performed using a Zeiss axioplan 2 micro- scope.Forbrightﬁeld picturesa DICﬁlterwasused, andforﬂuorescence images the appropriate ﬂuorescence ﬁlter was used. A Zeiss AxioCam colorcameradrivenbyAxioVisionv4.8.2softwarewasappliedforimage acquisition. Genome-wide RNAi screen For the genome-wide RNAi screen, RNAi treatment was performed as described above. Two L4 stage worms were added to each well con- taining different RNAi clones. The plates were then incubated for 8 d at 25. After the ﬁrst 24 hr period, the two P generation animals were removed. The F1 progeny was observed. Positive clones were retested in three independent experiments. The level of suppression was deter- mined by comparison with a negative control (let-418 worms on HT115 bacteria containing empty L4440 vector) and was shown to be statistically signiﬁcant (P , 0.001, Fisher’s exact test, data not shown). In our study, we show that LET-418 is required for the transition to postembryonic development. let-418 mutants stop their development at the L1 stage and show no divisions of the germ cells and the blast cells. let-418 L1 larvae look superﬁcially similar to L1 diapaused larvae, but differ in their survival rate. A genome-wide RNAi screen identiﬁed 29 suppressors of let-418 larval arrest, the majority of which encoded chromatin factors. All of the suppressors, except the histone methylase- encoding gene, set-26, suppressed both the developmental defect and the ectopic expression of germline genes associated with the develop- mental arrest of the let-418 mutant. Finally, we show that a subset of the suppressors, together with LET-418, regulate common target genes, including germline genes and DAF-16 targets. We propose that an epigenetic network is controlling the transition to postembryonic development by acting on common target genes. A Network of Chromatin Factors Is Regulating the Transition to Postembryonic Development in Caenorhabditis elegans LET-418, together with the krüppel-like protein MEP-1 and the histone deacetylase HDA-1, is also a member of the so-called MEC complex, which is required for postembryonic development and repression of the germline ex- pression program in somatic tissues (Passannante et al. 2010). Fur- thermore, LET-418, together with the histone H3K4 demethylase SPR-5/LSD1, prevents somatic reprogramming of the germline stem cells (Käser-Pébernard et al. 2014). Although studies in various organisms have revealed important developmental functions of Mi2 proteins and NuRD, the regulatory networks to which they contribute are not understood. RNA isolation The following strains were used in this study: wild-type (var. Bristol), let-418(n3536ts) (FR843), ayIs6[pBH47.70(hlh-8p::gfp) + pMH86(dpy- 20(+))] (PD4666), unc-119(e2498::Tc1); wIs51[scmp::gfp + unc-119p:: unc-119(+)] (JR667), bnIs1[pie-1p::gfp::pgl-1 + unc-119p::unc-119(+)] (SS747), bnIs1[pie-1p::gfp::pgl-1 + unc-119p::unc-119(+)];let-418(n3536ts) (FR1195), set-26(tm2467) (FR1496), set-9(n4949) (MT16426), set- 26(tm2467);let-418(n3536ts) (FR1499), set-9(n4949);let-418(n3536ts) (FR1500), set-26(tm2467); set-9(n4949);let-418(n3536ts) (FR1506), nurf-1(n4295) (MT13649), and nurf-1(n4295);let-418(n3536ts) (FR1495). All C. elegans strains were maintained at 20 using standard conditions (Brenner 1974). Synchronized L4 stage animals were treated with RNAi at 25 as de- scribed above. After 24 hr, next generation embryos were collected in M9 solution from the gravid adults by hypochlorite treatment and were kept at 25 for 10 hr. Synchronized L1 worms were then put back onto the relevant RNAi plates for 3 hr at 25. After incubation, larvae were harvested and RNA was isolated using a QIAGEN RNeasy Mini kit, according to the manufacturer’s protocol, combined with Precellys 24 0.5 mm glass beads to break open the animals. Quantitative real-time PCR (qRT-PCR) cDNA was synthesized from total RNA using the QuantiTech Reverse Transcription Kit (QIAGEN). SensiFast SYBR No-ROX Kit (Bioline) was used for qPCR with a Corbett Rotor-Gene 6000 machine driven by Rotor-Gene 6000 v1.7 software. The primers were designed using Primer3 online software (http://bioinfo.ut.ee/primer3-0.4.0/primer3/). At least one primer of each primer pair aligned to an exon–exon junc- tion. Primers designed for ama-1 and act-1 genes were used for nor- malization. The efﬁciency of the primer pairs was tested by setting a standard curve using a serial dilution of cDNA, and the speciﬁcity of the primers was monitored by analyzing the melting curves of each re- action. Data from triplicate reactions were analyzed using a 22DDCt method. Biological replicates were used to conﬁrm the results. For P granule and blast cell screens To determine the numberof P granule-containingsomatic cells, M cells, or V cells, 10 each of wild-type and let-418(n3536) L4 stage animals carrying the appropriate transgene were transferred to RNAi plates and kept at 25. Adults were removed after 24 hr of incubation and 40 F1 progeny were analyzed after 48 hr to determine the number of ﬂuores- cent cells. The Mann–Whitney U-test was used to determine the sig- nificance level. MATERIALS AND METHODS RNAi treatment A previously described feeding technique was applied for RNAi treat- ment during the study (Timmons and Fire 1998). For the screen and for later experiments, RNAi clones from the Ahringer library were used (Kamath and Ahringer 2003). During the screen 12-well agar plates were used, whereas during other RNAi experiments 5 cm-wide petri dishes were applied. In all cases, HT115 bacteria containing empty L4440 vector were used as control. Postembryonic development in C. elegans is dependent on nutrient availability (Baugh 2013). In the absence of food, freshly hatched L1 (ﬁrst larval stage) larvae undergo a developmental arrest (diapause) that is dependent on DAF-16/FOXO (Baugh and Sternberg 2006; Fukuyama et al. 2006). In diapaused larvae, germ cells and blast cells do not divide. It was shown recently that DAF-16/FOXO cell- nonautonomously controls the activity of both the dbl-1/TGFb and the daf-12/NHR signaling pathways, which both promote postembry- onic development (Kaplan et al. 2015). MATERIALS AND METHODS Culture conditions and C. elegans strains A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors g Freshly hatched wild-type larvae possess 10 pairs of seam cells. All of them, except the most anterior ones (H0), are blast cells that will divide during development and give rise to different cell types, such as hypodermal cells, neurons, and glial cells (Sulston and Horvitz 1977) (Figure 1A, e and f). We analyzed the developmental pattern of the seam cells using the reporter gene scm::gfp (Terns et al. 1997) and found that, in let-418 L1 larvae, the 10 pairs of V cells do not divide (Figure 1A, g). Furthermore, freshly hatched wild-type L1 larvae have a single M cell which, during further development, will give rise to two coelomo- cytes, 14 body wall muscle cells, and two sex myoblast cells (Figure 1A, i and j) (Sulston and Horvitz 1977). Using the hlh-8::gfp marker, which is expressed in all undifferentiated cells of the M lineage (Harfe et al. 1998), we observed a mitotic arrest of the M cell in let-418 mutants (Figure 1A, k). In summary, our data suggest that, in arrested let-418 L1 larvae, blast cells remain mitotically arrested. This growth arrest is similar to the L1 diapause, a dormant state that freshly hatched wild- type L1 larvae enter if no food is encountered (Baugh and Sternberg 2006; Fukuyama et al. 2006), and where the Z, V, and M cells are also mitotically arrested (Figure 1A, d, h, and l). Blast cell quiescence in L1 diapaused larvae is dependent on the DAF-16/FOXO transcription factor (Baugh and Sternberg 2006). In the absence of daf-16 activity, blast cells are able to divide in starved L1 larvae. However, In let-418 mutants, blast cell quiescence seems not to depend on DAF-16, as M cell division in daf-16;let-418 double mutant worms could not be To better understand the role of LET-418 during the transition to postembryonic development and uncover its regulatory network, we performed a genome-wide RNAi screen to identify suppressors of L1 arrest. Out of the 16,757 genes of the Ahringer library (Timmons et al. 2001; Kamath et al. 2003), 29 suppressors were identiﬁed (Table 1). Among the 29 suppressor genes, 24 are known to be involved in chro- matin regulation. At least 15 of them encode proteins that belong to activating chromatin complexes, such as the ISWI/NURF, SWR1, NuA4, KAT8/MOF, and COMPASS complexes (Table 1). p The proteins NURF-1 and ISW-1 are homologs of the ISWI complex members. LET-418 promotes postembryonic development The progeny of temperature-sensitive let-418(n3536) mutants, grown at the restrictive temperature of 25, fail to develop past the ﬁrst larval (L1) stage. Examination of the mutant phenotype revealed that the two primordial germ cells (PGCs) Z2/Z3 do not divide in let-418 L1 larvae (Figure 1A, c), whereas proliferation starts after hatching in wild-type animals (Figure 1A, a and b). In freshly hatched wild-type larvae, other cells, called blast cells, will also start to divide and differentiate to generate adult structures. To determine if these blast cells were also mitotically arrested in let-418 mutants, we examined the V (seam) and the M (mesoblast) cell lineages. ModEncode analysis Interpreted ChIP-seq data for each gene were downloaded from the modENCODE C. elegans online database (http://www.modencode.org) (Celniker et al. 2009). The list of affected genes was extracted using PAVIS peak annotation online software (http://manticore.niehs.nih. gov/pavis2/). Gene ontology (GO) analyses were performed using the DAVID bioinformatics database (Huang et al. 2009a,b). P-values were determined using Fisher’s exact test. In addition to blast cell quiescence, starved L1 larvae exhibit an extended survival rate in the absence of food and, when returned to food, can recover from this diapause state and resume their develop- ment. We measured the survival rate of arrested let-418 L1 larvae. Both fed and starved let-418 L1 mutants showed a signiﬁcantly decreased survival rate as compared to starved wild-type L1 larvae (Figure 1B). However, starved let-418 L1 larvae exhibited a better survival rate than fed let-418 worms (Figure 1B), indicating that starvation in- creases the resistance of let-418 mutants. We also tested if a lack of daf-16 activity would improve the survival rate of let-418 in fed con- ditions, but we could not observe any effect (Figure S1D). Absence of daf-16 activity prevents the occurrence of some defects in the let-418 larvae, but not the survival rate. Next, we determined the recovery rate of let-418(ts) mutants by returning them to the permissive tem- perature after different time points. Interestingly, starved let-418(ts) L1 larvae recovered better than fed let-418(ts) mutants, which almost completely lost their potential to resume development after 3 d at restrictive temperature (Figure 1B) and exhibited a hlh-8::gfp expres- sion pattern comparable to starved wild-type L1 larvae (Figure S1, B and C). Obviously, starvation protects let-418 worms and preserves their developmental potential. In summary, our data demonstrated that let-418 L1 larvae stop postembryonic development in a process that looks superﬁcially similar, but is not identical to, that observed in a starvation-induced L1 diapause. Starvation assay Synchronized L4 stage animals were transferred to 25. After 24 hr, the F1 generation of synchronized embryos was collected and equal 344 \| P. Erdelyi et al. statistical analysis, we used the Student’s t-test to compare the relative mRNA level of different genetic variants. detected (Figure S1A). However, after close examination of the M cell morphology in let-418 single and daf-16;let-418 double mutants, we observed that a large proportion of let-418 mutant animals were losing the expression of the M cell marker hlh-8::gfp after 3 d of developmental arrest, whereas daf-16;let-418 double mutants or starved wild-type an- imals did not (Figure S1, B and C). This observation suggested that an absence of daf-16 does not suppress mitotic quiescence, but instead prevents the occurrence of some other defects in the let-418 mutant larvae, as indicated by the loss of expression of the M cell marker (Figure S1, B and C). Volume 7 February 2017 \| Chromatin Factors in C. elegans Data availability Further analysis of daf-16/let-418 interaction is provided in Supplemen- tal Material, Figure S1. Suppression of let-418 developmental arrest by set-26 is described in Figure S2. Suppression quantiﬁcation of let-418- associated ectopic P granule expression, and M cell and V cell mitotic arrest, are provided in Figure S3, Figure S4, and Figure S5, respectively. LET-418 and DAF-16 common target genes are presented in Figure S6. Figure S7 shows the expression levels of selected DAF-16 targets measured in wild-type, let-418, daf-16, and daf-16;let-418 mutant backgrounds. A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors Larvae were allowed to hatch at 25 in the presence or absence of food; starved larvae were returned to food following the indicated number of days. All larvae were maintained at 15. wt, wild-type. through chromatin remodeling, histone variant deposition, and histone acetylation [reviewed in Lu et al. (2009)]. Genetic interaction of let-418 with components of these two complexes suggests that LET-418 could prevent the establishment of active chromatin at various loci in the genome. test if both genes act as suppressors, we generated set-26;let-418 and set-9;let-418 double mutants as well as set-26;set-9;let-418 triple mu- tants. We found that only the mutation in set-26, but not in set-9, could suppress the L1 arrest phenotype of let-418 (Figure S2). Consistently, the rate of suppression in the set-26;let-418 double mutants was iden- tical to that of the triple mutants set-26;set-9;let-418, suggesting that only set-26, but not set-9, is a let-418 suppressor (Table 1). The reason that we identiﬁed set-9 as a suppressor was likely due to the fact that set-9(RNAi) cross-inhibited set-26 expression. set-26 encodes a H3K9 methyltransferase that is involved in the transgenerational sterility of spr-5 mutants lacking H3K4 demethylase activity (Greer et al. 2014). Some interesting aspects of the set-26 suppression effect will be further discussed below. hcf-1, dpy-30, wdr-5.1, cfp-1, and set-2 encode worm homologs of the COMPASS complex (Li and Kelly 2011; Shilatifard 2012). The COMPASS complex is responsible for H3K4 methylation, which cor- relates with transcriptional activity and might antagonize let-418 func- tion during development. mes-2/3/6 encode members of a worm PRC2 repressive complex, which is considered to be a regulator of pluripotency and differentiation in both mammals and C. elegans (Boyer et al. 2006; Lee et al. 2006; Yuzyuk et al. 2009). Absence of let-418 activity could lead to mislocal- ization of Polycomb proteins and the repression of inappropriate genes, and the suppression effect obtained by deactivating PRC2 components may indicate the need to remove the repression of these genes to allow development in a let-418 background (see also Discussion). We also identiﬁed the two genes, sumv-1 and sumv-2, whose protein products are members of a putative worm KAT8/MOF his- tone acetyltransferase complex (Rea et al. 2007). This chromatin- activating complex is known to play a role during vulval development (Yucel et al. 2014). Among the RNAi suppressors, we found two genes, set-26 and set-9, that showed 97% identity at the level of their nucleotide sequence. A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors Reduction of the functions of isw-1 and nurf-1 has already been shown to suppress the larval-lethal phenotype of mep-1(q660) and let-418(n3536) (Andersen et al. 2006). Our data conﬁrmed these results. The proteins encoded by zhit-1, arp-6, C17E4.6, mrg-1, and gﬂ-1 are homologous to members of the mammalian SWR1 complex (Lu et al. 2009), which carries out the incorporation of histone variants such as H2A.Z (Mizuguchi et al. 2004). Depletion of members of the SWR1 complex or of HTZ-1, the C. elegans homolog of H2A.Z, suppresses the let-418 phenotype. Moreover, MRG-1 and GFL-1 are shared members of the NuA4 acetylation complex, which also includes the products of the suppressor gene homologs ZK1127.3 and MYS-4. Both the SWR1 and NuA4 complexes have been shown to exhibit antisilencing activity 345 Volume 7 February 2017 \| Chromatin Factors in C. elegans re 1 let-418 mutants arrest as L1. (A) Primordial germ cells and blast cells are mitotically arrested in let-418 mutants. a–d show differential ference contrast microscopy images of wt (a, b, and d) and let-418 (c) developing larvae. e–h show V lineage and division pattern in wt (e, f, h) and let-418 (g) revealed by the reporter construct scm::gfp. (3d = days). i–l display mesoblast (M) lineage and division pattern in wt (i, j, and d let-418 (k) larvae visualized by the reporter construct hlh-8::gfp. (B) LET-418 is required to survive and recover from starvation. Recovery and val assays of indicated genotypes. Larvae were allowed to hatch at 25 in the presence or absence of food; starved larvae were returned to following the indicated number of days. All larvae were maintained at 15. wt, wild-type. Figure 1 let-418 mutants arrest as L1. (A) Primordial germ cells and blast cells are mitotically arrested in let-418 mutants. a–d show differential interference contrast microscopy images of wt (a, b, and d) and let-418 (c) developing larvae. e–h show V lineage and division pattern in wt (e, f, and h) and let-418 (g) revealed by the reporter construct scm::gfp. (3d = days). i–l display mesoblast (M) lineage and division pattern in wt (i, j, and l) and let-418 (k) larvae visualized by the reporter construct hlh-8::gfp. (B) LET-418 is required to survive and recover from starvation. Recovery and survival assays of indicated genotypes. 346 \| P. Erdelyi et al. A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors To The autosome-speciﬁc H3K36 methyltransferase, MES-4, was also found among the suppressors (Furuhashi et al. 2010; Rechtsteiner et al. 346 \| P. Erdelyi et al. Chromatin factors are shown in bold. SWR1, Swi2/Snf2-related ATPase 1; MYST, Moz, Ybf2/Sas3, Sas2, Tip60; KAT, lysine (K) acetyltransferase; NLS, nuclear localisation signal; MOF, males absent on the ﬁrst; PHD, plant homeodomain; SET, Su(var)3-9, Enhancer-of-zeste and Trithorax; LSY, laterally symmetric; MATH, Meprin and TRAF-Homology. aStrong suppression, larvae reach L4/adulthood. bWeak suppression, larvae bypass L1 arrest. cMiddle suppression, larvae reach L2/L3 stage. A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors n Table 1 let-418 suppressor identities Complex Gene Name Human Ortholog ISWI/NURF isw-1a SMARCA1, SMARCA5 nurf-1b CECR2 SWR1 C17E4.6b VPS72 zhit-1b ZNHIT1 arp-6b ACTR6 htz-1c H2AFV, H2AFZ SWR1/NuA4 gﬂ-1b YEATS4 mrg-1a MRG15 NuA4 zk1127.3c — mys-4b MYST3 COMPASS cfp-1b CXXC1 dpy-30b DPY-30 wdr-5.1b WDR5, WDR5B hcf-1b HCF-1 set-2b SETD1A, SETD1B Polycomb mes-3b — mes-6b EED mes-2b EZH2 isoform b KAT8/MOF-like sumv-1c INO80D sumv-2c — Other C06A5.3c PSIP1, HDGFL1, HDGF mes-4a ASH1L, EZH1, EZH2 set-26c KMT2E, SETD5 lex-1c ATAD2, ATAD2B lsl-1b ZBTB20, ZBTB45, ZBTB7C ZFP57, ZNF296 math-33c USP-7 T19B4.5c — gtbp-1b G3BP1, G3BP2 M03C11.3c — n Table 1 let-418 suppressor identities Complex Gene Name Human Ortholog Description ISWI/NURF isw-1a SMARCA1, SMARCA5 ATPase component of a nucleosome remodeling factor (NURF)-like complex nurf-1b CECR2 Acts with isw-1 in vulval development SWR1 C17E4.6b VPS72 Gene expression regulator zhit-1b ZNHIT1 HIT-type zinc ﬁnger protein arp-6b ACTR6 Actin-related protein htz-1c H2AFV, H2AFZ Regulate gene expression with SWR1 in pharynx SWR1/NuA4 gﬂ-1b YEATS4 Transcription factor mrg-1a MRG15 Chromodomain-containing protein, promotes cell-proliferation NuA4 zk1127.3c — Unknown mys-4b MYST3 MYST family histone acetyltransferase COMPASS cfp-1b CXXC1 CFP1 (CpG-binding protein, CXXC Finger Protein 1) homolog dpy-30b DPY-30 Hermaphrodite dosage compensation and normal male development wdr-5.1b WDR5, WDR5B WD40 repeat-containing proteins, regulate H3K4 methylation levels hcf-1b HCF-1 Regulates cell division and mitotic histone modiﬁcation set-2b SETD1A, SETD1B Histone H3K4 methyltransferase, germline development, postembryonic development Polycomb mes-3b — Required maternally for normal germline development and for anteroposterior patterning mes-6b EED Required maternally for normal germline development and for anteroposterior patterning mes-2b EZH2 isoform b Required maternally for normal germline development and for anteroposterior patterning KAT8/MOF-like sumv-1c INO80D Encodes a protein with similarity to the KAT8 NLS3 nonenzymatic subunit of the mammalian KAT8/MOF histone acetyltransferase complex sumv-2c — Encodes a protein with similarity to the KAT8 NLS3 nonenzymatic subunit of the mammalian KAT8/MOF histone acetyltransferase complex Other C06A5.3c PSIP1, HDGFL1, HDGF Involved in reproduction, PWWP domain-containing protein mes-4a ASH1L, EZH1, EZH2 SET domain-containing protein, required maternally for normal germline development set-26c KMT2E, SETD5 PHD-zinc ﬁnger and a SET domain lex-1c ATAD2, ATAD2B Contains an ATPase domain and a bromodomain, positively regulate expression of repetitive sequences lsl-1b ZBTB20, ZBTB45, ZBTB7C, ZFP57, ZNF296 LSY-2-like math-33c USP-7 Encodes a protein with a meprin-associated Traf homology (MATH) domain T19B4.5c — Unknown gtbp-1b G3BP1, G3BP2 Predicted to have nucleotide-binding activity and nucleic acid-binding activity M03C11.3c — Unknown Chromatin factors are shown in bold SWR1 Swi2/Snf2-related ATPase 1; MYST Moz Ybf2/Sas3 Sas2 Tip60; KAT lysine (K) acetyltransferase; NLS nucl n Table 1 let-418 suppressor identities ATPase component of a nucleosome remodeling factor (NURF)-like complex Acts with isw-1 in vulval development Gene expression regulator HIT-type zinc ﬁnger protein Actin-related protein Regulate gene expression with SWR1 in pharynx Transcription factor Chromodomain-containing protein, promotes cell-proliferation Unknown MYST family histone acetyltransferase CFP1 (CpG-binding protein, CXXC Finger Protein 1) homolog Hermaphrodite dosage compensation and normal male development WD40 repeat-containing proteins, regulate H3K4 methylation levels Regulates cell division and mitotic histone modiﬁcation Histone H3K4 methyltransferase, germline development, postembryonic development Required maternally for normal germline development and for anteroposterior patterning Required maternally for normal germline development and for anteroposterior patterning Required maternally for normal germline development and for anteroposterior patterning Encodes a protein with similarity to the KAT8 NLS3 nonenzymatic subunit of the mammalian KAT8/MOF histone acetyltransferase complex Encodes a protein with similarity to the KAT8 NLS3 nonenzymatic subunit of the mammalian KAT8/MOF histone acetyltransferase complex Involved in reproduction, PWWP domain-containing protein SET domain-containing protein, required maternally for normal germline development PHD-zinc ﬁnger and a SET domain Contains an ATPase domain and a bromodomain, positively regulate expression of repetitive sequences LSY-2-like Encodes a protein with a meprin-associated Traf homology (MATH) domain Unknown P di d h l id bi di i i d l i id bi di i i Encodes a protein with a meprin-associated Traf homology (MATH) domain Unknown Predicted to have nucleotide-binding activity and nucleic acid-binding activity Unknown Chromatin factors are shown in bold. A genome-wide RNAi screen for suppressors of the let- 418 L1 arrest phenotype mostly identiﬁed chromatin factors SWR1, Swi2/Snf2-related ATPase 1; MYST, Moz, Ybf2/Sas3, Sas2, Tip60; KAT, lysine (K) acetyltransferase; NLS, nuclear localisation signal; MOF, males absent on the ﬁrst; PHD, plant homeodomain; SET, Su(var)3-9, Enhancer-of-zeste and Trithorax; LSY, laterally symmetric; MATH, Meprin and TRAF-Homology. aStrong suppression, larvae reach L4/adulthood. bWeak suppression, larvae bypass L1 arrest. cMiddle suppression, larvae reach L2/L3 stage. 2010). MES-4 regulates active chromatin and might be required for the transcription of genes that are induced in the absence of let-418. embryonic polarity (McCloskey and Kemphues 2012). Absence of math-33 activity could destabilize proteins responsible for the let-418 phenotype. lsl-1 encodes a zinc ﬁnger transcription factor whose function is unknown, and gtbp-1 was shown to be associated with stress granules in human cells (Jedrusik-Bode et al. 2013). M03C11.3 and T19B4.5 encode gene products whose functions are not known yet. lex-1 encodes a bromodomain-containing ATPase, which was shown to promote gene expression in the context of repetitive se- quences (Tseng et al. 2007). LEX-1 might also be involved in the expression of genes that are induced in the absence of let-418. C06A5.3 encodes a protein with a PWWP domain that interacts with methylated histone H4 at lysine 20. This speciﬁc histone mod- iﬁcation is known to recruit cell cycle checkpoint proteins (Wang et al. 2009). In the absence of let-418 activity, the C06A5.3 protein might interact abnormally with checkpoint proteins that could block the blast cell cycle. Insummary, we isolated mainly chromatin factors in oursuppressor screen, suggesting that there is a need to modify the expression of a large number of target genes to suppress let-418-associated defects. Suppression of let-418-associated somatic P granules While P granules appear only in the PGCs Z2/Z3 of wild-type L1 animals, let-418 mutants also show ectopic P granule components around somatic nuclei (Unhavaithaya et al. 2002). Using the reporter construct pie-1p::gfp::pgl-1 to monitor the presence of P granules, let- 418 mutants show an average of 12 P granule-positive somatic cells The remaining ﬁve genes are not known to be directly associated with chromatin function. math-33 participates in protein metab- olism. It was ﬁrst identiﬁed as a cde (cosuppression defective) gene (Robert et al. 2005). More recently, it was shown to be part of the deubiquitylation machinery required for the establishment of Volume 7 February 2017 \| Chromatin Factors in C. elegans \| 347 suppressor genes encode chromatin-associated proteins, we tested if Figure 2 Suppression of ectopic P granule expression. (A) Perinuclear P granules are observed in somatic cells of let-418 and set-26(RNAi) worms, but are missing in animals that are fed with mes-4 or htz-1 RNAi. Presence of the P granule was revealed by the reporter construct pgl-1:: gfp. Bar, 20 mm. (B) pgl-1 mRNA level is decreased in htz-1;let-418 but not in set-26;let-418 animals. mRNA level of pgl-1 was measured by qRT-PCR and represented as fold induction of mRNA expression vs. wt. Total mRNA was isolated from wt and let-418 L1 animals treated with the indicated suppressor RNAi. indicates a P value # 0.001 mRNA, messenger RNA; qRT-PCR, quantitative real-time polymerase chain re- action; RNAi, RNA interference; wt, wild-type. (Figure S3). We tested if our suppressor RNAis were able to reduce the ectopic expression of the reporter construct in a let-418 background. Of the 29 suppressors, 28 signiﬁcantly decreased the number of cells show- ing ectopic P granules in let-418 mutants (Figure 2A and Figure S3). g p g ( g g ) Interestingly, let-418 animals fed on set-26 RNAi grow to the L2/L3 stage, but a large number of P granule-containing cells are still detected (Figure 2A). In contrast, let-418 animals treated with htz-1 or mes-4 RNAi develop at least to the same stage, if not further, but show no ectopic P granules (Figure 2A). To further analyze the expression of genes encoding P granule components in let-418;set-26(RNAi) animals, we measured the mRNA level of pgl-1. Consistent with the pie-1p::gfp:: pgl-1 reporter analysis, the pgl-1 expression level was elevated in let-418 L1 larvae, but we observed no reduction in let-418;set-26 animals (Figure 2B). Suppression of the let-418-associated mitotic arrest of the blast cells To approach this question, we selected a subset of germline genes that were found in our list of upregulated genes and measured their transcriptional activity in let-418, nurf-1, htz-1, let- 418;nurf-1, and let-418;htz-1 backgrounds (Passannante et al. 2010). As expected, deps-1, pgl-1 (encoding P granule components), and pie-1 (encoding the germline determinant PIE-1) mRNA levels were suppressor genes encode chromatin-associated proteins, we tested if they could have an effect on the transcriptional activity of some let- 418 target genes. To approach this question, we selected a subset of germline genes that were found in our list of upregulated genes and measured their transcriptional activity in let-418, nurf-1, htz-1, let- 418;nurf-1, and let-418;htz-1 backgrounds (Passannante et al. 2010). As expected, deps-1, pgl-1 (encoding P granule components), and pie-1 (encoding the germline determinant PIE-1) mRNA levels were Suppression of let-418-associated somatic P granules In contrast to most of the suppressors that restored development and restricted P granules to the germline, lack of set-26 activity only sup- pressed developmental arrest. This result shows that the presence of ectopic P granules does not prevent the onset of postembryonic develop- ment, and suggests that the regulation of development and repression of germline gene expression in somatic cells could be controlled by distinct functions of LET-418. Another interpretation of this result could be that the absence of set-26 activity sets up a sensitized background for ectopic expression of P granule components. Consistent with this hypothesis is the slight pgl-1 overexpression observed in set-26(RNAi) (Figure 2B). Suppression of the let-418-associated mitotic arrest of the blast cells We followed the division pattern of the M cell using hlh-8::gfp, which is expressed only in undifferentiated cells of the M cell lineage (Harfe et al. 1998). In L1 arrested let-418(n3536) animals, the M cell did not divide (Figure 1A, k). When let-418 mutant animals were treated with the suppressor RNAis, we detected divisions of the M blast cell in all cases (Figure S4), although the level of M cell divisions did not fully corre- spond to the observed developmental stage of the worms. Thus, our data suggest that M cell division and larval growth are uncoupled in let- 418 animals. A delay in M cell division compared to the growth rate was previously reported in a daf-2 mutant lacking insulin receptor activity (Chen and Baugh 2014). Figure 2 Suppression of ectopic P granule expression. (A) Perinuclear P granules are observed in somatic cells of let-418 and set-26(RNAi) worms, but are missing in animals that are fed with mes-4 or htz-1 RNAi. Presence of the P granule was revealed by the reporter construct pgl-1:: gfp. Bar, 20 mm. (B) pgl-1 mRNA level is decreased in htz-1;let-418 but not in set-26;let-418 animals. mRNA level of pgl-1 was measured by qRT-PCR and represented as fold induction of mRNA expression vs. wt. Total mRNA was isolated from wt and let-418 L1 animals treated with the indicated suppressor RNAi. indicates a P value # 0.001 mRNA, messenger RNA; qRT-PCR, quantitative real-time polymerase chain re- action; RNAi, RNA interference; wt, wild-type. Freshly hatched larvae possess 10 pairs of seam cells. We monitored the developmental fate of the seam cell lineage using animals carrying a scm::gfp transgene (Terns et al. 1997). let-418 L1 larvae exhibited no seam cell division (Figure 1A, g and Figure S5); However, inactivation of all 29 suppressor genes by RNAi resulted in seam cell divisions to different extents (Figure S5). Similar to our previous observations re- garding the M cell divisions, the number of seam cells was lower than expected from the apparent developmental stage of the animal (Figure S5). Altogether, our results show that RNAi of the suppressors can restore blast cell division in a let-418 mutant background to some extent, although none of the worms could reach adulthood. suppressor genes encode chromatin-associated proteins, we tested if they could have an effect on the transcriptional activity of some let- 418 target genes. DISCUSSION In the absence of let-418 maternal gene activity, let-418 mutants do not initiate postembryonic development. The PGCs and blast cells do not divide. Overall, arrested let-418 worms resemble L1 diapaused larvae, except that they do not survive and recover as well as starved larvae that have been returned to food. By performing a genome-wide RNAi sup- pressor screen, we could show that let-418 developmental arrest is de- pendent mainly on genes encoding chromatin factors and on a few other genes exhibiting various functions. All the suppressors identiﬁed in this study, except the histone methyltransferase-encoding gene set- 26, are able to suppress not only the developmental defects but also the somatic expression of germline-speciﬁc genes associated with the lack of let-418 activity. In contrast, set-26 suppresses developmental defects but not the somatic expression of germline genes. Figure 3 set-26 suppression of M cell mitotic arrest. (A) Inactivation of set-26 in let-418 mutant background allows the M cell to divide. How- ever, set-26;let-418 animals show reduced numbers of muscle cell precursors compared to wt. Numbers in brackets correspond to the number of germ cells indicative of the developmental stage. (B) M cell division pattern is revealed by the reporter construct hlh-8::gfp. De- veloping L1 shows BWM cell precursors (ﬁlled arrowheads) and L4 larva show SM precursors (open arrowheads). SM precursor cells are mislocalized in set-26;let-418 animals. BWM, body wall muscle; CC, coelomocytes; eL1, early L1 with four germ cells (Z1–4); SM, sex myoblast; wt, wild-type. increased in let-418 larvae (Figure 4). Their transcriptional induc- tions were signiﬁcantly lowered when htz-1 or nurf-1 were depleted by RNAi (Figure 4A), indicating that nurf-1 and htz-1 are required for the upregulation of these three genes in let-418 mutants. Fur- thermore, we took advantage of available ChIP-seq data for LET- 418, NURF-1, and HTZ-1 that were generated by the modENCODE consortium using L3 larvae as starting material (Gerstein et al. 2010). We compared their list of target genes to identify common direct targets. The analysis showed that LET-418 and HTZ-1, as well as LET-418 and NURF-1, were sharing a signiﬁcant number of common target genes in L3 larvae (P , 0.001) (Figure 5). GO anal- ysis of the common targets revealed that they are enriched in genes with predicted function in embryonic and postembryonic develop- ment, as well as reproductive developmental processes (Huang et al. 2009a,b). Volume 7 February 2017 \| Chromatin Factors in C. elegans LET-418, NURF-1, and HTZ-1 are regulating common target genes Nevertheless, absence of daf-16 seems to prevent the occurrence of some other defects in the let-418 mutant larvae, as indicated by the loss of the M cell marker expression (Figure S1). This suggests that DAF-16 is involved in both types of developmental arrest, one due to starvation and the other due to an absence of let-418 activity. Therefore, we were interested to know if LET-418 and DAF-16 could share common reg- ulated genes. Indeed, a comparison between the transcription proﬁles of daf-16 and let-418 mutants revealed a signiﬁcant number of com- monly regulated targets that, following GO analysis, belong to gene categories involved in metabolic pathways (Figure S6, A and B) (Passannante et al. 2010; Kaplan et al. 2015). We selected three of these common genes, namely dct-3, W08A12.4, and fbxa-165 (Figure S7), and tested if they are also coregulated by a subset of our suppressors. We compared the transcriptional activities of these three genes in let- 418, nurf-1, htz-1, or set-26 animals with that of let-418;htz-1, let-418; nurf-1, or let-418;set-26 animals. The mRNA levels of dct-3, W08A12.4, and fbxa-165 were increased in let-418 mutants, and this induction was dependent on nurf-1, htz-1, and set-26 activity (Figure 4B). This indi- cates that LET-418, together with other chromatin-associated factors, is controlling the transcription of a subset of DAF-16 targets that may be involved in larval development. LET-418, NURF-1, and HTZ-1 are regulating common target genes In a previous study, we showed that the upregulated genes in let-418 L1 larvae were enriched in germline-speciﬁc genes, including those encod- ing P granule components (Passannante et al. 2010). Since many of our 348 \| P. Erdelyi et al. Figure 3 set-26 suppression of M cell mitotic arrest. (A) Inactivation of set-26 in let-418 mutant background allows the M cell to divide. How- ever, set-26;let-418 animals show reduced numbers of muscle cell precursors compared to wt. Numbers in brackets correspond to the number of germ cells indicative of the developmental stage. (B) M cell division pattern is revealed by the reporter construct hlh-8::gfp. De- veloping L1 shows BWM cell precursors (ﬁlled arrowheads) and L4 larva show SM precursors (open arrowheads). SM precursor cells are mislocalized in set-26;let-418 animals. BWM, body wall muscle; CC, coelomocytes; eL1, early L1 with four germ cells (Z1–4); SM, sex myoblast; wt, wild-type. Nevertheless, absence of daf-16 seems to prevent the occurrence of some other defects in the let-418 mutant larvae, as indicated by the loss of the M cell marker expression (Figure S1). This suggests that DAF-16 is involved in both types of developmental arrest, one due to starvation and the other due to an absence of let-418 activity. Therefore, we were interested to know if LET-418 and DAF-16 could share common reg- ulated genes. Indeed, a comparison between the transcription proﬁles of daf-16 and let-418 mutants revealed a signiﬁcant number of com- monly regulated targets that, following GO analysis, belong to gene categories involved in metabolic pathways (Figure S6, A and B) (Passannante et al. 2010; Kaplan et al. 2015). We selected three of these common genes, namely dct-3, W08A12.4, and fbxa-165 (Figure S7), and tested if they are also coregulated by a subset of our suppressors. We compared the transcriptional activities of these three genes in let- 418, nurf-1, htz-1, or set-26 animals with that of let-418;htz-1, let-418; nurf-1, or let-418;set-26 animals. The mRNA levels of dct-3, W08A12.4, and fbxa-165 were increased in let-418 mutants, and this induction was dependent on nurf-1, htz-1, and set-26 activity (Figure 4B). This indi- cates that LET-418, together with other chromatin-associated factors, is controlling the transcription of a subset of DAF-16 targets that may be involved in larval development. DISCUSSION elegans 349 in ES cells, LET-418 could be responsible for the restriction of gene ession and to ensure responsiveness to developmental cues such as proteins are required for the proper localization of MES-4 exclu to the autosomes (Fong et al. 2002; Bender et al. 2004), and deple re 4 LET-418, NURF-1, and HTZ-1 regulate common target genes. (A) Upregulation of the germline gene expression in let-418 depe F-1 and HTZ-1. mRNA levels of deps-1, pie-1, and pgl-1 were measured by qRT-PCR and represented as fold induction of mRNA expr wt. Total mRNA was isolated from wt and let-418 L1 animals treated with the indicated suppressor RNAi. ama-1 was used to normal egulation of DAF-16 target expression in let-418 depends on NURF-1 and HTZ-1. mRNA levels of dct-3, W08A12.4, and fbxa-165 sured by qRT-PCR and represented as fold induction of mRNA expression vs. wt. Total mRNA was isolated from wt and let-418 L1 a ted with the indicated suppressor RNAi. ama-1 was used to normalize. mRNA, messenger RNA; qRT-PCR, quantitative real-time polym n reaction; RNAi, RNA interference; wt, wild-type. Figure 4 LET-418, NURF-1, and HTZ-1 regulate common target genes. (A) Upregulation of the germline gene expression in let-418 depends on NURF-1 and HTZ-1. mRNA levels of deps-1, pie-1, and pgl-1 were measured by qRT-PCR and represented as fold induction of mRNA expression vs. wt. Total mRNA was isolated from wt and let-418 L1 animals treated with the indicated suppressor RNAi. ama-1 was used to normalize. (B) Upregulation of DAF-16 target expression in let-418 depends on NURF-1 and HTZ-1. mRNA levels of dct-3, W08A12.4, and fbxa-165 were measured by qRT-PCR and represented as fold induction of mRNA expression vs. wt. Total mRNA was isolated from wt and let-418 L1 animals treated with the indicated suppressor RNAi. ama-1 was used to normalize. mRNA, messenger RNA; qRT-PCR, quantitative real-time polymerase chain reaction; RNAi, RNA interference; wt, wild-type. proteins are required for the proper localization of MES-4 exclusively to the autosomes (Fong et al. 2002; Bender et al. 2004), and depletion of the polycomb proteins MES-2/3/6 could lead to a chromosome-wide relocalization of MES-4, resulting in decreased MES-4 activity at LET- 418 target genes (Boyer et al. 2006; Lee et al. 2006; Yuzyuk et al. 2009). DISCUSSION Unfortunately, the germline genes deps-1, pgl-1, and pie-1 that we tested for their transcriptional activity in let-418, nurf-1, and htz-1 backgrounds do not ﬁgure in this list. Either they are not direct targets or they are not bound by LET-418, NURF-1, and HTZ-1 at the L3 stage, or both, suggesting a dynamic binding of targets during development. These results indicate that the chromatin proteins LET-418, HTZ-1, and NURF-1 are likely recruited to common tar- get genes, consistent with the idea that these chromatin factors are together regulating important developmental processes. In this paper, we describe that a failure of let-418 worms to initiate postembryonic development was mainly dependent on genes encoding chromatin factors that are members of multi-protein complexes known to activate transcription. Recently, we found that LET-418, together with the histone demethylase SPR-5, blocks the activity of the COMPASS complex to ensure germ cell fate maintenance (Käser- Pébernard et al. 2014). A similar mechanism could function at the transition between embryonic and postembryonic development in the absence of food. In such a model (Figure 6), LET-418 could control the access of activator complexes, such as COMPASS, but also ISW/NURF, SWR1, KAT8/MOF, and NuA4, to their target genes, thereby maintaining the quiescence and pluripotency of blast cells. Consistent with this hypothesis, we observed that upregulation of some known LET-418-regulated genes, such as the germline genes pie-1, deps-1, and pgl-1, and a subset of DAF-16 targets, is dependent on NURF-1 or HTZ-1 (Figure 4) (Passannante et al. 2010). Altogether, our data suggest that LET-418 could organize the chromatin structure at target genes, the regulation of which is crucial for developmental transitions. In mammalian cells, Mi2 together with other NuRD components restricts the activity of genes to allow the transition of ES cells from the pluripotent state to lineage commitment (Reynolds et al. 2012, 2013). It is proposed that this regulation is necessary to keep cells responsive to developmental cues. In worms, postembryonic development consists essentially of PGC and blast cell division and differentiation. Similar to Starved L1 larvae and arrested let-418 mutants look superﬁcially similar. However, blast cell quiescence in starved larvae, but not in let-418 animals, depends on DAF-16 (Baugh and Sternberg 2006). Volume 7 February 2017 \| Chromatin Factors in C. 350 \| P. Erdelyi et al. DISCUSSION An alternative way to interpret the suppressor activity of mes-2/3/6 is that LET-418 plays a role in the recruitment or localization of Mi2 in ES cells, LET-418 could be responsible for the restriction of gene expression and to ensure responsiveness to developmental cues such as the presence of food. Among our suppressors, we also found genes encoding members of the Polycomb complex, which, like LET-418, is known to repress tran- scription during development. One possible way to interpret this result is that the suppression effect may occur through MES-4; Polycomb 350 \| P. Erdelyi et al. Figure 5 LET-418 and HTZ-1, and LET-418 and NURF-1, share a signiﬁcant number of common direct targets (P-value , 0.001). P-values were determined by Fisher’s exact test. Gene ontology categories that are highly enriched for LET-418/HTZ-1 (A) and LET-418/NURF-1 (B) common targets are shown. The percentage of enrichment is indicated by the bars. Figure 5 LET-418 and HTZ-1, and LET-418 and NURF-1, share a signiﬁcant number of common direct targets (P-value , 0.001). P-values were determined by Fisher’s exact test. Gene ontology categories that are highly enriched for LET-418/HTZ-1 (A) and LET-418/NURF-1 (B) common targets are shown. The percentage of enrichment is indicated by the bars. Figure 5 LET-418 and HTZ-1, and LET-418 and NURF-1, share a signiﬁcant number of common direct targets (P-value , 0.001). P-values were determined by Fisher’s exact test. Gene ontology categories that are highly enriched for LET-418/HTZ-1 (A) and LET-418/NURF-1 (B) common targets are shown. The percentage of enrichment is indicated by the bars. Polycomb proteins (Figure 6). This is supported by our recent ﬁnd- ings that LET-418 interacts with MES-2 and can recruit it to the chromatin (Käser-Pébernard et al. 2016). Absence of let-418 activity could lead to mistargeting of MES-2 and the repression of genes that are not normally bound by the Polycomb complex, thereby resulting in a developmental arrest of let-418 animals. Depletion of PRC2 components may remove this repression and allow development in a let-418 background. This model is also supported by the ﬁnding that, in ES cells, Polycomb repressive complex 2 is recruited by NuRD to speciﬁc target genes to direct their repression (Reynolds et al. 2011). set-26 was the only suppressor gene that did not suppress the so- matic expression of germline genes in let-418 larvae. set-26 is expressed in most, if not all, somatic cells (Ni et al. DISCUSSION 2011) and the set-26 gene Figure 6 Model for the action of LET-418 and the suppressors. In the presence of food, larval devel- opment is initiated and LET-418, together with a network of other chromatin factors, is responsible for modulating the transcription of germline genes, a subset of DAF-16 targets, and some other prode- velopmental arrest genes. This network of chromatin factors includes activator proteins that might com- pete for regulation at the promoter of target genes and repressors whose localization could be de- termined by LET-418. MES-2 was shown recently to interact with LET-418 (Käser-Pébernard et al. 2016). Volume 7 February 2017 \| Chromatin Factors in C. elegans \| 351 Figure 6 Model for the action of LET-418 and the suppressors. In the presence of food, larval devel- opment is initiated and LET-418, together with a network of other chromatin factors, is responsible for modulating the transcription of germline genes, a subset of DAF-16 targets, and some other prode- velopmental arrest genes. This network of chromatin factors includes activator proteins that might com- pete for regulation at the promoter of target genes and repressors whose localization could be de- termined by LET-418. MES-2 was shown recently to interact with LET-418 (Käser-Pébernard et al. 2016). Volume 7 February 2017 \| Chromatin Factors in C. elegans \| 351 351 Eshed, Y., S. F. Baum, and J. L. Bowman, 1999 Distinct mechanisms pro- mote polarity establishment in carpels of Arabidopsis. Cell 99: 199–209. product methylates lysine 9 of histone H3 in vitro (Greer et al. 2014), which is usually associated with chromatin compaction and gene re- pression (Li et al. 2007; Bannister and Kouzarides 2011). One possible interpretation is that perturbation of the H3K9 methylation level and localization of this modiﬁcation could affect chromatin structure, thereby leading to the reactivation of crucial prodevelopmental genes in let-418;set-26 animals. However, such a putative perturbation of the H3K9 methylation pattern would not repress the germline transcrip- tion program in somatic cells. Fong, Y., L. Bender, W. Wang, and S. Strome, 2002 Regulation of the different chromatin states of autosomes and X chromosomes in the germ line of C. elegans. Science 296: 2235–2238. line of C. elegans. Science 296: 2235–2238. Fukuyama, M., A. E. Rougvie, and J. H. Rothman, 2006 C. elegans DAF-18/ PTEN mediates nutrient-dependent arrest of cell cycle and growth in the germline. Curr. Biol. 16: 773–779. Furuhashi, H., T. Takasaki, A. Rechtsteiner, T. LITERATURE CITED Kaplan, R. E., Y. Chen, B. T. Moore, J. M. Jordan, C. S. Maxwell et al., 2015 dbl-1/TGF-beta and daf-12/NHR signaling mediate cell- nonautonomous effects of daf-16/FOXO on starvation-induced developmental arrest. PLoS Genet. 11: e1005731. Andersen, E. C., X. Lu, and H. R. Horvitz, 2006 C. elegans ISWI and NURF301 antagonize an Rb-like pathway in the determination of multiple cell fates. Development 133: 2695–2704. Bannister, A. J., and T. Kouzarides, 2011 Regulation of chromatin by histone modiﬁcations. Cell Res. 21: 381–395. Käser-Pébernard, S., F. Müller, and C. Wicky, 2014 LET-418/Mi2 and SPR- 5/LSD1 cooperatively prevent somatic reprogramming of C. elegans l ll ll Käser-Pébernard, S., F. Müller, and C. Wicky, 2014 LET-418/Mi2 and SPR- 5/LSD1 cooperatively prevent somatic reprogramming of C. elegans germline stem cells. Stem Cell Rep. 2: 547–559. Baugh, L. R., 2013 To grow or not to grow: nutritional control of devel- opment during Caenorhabditis elegans L1 arrest. Genetics 194: 539–555 Käser-Pébernard, S., C. Pfefferli, C. Aschinger, and C. Wicky, 2016 Fine-tuning of chromatin composition and Polycomb recruitment by two Mi2 homologues during C. elegans early embryonic development. Epigenetics Chromatin 9: 39. Baugh, L. R., and P. W. Sternberg, 2006 DAF-16/FOXO regulates tran- scription of cki-1/Cip/Kip and repression of lin-4 during C. elegans L1 arrest. Curr. Biol. 16: 780–785. Bender, L. B., R. Cao, Y. Zhang, and S. Strome, 2004 The MES-2/MES-3/ MES-6 complex and regulation of histone H3 methylation in C. elegans. Curr. Biol. 14: 1639–1643. Kashiwagi, M., B. A. Morgan, and K. Georgopoulos, 2007 The chromatin remodeler Mi-2b is required for establishment of the basal epidermis and normal differentiation of its progeny. Development 134: 1571–1582. Boyer, L. A., K. Plath, J. Zeitlinger, T. Brambrink, L. A. Medeiros et al., 2006 Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature 441: 349–353. Kehle, J., D. Beuchle, S. Treuheit, B. Christen, J. A. Kennison et al., 1998 dMi-2, a hunchback-interacting protein that functions in Polycomb repression. Science 282: 1897–1900. Brenner, S., 1974 The genetics of Caenorhabditis elegans. Genetics 77: 71– 94. Lee, T. I., R. G. Jenner, L. A. Boyer, M. G. Guenther, S. S. Levine et al., 2006 Control of developmental regulators by Polycomb in human embryonic stem cells. Cell 125: 301–313. Celniker, S. E., L. A. L. Dillon, M. B. Gerstein, K. C. Gunsalus, S. Henikoff et al., 2009 Unlocking the secrets of the genome. Nature 459: 927–930. Chen, Y., and L. R. DISCUSSION Li, H. Kimura et al., 2010 Trans-generational epigenetic regulation of C. elegans primordial germ cells. Epigenetics Chromatin 3: 15. Only ﬁve suppressors were identiﬁed that did not belong to chro- matin factors. As part of the deubiquitylation machinery, MATH-33 might alter the metabolism of protein turnover to an extent that is sufﬁcient to partially suppress the developmental defects of let-418 (McCloskey and Kemphues 2012). Genes involved in protein metabo- lism have been found to suppress defects associated with mutations in another transcriptional repressor, lin-35 (Polley and Fay 2012). This suggests that altering the amounts of some key proteins in the process might be sufﬁcient to trigger the onset of postembryonic development. Only ﬁve suppressors were identiﬁed that did not belong to chro- matin factors. As part of the deubiquitylation machinery, MATH-33 might alter the metabolism of protein turnover to an extent that is sufﬁcient to partially suppress the developmental defects of let-418 (McCloskey and Kemphues 2012). Genes involved in protein metabo- lism have been found to suppress defects associated with mutations in another transcriptional repressor, lin-35 (Polley and Fay 2012). This suggests that altering the amounts of some key proteins in the process might be sufﬁcient to trigger the onset of postembryonic development. Altogether,oursuppressorscreenuncoveredanetworkofchromatin factors that regulate the transition from embryonic to postembryonic development.Thistransitioninvolvescrucialdecisions,suchasexitfrom thecellcycleoftheblastcells,cellproliferation,anddifferentiation.Strict control of these processes is critical during development, but also to prevent aberrant proliferation and differentiationduringtumorigenesis. Analyzing how these various chromatin factors, together with LET-418, are recruited to key developmental genes will contribute to our under- standing of the functional importance of chromatin in the response to developmental cues. Gerstein, M. B., Z. J. Lu, E. L. Van Nostrand, C. Cheng, B. I. Arshinoff et al., 2010 Integrative analysis of the Caenorhabditis elegans genome by the modENCODE project. Science 330: 1775–1787. Greer, E. L., S. E. Beese-Sims, E. Brookes, R. Spadafora, Y. Zhu et al., 2014 A histone methylation network regulates transgenerational epigenetic memory in C. elegans. Cell Rep. 7: 113–126. Harfe, B. D., A. Vaz Gomes, C. Kenyon, J. Liu, M. Krause et al., 1998 Analysis of a Caenorhabditis elegans Twist homolog identiﬁes conserved and divergent aspects of mesodermal patterning. Genes Dev. 12: 2623–2635. Altogether,oursuppressorscreenuncoveredanetworkofchromatin factors that regulate the transition from embryonic to postembryonic development.Thistransitioninvolvescrucialdecisions,suchasexitfrom thecellcycleoftheblastcells,cellproliferation,anddifferentiation.Strict control of these processes is critical during development, but also to prevent aberrant proliferation and differentiationduringtumorigenesis. DISCUSSION Analyzing how these various chromatin factors, together with LET-418, are recruited to key developmental genes will contribute to our under- standing of the functional importance of chromatin in the response to developmental cues. Huang, D. W., B. T. Sherman, and R. A. Lempicki, 2009a Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 37: 1–13. Huang, D. W., B. T. Sherman, and R. A. Lempicki, 2009b Systematic and integrative analysis of large gene lists using DAVID bioinformatics re- sources. Nat. Protoc. 4: 44–57. Jedrusik-Bode, M., M. Studencka, C. Smolka, T. Baumann, H. Schmidt et al., 2013 The sirtuin SIRT6 regulates stress granule formation in C. elegans and mammals. J. Cell Sci. 126: 5166–5177. Kaji, K., I. M. Caballero, R. Macleod, J. Nichols, V. A. Wilson et al., 2006 The NuRD component Mbd3 is required for pluripotency of embryonic stem cells. Nat. Cell Biol. 8: 285–292. ACKNOWLEDGMENTS Kaji, K., J. Nichols, and B. Hendrich, 2007 Mbd3, a component of the NuRD co-repressor complex, is required for development of pluripotent cells. Development 134: 1123–1132. We thank L. Bulliard for technical assistance and F. Müller for critically reading the manuscript. Some strains were provided by the Mitani Lab at the Tokyo Women’s Medical University School of Medicine (Japan), and by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health Ofﬁce of Re- search Infrastructure Programs (P40 OD010440). Kamath, R. S., and J. Ahringer, 2003 Genome-wide RNAi screening in Caenorhabditis elegans. Methods 30: 313–321. Kamath, R. S., A. G. Fraser, Y. Dong, G. Poulin, R. Durbin et al., 2003 Systematic functional analysis of the Caenorhabditis elegans genome using RNAi. Nature 421: 231–237. LITERATURE CITED Baugh, 2014 Ins-4 and daf-28 function redundantly to regulate C. elegans L1 arrest. Dev. Biol. 394: 314–326. Li, B., M. Carey, and J. L. Workman, 2007 The role of chromatin during transcription. Cell 128: 707–719. g g De Vaux, V., C. Pfefferli, M. Passannante, K. Belhaj, A. von Essen et al., Li, T., and W. G. Kelly, 2011 A role for Set1/MLL-related components in epigenetic regulation of the Caenorhabditis elegans germ line. PLoS Genet. 7: e1001349. De Vaux, V., C. Pfefferli, M. Passannante, K. Belhaj, A. von Essen et al., 2013 The Caenorhabditis elegans LET-418/Mi2 plays a conserved role in lifespan regulation. Aging Cell 12: 1012–1020. 352 \| P. Erdelyi et al. Robert, V. J., T. Sijen, J. van Wolfswinkel, and R. H. Plasterk, 2005 Chromatin and RNAi factors protect the C. elegans germline against repetitive se- quences. Genes Dev. 19: 782–787. Lu, P. Y., N. Levesque, and M. S. Kobor, 2009 NuA4 and SWR1-C: two chromatin-modifying complexes with overlapping functions and components. Biochem. Cell Biol. 87: 799–815. p McCloskey, R. J., and K. J. Kemphues, 2012 Deubiquitylation machinery is required for embryonic polarity in Caenorhabditis elegans. PLoS Genet. 8: e1003092. q Shilatifard, A., 2012 The COMPASS family of histone H3K4 methylases: mechanisms of regulation in development and disease pathogenesis. Annu. Rev. Biochem. 81: 65–95. McDonel, P., I. Costello, and B. Hendrich, 2009 Keeping things quiet: roles of NuRD and Sin3 co-repressor complexes during mammalian development. Int. J. Biochem. Cell Biol. 41: 108–116. Solari, F., and J. Ahringer, 2000 NURD-complex genes antagonise Ras- induced vulval development in Caenorhabditis elegans. Curr. Biol. 10: 223–226. Sulston, J. E., and H. R. Horvitz, 1977 Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56: 110–156. Mizuguchi, G., X. Shen, J. W. Landry, W. Wei-Hua, S. Sen et al., 2004 ATP- driven exchange of histone H2AZ variant catalyzed by SWR1 chromatin remodeling complex. Science 303: 343–348. Terns, R. M., P. Kroll-Conner, J. Zhu, S. Chung, and J. H. Rothman, 1997 A deﬁciency screen for zygotic loci required for establishment and patterning of the epidermis in Caenorhabditis elegans. Genetics 146: 185–206. Ni, Z., A. Ebata, E. Alipanahiramandi, and S. S. Lee, 2011 Two SET domain containing genes link epigenetic changes and aging in Caenorhabditis elegans. Aging Cell 11: 315–325. Ogas, J., J. C. Cheng, Z. R. Sung, and C. Somerville, 1997 Cellular differ- entiation regulated by gibberellin in the Arabidopsis thaliana pickle mutant. LITERATURE CITED Science 277: 91–94. Timmons, L., and A. Fire, 1998 Speciﬁc interference by ingested dsRNA. Nature 395: 854. Timmons, L., D. L. Court, and A. Fire, 2001 Ingestion of bacterially ex- pressed dsRNAs can produce speciﬁc and potent genetic interference in Caenorhabditis elegans. Gene 263: 103–112. Ogas, J., S. Kaufmann, J. Henderson, and C. Somerville, 1999 PICKLE is a CHD3 chromatin-remodeling factor that regulates the transition from embryonic to vegetative development in Arabidopsis. Proc. Natl. Acad. Sci. USA 96: 13839–13844. Tseng, R. J., K. R. Armstrong, X. Wang, and H. M. Chamberlin, 2007 The bromodomain protein LEX-1 acts with TAM-1 to modulate gene ex- pression in C. elegans. Mol. Genet. Genomics 278: 507–518. Passannante, M., C. O. Marti, C. Pfefferli, P. S. Moroni, S. Kaeser-Pebernard et al., 2010 Different Mi-2 complexes for various developmental func- tions in Caenorhabditis elegans. PLoS One 5: e13681. Unhavaithaya, Y., T. H. Shin, N. Miliaras, J. Lee, T. Oyama et al., 2002 MEP-1 and a homolog of the NURD complex component Mi-2 act together to maintain germline-soma distinctions in C. elegans. Cell 111: 991–1002. Polley, S. R., and D. S. Fay, 2012 A network of genes antagonistic to the LIN-35 retinoblastoma protein of Caenorhabditis elegans. Genetics 191: 1367–1380. von Zelewsky, T., F. Palladino, K. Brunschwig, H. Tobler, A. Hajnal et al., 2000 The C. elegans Mi-2 chromatin-remodelling proteins function in vulval cell fate determination. Development 127: 5277–5284. Rea, S., G. Xouri, and A. Akhtar, 2007 Males absent on the ﬁrst (MOF): from ﬂies to humans. Oncogene 26: 5385–5394. Rechtsteiner, A., S. Ercan, T. Takasaki, T. M. Phippen, T. A. Egelhofer et al., 2010 The histone H3K36 methyltransferase MES-4 acts epigenetically to transmit the memory of germline gene expression to progeny. PLoS Genet. 6: e1001091. Wang, Y., B. Reddy, J. Thompson, H. Wang, K. Noma et al., 2009 Regulation of Set9-mediated H4K20 methylation by a PWWP domain protein. Mol. Cell 33: 428–437. Yoshida, T., I. Hazan, J. Zhang, S. Y. Ng, T. Naito et al., 2008 The role of the chromatin remodeler Mi-2beta in hematopoietic stem cell self- renewal and multilineage differentiation. Genes Dev. 22: 1174–1189. Reynolds, N., M. Salmon-Divon, H. Dvinge, A. Hynes-Allen, G. Balasooriya et al., 2011 NuRD-mediated deacetylation of H3K27 facilitates re- cruitment of Polycomb Repressive Complex 2 to direct gene repression. EMBO J. 31: 593–605. Yucel, D., M. Hoe, E. Llamosas, S. Kant, C. Volume 7 February 2017 \| Chromatin Factors in C. elegans \| 353 Communicating editor: K. S. McKim LITERATURE CITED Jamieson et al., 2014 SUMV-1 antagonizes the activity of synthetic multivulva genes in Caenorhabditis elegans. Dev. Biol. 392: 266–282. Reynolds, N., P. Latos, A. Hynes-Allen, R. Loos, D. Leaford et al., 2012 NuRD suppresses pluripotency gene expression to promote tran- scriptional heterogeneity and lineage commitment. Cell Stem Cell 10: 583–594. Yuzyuk, T., T. H. Fakhouri, J. Kiefer, and S. E. Mango, 2009 The polycomb complex protein mes-2/E(z) promotes the transition from develop- mental plasticity to differentiation in C. elegans embryos. Dev. Cell 16: 699–710. Reynolds, N., A. O’Shaughnessy, and B. Hendrich, 2013 Transcriptional repressors: multifaceted regulators of gene expression. Development 140: 505–512. Communicating editor: K. S. McKim Communicating editor: K. S. McKim Volume 7 February 2017 \| Chromatin Factors in C. elegans \| 353
https://openalex.org/W2952145734	https://www.nature.com/articles/s41467-019-09361-0.pdf	English	null	The parasitic worm product ES-62 normalises the gut microbiota/bone marrow axis in inflammatory arthritis	bioRxiv (Cold Spring Harbor Laboratory)	2,018	cc-by	15,448	1 Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow G4 0RE, UK. 2 Institute of Infection, Immunity and Inﬂammation, University of Glasgow, Glasgow G12 8TA, UK. Correspondence and requests for materials should be addressed to M.M.H. (email: Margaret. Harnett@glasgow.ac.uk) or to W.H. (email: w.harnett@strath.ac.uk) ARTICLE NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications Results However, it is unclear whether tissue-resident or blood-borne parasitic worms can mediate these effects via modulation of the host microbiome and if so, which mechanisms they utilise. y That helminths can ameliorate chronic inﬂammatory disorders has often been attributed to their capacity to excrete or secrete molecules (ES) that exert immunoregulation2. Amongst the best characterised ES products is ES-62, a phosphorylcholine (PC)-containing glycoprotein secreted by the ﬁlarial nematode Acanthocheilonema viteae that we have shown to prevent initia- tion and progression of pathology in mouse models of certain allergic (asthma, contact dermatitis) and autoimmune (RA, SLE) inﬂammatory diseases1,2,15–20. Collectively, our studies have identiﬁed a unifying mechanism of action that allows effective protection irrespective of the inﬂammatory phenotype: thus, by subverting TLR4 signalling to downregulate aberrant MyD88- responses, ES-62 homeostatically resets the regulatory:effector immune cell balance, primarily to restore levels of IL-10+ regulatory B cells and suppress pathological IL-17-driven inﬂammation1,2,15–21. In both experimental models of RA and human disease, perturbation of the microbiota has been shown to disrupt the balance of pathogenic Th17 cells and the counter- regulatory Bregs and Tregs that act to homeostatically resolve inﬂammation9,10,22. Thus, our aim here was to investigate whe- ther the anti-inﬂammatory actions of ES-62 reﬂected an ability to impact on the microbiome. We now show that whilst joint disease in the collagen-induced arthritis (CIA) mouse model An overview of the microbiota at the phylum level shows substantial changes between healthy Naive mice and those with arthritis (PBS; Fig. 1a). Firmicutes and Bacteroidetes are the predominant phyla in all groups, but CIA mice in particular, exhibit outgrowths of Firmicutes and Proteobacteria in the ileum, whereas they demonstrate decreased levels of Firmicutes with a compensatory outgrowth of Bacteroidetes, in the colon (Fig. 1a). ES-62 essentially helps maintain the healthy microbiome diversity of Naive mice, which was reduced in CIA mice (Fig. 1a). Deeper analysis illustrates the differential diversity signatures of healthy and arthritic mice, as well as the impact of ES-62 on a global scale (Fig. 1b; Supplementary Table 1). As a consequence of our pooling strategy, we have refrained from discussion of species- level changes as large-scale population studies would be required to address this with complete conﬁdence. Results ES-62 normalises the gut microbiome in protecting against CIA. ES-62 ameliorates CIA in terms of articular score and frailty, maintaining grip strength at a similar level to that of healthy, Naive (not subjected to CIA) DBA/1 mice (Supplemen- tary Figure 1). Commensal bacteria have increasingly been pro- posed to contribute to RA pathogenesis9,10 and decline in grip strength during ageing has been associated with changes in the gut microbiome23. In addition to being an indicator of frailty, grip strength is a predictor of a wide range of adverse health out- comes24, e.g., cardiovascular disease, which RA patients are at increased risk of developing25 and that are impacted by the microbiome9,26. Thus, to address whether ES-62-mediated pro- tection reﬂects modulation of the gut microbiota, a shotgun metagenomic approach was used to proﬁle bacteria populations present in the intestines of CIA mice. Initiation of RA (and CIA) pathogenesis is associated with disruption of the balance of effector:regulatory immune cells and so we characterised the bacterial changes pertaining during established arthritis in the ileum and colon; intestinal sites where the microbiome and the metabolic microenvironment play key roles in shaping Th17 and regulatory immune responses22,27,28. As there is well- documented variation in the microbiome amongst individuals due to a range of environmental factors, we adopted the strategy of pooling samples from mice with representative disease scores to minimise variation due to a factor we could control for, namely disease severity: speciﬁcally, we focused on those samples asso- ciated with a well-established severe level of disease in CIA mice (articular score: 7.00 ± 0.91) and clear protection against arthritis in ES-62-treated CIA mice (articular score: 0.50 ± 0.33) over three independent experiments. Nevertheless, gut, lung or oral dysbiosis has also been impli- cated in the aetiology of a wide range of autoimmune diseases, including musculoskeletal pathologies like rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE)9,10. Whether the protection afforded by GI helminths against these disorders similarly involves interaction with the microbiome is not clear but infection with Heligmosomoides polygyrus and Trichuris muris can result in increases in Lactobacillaceae and decreases in Prevotella species2,11,12, commensals reported to be dysregu- lated in RA patients9,10. In any case, helminth-mediated protec- tion against autoimmune disease is not limited to GI-tract parasites, with particularly striking examples of this involving ﬁlarial nematodes preventing development of RA13 and SLE14. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 P P arasitic helminths (worms) have evolved to modulate host immune and tissue repair responses to promote their survival by limiting inﬂammation that would otherwise drive their expulsion and cause pathology1. The recent eradica- tion of helminths (and other pathogens) appears to have resulted in over-activated immune systems and this provides a rationale for the increasing prevalence of allergic and autoimmune inﬂammatory disorders, as well as contributing to the rise in obesity and associated comorbidities2–4 in developing and urba- nised countries. Although genetic studies have identiﬁed gene variants associated with various inﬂammatory diseases, these alone do not appear to be strong risk factors, integration with environmental factors being required to trigger disease. Recog- nition of this has focused interest on the role of the microbiota3,5 and hence, on how helminths may regulate this in health and disease6,7. Indeed, commensal bacteria and gastrointestinal (GI) helminths appear to have evolved to reciprocally regulate the gut microbiome8 to homeostatically maintain immune system function2. Thus, GI helminths can induce regulatory responses to limit inﬂammation and promote intestinal barrier integrity, while intestinal bacteria play an essential role in training the immune system by impacting on stem and progenitor cells3,5. Certainly, there is increasing evidence from animal models that protection afforded by GI helminth infection against mucosal inﬂammatory disorders like asthma, inﬂammatory bowel disease and coeliac disease, involves modulation of the gut microbiota2,6,7. of RA is preceded by disturbance of the gut microbiome with accompanying intestinal inﬂammation and loss of barrier integ- rity, ES-62 acts to normalise the microbiome and maintain gut health. Furthermore, we report that prophylactic depletion of the gut microbiota with broad-spectrum antibiotics (ABX) reduces the consequent severity of arthritis in mice undergoing CIA and in addition, reduces the level of protection afforded by ES-62. These data therefore indicate that a normalised microbiome is required for the full induction of the immunoregulatory actions of ES-62. The parasitic worm product ES-62 normalises the gut microbiota bone marrow axis in inﬂammatory arthritis James Doonan1, Anuradha Tarafdar 2, Miguel A. Pineda2, Felicity E. Lumb1, Jenny Crowe 2, Aneesah M. Khan2, Paul A. Hoskisson 1, Margaret M. Harnett2 & William Harnett1 James Doonan1, Anuradha Tarafdar 2, Miguel A. Pineda2, Felicity E. Lumb1, Jenny Crowe 2, Aneesah M. Khan2, Paul A. Hoskisson 1, Margaret M. Harnett2 & William Harnett1 The human immune system has evolved in the context of our colonisation by bacteria, viruses, fungi and parasitic helminths. Reﬂecting this, the rapid eradication of pathogens appears to have resulted in reduced microbiome diversity and generation of chronically activated immune systems, presaging the recent rise of allergic, autoimmune and metabolic disorders. Certainly, gastrointestinal helminths can protect against gut and lung mucosa inﬂammatory conditions by modulating the microbiome and suppressing the chronic inﬂammation associated with dysbiosis. Here, we employ ES-62, an immunomodulator secreted by tissue-dwelling Acanthocheilonema viteae to show that helminth-modulation of the gut microbiome does not require live infection with gastrointestinal-based worms nor is protection restricted to mucosal diseases. Speciﬁcally, subcutaneous administration of this deﬁned immunomodulator affords protection against joint disease in collagen-induced arthritis, a mouse model of rheumatoid arthritis, which is associated with normalisation of gut microbiota and prevention of loss of intestinal barrier integrity. 1 Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow G4 0RE, UK. 2 Institute of Infection, Immunity and Inﬂammation, University of Glasgow, Glasgow G12 8TA, UK. Correspondence and requests for materials should be addressed to M.M.H. (email: Margaret. Harnett@glasgow.ac.uk) or to W.H. (email: w.harnett@strath.ac.uk) 1 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Naive PBS lleum ES-62 Naive Colon Colon Colon PBS ES-62 Color key 0 0.4 Value 0.8 Actinobacteria Bacteroidetes Deferribacteres Firmicutes Proteobacteria Naive PBS ES-62 Naive Bacteriodetes Bacteroidia Bacteroidaceae Porphyromonadaceae Prevotellaceae Rikenellaceae Proteobacteria Naive PBS 15 Naive Naive Firmicutes Clostridiaceae Lachnospiraceae 3 2 1 PBS PBS ES-62 ES-62 2.0 1.5 1.0 10 5 ES-62 Epsilonprotecbacteria Gammaproteobacteria Helicobacter Escherichia Colon lleum Amino acids and derivatives Carbohydrates Cell division and cell cycle Cell wall and capsule DNA Metabolism Membrance transport Miscellaneous Bacteroidetes Naive IIeum Colon 3 2 1 ES-62 Bacteroidetes Naive ES-62 1.5 1.0 0.5 Firmicutes Proteobacteria Bacteroidaceae Porphyromonadaceae Prevotellaceae Bacilli Clostridia Clostridiaceae Eubacteriaceae Lachnospiraceae Ruminococcaceae Firmicutes Proteobacteria Bacteroidaceae Porphyromonadaceae Prevotellaceae Bacilli Clostridia Clostridiaceae Eubacteriaceae Lachnospiraceae Ruminococcaceae Motility and chemotaxis Nitrogen metabolism Phages, prophages, transposable elements, plasmids Nucleosides and nucleotides Phosphorus metabolism Potassium metabolism Protein metabolism RNA metabolism Respiration Stress response Sulfur metabolism Virulence, disease and defense Colon Regulation and cell signaling Fatty acids, lipids, and isoprenoids Iron acquisition and metabolism Clustering-based subsystems Cofactors, vitamins, prosthetic groups,pigments PBS 1.5 Firmicutes 1.00 0.75 Naive PBS ES-62 Clostridia Clostridiales Clostridiaceae Lachnospiraceae Ruminococcaceae Faecalibacterium Blautia Erysipelotrichi Butyrivibrio Dorea Roseburia 1.0 0.5 ES-62 b c a d e f g h i Naive Colon Colon PBS ES-62 Color key 0 0.4 Value 0.8 cteria detes acteres es acteria Naive Firmicutes Clostridiaceae Lachnospiraceae PBS ES-62 2.0 1.5 1.0 lleum Firmicutes 1.00 0.75 Naive PBS ES-62 Clostridia Clostridiales Clostridiaceae Lachnospiraceae Ruminococcaceae Faecalibacterium Blautia Erysipelotrichi Butyrivibrio Dorea Roseburia b d f Color key 0 0.4 Value 0.8 b Naive PBS lleum ES-62 Actinobacte Bacteroide Deferribact Firmicutes Proteobact a b a Colon Naive Bacteriodetes Bacteroidia Bacteroidaceae Porphyromonadaceae Prevotellaceae Rikenellaceae PBS 1.5 1.0 0.5 ES-62 c d c Proteobacteria Naive PBS 15 10 5 ES-62 Epsilonprotecbacteria Gammaproteobacteria Helicobacter Escherichia Colon e f f e Naive 3 2 1 PBS ES-62 Amino acids and derivatives Carbohydrates Cell division and cell cycle Cell wall and capsule DNA Metabolism Membrance transport Miscellaneous Porp Motility and chemotaxis Nitrogen metabolism Phages, prophages, transposable elements, plasmids Nucleosides and nucleotides Phosphorus metabolism Potassium metabolism Protein metabolism RNA metabolism Respiration Stress response Sulfur metabolism Virulence, disease and defense Colon Regulation and cell signaling Fatty acids, lipids, and isoprenoids Iron acquisition and metabolism Clustering-based subsystems Cofactors, vitamins, prosthetic groups,pigments g g 1 Bacteroidetes Naive IIeum 3 2 1 ES-62 Po Firmicutes Proteobacteria Bacteroidaceae Porphyromonadaceae Prevotellaceae Bacilli Clostridia Clostridiaceae Eubacteriaceae Lachnospiraceae Ruminococcaceae h i Colon Bacteroidetes Naive ES-62 1.5 1.0 0.5 Firmicutes Proteobacteria Bacteroidaceae Porphyromonadaceae Prevotellaceae Bacilli Clostridia Clostridiaceae Eubacteriaceae Lachnospiraceae Ruminococcaceae i 3 i i h Perturbation of the microbiome was also observed in the ileum of CIA mice and again, exposure to ES-62 normalised this towards the healthy community (Fig. Results Nevertheless, drilling down on the modulation of the Gram-negative Bacteroidetes phylum reveals differential signatures throughout each of the predominant Bacteroides, Porphyromonas and Prevotella genera and the Rikenellaceae family between the colon contents of Naive and CIA mice, identifying those normalised by exposure to ES-62 (Fig. 1c). Similarly, in spite of the fact that differential proﬁles amongst the groups were observed throughout major genera (e.g., Bacillus, Staphyloccus, Streptococcus, Enterococcus and NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 2 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Moreover, whilst ES-62 increases serum IL-10 levels in CIA mice, this regulatory cytokine is found at similarly low levels in PBS-CIA, PBS-ABX and ES-62-ABX mice (Fig. 2g). At the same time, ES-62-mediated suppression of serum levels of IL-6, a cytokine that promotes B-cell (auto)immunity34, is lost following ABX treatment (Fig. 2h). Collectively, these ABX studies indicate that depletion of the microbiota interferes with generation of inﬂammatory mediators associated with CIA pathogenesis as well as with the loss of immunoregulatory elements (Bregs) normally contributing to resolution of chronic inﬂamma- tion. Importantly, they also indicate that ES-62-resetting of the effector:regulatory balance is dependent on an intact gut micro- biome and hence suggest that its restoration is a consequence of ES- 62-mediated normalisation of the microbiome dysbiosis associated with CIA. (Fig. 1g). Intriguingly, treatment of Naive, healthy mice with ES- 62 for the duration of the CIA model also promotes expansion of members of the Clostridiales order, again particularly the Lachnospiraceae, whilst decreasing abundance of Bacilli, the Bacteroidaceae and Porphyromonadaceae families of Bacteroi- detes and also, Proteobacteria in the ileum (Fig. 1h). Likewise, ES- 62 can also promote depletion of Bacteroidetes and expansion of Ruminococcaceae and Clostridiaceae families and the Clostridia class of Firmicutes in the colon of Naive mice (Fig. 1i). Importantly, ES-62’s modulation of the gut microbiota in the absence of the chronic inﬂammation associated with CIA suggests that it can act directly, and that there is not an absolute requirement for it to harness immunoregulatory mechanisms to maintain and ﬁne-tune microbiome homoeostasis. Antibiotics both ameliorate CIA and impact on ES-62 pro- tection. To address whether microbiome perturbation observed in CIA plays a role in the initiation and progression of inﬂam- matory arthritis, we investigated the effect of continuous exposure of mice to a cocktail of broad-spectrum antibiotics (ABX) administered from 1 week prior to initiation of CIA. Such ABX treatment had no obvious effect on overall health as, after a characteristic initial dip, there was no signiﬁcant difference in body weight amongst the CIA groups at cull (Supplementary Figure 2). Nevertheless, this regimen essentially eliminated the bacterial microbiota, irrespective of treatment group, whilst metagenomic analysis showed the residual gut community to be almost entirely comprising proteobacteria (Supplementary Fig- ure 2). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 As predicted from previous studies in several inﬂamma- tory arthritis models22,30,31, ABX treatment reduced the incidence (PBS, 65.2%; PBS-ABX, 36.3%, as measured by articular score ≥1) and severity of joint pathology in CIA mice, both in terms of articular score (Fig. 2a) and histopathology (Fig. 2b, c). In addi- tion, the protection afforded by ES-62 was reduced in ABX- treated animals (Fig. 2a). Thus, prophylactic administration of ABX resulted in an intermediate phenotype of CIA, irrespective of whether the mice were treated with PBS or ES-62. ES-62 also acts in CIA mice to suppress the functional maturation of osteoclasts (OC)20 that directly cause erosive joint damage. Interestingly, changes in the intestinal microbiome have been shown to impact on bone mass35,36 and we therefore investigated whether the intermediate phenotype of joint pathology occurring in ABX-treated CIA mice reﬂected modula- tion of osteoclastogenesis resulting from perturbation of the gut microbiota. Exposure of CIA mice to ES-62 restored the numbers of OC differentiated from bone marrow (BM) progenitors ex vivo to Naive levels (Fig. 3a) and blocked their fusion to large, active multinucleated cells (Fig. 3a–c) that resorb bone20. Consistent with its amelioration of CIA pathology, ABX administration resulted in a decrease in large multinucleated OCs and a corresponding increase in total numbers of BM-derived OCs from CIA mice (Fig. 3a–c). Under these conditions of microbiota depletion, the ability of ES-62 to inhibit generation of large multinucleated OCs is lost (Fig. 3b): rather, and perhaps consistent with the increased joint inﬂammation exhibited by these mice, the ES-62-ABX group exhibits the largest multi- nucleated cells of the ABX-treated cohorts, displaying signiﬁ- cantly lower numbers of OCs but larger, multinucleated cells than Naive-ABX mice. ES-62 rewires osteoclastogenesis by modulating the RANK/OPG bone remodelling axis20 and this is evidenced again here by its ability to signiﬁcantly decrease RANK expression and increase (albeit not signiﬁcantly) expression of the decoy receptor, OPG relative to that seen in BM of PBS-CIA mice (Fig. 3d, e). This axis is also targeted by ABX treatment, with the elevated RANK expression observed in PBS-CIA BM being lost following ABX treatment. In addition, the elevated OPG ES-62-mediated protection against CIA is associated with restoration of the homoeostatic balance of regulatory:effector B- cell responses via downregulation of aberrant MyD88 signalling2,21. Typically, ES-62 reduces pathogenic anti-collagen type II (CII) IgG2a, but not IgG1 antibody production32,33. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Fig. 1 ES-62 normalises the microbiome during CIA towards a Naive phenotype. The composition of bacterial phyla present in the ileum and colon of Naive, PBS and ES-62-treated CIA animals, presenting proportion values as pie charts (a) from a single representative experiment using pooled samples from three mice in each condition. b Heatmap analysis of all bacteria present in the colon of Naive, PBS and ES-62-treated CIA animals from this representative model (articular scores; naive, 0; PBS-CIA, 6.3 ± 2.6; ES-62-CIA, 0; p < 0.05 for PBS-CIA versus Naive and ES-62-CIA mice) is shown. The proﬁle of bacterial abundance, is provided in Supplementary Table 1 for clarity. c–g Metagenonic analysis was performed on three independent models: in each model, samples from three representative mice/group were pooled in each individual experiment and the mice selected exhibited the following articular scores: Naive, 0; PBS-CIA, 7.00 ± 0.91; ES-62-CIA, 0.50 ± 0.33, where p < 0.001 for the PBS-CIA versus Naive and ES-62-CIA cohorts and n = 9 mice for all groups over the three independent models. Statistically signiﬁcant changes (two-way ANOVA with LSD Fishers multiple comparisons) observed over the three independent experiments (mean ± SEM, n = 3) between the PBS-CIA and ES-62-CIA groups in Bacteroidetes (c; Bacteroidaceae; PBS versus Naive or ES-62, p < 0.05), Firmicutes (d; Clostridiales; PBS versus Naive, p < 0.01 and PBS versus ES-62, p < 0.05) and Proteobacteria (e; Epsilonproteobacteria; PBS versus Naive or ES-62, p < 0.05) in the colon and Firmicutes in the ileum (f; Clostridiaceae; PBS versus ES-62, p < 0.05) as well as functional metagenomics of the colon (g; Phages; Naive versus PBS, p < 0.01, Protein metabolism; PBS versus Naive or ES-62, p < 0.01) are presented as heatmaps with changes in bacterial populations in PBS- and ES-62-treated CIA animals normalised to Naive controls. Ileum (h) and colon (i) faecal matter from Naive or ES-62-treated Naive animals was analysed for changes in bacterial populations in a single experiment using samples pooled from three animals per group and displayed as heatmaps, where the ES-62-treated samples are normalised to the Naive controls antibodies (Fig. 2e). Consistent with the effects of ABX on effector B-cell responses, analysis of splenic IL-10+CD19+ Bregs showed that both the decrease occurring during CIA (PBS-CIA) and also the maintenance of healthy levels in ES-62-CIA mice2,21 were lost in ABX-treated animals (Fig. 2f). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 1a). ES-62 clearly promoted growth of the Clostridiales, again particularly the Clostridaceae and Lachnospiraceae families, generally even beyond the levels found in healthy mice (Fig. 1f ). In the context of the relative paucity of bacteria in the ileum relative to the colon, this outgrowth perhaps explains why ES-62 increases the overall species diversity observed in the ileum but not the colon of CIA mice (Supplementary Figure 2). In addition, reﬂecting that CIA perturbs and ES-62 normalises gut bacteria, functional metage- nomic analysis showed that ES-62 generally acted to normalise the metabolic capacity of the colonic microbiome in CIA mice Clostridium) of Gram-positive Firmicutes (Fig. 1b; Supplemen- tary Table 1), the most dramatic changes with established CIA occurred within the Clostridiales order. In particular, decreases in the Clostridiaceae and the Lachnospiraceae (Fig. 1d) families were noted with ES-62 promoting maintenance of the Ruminococcus, Faecalibacterium and Blauti genera and the family Erysipelo- trichaceae in addition to the butyrate-producing genera Dorea and Roseburia, the latter having been implicated in gut health and inﬂammation homoeostasis29. In terms of the Proteobacteria, CIA was associated with outgrowth of members of the Helicobacter (Epsilonproteobacteria) and Escherichia (Gamma- proteobacteria) genera and again this was normalised by ES-62 (Fig. 1e). 3 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 3 NS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 5 PBS Naive PBS ES-62 PBS+ABX ES-62 ES-62+ABX 4 3 2 1 0 5 0.8 2.5 2.0 Anti-CII IgG1 (OD) 1.5 1.0 0.5 0.0 PBS PBS ES-62 ES-62 0.6 Anti-CII IgG2a (OD) ABX None 0.4 0.2 0.0 PBS PBS ABX ABX ES-62 ES-62 4 Pathological score 3 2 1 0 0.8 400 160 120 80 40 0 350 300 250 IL-10 pg/ml IL-6 pg/ml 200 150 0.6 0.4 % IL-10+ CD19+ cells 0.2 0.0 Naive Naive PBS PBS PBS PBS PBS PBS ABX ABX ABX ES-62 ES-62 ES-62 ES-62 ES-62 ES-62 PBS PBS ABX ES-62 ES-62 20 22 24 Day 26 28 30 Articular score ** * # * * * * * * # a b c d e f g h Fig. 2 Antibiotic treatment results in an intermediate disease phenotype. a Data are presented as articular scores (mean ± SEM) and are pooled from th independent experiments (PBS; n = 23, ES-62; n = 19, PBS-ABX; n = 22 and ES-62-ABX; n = 22). b H&E staining (scale bars represent 200 µm) of hi paws, from representative mice from each treatment group (clinical articular scores of individual paws were: PBS ≥3 (n = 9), ES-62 = 0 (n = 7), PBS ABX = 0 (n = 3) and ES-62-ABX ≥3 (n = 3)) selected to compare joint pathology occurring in the similarly scored ES-62-CIA and PBS-ABX (score 0) a ES-62-ABX and PBS-CIA (articular score ≥3) mice. c Blind scoring of pathology exhibited in joint sections imaged in (b). d, e Levels of collagen II (C speciﬁc IgG2a (d; PBS; n = 18, ES-62; n = 13, PBS-ABX; n = 16 and ES-62-ABX; n = 14) and IgG1 (e; PBS; n = 15, ES-62; n = 10, PBS-ABX; n = 16 and ES-6 ABX; n = 14) antibodies in serum were determined by ELISA. f Splenic B regulatory cells (CD19+IL-10+ cells) were determined by ﬂow cytometry (Nai n = 13, PBS; n = 21, ES-62; n = 14, Naive-ABX; n = 8 PBS-ABX; n = 14 and ES-62-ABX; n = 15). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 b H&E staining (scale bars represent 200 µm) of hind paws, from representative mice from each treatment group (clinical articular scores of individual paws were: PBS ≥3 (n = 9), ES-62 = 0 (n = 7), PBS- ABX = 0 (n = 3) and ES-62-ABX ≥3 (n = 3)) selected to compare joint pathology occurring in the similarly scored ES-62-CIA and PBS-ABX (score 0) and ES-62-ABX and PBS-CIA (articular score ≥3) mice. c Blind scoring of pathology exhibited in joint sections imaged in (b). d, e Levels of collagen II (CII)- speciﬁc IgG2a (d; PBS; n = 18, ES-62; n = 13, PBS-ABX; n = 16 and ES-62-ABX; n = 14) and IgG1 (e; PBS; n = 15, ES-62; n = 10, PBS-ABX; n = 16 and ES-62- ABX; n = 14) antibodies in serum were determined by ELISA. f Splenic B regulatory cells (CD19+IL-10+ cells) were determined by ﬂow cytometry (Naive; n = 13, PBS; n = 21, ES-62; n = 14, Naive-ABX; n = 8 PBS-ABX; n = 14 and ES-62-ABX; n = 15). g, h IL-10 (g: PBS; n = 6, ES-62; n = 5, PBS-ABX; n = 6 and ES-62-ABX; n = 6) and IL-6 (h: PBS; n = 24, ES-62; n = 10, PBS-ABX; n = 12 and ES-62-ABX; n = 13) concentrations in serum were determined by ELISA. Statistics: all results are presented as mean ± SEM and each symbol represents an individual mouse; data are from one (g) or pooled from two or three independent experiments. Statistical signiﬁcance was determined using two-way ANOVA (a) or one-way ANOVA (c–h) with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05 and p < 0.01 and #p < 0.05 (PBS vs. PBS + ABX [a] and ES-62-ABX vs. Naive [f]) involvement in the ability of ES-62 to modulate pathogenic osteoclastogenesis in CIA. expression observed in ES-62-conditioned OCs was abrogated following ABX treatment (Fig. 3d, e): these changes would result in similar levels of RANKL-driven osteoclastogenesis consistent with the intermediate CIA phenotype observed in PBS- and ES- 62-ABX animals. Supporting these ABX-changes in osteoclasto- genesis and consequently bone damage, relative to Naive mice, PBS-ABX and ES-62-ABX animals exhibit similar grip strengths (70.8 ± 5.4% and 58.8 ± 6.0%, respectively) that are intermediate to those displayed by PBS- (42.3 ± 6.2%) and ES-62-CIA mice (90.1 ± 9.6%). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Reﬂecting the ABX- driven amelioration of CIA pathology, anti-CII Ig2a levels in PBS- ABX mice are not signiﬁcantly different from those in ES-62-CIA mice, whilst those in ES-62-ABX mice are not signiﬁcantly reduced relative to those in PBS-CIA mice (Fig. 2d). No differences were detected amongst any of the groups in terms of anti-CII IgG1 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 4 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 b H&E staining (scale bars represent 200 µm) of hind paws, from representative mice from each treatment group (clinical articular scores of individual paws were: PBS ≥3 (n = 9), ES-62 = 0 (n = 7), PBS- ABX = 0 (n = 3) and ES-62-ABX ≥3 (n = 3)) selected to compare joint pathology occurring in the similarly scored ES-62-CIA and PBS-ABX (score 0) and ES-62-ABX and PBS-CIA (articular score ≥3) mice. c Blind scoring of pathology exhibited in joint sections imaged in (b). d, e Levels of collagen II (CII)- speciﬁc IgG2a (d; PBS; n = 18, ES-62; n = 13, PBS-ABX; n = 16 and ES-62-ABX; n = 14) and IgG1 (e; PBS; n = 15, ES-62; n = 10, PBS-ABX; n = 16 and ES-62- ABX; n = 14) antibodies in serum were determined by ELISA. f Splenic B regulatory cells (CD19+IL-10+ cells) were determined by ﬂow cytometry (Naive; n = 13, PBS; n = 21, ES-62; n = 14, Naive-ABX; n = 8 PBS-ABX; n = 14 and ES-62-ABX; n = 15). g, h IL-10 (g: PBS; n = 6, ES-62; n = 5, PBS-ABX; n = 6 and ES-62-ABX; n = 6) and IL-6 (h: PBS; n = 24, ES-62; n = 10, PBS-ABX; n = 12 and ES-62-ABX; n = 13) concentrations in serum were determined by ELISA. Statistics: all results are presented as mean ± SEM and each symbol represents an individual mouse; data are from one (g) or pooled from two or three independent experiments. Statistical signiﬁcance was determined using two-way ANOVA (a) or one-way ANOVA (c–h) with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05 and *p < 0.01 and #p < 0.05 (PBS vs. PBS + ABX [a] and ES-62-ABX vs. Naive [f]) Fig. 2 Antibiotic treatment results in an intermediate disease phenotype. a Data are presented as articular scores (mean ± SEM) and are pooled from three independent experiments (PBS; n = 23, ES-62; n = 19, PBS-ABX; n = 22 and ES-62-ABX; n = 22). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 g, h IL-10 (g: PBS; n = 6, ES-62; n = 5, PBS-ABX; n = 6 a ES-62-ABX; n = 6) and IL-6 (h: PBS; n = 24, ES-62; n = 10, PBS-ABX; n = 12 and ES-62-ABX; n = 13) concentrations in serum were determined by ELIS Statistics: all results are presented as mean ± SEM and each symbol represents an individual mouse; data are from one (g) or pooled from two or th independent experiments. Statistical signiﬁcance was determined using two-way ANOVA (a) or one-way ANOVA (c–h) with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05 and p < 0.01 and #p < 0.05 (PBS vs. PBS + ABX [a] and ES-62-ABX vs. Naive [f]) NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 ARTICL 5 PBS Naive PBS ES-62 PBS+ABX ES-62 ES-62+ABX 4 3 2 1 0 ABX None 20 22 24 Day 26 28 30 Articular score ** * # a b 5 PBS PBS+ABX ES-62 ES-62+ABX 4 3 2 1 0 ABX None 20 22 24 Day 26 28 30 Articular score * # a b Naive PBS ES-62 ABX None b b a None ABX 0.8 0.6 Anti-CII IgG2a (OD) 0.4 0.2 0.0 PBS PBS ABX ES-62 ES-62 * d 5 4 Pathological score 3 2 1 0 PBS PBS ABX ES-62 ES-62 * * c 2.5 2.0 Anti-CII IgG1 (OD) 1.5 1.0 0.5 0.0 PBS PBS ES-62 ES-62 ABX e d c 400 350 300 250 IL-10 pg/ml 200 150 PBS PBS ABX ES-62 ES-62 * g 0.8 0.6 0.4 % IL-10+ CD19+ cells 0.2 0.0 Naive Naive PBS PBS ABX ES-62 ES-62 * # f 160 120 80 40 0 IL-6 pg/ml PBS PBS ABX ES-62 ES-62 * h h g ABX Fig. 2 Antibiotic treatment results in an intermediate disease phenotype. a Data are presented as articular scores (mean ± SEM) and are pooled from three independent experiments (PBS; n = 23, ES-62; n = 19, PBS-ABX; n = 22 and ES-62-ABX; n = 22). NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 3 ES-62 requires the gut microbiome to protect the bone remodelling axis. a–c Osteoclasts (OCs) were differentiated from bone marrow obtained at cull and cultured for 5 days before differentiation, in terms of numbers (a) and size (b) of OCs, was measured using ImageJ analysis software and data were normalised as a percentage of Naive controls with representative images (c) provided (Naive; n = 11, PBS; n = 11, ES-62; n = 6, Naive + ABX; n = 2, PBS + ABX; n = 3, ES-62 + ABX; n = 3). Whole bone marrow was used to quantify RANK (d; PBS; n = 20, ES-62; n = 13, PBS + ABX; n = 15, ES-62 + ABX; n = 15) and OPG (e; PBS; n = 17, ES-62; n = 11, PBS + ABX; n = 16, ES-62 + ABX; n = 14) mRNA levels using qRT-PCR, and fold change was calculated following normalisation to Naive controls. Statistics: all data are presented as mean ± SEM. In a and b, each symbol represents experimental replicates and in d and e each symbol represents individual mice and data are pooled from three independent experiments. Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001 Supplementary Table 1) that have been implicated in main- tenance of epithelial integrity and gut health29. Strikingly, mice with CIA display severe gut pathology and inﬂammation, the levels of which directly correlate with severity of CIA (Fig. 4a). Moreover, ES-62 treatment protects against such gut damage, speciﬁcally the thickening and generation of stubby villi in the ileum (Fig. 4b, c) and hole-like lesions in the colon (Fig. 4b. d). The physiological relevance of these colon lesions is unclear but they may reﬂect the suppression of mucus production by goblet cells in response to bacteria39 or alternatively, attachment/effa- cement lesions induced by pathogenic bacteria, including E. coli40. The gut pathology in PBS-CIA mice was diminished in those administered ABX (Fig. 4b–d), suggesting that the associated reduction in chronic gut (and consequently systemic) inﬂammation contributes to the amelioration of CIA in PBS-ABX mice. Of note, ES-62-ABX mice displayed increased pathology and inﬂammatory cell inﬁltration of the gut tissue compared with ES-62-CIA mice (Fig. 4b), ﬁndings again consistent with chronic gut inﬂammation promoting disease severity. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 both ileum villi thickness and colonic lesions following immunisation (Fig. 5a–c). Interestingly, the ileum pathology occurring during the initiation stage in PBS-CIA mice appeared to resolve by the end of the preclinical phase (day 21), although thickening and shortening of the villi was again induced by the booster CII immunisation. This pattern was not the case for the colon lesions, which peaked at day 21 and were maintained throughout active disease in PBS-CIA mice. As opposed to CIA mice, ileum integrity was maintained throughout all phases of disease, whilst the high levels of colon lesions generated by day 21 of the preclinical phase were reduced in ES-62-treated animals (Fig. 5b, c). Supplementary Table 1) that have been implicated in main- tenance of epithelial integrity and gut health29. Strikingly, mice with CIA display severe gut pathology and inﬂammation, the levels of which directly correlate with severity of CIA (Fig. 4a). Moreover, ES-62 treatment protects against such gut damage, speciﬁcally the thickening and generation of stubby villi in the ileum (Fig. 4b, c) and hole-like lesions in the colon (Fig. 4b. d). The physiological relevance of these colon lesions is unclear but they may reﬂect the suppression of mucus production by goblet cells in response to bacteria39 or alternatively, attachment/effa- cement lesions induced by pathogenic bacteria, including E. coli40. The gut pathology in PBS-CIA mice was diminished in those administered ABX (Fig. 4b–d), suggesting that the associated reduction in chronic gut (and consequently systemic) inﬂammation contributes to the amelioration of CIA in PBS-ABX mice. Of note, ES-62-ABX mice displayed increased pathology and inﬂammatory cell inﬁltration of the gut tissue compared with ES-62-CIA mice (Fig. 4b), ﬁndings again consistent with chronic gut inﬂammation promoting disease severity. Supporting our metagenomics data, qPCR analysis showed that ES-62 prevents the enrichment of Bacteroidetes and maintains levels of Firmicutes in the colon of mice during established CIA (Fig. 5d, e). However, analysis at the various phases of the disease suggested a more dynamic situation: for example, following the primary immunisation with CII/CFA, there is a signiﬁcant decrease in levels of both Bacteroidetes and Firmicutes, with the rise in Bacteroidetes evident in established arthritis only occurring following the booster immunisation (Fig. 5d, e). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 a c d e b 800 600 400 200 0 3 2 1 0 30 20 10 0 OC size * * * * * * Size of OCs (µm2) None Naive PBS ES-62 ABX RANK mRNA OPG mRNA 200 OC number * * *** * Number of OCs 150 100 50 0 Naive PBS ES-62 Naive PBS ABX ES-62 Naive PBS ES-62 PBS ES-62 PBS ES-62 PBS ES-62 PBS ES-62 ABX ABX Naive PBS ABX ES-62 Fig. 3 ES-62 requires the gut microbiome to protect the bone remodelling axis. a–c Osteoclasts (OCs) were differentiated from bone marrow obtained at cull and cultured for 5 days before differentiation, in terms of numbers (a) and size (b) of OCs, was measured using ImageJ analysis software and data were normalised as a percentage of Naive controls with representative images (c) provided (Naive; n = 11, PBS; n = 11, ES-62; n = 6, Naive + ABX; n = 2, PBS + ABX; n = 3, ES-62 + ABX; n = 3). Whole bone marrow was used to quantify RANK (d; PBS; n = 20, ES-62; n = 13, PBS + ABX; n = 15, ES-62 + ABX; n = 15) and OPG (e; PBS; n = 17, ES-62; n = 11, PBS + ABX; n = 16, ES-62 + ABX; n = 14) mRNA levels using qRT-PCR, and fold change was calculated following normalisation to Naive controls. Statistics: all data are presented as mean ± SEM. In a and b, each symbol represents experimental replicates and in d and e each symbol represents individual mice and data are pooled from three independent experiments. Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001 a 200 OC number * * * * Number of OCs 150 100 50 0 Naive PBS ES-62 Naive PBS ABX ES-62 b 800 600 400 200 0 OC size * * * * Size of OCs (µm2) Naive PBS ES-62 Naive PBS ABX ES-62 b a a e 30 20 10 0 * OPG mRNA PBS ES-62 PBS ES-62 ABX d 3 2 1 0 * ** RANK mRNA PBS ES-62 PBS ES-62 ABX c None Naive PBS ES-62 ABX d c e Fig. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Collectively, these data not only support a role for the microbiome in osteoclastogenesis but also implicate its ES-62 protects against gut pathology in CIA. Intestinal dys- biosis has been associated with loss of gut integrity and chronic inﬂammation in conditions such as obesity that promote auto- immune disorders like RA and cardiovascular comorbidities37,38. Our metagenomic analysis showed that established CIA was associated with gut microbiome changes, in particular with reduction in butyrate-producing bacteria (Fig. 1b, d; 5 5 ARTICLE NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 b Representative H&E images of ileum (top row), colon (middle row) and PAS-stained colon (bottom row) sections displaying scale bar (200 µm) and articular scores of the mice are shown on the images. c Quantitative analysis of ileum villi thickness, where symbols represent mean values of replicate sections as measured using ImageJ analysis software (n = 3/group with 3–5 replicates/animal and data are from two independent experiments). d Lesions were enumerated per colon section of individual mice using ImageJ imaging software (Naive; n = 6, PBS; n = 8, ES-62; n = 7, Naive + ABX; n = 6, PBS + ABX; n = 7, ES-62 + ABX; n = 7). Statistics: data are pooled from two independent experiments. Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001, whilst in (c), PBS-ABX is not signiﬁcantly different from the Naive group (both Foxp3+ Tregs and IL-10+ Tr1 cells41,42) which by suppressing and resolving chronic inﬂammation, could contri- bute to the maintenance of gut barrier integrity. However, whilst we ﬁnd that the induction of CIA is accompanied by loss of splenic Tr1 cells, the predominant regulatory T-cell subset defective in CIA43, ES-62 does not restore their levels back towards those of Naive mice (Fig. 5g). These data are consistent with our previous ﬁndings that protection against CIA afforded by ES-62 is not associated with induction of either Treg or Tr1 cell responses21. Moreover, ABX treatment does not modulate either the CIA-associated depletion of Tr1 cells or the failure of ES-62 to induce their restoration, suggesting their development in this model is not impacted by the status of the microbiome (Fig. 5g). Thus, collectively these data indicate that ES-62 likely promotes Butyrivibrio species to maintain butyrate-mediated protection of intestinal epithelium physiology and barrier integrity44. The dynamic changes observed in the gut during the initiation phase further questions whether CIA-associated autoimmune (Th17) inﬂammation is responsible for triggering gut pathology and consequently, perturbation of the microbiome or vice versa. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 a b c d Naive Naive CIA 6 CIA 6 CIA 9 CIA 9 CIA 12 CIA 12 4 PBS ES-62 ABX PBS ABX ES-62 0 1 2 2 0 0 3 2 250 120 Colon lesions 80 40 0 * * ** * * 200 Villi thickness (µm) 150 100 50 Naive PBS ES-62 Naive PBS ES-62 ABX Naive PBS ES-62 Naive PBS ES-62 ABX 0 1 5 Ileum Ileum Colon 50 µm 50 µm 50 µm 50 µm 200 µm 200 µm 200 µm 1000 µm 500 µm 200 µm 200 µm 200 µm 200 µm 200 µm 200 µm Fig. 4 CIA is accompanied by microbiome-dependent gut pathology. a Representative H&E images (scale bar representing 50 µm) of ileum sections of Naive and PBS-CIA mice with differential disease severity at cull (mouse articular score as indicated). b Representative H&E images of ileum (top row), colon (middle row) and PAS-stained colon (bottom row) sections displaying scale bar (200 µm) and articular scores of the mice are shown on the images. c Quantitative analysis of ileum villi thickness, where symbols represent mean values of replicate sections as measured using ImageJ analysis software (n = 3/group with 3–5 replicates/animal and data are from two independent experiments). d Lesions were enumerated per colon section of individual mice using ImageJ imaging software (Naive; n = 6, PBS; n = 8, ES-62; n = 7, Naive + ABX; n = 6, PBS + ABX; n = 7, ES-62 + ABX; n = 7). Statistics: data are pooled from two independent experiments. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 By contrast, ES-62 acts to maintain healthy levels of Firmicutes in CIA mice throughout disease, but most strongly during the initiation and preclinical phases and in addition, inhibits the CIA-induced increase in Bacteroidetes during established disease. Moreover, although ES-62 promoted enrichment of butyrate- producing bacteria (Butyrivibrio) at all stages of disease, this was most evident in the initiation phase (Fig. 5f). Butyrate has been reported to promote the development of regulatory T cells To further address the role of gut pathology in CIA, we analysed the ileum and colon tissue from Naive mice and those undergoing CIA (treated with PBS or ES-62) at key points during initiation and progression of disease: (i) Naive mice, (ii) the breaking of tolerance and initiation of disease, following immunisation with CII/CFA (≤day 14), (iii) preclinical (day 21, prior to booster immunisation with CII) and (iv) established disease (≥day 28; articular score: PBS–5.2 ± 0.8; ES-62–0.9 ± 1.24). This analysis revealed the presence of gut pathology in PBS-CIA mice during the initiation phase with an increase in URE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 6 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001, whilst in (c), PBS-ABX is not signiﬁcantly different from the Naive group a b Naive Naive CIA 6 CIA 6 CIA 9 CIA 9 CIA 12 CIA 12 4 PBS ES-62 ABX PBS ABX ES-62 0 1 2 2 0 0 3 2 0 1 5 Ileum Ileum Colon 50 µm 50 µm 50 µm 50 µm 200 µm 200 µm 200 µm 1000 µm 500 µm 200 µm 200 µm 200 µm 200 µm 200 µm 200 µm a b c 250 * * 200 Villi thickness (µm) 150 100 50 Naive PBS ES-62 Naive PBS ES-62 ABX d 120 Colon lesions 80 40 0 * * ** Naive PBS ES-62 Naive PBS ES-62 ABX c ABX Fig. 4 CIA is accompanied by microbiome-dependent gut pathology. a Representative H&E images (scale bar representing 50 µm) of ileum sections of Naive and PBS-CIA mice with differential disease severity at cull (mouse articular score as indicated). b Representative H&E images of ileum (top row), colon (middle row) and PAS-stained colon (bottom row) sections displaying scale bar (200 µm) and articular scores of the mice are shown on the images. c Quantitative analysis of ileum villi thickness, where symbols represent mean values of replicate sections as measured using ImageJ analysis software (n = 3/group with 3–5 replicates/animal and data are from two independent experiments). d Lesions were enumerated per colon section of individual mice using ImageJ imaging software (Naive; n = 6, PBS; n = 8, ES-62; n = 7, Naive + ABX; n = 6, PBS + ABX; n = 7, ES-62 + ABX; n = 7). Statistics: data are pooled from two independent experiments. Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001, whilst in (c), PBS-ABX is not signiﬁcantly different from the Naive group Fig. 4 CIA is accompanied by microbiome-dependent gut pathology. a Representative H&E images (scale bar representing 50 µm) of ileum sections of Naive and PBS-CIA mice with differential disease severity at cull (mouse articular score as indicated). NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 y precedes and shapes the autoimmune inﬂammation rthritis, we ﬁnd that CIA-associated perturbation of the microbiota is evident as early as day 6 (post CII ation): this is broadly normalised by exposure to ES- the worm product modulating levels of potentially nic bacteria (such as the Enterobacteriaceae family of proteobacteria) to below those found in Naive animals A h C A d b h d ileum villi particularly evident at this early time poin protected against by ES-62 (Fig. 6b). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Interestingly, therefore, whilst conﬁrming our ﬁndings that dysbiosis and loss of intestinal barrier integrity are established by day 14 in the CIA model, a very recent report45 shows that such breakdown in gut function precedes, and is required for, CIA-associated IL-17 responses as increases in IL-17 can only be determined in the small intestine from day 14 whilst those in the MLN occurred even later, being present during the established disease phase (by day 35, but not detected on day 14), and that ABX-depletion of the microbiota inhibits such IL-17 production and severity of joint disease. Moreover, these effects appeared not to require (any) inﬂammation for initiation as they could not be replicated by sham immunisation with CFA45. Consistent with this idea that dysbiosis and associated gut 7 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunicatio ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 These cha function occur in the absence of the systemic in (serum IL-6 cytokine and CII-speciﬁc IgG2a antibodi present around the limit of detection of the ELISA with no differences observed amongst Naive, PBS-C 62-CIA groups), known to be targeted by ES-62 in C h l b d Naive a b c d e f g Villi thickness Bacteriodetes Butyrivibrio PBS ES-62 PBS ES-62 PBS ES-62 Firmicutes PBS ES-62 Colon lesions Ileum Colon Initiation Pre-clinical Disease Initiation Pre-clinical Disease Initiation Pre-clinical Disease Initiation Pre-clinical Disease Initiation Pre-clinical Disease 50 75 100 125 150 * * * * PBS ES-62 Initiation Pre-clinical Disease * % of naive 50 0 100 150 200 % of naive 0 100 300 0.0 0.1 0.2 0.3 0.4 % IL-10+CD3+ T cells 200 % of naive 50 0 100 150 % of naive 300 400 0 100 200 % of naive Naive PBS ES-62 Naive PBS ABX ES-62 Naive a Ileum Colon Initiation a Pre-clinical Disease c Colon lesions PBS ES-62 Initiation Pre-clinical Disease * 300 400 0 100 200 % of naive b Villi thickness PBS ES-62 Initiation Pre-clinical Disease 50 75 100 125 150 ** % of naive b c e Firmicutes PBS ES-62 Initiation Pre-clinical Disease * 50 0 100 150 % of naive d Bacteriodetes PBS ES-62 Initiation Pre-clinical Disease * 50 0 100 150 200 % of naive d e g * * 0.0 0.1 0.2 0.3 0.4 % IL-10+CD3+ T cells Naive PBS ES-62 Naive PBS ABX ES-62 f Butyrivibrio PBS ES-62 Initiation Pre-clinical Disease 0 100 300 200 % of naive f g g ABX pathology precedes and shapes the autoimmune inﬂammation driving arthritis, we ﬁnd that CIA-associated perturbation of the colon microbiota is evident as early as day 6 (post CII immunisation): this is broadly normalised by exposure to ES- 62, with the worm product modulating levels of potentially pathogenic bacteria (such as the Enterobacteriaceae family of Gamma proteobacteria) to below those found in Naive animals (Fig. 6a). Accompanying the CIA-dysbiosis, there is damage to both the ileum and colon, with shortening and thickening of the ileum villi particularly evident at this early time point and this is protected against by ES-62 (Fig. 6b). ARTICLE p Although the precise details of how gut microbiome perturba- tion occurs and drives chronic autoimmune inﬂammation and joint destruction in CIA remain to be deﬁned, commensal bacteria play key roles in educating the immune system and BM progenitors and hence, loss of microbiome diversity can impact on both inﬂammation and bone homoeostasis46,47. Such training involves interactions of Pattern Recognition Receptors (PRR; e.g., TLRs and NODs) with the gut microbiome2,3,48. Consistent with this, ES-62 can rewire BM progenitors and stromal cells from CIA mice to an anti-inﬂammatory, regulatory or tissue repair phenotype2,16,20,49–51 by subverting TLR4 signalling to prevent the upregulation of the key TLR signal transducer, MyD88 observed during chronic inﬂammation52. Interestingly, therefore, given the increased incidence and severity of CIA in ES-62-ABX mice, we show that ES-62 dampening of aberrant MyD88 expression is abolished by ABX treatment (Fig. 6d). Commensal bacteria can also dynamically educate OC progenitor (OCP) maturation46,47: as OCs normally act in concert with osteoblasts (OBs) to homeostatically maintain healthy bone, the enhanced functional capacity of OCs elicited by commensal bacteria may actually render hosts more susceptible to bone damage during chronic inﬂammation, a condition that promotes bone remodelling, particularly as commensal bacteria also appear to act to suppress OB function46,47. Reﬂecting the dynamic and differential changes in the colonic microbiota, gut inﬂammation and pathology, there is a strong increase in the monocyte populations containing OCPs during the early inﬂammatory initiation phase of CIA that has resolved by the end of the preclinical phase only to increase again in the established phase of arthritis following the booster CII immunisation. Consistent with our data that ES-62 does not fundamentally suppress OC differentiation, there are no differences in the percentage of monocytes between the PBS-CIA and ES-62-CIA groups (Fig. 6e). However, following both primary and booster immunisations, BM from CIA mice shows enhanced capacity for functional OC maturation that is suppressed by ES-62 (Fig. 6f, g). The normalisation of the gut microbiota by ES-62 may actually result in a dual pronged mechanism by which butyrate, in addition to its local gut-protecting actions, could also impact systemically to protect more directly against joint pathology. Consistent with this, butyrate has been reported to suppress osteoclastogenesis58, and by protecting against pathological bone loss, to regulate bone mass35. ARTICLE Collectively, these ﬁndings indicate that ES-62 primarily acts directly to normalise the CIA-associated gut dysbiosis that both triggers mucosal inﬂammation and initiates the mesenteric lymph node (MLN) IL-17 responses that drive autoimmunity, rather than indirectly modulating the dysbiosis and gut pathology arising from the systemic inﬂamma- tion that results from the CIA-immunisation protocol. g y Potentially pathogenic changes in the gut microbiome have been described following analysis of stool faeces from patients with new-onset disease (enrichment of Prevotella copri10,53), as well as those with established arthritis (changes in the Clos- tridiaceae, Coriobacteriaceae and Lachnospiraceae families54,55). Perhaps reﬂecting this, our metagenomic analysis of CIA mice during the established disease phase shows ES-62’s protective actions to be most strongly associated with maintenance of the Clostridiaceae, Lachnospiraceae and Ruminococcaceae families, particularly those associated with butyrate production, e.g., Ruminococcus. Administration of butyrate was previously found to ameliorate CIA severity, notably in terms of reduction in each of inﬂammatory cell inﬁltration of the joint, pannus formation and cartilage and bone destruction56. By contrast, butyrate exa- cerbated antibody-induced arthritis, a model which bypasses the initiation and adaptive immunity phases of disease, suggesting that butyrate needs to act during these preclinical phases to exhibit its protective actions56. Possibly consistent with this, whilst ES-62 protects against depletion of butyrate-producing species in mice with established disease, we also ﬁnd Butyrivibrio to be most enriched by ES-62 in the CIA initiation phase. Col- lectively, these data suggest that depletion of butyrate-producing bacteria associated with onset of CIA may contribute to the gut pathology promoting and perpetuating the inﬂammation that drives immune tolerance breakdown and consequent auto- immune joint disease. Certainly, butyrate is known to regulate gut barrier integrity44 and goblet cell production of MUC257. Indeed, we ﬁnd microbiome dysbiosis and gut pathology to precede joint disease onset, being observed within 6 days of primary CII immunisation and in apparent absence of systemic inﬂammation. Furthermore, the loss of colon barrier integrity peaks by the end of the CIA preclinical phase (d21) and such gut pathology is accompanied by dynamic changes in the microbiome of CIA mice as evidenced by the decrease in colon abundance of Bac- teroidetes and Firmicutes during the early initiation phase of disease and the enrichment of the former during established arthritis. ARTICLE Fig. 5 ES-62 protects against gut pathology occurring prior to onset of arthritis. a Representative H&E images (scale bars 200 µm) of ileum and colon of mice culled during the Naive, initiation (post immunisation ≤d14), preclinical (d21 prior to challenge) and disease (established arthritis d ≥28, articular score PBS, 5.2 ± 0.8; ES-62–0.9 ± 1.24) phases of CIA. Changes in villus thickness (b; Initiation; PBS, n = 5, ES-62, n = 6, Pre-Clinical; PBS and ES-62, n = 3, Disease; PBS, n = 4, ES-62, n = 5 mice) and number of colon lesions (c; Initiation; PBS and ES-62, n = 12 images spanning six mice each, Pre-Clinical; PBS, n = 6 [3 mice], ES-62, n = 4 [2 mice]; Disease; PBS, n = 8 [4 mice], ES-62, n = 10 [5 mice]) were quantiﬁed using ImageJ analysis software with data normalised to values of Naive mice and presented as mean ± SEM. Statistical signiﬁcance was determined using two-way ANOVA comparing PBS and ES-62 treatment at each time point where p < 0.01 and p < 0.001. d–f Changes in Bacteroidetes (d), Firmicutes (e) and Butyrivibrio (f) populations in the colon faecal matter of Naive, PBS- and ES-62-treated animals were measured by qPCR, normalised to total bacterial content and presented as % Naive mice. Data are presented as mean ± SEM values from individual animals (Initiation; n = 6, Pre-Clinical; n = 3, Disease; n = 10) and statistical signiﬁcance determined using two-way ANOVA with LSD Fishers multiple comparisons where p < 0.05, p < 0.01 and p < 0.001. g Tr1 cells (CD3+IL-10+ cells) were determined by ﬂow cytometric analysis of splenocytes from individual mice (Naive; n = 4, PBS; n = 7, ES-62; n = 4, Naive-ABX; n = 4 PBS-ABX; n = 6 and ES-62-ABX; n = 7) from a single experiment. Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons where p < 0.05 distal to the gut, such as the joints. Collectively, our data suggest that ES-62-mediated protection reﬂects broad normalisation of the gut dysbiosis observed in arthritic mice. exposure to ES-62 (Fig. 6c). NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 These changes in gut function occur in the absence of the systemic inﬂammation (serum IL-6 cytokine and CII-speciﬁc IgG2a antibodies were only present around the limit of detection of the ELISAs employed, with no differences observed amongst Naive, PBS-CIA and ES- 62-CIA groups), known to be targeted by ES-62 in CIA (Fig. 2). By contrast, the late microbiota-driven IL-17 responses developed in the MLN in the clinical phase of CIA45 were suppressed by ileum villi particularly evident at this early time point and this is protected against by ES-62 (Fig. 6b). These changes in gut function occur in the absence of the systemic inﬂammation (serum IL-6 cytokine and CII-speciﬁc IgG2a antibodies were only present around the limit of detection of the ELISAs employed, with no differences observed amongst Naive, PBS-CIA and ES- 62-CIA groups), known to be targeted by ES-62 in CIA (Fig. 2). By contrast, the late microbiota-driven IL-17 responses developed in the MLN in the clinical phase of CIA45 were suppressed by ileum villi particularly evident at this early time point and this is protected against by ES-62 (Fig. 6b). These changes in gut function occur in the absence of the systemic inﬂammation (serum IL-6 cytokine and CII-speciﬁc IgG2a antibodies were only present around the limit of detection of the ELISAs employed, with no differences observed amongst Naive, PBS-CIA and ES- 62-CIA groups), known to be targeted by ES-62 in CIA (Fig. 2). By contrast, the late microbiota-driven IL-17 responses developed in the MLN in the clinical phase of CIA45 were suppressed by NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 8 ARTICLE Intriguingly, we ﬁnd spikes of functional maturation of OCs in the initiation and established phases of CIA that are prevented by ES-62 and are associated with its enrichment of Butyrivibrio. The mechanisms by which NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 e Initiation; n = 6, Pre-Clinical; n = 3, Disease; n = 4) or mean ± SD om experimental replicates (f; Initiation; n = 18, Pre-Clinical; n = 9, Disease; PBS: n = 12 and ES-62; n = 15). Statistical signiﬁcance was determined usin wo- (e, f) or one-way (d) ANOVA with LSD Fishers multiple comparisons and indicated by asterisks, p < 0.05, p < 0.01 and *p < 0.001 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361 Proteobacteria Naive PBS ES-62 1 a 2 3 4 Enterobacteriaceae Lactobacillaceae Segmented filamentous Lachnospiraceae PBS Ileum ES-62 Colon b 200 µm 200 µm 200 µm 200 µm Naive PBS ES-62 2.0 *** 1.5 % IL-17+ MLN cells 1.0 0.5 0.0 c b b c a % of monocytes Initiation Pre-clinical Disease 0 100 50 150 200 % of naive 250 PBS ES-62 e Naive PBS ES-62 Naive PBS ABX ES-62 2.0 * 1.5 MyD88 mRNA 1.0 0.5 0.0 d d e PBS OC size ES-62 * * Initiation Pre-clinical Disease 0 200 400 % of naive 600 f PBS ES-62 Initiation Pre-clinical Disease g f g Fig. 6 ES-62 modulates the gut-bone marrow axis during the early phases of CIA. a Changes in the indicated bacterial populations in the colon faecal matter of Naive, PBS- or ES-62-treated animals were measured at day 6 post CII/CFA-immunisation by qPCR and data were normalised to total bacterial content and presented as fold change compared to Naive mice (Naive, n = 3; PBS-CIA, n = 3 and ES-62, n = 2). b Representative H&E images of ileum and colon pathology of mice culled at day 6 of CIA. Scale bars are 200 µm. c IL-17+ lymphocytes were determined by ﬂow cytometric analysis of MLN cells from individual mice (naïve, n = 5: PBS, n = 5 and ES-62, n = 5). Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multiple comparisons and signiﬁcance indicated by asterisks, p < 0.05 and *p < 0.001. d Whole bone marrow was used to quantify MyD88 mRNA levels using qRT-PCR and normalised to naive controls. e The proportion of monocytes (CD3-B220-Ter119−Ly6G−Ly6C+) in bone marrow was measured by ﬂow cytometry and normalised to those in Naive control mice. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Proteobacteria Naive PBS PBS ES-62 Ileum Naive % of monocytes PBS ES-62 ES-62 2.0 *** 1.5 % IL-17+ MLN cells 1.0 0.5 0.0 Colon 1 a b c Naive PBS PBS ES-62 Naive PBS ABX OC size ES-62 ES-62 PBS ES-62 2.0 * * * 1.5 MyD88 mRNA 1.0 0.5 0.0 Initiation Pre-clinical Disease Initiation Pre-clinical Disease 0 200 400 % of naive 600 Initiation Pre-clinical Disease 0 100 50 150 200 % of naive 250 PBS ES-62 d e f g 2 3 4 Enterobacteriaceae Lactobacillaceae Segmented filamentous Lachnospiraceae 200 µm 200 µm 200 µm 200 µm ig. 6 ES-62 modulates the gut-bone marrow axis during the early phases of CIA. a Changes in the indicated bacterial populations in the colon faeca matter of Naive, PBS- or ES-62-treated animals were measured at day 6 post CII/CFA-immunisation by qPCR and data were normalised to total bacter ontent and presented as fold change compared to Naive mice (Naive, n = 3; PBS-CIA, n = 3 and ES-62, n = 2). b Representative H&E images of ileum an olon pathology of mice culled at day 6 of CIA. Scale bars are 200 µm. c IL-17+ lymphocytes were determined by ﬂow cytometric analysis of MLN ce om individual mice (naïve, n = 5: PBS, n = 5 and ES-62, n = 5). Statistical signiﬁcance was determined using one-way ANOVA with LSD Fishers multip omparisons and signiﬁcance indicated by asterisks, p < 0.05 and *p < 0.001. d Whole bone marrow was used to quantify MyD88 mRNA levels usin RT-PCR and normalised to naive controls. e The proportion of monocytes (CD3-B220-Ter119−Ly6G−Ly6C+) in bone marrow was measured by ﬂow ytometry and normalised to those in Naive control mice. f, g Osteoclasts (OCs) were differentiated from bone marrow obtained at cull and cultured f days and size of OCs (f) was measured using ImageJ analysis software and normalised to those from Naive controls with representative images for eac isease stage in both treatment groups displayed (g). Statistics: data are presented as mean ± SEM values from individual animals (d; Naive; n = 4, PB = 8, ES-62; n = 6, Naive + ABX; n = 4, PBS + ABX; n = 8, ES-62 + ABX; n = 8. Discussion This study demonstrates that a deﬁned parasitic worm-derived product can impact on, and harness, the microbiome to exert its therapeutic effects against chronic inﬂammation in target organs 9 ARTICLE NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications Methods C ll i However, it is likely that at least in the dextran sodium sulphate (DSS) model of IBD that the severe physical damage to the intestinal epithelial barrier bypasses the capacity of ES-62 to normalise the microbiome, as the resultant leakage and aberrant bacterial colonisation would generate inﬂammation and pathology. In addition, loss of regulatory T-cell function appears necessary for pathogenesis in each of the T1D, MS and IBD models tested and ES-62 appears to lack the capacity to restore either Treg or Tr1 responses in all of the models of inﬂammatory diseases we have tested to date2,63. This is perhaps rather surprising given the well- documented induction of regulatory T cells by helminths2 and the capacity of T2-MZP Bregs, to induce Tr1 and Tregs64. However, we do not see signiﬁcant induction of T2-MZP Bregs, but rather a general upregulation of IL-10+ B cells in the CIA model21 where interestingly, we also fail to detect any impact of the microbiome on Tr1 responses. Thus, one possibility is that ES-62 only protects in models where (microbiota- regulated) Bregs can directly mediate immunoregulatory effects, such as CIA, MRL/Lpr-SLE and asthma models2. Consistent with this, whilst roles for Bregs have been proposed in MS and IBD (DSS) models, these appear to have been dependent on interac- tions with regulatory T cells and indeed, the IBD-T cell transfer models tested were performed in Rag1−/−mice that cannot develop B cells63. Flow cytometry. Spleen, lymph node (LN) and bone marrow (BM) cells were suspended in FACS buffer (2.5% BSA; 0.5 mM EDTA, in PBS) following red blood cell lysis (eBioscience, UK). BM cells were labelled with a cocktail of PE- labelled antibodies speciﬁc for CD3 (catalogue number: 100205), B220 (catalogue number: 103207) and Ter119 (catalogue number: 116207) to exclude analysis of lymphocytes and erythroid cell populations using a dump channel, and monocytes were identiﬁed by labelling with antibodies against CD11b (FITC; catalogue number: 101206), Ly6C (PerCP Cy5.5; catalogue number: 128011) and Ly6G (APC; catalogue number: 127613)20. Lymphocytes were labelled with antibodies speciﬁc for CD3 (FITC; catalogue number: 100305/6), CD19 (AF700; catalogue number: 115527/8), IL-10 (PE or APC; catalogue number: 505007/9) and IL-17 (PerCP5.5 or APC; catalogue number: 506915/6/9/20). For surface marker staining, antibodies were used at 0.2 µg/106 cells (1/100 dilution) except for anti- CD45, which was used at 1/200 dilution. For Fix/Perm staining of intracellular cytokines, 1 µg/106 cells (1/20 dilution) were used. Methods C ll i Collagen-induced arthritis. Male DBA/1 mice were purchased at 6–8 weeks of age (Envigo; Bicester, UK) and housed and maintained in the Central Research Facility of the University of Glasgow. All experiments were approved by, and conducted in accordance with, the Animal Welfare and Ethical Review Board of the University of Glasgow, UK Home Ofﬁce Regulations and Licences PPL P8C60C865, PIL I518666F7, PIL 1675F0C46 and PIL ICEBDB864. CIA was induced using bovine Collagen Type II (CII: 100 μg) emulsiﬁed with complete Freund’s adjuvant (MD Biosciences) injected intradermally on day 0. Mice were challenged with 200 μg of CII in PBS intraperitoneally on day 21. Animals were treated with PBS or puriﬁed endotoxin-free ES-62 (2 µg/injection) subcutaneously on days −2, 0 and 21, and joint inﬂammation and damage (articular score) were determined as described previously16,32,50. Grip strength was recorded as per the manufacturer’s instruction (Ugo Basile®, Italy) using a Gripometer, which measured the grip strength (peak force and time resistance) of the forelimbs of the mice. The animals were placed over a base plate and gripped a T-shaped grasping bar, which was connected to the peak ampliﬁer that automatically detects the animal’s response. Three measurements were taken and the average grip strength was calculated. In order to investigate the impact of gut microbiome perturbation on initiation and progression of inﬂammatory arthritis, animals were given drinking water containing [or not] a cocktail of antibiotics (500 mg/L Vancomycin, 1 g/L Neomycin and 1 g/L Metronidazole) to eliminate Gram-positive, Gram-negative and anaerobic microorganisms22 7 days prior to the induction of CIA and thereafter continuously throughout the experiment. Blood was sampled using endotoxin-free needles, and syringes and the resulting serum isolated and stored at −20 °C in endotoxin-free Eppendorf tubes. Paw, ileum and colon tissue was ﬁxed in 4% paraformaldehyde; ileum and colon faecal contents were collected in sterile RNAlater (Sigma, UK) and stored at −80 °C. Endotoxin- free ES-62 was puriﬁed from spent culture medium as described in detail previously32. Given that normalisation of the gut microbiome appears cen- tral to ES-62 resetting of immunoregulation in CIA, at ﬁrst sight it may appear rather counter-intuitive that the worm product fails to afford protection63 in mouse models of type-1 diabetes (T1D), multiple sclerosis (MS) and inﬂammatory bowel disease (IBD); conditions which have each been reported to exhibit patterns of microbiome dysbiosis. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Finally, TLR4/MyD88 signalling in RA had previously been attributed solely to recognition of DAMPs in the joint65 and thus collectively, our ﬁndings shed new light on its pathogenic roles in initiation and progression of disease as well as emphasise its potential as a therapeutic target in RA. In particular, they underscore the complex and central role of TLR4/MyD88 sig- nalling in regulating the gut-bone marrow axis in musculoskeletal homoeostasis and its dysregulation resulting in systemic inﬂam- mation, breaking of tolerance, aberrant osteoclastogenesis and consequently joint destruction in arthritis. Moreover, they suggest that ES-62 may achieve its protective effects in CIA by directly targeting this key regulatory node to normalise microbiome dysbiosis and associated gut pathology in order to rebalance the gut-bone marrow axis and limit aberrant inﬂammation and joint damage, by consequently homeostatically restoring levels of Bregs and resetting osteoclastogenesis. Thus, by exploiting ES-62 as a unique tool to dissect pathogenic and protective microbial sig- natures in CIA, we could potentially understand how to elicit homoeostatic regulation of the gut and resolve inﬂammation in autoimmune inﬂammatory arthritis. However, in terms of CIA, our current working model is that the induced dysbiosis may impact on MyD88-integrated micro- biota-sensing intestinal epithelial-dendritic cell interactions61,62 to drive gut pathology and loss of barrier integrity and to train inﬂammation and (auto)immune responses that in turn impact on the microbiome and perpetuate chronic pathology. ES-62 exploits TLR4 to sense (aberrant) MyD88 signalling2 allowing such changes in the microbiome to be detected by the helminth product and enabling it to dynamically stabilise intestinal epi- thelial barrier function and integrity and hence ﬁne-tune bacterial species diversity and abundance to homeostatically maintain a healthy microbiome. It is interesting therefore in this context that systemic Th17 differentiation and consequent autoimmune arthritis occurring in the IL1rn−/−model of RA is dependent on TLR49,30. Moreover, the accompanying dysbiosis, that on faecal transfer can confer arthritis-predisposing Th17 inﬂammation in wild-type mice, is also regulated by TLR430. Furthermore, inter- estingly, 11/44 taxa disrupted in IL1rn−/−mice were normalised in IL1rn−/−TLR4−/−animals and these included Ruminococcus species, which are also promoted by ES-6230. In mechanistic terms, Breg levels—reset by ES-62 in CIA— have also been reported to be regulated by the microbiome in other models of autoimmune arthritis22. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Indeed, consistent with the proposal that Bregs are homeostatically induced to resolve dysbiosis-induced inﬂammation in autoimmune arthritis, as indicated by disruption of the process by ABX treatment22, we have similarly found the ES-62-mediated restoration of IL-10+ B cells, as well as the suppression of pathogenic anti-CII antibody and IL-6 production, in CIA to be compromised by such per- turbation of the microbiome. Collectively, these ﬁndings suggest that by targeting MyD88 to maintain microbiome homoeostasis, ES-62 may subsequently achieve its immunoregulatory effects, which in turn, by dampening and resolving chronic inﬂamma- tion, further contribute to the normalisation of intestinal barrier integrity and the gut microbiome. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 f, g Osteoclasts (OCs) were differentiated from bone marrow obtained at cull and cultured for 5 days and size of OCs (f) was measured using ImageJ analysis software and normalised to those from Naive controls with representative images for each disease stage in both treatment groups displayed (g). Statistics: data are presented as mean ± SEM values from individual animals (d; Naive; n = 4, PBS; n = 8, ES-62; n = 6, Naive + ABX; n = 4, PBS + ABX; n = 8, ES-62 + ABX; n = 8. e Initiation; n = 6, Pre-Clinical; n = 3, Disease; n = 4) or mean ± SD from experimental replicates (f; Initiation; n = 18, Pre-Clinical; n = 9, Disease; PBS: n = 12 and ES-62; n = 15). Statistical signiﬁcance was determined using two- (e, f) or one-way (d) ANOVA with LSD Fishers multiple comparisons and indicated by asterisks, p < 0.05, p < 0.01 and p < 0.001 densities within a particular microbial niche. Thus, by coordi- nating gene expression responses both within and across species, quorum sensing molecules shape the microbial community and its interaction with the host, particularly with respect to virulence and pathogenesis;59,60 our future plans therefore encompass determining whether ES-62 can modulate bacterial growth directly as a starting point in investigating whether quorum- sensing activities contribute to its ability to normalise microbiome dysbiosis and maintain gut homoeostasis. ES-62 orchestrates such maintenance of the complex homo- eostasis of the gut-bone marrow axis are not clear. However, as ES-62 can modulate the gut microbiota in Naive healthy mice, it appears that such ﬁne-tuning of the microbiome can occur directly, rather than indirectly via potential anti-inﬂammatory and immunoregulatory actions that suppress gut pathology, to protect against dysbiosis. It is therefore possible that ES-62 may act in a manner analogous to quorum-sensing molecules59,60, which are produced by bacteria and signal to regulate population URE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 10 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 4767992.3, 4767991.3, 4767990.3, 4767989.3, 4767988.3, 4767987.3, 4767986.3, 4738191.3, 4738190.3, 4738025.3, 4737887.3, 4737053.3 and 4737052.3. qPCR was used to validate changes in bacterial populations using primers speciﬁc for Bacteroidetes (forward: GTTTAATTCGATGATACGCGAG, reverse: TTAAGCC GACACCTCACGG)71, Firmicutes (forward: GGAGCATGTGGTTTAATTCGAA GCA, reverse: AGCTGACGACAACCATGCAC)71, Butyrivibrio (forward: GCGAAGAAGTATTTCGGTAT, reverse: CCAACACCTAGTATTCATC)72, Proteobacteria (forward: TCGTCAGCTCGTGTCGTGA, reverse: CGTAAGGGC CATGATG)73, Enterobacteriaceae (forward: GTGCCAGCAGC CGCGGTAA, reverse: GCCTCAAGGGCACAACCTCCAAG)74, Lactobacillaceae (forward: TGGAAACAGGTGCTAATACCG, reverse: GTCCATTGTGGAAGATTCCC)75, segmented ﬁlamentous bacteria (forward: GACGCTGAGGCATGAGAGCAT, reverse: GACGGCACGG ATTGTTATTCA)76, Lachnospiraceae (forward: CGGTACCTGACTAAGAAGC, reverse: AGTTTCATTCTTGCGAACG)77 and were normalised to the total levels of bacteria using pan-bacterial primers (forward: CGGTGAATACGTTCCCGG, reverse: TACGGCTACCTTGTTACGACTT)77. Histology. Ileum, colon and joint (paw) tissues from individual mice in each treatment group were ﬁxed in 4% paraformaldehyde for 24 h before gut tissues were embedded in OCT and paw joints were decalciﬁed and subsequently parafﬁn embedded. Parafﬁn sections (6 μm) and OCT cryosections (9–10 μm) were pre- pared and standard H&E histological staining was performed on all tissues for identiﬁcation of morphological changes16,20,50. Ileum (villi thickness) and colon (number of lesions) pathology was quantitated by ImageJ analysis. Joint pathology was scored according to the grading system of 0 for no inﬂammation, 1 for mild inﬂammation, pannus formation and bone damage, up to a score of 4 representing a high level of inﬂammation, pannus inﬁltration and bone and cartilage destruc- tion, as previously described67. Osteoclast differentiation. OCs were differentiated from BM obtained from the hind limbs of experimental animals as previously described20. Brieﬂy, following removal of adherent cells, BM cells were cultured in αMEM medium supplemented with 30 ng/ml M-CSF and 50 ng/ml RANKL (Peprotech, London, UK) and then assessed for OC differentiation by TRAP staining (Leukocyte Acid Phosphatase Kit, Sigma-Aldrich, UK) on day 5. Images were obtained using an EVOS FL Auto Cell Imaging System. TRAP+ cells with ≥3 nuclei were enumerated, and ImageJ soft- ware was used to calculate the average size of multinucleated OCs per ﬁeld of view (FoV)20. Statistics. All data were analysed using GraphPad Prism 6 software using one- or two-way ANOVA with Fishers LSD post-tests for parametric data or Kruskal–Wallis test and Dunn’s post test for non-parametric data. Unsupervised hierarchical clustering with Euclidean distance was performed on the colon sam- ples using the heatmap.2 function of the gplots package in R. Supervised heatmaps were generated using GraphPad Prism 7 software. Indicators of signiﬁcance include p < 0.05, p < 0.01 and p < 0.001. Serum cytokine and antibody ELISAs. References and helminth parasites. Parasite Immunol. 38, 5–11 (2016). 7. Giacomin, P., Agha, Z. & Loukas, A. Helminths and intestinal ﬂora team up to improve gut health. Trends Parasitol. 32, 664–666 (2016). l b h l h g to improve gut health. Trends Parasitol. 32, 664–666 (2016). 8. Gause, W. C. & Maizels, R. M. Macrobiota - helminths as active participants and partners of the microbiota in host intestinal homeostasis. Curr. Opin. Microbiol. 32, 14–18 (2016). 9. Abdollahi-Roodsaz, S., Abramson, S. B. & Scher, J. U. The metabolic role of the gut microbiota in health and rheumatic disease: mechanisms and interventions. Nat. Rev. Rheumatol. 12, 446–455 (2016). Genomic DNA from the ileum and colon faecal matter was puriﬁed using QIAamp DNA Stool Mini Kit (Qiagen, Germany) and stored at −20 °C. For shotgun metagenomic analysis using the Ion Torrent PGM™platform, samples from three individual mice per group were pooled and between 10 and 100 ng of the pooled DNA was fragmented (NEB Fast DNA Fragmentation & Library Prep Set for Ion Torrent, NEB Inc, UK) and barcoded (IonXpress Barcode Adapters Kit, ThermoFisher Scientiﬁc, UK). Barcoded libraries were quantiﬁed using a Qubit Fluorometer (ThermoFisher Scientiﬁc, UK) and bioanalyser (High Sensitivity DNA analysis Kit, Agilent, UK). Up to three barcoded libraries were combined per Ion 316™Chip Kit v2 following library preparation using the Ion PGM™Hi-Q™View OT2 and Ion PGM™Hi-Q™View Sequencing Kits (ThermoFisher Scientiﬁc, UK). Data were extracted as FASTQ ﬁles and analysed using MG-RAST to generate taxonomic data from sequencing reads70. The number of reads per phylum, class, order, family, genera or species of interest were expressed as a composition of all bacteria present to normalise for variation between sequencing runs. Sequencing runs can be accessed using MG-RAST IDs; mgm4777616.3, 4777615.3, 4777614.3, 4777613.3, 4777481.3, 4777480.3, 4777479.3, 4777478.3, 4767994.3, 4767993.3, 10. Clemente, J. C., Manasson, J. & Scher, J. U. The role of the gut microbiome in systemic inﬂammatory disease. BMJ (Clin. Res. Ed.) 360, j5145 (2018). 11. Grencis, R. K., Humphreys, N. E. & Bancroft, A. J. Immunity to gastrointestinal nematodes: mechanisms and myths. Immunol. Rev. 260, 183–205 (2014). 12. Reynolds, L. A. et al. Commensal-pathogen interactions in the intestinal tract: lactobacilli promote infection with, and are promoted by, helminth parasites. Gut Microbes 5, 522–532 (2014). 13. Panda, A. K., Ravindran, B. & Das, B. K. Data availability The data supporting the ﬁndings are all contained within the article and Supplementary Information ﬁles. In addition, the metagenomic data are available from the MG-RAST database using the accession numbers provided in the Methods section. Other primary data ﬁles are available from the corresponding authors on reasonable request. qRT-PCR. BM cells (106) were lysed in RLT lysis buffer prior to mRNA extraction using RNeasy Plus Mini kit (Qiagen, Germany) according to the manufacturer’s instructions. The High Capacity cDNA Reverse Transcriptase kit (Applied Bio- systems, Life Technology, UK) was used to generate cDNA for use with StepOne Plus™real-time PCR system (Applied Biosystems, UK) and KiCqStart® qPCR Ready Mix (Sigma-Aldrich). Pre-designed KiCqStart™primers (Sigma-Aldrich) were purchased to evaluate RANK (tnfrsf11a; forward: GAAATAAGGA GTCCTCAGGG, reverse: GAAATAAGGAGTCCTCAGGG), OPG (tnfrsf11b; forward: GAAGATCATCCAAGACATTGAC, reverse: TCCTCCATAAACTG AGTAGC), MyD88 (myd88; forward: GAAGATCATCCAAGACATTGAC, reverse: TCCTCCATAAACTGAGTAGC) and β-actin (actb; forward: GATGTAT GAAGGCTTTGGTC, reverse: TGTGCACTTTTATTGGTCTC). Data were nor- malised to the reference gene β-actin to obtain the ΔCT values that were used to calculate the fold change from the ΔΔCT values following normalisation to bio- logical control group. Received: 23 July 2018 Accepted: 7 March 2019 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 Interleukin-6 (IL-6) and IL-10 expression was measured by ELISA according to the manufacturer’s instructions (BD Bios- ciences, Oxford, UK). For determination of collagen type II (CII)-speciﬁc IgG1 and IgG2a antibodies in serum32, high-binding 96-well ELISA plates were coated with CII (5 µg/ml) overnight at 4 °C before washing and blocking with BSA/PBS. Serum was diluted 1:100 and then serially diluted threefold until 1:218,700 and incubated with HRP-conjugated goat anti-mouse IgG1 or IgG2a (1:10,000) in 10% FBS/PBS prior to developing with TMB and 2 M sulphuric acid and read at an optical density of 450 nm. Reporting Summary. Further information on experimental design is available in the Nature Research Reporting Summary linked to this article. References 1. Harnett, W. Secretory products of helminth parasites as immunomodulators. Mol. Biochem. Parasitol. 195, 130–136 (2014). 1. Harnett, W. Secretory products of helminth parasites as immunomodulators. Mol. Biochem. Parasitol. 195, 130–136 (2014). 2. Harnett, M. M. & Harnett, W. Can parasitic worms cure the modern world’s ills? Trends Parasitol. 33, 694–705 (2017). 2. Harnett, M. M. & Harnett, W. Can parasitic worms cure the modern world’s ills? Trends Parasitol. 33, 694–705 (2017). ( ) 3. Bach, J. F. The hygiene hypothesis in autoimmunity: the role of pathogens and commensals. Nat. Rev. Immunol. 18, 105–120 (2018). 4. de Ruiter, K. et al. Helminths, hygiene hypothesis and type 2 diabetes. Parasite Immunol. 39, https://doi.org/10.1111/pim.12404 (2017). ( ) 3. Bach, J. F. The hygiene hypothesis in autoimmunity: the role of pathogens and commensals. Nat. Rev. Immunol. 18, 105–120 (2018). Metagenomics. Given the well-documented variation amongst individual animals that typically exhibit differential scores/affected joints in the CIA model and the range of environmental parameters impacting on the microbiome68,69, we attempted to minimise variation by pooling samples from three mice of similar disease score/group from each independent CIA model rather than analysing each individual mouse. Furthermore, pooling of samples also reduces variation in sample and library preparation, whilst barcoding of these pooled samples allowed analysis of the different treatment groups on a single chip and provided internal controls for comparison of treatment groups within each experiment. This process was repeated for three independent CIA models (involving nine repre- sentative naive, PBS-CIA [articular score: 7.00 ± 0.91] and ES-62-CIA [articular score: 0.50 ± 0.33, where **p < 0.001 for the PBS-CIA versus naive and ES-62-CIA cohorts] mice overall), further reducing the impact of disease score and individual variation. , ( ) 4. de Ruiter, K. et al. Helminths, hygiene hypothesis and type 2 diabetes. Parasite Immunol. 39, https://doi.org/10.1111/pim.12404 (2017). 5. Blander, J. M., Longman, R. S., Iliev, I. D., Sonnenberg, G. F. & Artis, D. Regulation of inﬂammation by microbiota interactions with the host. Nat. Immunol. 18, 851–860 (2017). 6. Zaiss, M. M. & Harris, N. L. Interactions between the intestinal microbiome and helminth parasites. Parasite Immunol. 38, 5–11 (2016). 6. Zaiss, M. M. & Harris, N. L. Interactions between the intestinal microbiome and helminth parasites. Parasite Immunol. 38, 5–11 (2016). 7. Giacomin, P., Agha, Z. & Loukas, A. Helminths and intestinal ﬂora team up to improve gut health. Trends Parasitol. 32, 664–666 (2016). Methods C ll i Gating strategies are shown in Supplementary Figure 3. All antibodies were purchased from BioLegend, UK. Fixable viability stain (APC-ef780; ThermoFisher Scientiﬁc, UK) was used to select for live cells and for analysis of IL-10+ regulatory B cells (Bregs), lympho- cytes were stimulated with PMA, ionomycin, Brefeldin A and LPS as described previously18,21. Data were acquired using a FACS Canto or BD LSRII ﬂow cyt- ometer and populations were gated as described previously using isotype and ﬂuorescence minus one (FMO) controls using FlowJo, LLC analysis software (Tree Star/BD)18,20,21,50,66. 11 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications URE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-019-09361-0 type 1 cells and reduction of collagen-induced arthritis. Arthritis Res. Ther. 14, R32 (2012). type 1 cells and reduction of collagen-induced arthritis. Arthritis Res. Ther. 14, R32 (2012). 15. Melendez, A. J. et al. Inhibition of FcepsilonRI-mediated mast cell responses by ES-62, a product of parasitic ﬁlarial nematodes. Nat. Med. 13, 1375–1381 (2007). 44 44. Mathewson, N. D. et al. Gut microbiome-derived metabolites modulate intestinal epithelial cell damage and mitigate graft-versus-host disease. Nat. Immunol. 17, 505–513 (2016). 16. Pineda, M. A. et al. The parasitic helminth product ES-62 suppresses pathogenesis in collagen-induced arthritis by targeting the interleukin-17-producing cellular network at multiple sites. Arthritis Rheum. 64, 3168–3178 (2012). 45. Jubair, W. K. et al. Modulation of inﬂammatory arthritis in mice by gut microbiota through mucosal inﬂammation and autoantibody generation. Arthritis Rheumatol. 70, 1220–1233 (2018). p 17. Rzepecka, J. et al. The helminth product, ES-62, protects against airway inﬂammation by resetting the Th cell phenotype. Int. J. Parasitol. 43, 211–223 (2013). 46. Novince, C. M. et al. Commensal gut microbiota immunomodulatory actions in bone marrow and liver have catabolic effects on skeletal homeostasis in health. Sci. Rep. 7, 5747 (2017). 18. Rodgers, D. T. et al. The parasitic worm product ES-62 targets myeloid differentiation factor 88-dependent effector mechanisms to suppress antinuclear antibody production and proteinuria in MRL/lpr mice. Arthritis Rheumatol. 67, 1023–1035 (2015). 47. Charles, J. F., Ermann, J. & Aliprantis, A. O. The intestinal microbiome and skeletal ﬁtness: Connecting bugs and bones. Clin. Immunol. 159, 163–169 (2015). 19. Coltherd, J. C. et al. The parasitic worm-derived immunomodulator, ES-62 and its drug-like small molecule analogues exhibit therapeutic potential in a model of chronic asthma. Sci. Rep. 6, 19224 (2016). 48. Christ, A. et al. Western diet triggers NLRP3-dependent innate immune reprogramming. Cell 172, 162–175 e114 (2018). p 20. Doonan, J. et al. Protection against arthritis by the parasitic worm project ES- 62, and its drug-like small molecule analogues, is associated with inhibition of osteoclastogenesis. Front. Immunol. 9, 1016 (2018). 49. Pineda, M. A., Eason, R. J., Harnett, M. M. & Harnett, W. From the worm to the pill, the parasitic worm product ES-62 raises new horizons in the treatment of rheumatoid arthritis. Lupus 24, 400–411 (2015). 21. Rodgers, D. T. et al. ARTICLE Protection against collagen-induced arthritis in mice afforded by the parasitic worm product, ES-62, is associated with restoration of the levels of interleukin-10−producing B cells and reduced plasma cell inﬁltration of the joints. Immunology 141, 457–466 (2014). 50. Pineda, M. A., Rodgers, D. T., Al-Riyami, L., Harnett, W. & Harnett, M. M. ES-62 protects against collagen-induced arthritis by resetting interleukin-22 toward resolution of inﬂammation in the joints. Arthritis Rheumatol. 66, 1492–1503 (2014). 51. Harnett, M. M. et al. From Christian de Duve to Yoshinori Ohsumi: more to autophagy than just dining at home. Biomed. J. 40, 9–22 (2017). 22. Rosser, E. C. et al. Regulatory B cells are induced by gut microbiota-driven interleukin-1beta and interleukin-6 production. Nat. Med. 20, 1334–1339 (2014). p 23. Buigues, C. et al. Effect of a prebiotic formulation on frailty syndrome: a randomized, double-blind clinical trial. Int. J. Mol. Sci. 17, E932 (2016). 52. Li, J., Wang, X., Zhang, F. & Yin, H. Toll-like receptors as therapeutic targets for autoimmune connective tissue diseases. Pharmacol. Ther. 138, 441–451 (2013). 24. Celis-Morales, C. A. et al. Associations of grip strength with cardiovascular, respiratory, and cancer outcomes and all cause mortality: prospective cohort study of half a million UK Biobank participants. BMJ (Clinical research ed) 361, k1651 (2018). 53. Maeda, Y. et al. Dysbiosis contributes to arthritis development via activation of autoreactive T cells in the intestine. Arthritis Rheumatol. 68, 2646–2661 (2016). 25. Charles-Schoeman, C. Cardiovascular disease and rheumatoid arthritis: an update. Curr. Rheumatol. Rep. 14, 455–462 (2012). 54. Zhang, X. et al. The oral and gut microbiomes are perturbed in rheumatoid arthritis and partly normalized after treatment. Nat. Med. 21, 895–905 (2015). 26. Brown, J. M. & Hazen, S. L. The gut microbial endocrine organ: bacterially derived signals driving cardiometabolic diseases. Annu. Rev. Med. 66, 343–359 (2015). 55. Chen, J. et al. An expansion of rare lineage intestinal microbes characterizes rheumatoid arthritis. Genome Med. 8, 43 (2016). 56. Mizuno, M., Noto, D., Kaga, N., Chiba, A. & Miyake, S. The dual role of short fatty acid chains in the pathogenesis of autoimmune disease models. PLoS ONE 12, e0173032 (2017). 27. Grigg, J. B. & Sonnenberg, G. F. Host-microbiota interactions shape local and systemic inﬂammatory diseases. J. Immunol. 198, 564–571 (2017). 57. Burger-van Paassen, N. et al. The regulation of intestinal mucin MUC2 expression by short-chain fatty acids: implications for epithelial protection. Biochem. J. 420, 211–219 (2009). y y 28. ARTICLE Zeng, H. & Chi, H. Metabolic control of regulatory T cell development and function. Trends Immunol. 36, 3–12 (2015). 29. Tamanai-Shacoori, Z. et al. Roseburia spp.: a marker of health? Future Microbiol. 12, 157–170 (2017). 58. Rahman, M. M. et al. Two histone deacetylase inhibitors, trichostatin A and sodium butyrate, suppress differentiation into osteoclasts but not into macrophages. Blood 101, 3451–3459 (2003). 30. Rogier, R. et al. Aberrant intestinal microbiota due to IL-1 receptor antagonist deﬁciency promotes IL-17- and TLR4-dependent arthritis. Microbiome 5, 63 (2017). p g 59. Verbeke, F. et al. Peptides as quorum sensing molecules: measurement techniques and obtained levels in vitro and in vivo. Front. Neurosci. 11, 183 (2017). 31. Rogier, R. et al. Alteration of the intestinal microbiome characterizes preclinical inﬂammatory arthritis in mice and its modulation attenuates established arthritis. Sci. Rep. 7, 15613 (2017). 60. Abisado, R. G., Benomar, S., Klaus, J. R., Dandekar, A. A. & Chandler, J. R. Bacterial quorum sensing and microbial community interactions. mBio 9, e01749–18 (2018). p 32. McInnes, I. B. et al. A novel therapeutic approach targeting articular inﬂammation using the ﬁlarial nematode-derived phosphorylcholine- containing glycoprotein ES-62. J. Immunol. 171, 2127–2133 (2003). 61. Rescigno, M. The intestinal epithelial barrier in the control of homeostasis and immunity. Trends Immunol. 32, 256–264 (2011). 33. Harnett, M. M. et al. The phosphorycholine moiety of the ﬁlarial nematode immunomodulator ES-62 is responsible for its anti-inﬂammatory action in arthritis. Ann. Rheum. Dis. 67, 518–523 (2008). 62. Rescigno, M. Dendritic cell-epithelial cell crosstalk in the gut. Immunol. Rev. 260, 118–128 (2014). 63. Doonan, J. et al. Failure of the anti-inﬂammatory parasitic worm product ES-62 to provide protection in mouse models of type i diabetes, multiple sclerosis, and inﬂammatory bowel disease. Molecules 23, https://doi.org/ 10.3390/molecules23102669 (2018). 34. Maseda, D., Bonami, R. H. & Crofford, L. J. Regulation of B lymphocytes and plasma cells by innate immune mechanisms and stromal cells in rheumatoid arthritis. Expert Rev. Clin. Immunol. 10, 747–762 (2014). 64. Rosser, E. C. & Mauri, C. Regulatory B cells: origin, phenotype, and function. Immunity 42, 607–612 (2015). p 35. Lucas, S. et al. Short-chain fatty acids regulate systemic bone mass and protect from pathological bone loss. Nat. Commun. 9, 55 (2018). from pathological bone loss. Nat. Commun. 9, 55 (2018). 36. Sjogren, K. et al. The gut microbiota regulates bone mass in mice. J. Bone Mineral. Res.: Off. J. Am. Soc. Bone Mineral. Res. ARTICLE 27, 1357–1367 (2012). 65. Abdollahi-Roodsaz, S., van de Loo, F. A. & van den Berg, W. B. Trapped in a vicious loop: Toll-like receptors sustain the spontaneous cytokine production by rheumatoid synovium. Arthritis Res. Ther. 13, 105 (2011). 37. Jin, C., Henao-Mejia, J. & Flavell, R. A. Innate immune receptors: key regulators of metabolic disease progression. Cell. Metab. 17, 873–882 (2013). y y 66. Lumb, F. E. et al. Dendritic cells provide a therapeutic target for synthetic small molecule analogues of the parasitic worm product, ES-62. Sci. Rep. 7, 1704 (2017). 38. Versini, M., Jeandel, P. Y., Rosenthal, E. & Shoenfeld, Y. Obesity in autoimmune diseases: not a passive bystander. Autoimmun. Rev. 13, 981–1000 (2014). 67. Rzepecka, J. et al. Prophylactic and therapeutic treatment with a synthetic analogue of a parasitic worm product prevents experimental arthritis and inhibits IL-1β production via NRF2-mediated counter-regulation of the inﬂammasome. J. Autoimmun. 60, 59–73 (2015). 39. Kim, J. J. & Khan, W. I. Goblet cells and mucins: role in innate defense in enteric infections. Pathogens 2, 55–70 (2013). 40. Golan, L., Gonen, E., Yagel, S., Rosenshine, I. & Shpigel, N. Y. Enterohemorrhagic Escherichia coli induce attaching and effacing lesions and hemorrhagic colitis in human and bovine intestinal xenograft models. Dis. Model Mech. 4, 86–94 (2011). 68. Reese, A. T. & Dunn, R. R. Drivers of microbiome biodiversity: a review of general rules, feces, and ignorance. mBio 9, e01294–18 (2018). 41. Arpaia, N. et al. Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature 504, 451–455 (2013). 69. Franklin, C. L. & Ericsson, A. C. Microbiota and reproducibility of rodent models. Lab. Anim. (NY) 46, 114–122 (2017). 70. Meyer, F. et al. The metagenomics RAST server - a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC Bioinforma. 9, 386 (2008). 42. Kaisar, M. M. M., Pelgrom, L. R., van der Ham, A. J., Yazdanbakhsh, M. & Everts, B. Butyrate conditions human dendritic cells to prime type 1 regulatory T cells via both histone deacetylase inhibition and G protein- coupled receptor 109A signaling. Front. Immunol. 8, 1429 (2017). g y y coupled receptor 109A signaling. Front. Immunol. 8, 1429 (201 71. Yang, Y. W. et al. Use of 16S rRNA gene-targeted group-speciﬁc primers for real-time PCR Analysis of predominant bacteria in mouse feces. Appl. Environ. Microbiol. 81, 6749–6756 (2015). 43. Carter, N. A., Rosser, E. C. & Mauri, C. References Rheumatoid arthritis patients are free of ﬁlarial infection in an area where ﬁlariasis is endemic: comment on the article by Pineda et al. Arthritis Rheum. 65, 1402–1403 (2013). 14. Panda, A. K. & Das, B. K. Absence of ﬁlarial infection in patients of systemic lupus erythematosus (SLE) in ﬁlarial endemic area: a possible protective role. Lupus 23, 1553–1554 (2014). 12 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunicatio NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunicatio Acknowledgements The work was funded by an award to M.M.H., W.H. and P.A.H. from Arthritis Research UK (21133). The work was funded by an award to M.M.H., W.H. and P.A.H. from Arthritis Research UK (21133). Competing interests: The authors declare no competing interests. Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ 74. Hammer, A. M. et al. The effects of alcohol intoxication and burn injury on the expression of claudins and mucins in the small and large intestines. Shock (Augusta, Ga) 45, 73–81 (2016). Journal peer review information: Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. g 75. Rohani, M., Noohi, N., Talebi, M., Katouli, M. & Pourshaﬁe, M. R. Highly heterogeneous probiotic Lactobacillus species in healthy iranians with low functional activities. PLoS ONE 10, e0144467 (2015). Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 76. Geuking, M. B. et al. Intestinal bacterial colonization induces mutualistic regulatory T cell responses. Immunity 34, 794–806 (2011). 77. Rinttila, T., Kassinen, A., Malinen, E., Krogius, L. & Palva, A. Development of an extensive set of 16S rDNA-targeted primers for quantiﬁcation of pathogenic and indigenous bacteria in faecal samples by real-time PCR. J. Appl. Microbiol. 97, 1166–1177 (2004). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. ARTICLE Interleukin-10 produced by B cells is crucial for the suppression of Th17/Th1 responses, induction of T regulatory 13 Additional information 72. Boeckaert, C. et al. Accumulation of trans C18:1 fatty acids in the rumen after dietary algal supplementation is associated with changes in the Butyrivibrio community. Appl. Environ. Microbiol. 74, 6923–6930 (2008). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 019-09361-0. y pp 73. Bacchetti De Gregoris, T., Aldred, N., Clare, A. S. & Burgess, J. G. Improvement of phylum- and class-speciﬁc primers for real-time PCR quantiﬁcation of bacterial taxa. J. Microbiol. Methods 86, 351–356 (2011). Author contributions J.D., A.T., M.A.P., F.L., J.C. and A.M.K. performed the experiments for the study designed by M.M.H., W.H. and P.A.H. J.D. and F.L. manufactured ES-62. M.M.H., W.H. and J.D. wrote the paper and all authors were involved in reviewing and revising the paper and have approved the ﬁnal version. NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunications 14 NATURE COMMUNICATIONS \| (2019) 10:1554 \| https://doi.org/10.1038/s41467-019-09361-0 \| www.nature.com/naturecommunicatio
https://openalex.org/W2134439625	https://pssjournal.biomedcentral.com/track/pdf/10.1186/1754-9493-8-2	English	null	Hospital costs associated with surgical morbidity after elective colorectal procedures: a retrospective observational cohort study in 530 patients	Patient safety in surgery	2,014	cc-by	6,122	Abstract Background: Postoperative complications contribute to morbidity and mortality. This study assessed the impact of surgical complications on healthcare resource utilization for patients undergoing elective colorectal procedures. Method: Data were obtained on 530 consecutive colorectal operations performed from January 2010 to January 2011. Patient demographics, type of procedure, surgical complications classified as Clavien 1–5, length of stay, 60-day readmission rate, and hospital costs were recorded. Results: Seventy-five percent of the operations were associated with malignancy, and 26% were pelvic procedures. Thirty-five percent of the patients developed at least one complication, 21% of the complications did not require intervention. The readmission rate was 7.4%. Nine patients died during 60-day post discharge follow up.Median length of stay was 9 (3–34) days in uncomplicated and 16 (4–205) days in complicated cases. Occurrence of any complication at index admission increased total hospital costs 2.1-fold (EUR 25,680 vs. EUR 12,405), with the largest cost differential attributed to wound dehiscence and/or suture line failure requiring reoperation. These increases were primarily due to prolonged hospitalization and ICU expenditures. Readmission resulted in a further increase to an average cost of EUR 12,585 per re-admitted patient.Multivariate analysis showed that BMI > 25, obesity, operation complexity and surgeon significantly affected the risk for complication. Also, hospital costs were significantly increased by any postoperative complications, reoperations, high complexity of surgical procedures and high comorbidity index. Conclusions: Reducing morbidity after colorectal procedures improves quality of care and patient safety, and may also substantially reduce hospital costs and increase the efficiency of resource utilization. Keywords: Elective colorectal procedures, Outcomes, Costs Keywords: Elective colorectal procedures, Outcomes, Costs © 2014 Zoucas and Lydrup; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. * Correspondence: Evita.Zoucas@med.lu.se Department of Surgery, Skåne University Hospital, Lund/Malmö, Sweden Hospital costs associated with surgical morbidity after elective colorectal procedures: a retrospective observational cohort study in 530 patients Evita Zoucas* and Marie-Louise Lydrup Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Open Access Study population and methods We conducted an observational retrospective cohort study at Skåne University Hospital, in southern Sweden. The hos- pital serves as a tertiary care referral center with a catch- ment of 2,1 miljon (approx. 1/5 of the total population). The colorectal unit is by case volume the largest in the country. Data from the operative database for elective colo- rectal procedures (ORBIT) was merged with the hospital internal accounting data and the hospital internal account- ing database (FINN). All in- and out-patient records from January 2010 to April 2011 were systematically reviewed. All patients who underwent elective laparoscopic (5 cases) or open abdominal colorectal procedures during the period 1 January 2010 to 31 December 2010 were included. The patients were included in the unit’s standard enhanced re- covery after surgery clinical care protocol which is based on the concensus protocol described by Fearon et al. [9]. Colo- rectal cases involving out-patient procedures during that period were excluded. Total hospital costs were regarded as estimated ex- penses accrued from indirect and direct patient care during a hospitalization period. Direct costs comprised all costs for providing the care, such as physician and nursing staff salaries and expenditures on medications and other medical supplies. Overhead costs included the financing of administrative infrastructure and fixed facilities. Clinical data on patient demographic characteristics, comorbid status, disease, and type of surgical procedure are shown in Table 1. The patients were stratified into four groups according to their body mass index (BMI), considering values > 25 to indicate overweight and > 30 obesity. Comorbid status was assessed using the Deyo- Charlson index [10]. Index surgical procedures were ar- bitrarily divided into four groups based on the complex- ity of each procedure. Operating time, blood loss, and attending surgeon (14 specialists in total) were noted, along with length of hospital stay, number and type of complications, unplanned return to surgery (reopera- tion), and readmission. Complications occurring during The estimation of cost was based on the following units: the number of days for in-hospital care, total mi- nutes under aneasthesia and total hours in the post- op- erative care unit or/and ICU. In the present system costs for salaries, medication and medical supplies were incor- porated in the above units. Total costs accrued during the index hospitalization period and during readmission were recorded separately. Introduction increased health care costs and resource utilization [6,7]. In support of those observations, Soop et al. [8] noted that ad- verse effects were common and caused significant con- sumption of healthcare resources in Swedish hospitals, although hospital costs per se were not reported. Frequent peri-operative complications as infections and haemorrhage are potentially preventable, thus the control of cost in surgi- cal patients may be inherently associated with provider outcomes. Increasing public scrutiny of the quality of care provided by hospitals has prompted studies of the documentation, man- agement, and prevention of complications [1-3]. In that context, safety and quality have become prominent criteria in the evaluation of surgical care. The incidence of postop- erative complications in patients subjected to colorectal surgery has been shown to vary between 17% and 31% in investigations of both elective and emergent procedures [4,5]. In recent reports, postoperative complications The aim of the present review was to analyze the inci- dence, nature, and severity of postoperative complications after elective colorectal procedures performed at a tertiary Page 2 of 7 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 index admission until the patient was discharged or dur- ing 60-day post-operative follow-up for those discharged earlier were analyzed. The severity of each complication was classified according to Clavien et al. [11]. The most severe complications were those resulting in death (grade V). The severity of all other complications was defined by the morbidity inflicted. Complications were thus graded I (any deviation from normal postoperative course without the need for pharmacological treatment or intervention), II (requiring pharmacological treatment with eg antibiotics, b-blockers or blood tranfusions and total perenteral nutrition), IIIa (requiring surgical, endo- scopic or radiological intervention without general aneasthesia), IIIb (intervention under general aneasthe- sia) and IV (life-threatening, ICU) (Table 2). care center, and to explore the association between surgical outcomes and hospital costs. Study population and methods Costs for the two patients with Clavien grade I minor complications did not significantly affect total cost, while Clavien II complications although minor prolonged in- hospital care and thus affected costs. We analyzed ex- penditures on the following: nursing care in hospital wards, operative procedures, postoperative and intensive care units per index hospitalization period. Further wards, operative procedures, postoperative and intensive care units per index hospitalization period. Further Table 1 Patient characteristics (n = 530) Age, median (range) 68 (18–97) Female gender, n (%) 251 (47.2) BMI, n (%) < 20 39 (7.4) 20–25 280 (53.8) 26–30 158 (29.8) > 30 52 (9,8) Comorbidity index (Deyo-Charlson), median (range) 2 (0–7) Malignancy, n (%) 397 (75) Surgical complexity, n (%) (1) Small bowel, colon resection, colectomy 280 (52.8) (2) Sigmoid, recto-sigmoid resection 81 (15.3) (3) Rectum (resection, amputation) 144 (21.2) (4) Multivisceral resection 25 (4.7) Table 2 Outcome after colorectal surgery Outcome N Incidence (%) No complication 344 (64.9) Any complication 186 (35.1) Clavien grade I 2 II 37 IIIa 65 IIIb 56 IV 20 60-day mortality 9 (1.7) Surgery-related deaths 6 (1.1) Reoperation 55 (10.4) 60-day readmission 41 (7.4) Table 1 Patient characteristics (n = 530) Age, median (range) 68 (18–97) Female gender, n (%) 251 (47.2) BMI, n (%) < 20 39 (7.4) 20–25 280 (53.8) 26–30 158 (29.8) > 30 52 (9,8) Comorbidity index (Deyo-Charlson), median (range) 2 (0–7) Malignancy, n (%) 397 (75) Surgical complexity, n (%) (1) Small bowel, colon resection, colectomy 280 (52.8) (2) Sigmoid, recto-sigmoid resection 81 (15.3) (3) Rectum (resection, amputation) 144 (21.2) (4) Multivisceral resection 25 (4.7) Table 2 Outcome after colorectal surgery Outcome N Incidence (%) No complication 344 (64.9) Any complication 186 (35.1) Clavien grade I 2 II 37 IIIa 65 IIIb 56 IV 20 60-day mortality 9 (1.7) Surgery-related deaths 6 (1.1) Reoperation 55 (10.4) 60-day readmission 41 (7.4) Table 2 Outcome after colorectal surgery Page 3 of 7 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 subgroup analysis of costs was performed for the most common surgical complications. involving wound disruption, suture line failure, deep ab- scesses, or return to the operating theater. The presence of complications led to substantially elevated hospital costs (Table 3). Study population and methods Incrementally, increases in the costs of ward care and ICU were greater than increases in oper- ating room costs, representing longer care in surgical wards (days) or the ICU (hours) than length of operative procedure (minutes under aneaesthesia). Ward costs were raised 6- fold by failure of suture lines and 4,5-fold by wound dehiscence or deep wound infection, whereas increases in operation costs for those reasons were 2,5- and 1,5-fold, respectively (Table 4). Statistical analysis Descriptive statistics are shown as medians (interquartile range), means ± SE, and frequencies or percentages (95% Confidence interval) when appropriate. Comparisons of groups were carried out using Mann-Whitney’s test or Kruskal- Wallis test. Differences were considered signifi- cant if p < 0.05. Multivariate logistic regression was con- ducted when the response variable was the presence or absence of complication. The following model variables were pre-specified potential predictors: age, gender, BMI, comorbidity index, malignancy, complexity of pro- cedure, and attending surgeon.SPSS (SPSS Inc.USA) software was used for the statistical analyses. After adjustment for age, gender, and preoperative co- morbidity, multivariate analysis showed that surgical outcomes were independently associated with over- weight (BMI > 25), obesity and high complexity of the surgical procedures (Table 5). Outcomes of individual surgeons were significantly related to the frequency of complications in the above multivariate analysis (p < 0,0083) with odds ratio for complications varying be- tween 4,02 (95% confidence interval 1,59-10,10) and 0,51 (95% confidence interval 0,16-1,63). Statistical ana- lysis of variance also demonstrated that accumulation of costs was correlated with the comorbidity index, the complexity of surgical procedures, the presence of com- plications, and the need for reoperation (Table 6). More- over, complications remained associated with increased length of hospital stay and costs even after adjusting for surgical complexity. Results Five hundred thirty patients with a median age of 68 years underwent open or laparoscopic colorectal procedures dur- ing the 12-month study period. The majority of the patients were male, 29,8% were overweight and 9,8% were obese, but in general had few comorbid conditions. Three quarters of all surgical procedures were associated with malignancy, and 26% were pelvic procedures, which are considered to be major surgery (Table 1). One hundred eighty-six patients (35.1%) experienced at least one postoperative complica- tion, and the majority of those conditions (76%) were classi- fied as Clavien IIIa–IV . More than 40% of complications were severe (Clavien IIIb–IV), and 99% of all complications (Clavien II–IV) led to prolonged hospitalization. There were six surgery-related deaths, and a further three deaths occurred during the study period. The rates of reoperation and readmission were 10.4% and 7.4%, respectively (Table 2). The patients who were readmitted required fur- ther 465 days of hospital care at a total cost of EUR 515,899 (EUR 12,585/patient). Discussion In the present review, we found that postoperative com- plications were common after elective colorectal proce- dures, occurring in 35.1% of the patients who underwent abdominal surgery. To our knowledge, this is the first Swedish study to demonstrate the direct correlation Table 3 Outcome, length of stay, and total costs by type of complication Outcome N Length of stay in days Total costs (EUR) Median (range) Mean + _ SE No complication 344 9 (3–34) 12410 ± 384 Any complication 186 16 (4–205) 25680 ± 2289 Wound disruption 23 23* (7–50) 29230 ± 3022* Suture/Staple line dehiscence 21 30* (6–205) 47306 ±17194* Superficial site infection 30 15,5* (4–43) 21406 ± 2497* Deep wound infection 28 22.5* (6–8) 27145 ± 3118* Reoperation 55 28* (6–205) 38050 ± 2427* Surgery-related deaths 6 19* (5–52) 25102 ± 5985* p < 0.05 Mann–Whitney test. The most frequent complications were wound disrup- tion (23) and superficial (30) or deep (28) wound infec- tions. Failure of suture or staple line occurred in 21 of the 425 patients carrying such. One of the later cases was associated with a superficial wound infection and six with deep wound infections. Perioperative bleeding occurred in 13 cases. There were 32 urinary complica- tions (including urinary tract infection and bladder dys- function), 18 cardiovascular complications, two cases of pneumonia, two cases of graft necrosis, and five cases of central vein catheter sepsis. Bowel dysfunction or ob- struction was noted in 18 patients following index pro- cedure. Operation time for patients who developed complications and those without complications did not differ statistically. Discussion Table 3 Outcome, length of stay, and total costs by type of complication Length of stay was prolonged by 78% following com- plications and was prolonged even further in cases Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Page 4 of 7 Table 4 Analysis of costs by type of complication No complication Any complication Wound disruption Deep wound infection Superficial wound infection Suture line failure n = 346 n = 184 n = 23 n = 28 n = 30 n = 21 EUR EUR EUR EUR EUR EUR Ward care 3842 ± 118 9458 ± 1200 17991 ± 8043* 16531 ± 6142* 6647 ± 642* 23586 ± 9270* Surgery 5144 ± 125 8218 ± 467* 8385 ± 841* 8790 ± 932* 6723 ± 563* 13836 ± 3016* Postoperative/ ICU 754 ± 66 1949 ± 264* 2061 ± 564* 1569 ± 338* 1310 ± 30* 2022 ± 398* Values represent mean ± SE. p < 0.05 Mann–Whitney test. Table 4 Analysis of costs by type of complication to data published by Hawn et al. [19] and Tang et al. [20], and compare favourably to values reported in sev- eral earlier publications [21-23]. Male sex, obesity, spe- cific surgeons or hospitals, ostomy and total/subtotal colectomy were identified as risk factors for surgical site infection after colorectal surgery [22-25]. The circum- script number of cases present in our cohort did not allow subgroup analysis with regard to any technical as- pects associated with SSI. between surgical outcome and hospital costs in a con- secutive series of elective colorectal operations. The most frequent complications recorded as surgical site in- fection, wound disruption, suture line failure as well as peri-operative haemorrhage might be considered pro- vider related and potentially avoidable. Elective colorectal surgery has been shown to have a low mortality rate but a high morbidity rate ranging from 9% to 31%. Notably, use of fast-track protocols in colorectal cases has been found to improve the outcome, seen as a 50% reduction in the number of patients with complications after surgery [12-14]. Despite a target length of stay of 3 to 5 days in fast-track protocols, such as the one applied currently, patients without complica- tions in our series were hospitalized for a median of 9 days. Discussion Even though standard postoperative care routines were established, achieving continuous optimization of protocol adherence proved to be a challenge. Clearly, further prospective studies are needed to identify the rate of compliance with multimodal rehabilitation in the present unit. The incidence of anastomotic leak in elective cases has been found to vary between 1.4% and 7.9%, with the higher value noted in cases involving rectal procedures using an extraperitoneal suture line. Large series (> 800 patients) reported overall anastomotic leak rate under 4% [17,20,24,26]. Moreover, the risk of anastomotic leak has been ob- served to be increased by long operation time, low level of colorectal anastomosis, and comorbid conditions [27,28]. The absence of consensus regarding the defin- ition or the diagnosis of anastomotic leak makes it diffi- cult to compare the rates of anastomotic failure published in the literature. Here, we considered dehis- cence of the suture line in a blind colon or rectal seg- ment to be similar to an anastomotic leak in terms of morbidity, and we included patients with disruption of the distal stump after Hartmann’s procedure in this group. Both clinical and radiologic anastomotic leaks were registered. Yet, it is probable that the number of Surgical site infection (SSI) and anastomotic leak have been described as the most common adverse events in several investigations [15-17]. Internationally, SSI is con- sidered to be a major indicator of the quality of surgical care, and it has been estimated that 40% to 60% of SSIs are preventable [18]. The rates of superficial and deep wound infection noted in the present study are similar Table 5 Risk factors affecting outcome of colorectal surgery Variable p value Odds ratio for complication 95% Confidence interval Gender: male vs. female 0.483 1.004 0.672–1.500 Age > 75 years 0.667 1.109 0.690–1.775 Comorbidity: high vs. low 0.360 0.936 0.811–1.079 BMI > 30 < 0.0005 4.560 2.287-9.094 BMI 25–30 < 0.0005* 1.859 1.830–2.924 BMI 20–25 1,000 BMI < 20 0.639 1.226 0.532–2.870 surgical complexity: high vs. low < 0.00001* 1.826 1.449–2.302 p< 0,001. Multivariate logistic regression analysis. Table 5 Risk factors affecting outcome of colorectal surgery V i bl l Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Page 5 of 7 Table 6 Accumulation of costs (EUR) in relation to comorbidity score (1-6+), complexity of procedure (1–4), presence of complication or re-operation Comorbidity score N Mean EUR Std. Discussion Error 95% Confidence interval for Mean P-value Lower bound Upper bound 0 132 12760 677 11420 14100 1 41 28480 9340 9600 47360 2 208 15970 862 14270 17670 3 89 17220 1122 14990 19450 4 22 2,600 3442 16440 30750 0,000003 Kruskal-Wallis 5 15 16580 2166 11930 21230 6 23 24770 4302 15850 33690 Total 530 17060 884 15330 18800 Operationcomplexity 1 280 14810 1468 11920 17700 2 81 15680 1086 13520 17850 0,000001* Kruskal-Wallis 3 144 19700 1198 17340 22070 4 25 31590 3039 25320 37870 Total 530 17060 884 15330 18800 Complication absent 344 12410 384 11650 13160 present 186 25680 2289 21160 30200 0,000003* Mann–Whitney Total 530 17060 884 15330 18800 Re-operation absent 475 12880 402 10950 13970 present 55 38050 327 35380 42200 0,00001* Mann–Whitney Total 530 17060 884 15330 18800 Multivariable analysis of means. Table 6 Accumulation of costs (EUR) in relation to comorbidity score (1-6+), comp complication or re-operation Table 6 Accumulation of costs (EUR) in relation to comorbidity score (1-6+), complexity of procedure (1–4), presence of complication or re-operation Table 6 Accumulation of costs (EUR) in relation to comorbidity score (1-6+), complexity of procedure (1–4), presence of complication or re operation R) in relation to comorbidity score (1-6+), complexity of procedure (1–4), presence of also correlated with higher complication risks in our study, as has been reported by other researchers [32]. The rate of re-operation after colorectal surgery was deemed to be an independent factor for surgical quality. In the present review this factor compared favourably with contemporary studies [33]. radiologic anastomotic leaks was underestimated as asymptomatic patients did not undergo routine radio- logic investigation post-operatively. The incidence of midline wound disruption in our in- vestigation was high compared to rates of 0–2.7% after elective procedures in contemporary studies [5,17,29]. Suture technique monitored by the ratio of suture length to wound length (4–6/1) has proven to be the most im- portant factor affecting this variability [29], which indi- cates that midline dehiscence could easily be rectified by dissemination of appropriate knowledge and adherence to evidence-based techniques. The attending surgeon was also found to be an inde- pendent risk factor for complications in a fixed-variable statistical model. We regarded this model most appro- priate, because the surgeons in the present cases were not chosen at random, but were instead assigned to spe- cific operating lists according to their knowledge, train- ing, and experience. Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Yet, these skills were recognized as essential in the oc- currence of complications [38]. We identified a large dif- ference in the odds ratio for complication between individuals which indicated a distinct disparity within this group of surgeons. from consecutive patients treated at a single tertiary care unit during 2010, and hence the results might not be ap- plicable to other institutions or other time periods. In an attempt to eliminate the inaccuracy of administrative data in identifying complications, we conducted a sys- tematic review of the surgical records of all in- and out- patients in the index population and collected informa- tion on operations from the surgery-specific database. Unfortunately, the surgical unit did not have a struc- tured post-discharge surveillance program for SSIs, and thus there was probably an underestimation bias. Out- comes for discharged patients were evaluated 60 days postoperatively, and therefore long-term outcomes such as incisional hernias and aberrations in urinary and in- testinal functions were not registered. The hospital in- ternal accounting database was based on estimates correlated to diagnosis and did not permit detailed esti- mates of expenditure such as hourly physician cost or medication and medical supplies cost per patient. Never- theless, the deduction that postoperative complications are independently related to high hospital costs and pro- longed hospital stay remains true. Major part of the adverse events (wound infection, anastomotic failure, wound dehiscence, and periopera- tive bleeding) observed in our patients was avoidable, which is in accordance with contemporary studies [4,8]. These reports also found technical error to be the most common error type in surgery. Presumably, some pa- tients will always develop such complications, but, for any given patient, that should not be an expected out- come. It is also obvious, that, once identified, avoidable adverse outcomes, should be readily amenable to rectification. Our analysis showed that postoperative complications were associated with a substantial and statistically signifi- cant increase in total hospital costs, even after adjustment for type of surgery and comorbid conditions. The relative increase ranged from 207% for any complication to 380% for suture line failures, and, concomitantly, length of hos- pital stay was significantly prolonged. In conclusion, potentially avoidable complications fol- lowing elective colorectal surgery were frequent in the tertiary care unit we studied, and this resulted in a con- siderable increase in hospital costs and resource expen- ditures. Authors’ contributions Authors contributions EZMD, PhD and M-LL MD, PhD have both participated in the planning of the study, the collection of relevant data and preparation of the manuscript. Both authors read and approved the final manuscript. Both authors read and approved the final manuscript. Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Focus on the improvement of surgical outcomes contributes substantially to better quality of care and pa- tient safety and facilitates effective cost containment and resource utilization in healthcare. Nursing and postoperative or intensive care costs made the main contribution in this context, whereas the contri- bution from surgical operation costs was smaller than we had expected. Expenditures related to reoperations during index hospitalization,per se, could not be retrieved separ- ately from the database we used. Costs were further in- creased by readmissions, even if the currently observed readmission rate of 7.4% compares favorably with results obtained in earlier studies [39,40]. Acknowledgments g The authors wish to thank Mr Lars Wallgren, MSc, Lecturer, Dept of Statistics, Lund University, Lund, Sweden, for his contribution in the statistical analysis. Received: 28 October 2013 Accepted: 20 December 2013 Received: 28 October 2013 Accepted: 20 December 2013 Published: 3 January 2014 Competing interests Th h h The authors have no conflicts of interest in regard to this manuscript. These results were consistent with findings by other in- vestigators who have demonstrated a linear correlation be- tween surgical outcomes and accrued hospital costs. Robust evidence showed that programs for continuous quality improvement in surgical care, based on measure- ment and monitoring of outcome- and process- based qual- ity indicators, were effective in reducing post-operative morbidity, mortality and hospital costs [6,41-43]. In their analysis of 676 US hospitals, Fry et al. [43], indexed out- comes and efficiency to a reference set of cost effective hos- pitals and were able to demonstrate that inefficiency was the major contributor to excess cost. Colorectal procedures in the study population had a complication rate of 7, 7% in index and 16,3% in inefficient sites which resulted in 100% cost increase per incurred complication, a level that is sig- nificantly lower than the corresponding rates noted in the present review. As we identified that adverse outcomes re- sulted in very high average costs per event, it is clear that important cost savings could be achieved by reducing the occurrence of complications. Discussion In agreement with our findings, these aspects of surgeons have been found to have an impact on surgical outcome in numerous other studies [20,34,35]. Drolet et al. [36] observed decreased morbid- ity and mortality in patients who underwent elective colorectal procedures performed by experienced sur- geons, and Hubner and colleagues [22] noted that the surgeon constituted an independent risk factor for SSI after colon surgery. Surgical skills were hard to assess as the large number of assessment methods indicated [37]. Our results showed that age did not increase the risk of complications, which is contrary to previous findings and could be attributed to the fact that only elective cases were included [5,17]. Large studies have shown that > 10% weight loss prior to surgery and overweight constituted independent risk factors for morbidity after colorectal surgery [16,30]. As in the present report, sev- eral investigations [16,25,30,31] recognised that BMI above 25 and obesity were associated with increased risk for surgical complications. Complex procedures were Page 6 of 7 Page 6 of 7 Page 6 of 7 1. Bosma E, Veen EJ, Roukema JA: Incidence, nature and impact of error in surgery. Br J Surg 2011, 98:1654–1659. 2. Kable AK, Gibberd RW, Spigelman AD: Adverse events in surgical patients in Australia. Int J Qual Health Care 2002, 14:269–276. 3. Kirchhoff P, Clavien PL, Hahnloser D: Complications in colorectal surgery: risk factors and preventive strategies. Patient Saf Surg 2010, 4:5. 4. Healy MA, Shackford SR, Osler TM, Rogers FB, Burns E: Complications in surgical patients. Arch Surg 2002, 137:611–617. 1. Bosma E, Veen EJ, Roukema JA: Incidence, nature and impact of error in surgery. Br J Surg 2011, 98:1654–1659. 2. Kable AK, Gibberd RW, Spigelman AD: Adverse events in surgical patients in Australia. Int J Qual Health Care 2002, 14:269–276. 3. Kirchhoff P, Clavien PL, Hahnloser D: Complications in colorectal surgery: risk factors and preventive strategies. Patient Saf Surg 2010, 4:5. 4. Healy MA, Shackford SR, Osler TM, Rogers FB, Burns E: Complications in surgical patients. Arch Surg 2002, 137:611–617. 5. Longo WE, Virgo KS, Johnson FE, Oprian CA, Vernava AM, Wadr TP, Phelan MA, Henderson WG, Daley J, Khuri SF: Risk factors for morbidity and mortality after colectomy for colon cancer. Dis Colon Rectum 2000, 43:83–91. 6. Vaughan-Sarrazin M, Bayman L, Rasenthal G, Herderson W, Hendricks A, Cullen JJ: The business case for the reduction of surgical complications in VA hospitals. Surgery 2011, 149:474–483. Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 7. Khan NA, Quan H, Bugar JM, Lemaire JB, Brant R, Ghali WA: Association of postoperative complications with hospital costs and length of stay in a tertiary care center. J Gen Intern Med 2006, 21:177–180. 29. Millbourn D, Cengiz Y, Israelsson LA: Effect of stitch length on wound complications after closure of midline incisions: a randomized controlled trial. Arch Surg 2009, 144:1056–1059. 8. Soop M, Fryksmark U, Köster M, Haglund B: The incidence of adverse events in Swedish hospitals. A retrospective medical record review study. Int J Qual Health Care 2009, 21:285–291. 30. Veyrie N, Ata T, Muscari F, Couchard AC, Msika S, Hay JM, Figerhut A, Dziri C: Anastomotic leakage after elective right versus left colectomy for cancer. Prevalence and independent risk factors. J Am Coll Surg 2007, 205:785–793. 9. Fearon KC, Ljungqvist O, von Meyenfeldt M, Revhaug A, Dejong CH, Lassen K, Nygren J, Hansel J, Soop M, Andersen J, Kehlet H: Enhanced recovery after surgery: a consensus review of clinical care for patients undergoing colonic resection. Clin Nutr 2008, 24:466–477. 31. Merkow RP, Bilimoria KY, McCarter MD, Bentrem DJ: Effect of body mass index on short-term outcomes after colectomy for cancer. J Am Coll Surg 2009, 208:53–61. 32. Aust JB, Henderson W, Khuri S, Page CP: The impact of operative complexity on patient risk factors. Ann Surg 2005, 241:1024–1028. 10. Charlson ME, Pompei P, Ales K, MacKenzie CR: A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987, 40:373–383. 33. Van Westreenen HL, Ijpma FF, Wevers KP, Afzali H, Patijn GA: Reoperation after colorectal surgery is an independent predictor of the 1-year mortal- ity rate. Dis Colon Rectum 2011, 54:1438–1442. 11. Clavien PA, Barkun J, de Oliveira ML, Vauthey JN, Dindo D, Schulick RD, de Santibañes E, Perkolj J, Slankamenac K, Bassi C, Graf R, Vonlanthen R, Padbury R, Cameron JL, Makuuchi M: The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009, 250:187–196. 34. Prystowsky JB, Bordage G, Feinglass JM: Patient outcomes for segmental colon resection according to surgeon’s training, certification and experience. Surgery 2002, 132:663–672. 35. Bilimoria KY, Phillips JD, Rock CE, Hayman A, Prystowsky JB, Bentrem DJ: Effect of surgeon training, specialization and experience on outcomes for cancer surgery. A systematic review of the literature. Ann Surg Oncol 2009, 16:1799–1808. 12. References g g 5. Longo WE, Virgo KS, Johnson FE, Oprian CA, Vernava AM, Wadr TP, Phelan MA, Henderson WG, Daley J, Khuri SF: Risk factors for morbidity and mortality after colectomy for colon cancer. Dis Colon Rectum 2000, 43:83–91. 6. Vaughan-Sarrazin M, Bayman L, Rasenthal G, Herderson W, Hendricks A, Cullen JJ: The business case for the reduction of surgical complications in VA hospitals. Surgery 2011, 149:474–483. The present observational cohort study had a number of limitations. The data were derived retrospectively Page 7 of 7 Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Al-Rafaele WB, Parsons HM, Habermann EB, Kwaan M, Spencer MP, Henderson WG, Rothenberger DA: Operative outcomes beyond 30-day mortality. Ann Surg 2011, 253:947–952. 13. Varadhan KK, Neal KR, Dejong CHC, Fearson KCH, Ljungqvist O, Lobo DN: The enhanced recovery after surgery (ERAS) pathway for patients undergoing major elective open colorectal surgery: a meta-analysis of randomized controlled trials. Clin Nutr 2010, 29:434–440. 36. Drolet S, Mac Lean AR, Myers RP, Shaheen AAM, Dixon E, Buie WD: Elective resection of colon cancer by high-volume surgeons is associated with decreased morbidity. J Gastrointest Surg 2011, 15:541–550. 37. Van Hove PD, Tuijthof GJ, Verdaasdonk EG, Stassen LP, Dankelman J: Objective assessment of technical surgical skills. Br J Surg 2010, 97:972–987. 14. Wang G, Jiang ZW, Xu J, Gong JF, Bao Y, Xie LF, Li JS: Fast track rehabilitation program vs. conventional care after colorectal resection: a randomized clinical trial. World J Gastroenterol 2011, 17:671–676. 38. Mishriki SF: Surgical site infection. NICE forgot surgical skill. BMJ 2008, 337:a2579. 15. Kim J, Mittal R, Konyalian V, King J, Stamos M, Kumar R: Outcome analysis of patients undergoing colorectal resection for emergent and elective indications. Am Surg 2007, 73:991–993. 39. Wick EC, Shore AD, Hirose K, Ibrahim AM, Gearhart SL, Efron J, Weiner JP, Makary MA: Readmission rates and cost following colorectal surgery. Dis Colon Rectum 2011, 54:1475–1479. 16. Piessen G, Muscari F, Rivkine E, Idrissi MS, Lorimier G, Fingerhut A, Dziri C, Hay JM: Prevalence of and risk factors for morbidity after elective left colectomy. Arch Surg 2011, 146:1149–1155. 40. Messaris E, Sehgal R, Deiling S, Koltun WA, Stewart D, McKenna K, Poritz LS: Dehydration is the most common indication for readmission after diverting ileostomy creation. Dis Colon Rectum 2012, 55:175–180. 17. Sjo OH, Larsen S, Lunde OC, Nesbakken A: Short term outcome after emergency and elective surgery for colon cancer. Colorectal Dis 2009, 11:733–739. 41. Dimick JB, Chen SL, Taheri PA, Henderson WG, Khuri SF, Campbell DA: Hospital costs associated with surgical complications: a report from the private-sector national surgical quality improvement program. J Am Coll Surg 2004, 199:531–537. 18. Odom-Forren J: Preventing surgical site infections. Nursing 2006, 36:58–63. 19. Hawn M, Vick CC, Richman J, Holman W, Deierhoi RJ, Graham LA, 18. Odom-Forren J: Preventing surgical site infections. Nursing 2006, 36:58–63 19. Hawn M, Vick CC, Richman J, Holman W, Deierhoi RJ, Graham LA 19. Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Zoucas and Lydrup Patient Safety in Surgery 2014, 8:2 http://www.pssjournal.com/content/8/1/2 Hawn M, Vick CC, Richman J, Holman W, Deierhoi RJ, Graham LA, Henderson WG, Itani KMF: Surgical site infection prevention. Ann Surg 2011, 254:494–501. 42. Birkmeyer JD, Gust C, Dimick JB, Birkmeyer NJO, Skinner JS: Hospital quality and the cost of inpatient surgery in the United States. Ann Surg 2012, 255:1–5. 20. Tang R, Chemn HH, Wang YL, Changchien CR, Chen JS, Hsu KC, Chiang JM, Wang JY: Risk factors for surgical site infection after elective resection of the colon and rectum: a single-center prospective study of 2,809 con- secutive patients. Ann Surg 2001, 234:181–189. 43. Fry DE, Pine M, Jones BL, Meimban RJ: The impact of ineffective and inefficient care on the Wound infection, anastomotic failure, wound dehiscence, and perioperative bleeding excess costs of elective surgical procedures. J Am Coll Surg 2011, 212:779–786. 21. Pastor C, Artinyan A, Varma M, Kim E, Gibbs L, Garcia-Aguilar J: An increase in compliance with the surgical care improvement project measures does not prevent surgical site infection in colorectal surgery. Dis Colon Rectum 2010, 53:24–30. doi:10.1186/1754-9493-8-2 Cite this article as: Zoucas and Lydrup: Hospital costs associated with surgical morbidity after elective colorectal procedures: a retrospective observational cohort study in 530 patients. Patient Safety in Surgery 2014 8:2. 22. Hubner M, Diana M, Zanetti G, Eisenring M-C, Demartines N, Troillet N: Surgical site infections in colon surgery. the patient, the procedure, the hospital and the surgeon. Arch Surg 2011, 146:1240–1245. 23. Mac Kay GJ, Molloy RG, O’Dwyer PJ: C-reactive protein as a predictor of post-operative infective complications following elective colorectal resection. Colorectal Dis 2011, 13:583–587. 24. Buchs N, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B, Morel P: Incidence, consequences and risk factors for anastomotic dehiscence after colorec- tal surgery: a prospective monocentric study. Int J Colorectal Dis 2008, 23:265–270. Submit your next manuscript to BioMed Central and take full advantage of: 25. Imai E, Ueda M, Kanao K, Kubota T, Hasegawa H, Omae K, Kitajima M: Surgical site infection risk factors identified by multivariate analysis for patient undergoing laparoscopic, open colon, and gastric surgery. Am J Infect Control 2008, 36:727–731. • Convenient online submission • Thorough peer review 26. Platell C, Barwood N, Dorfmann G, Makin G: The incidence of anastomotic leaks in patients undergoing colorectal surgery. Colorectal Dis 2007, 9:71–79. 27. Konishi T, Watanabe T, Kishimoto J, Nagawa H: Risk factors for anastomotic leakage after surgery for colorectal cancer: results of prospective surveillance. J Am Coll Surg 2006, 202:439–444. 28. Lipska MA, Bisett IP, Paryry BR, Merrie AE: Anastomotic leakage after lower gastrointestinal anastomosis. ANZ J Surg 2006, 76:579–585. 28. Lipska MA, Bisett IP, Paryry BR, Merrie AE: Anastomotic leakage after lower gastrointestinal anastomosis. ANZ J Surg 2006, 76:579–585. Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W4213456994	https://www.biorxiv.org/content/biorxiv/early/2022/02/24/2022.02.24.481751.full.pdf	English	null	Cooperativity of myosin II motors in the non-regulated and regulated thin filaments investigated with high-speed AFM	bioRxiv (Cold Spring Harbor Laboratory)	2,022	cc-by	25,701	. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Cooperativity of myosin II motors in the non-regulated and regulated thin filaments 1 investigated with high-speed AFM 2 3 4 5 6 Oleg S. Matusovsky1 7 Alf Mansson2 8 Dilson E. Rassier1* 9 10 1 Department of Kinesiology and Physical Education, McGill University, Canada 11 12 2 Department of Chemistry and Biomedical Sciences, Linnaeus University, Kalmar, 13 Sweden. 14 15 16 17 Correspondence author: dilson.rassier@mcgill.ca 18 Article submitted to eLife 19 20 21 Keywords: myosin motors, actin filaments, motors cooperativity, HS-AFM 22 23 Running title: Cooperativity of myosin II motors 24 25 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is mad The copyright holder for this preprin this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Cooperativity of myosin II motors in the non-regulated and regulated thin filaments investigated with high-speed AFM 25 1 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Abstract 26 Skeletal myosins II are non-processive molecular motors, that work in ensembles to produce 27 muscle contraction while binding to the actin filament. Although the molecular properties of 28 myosin II are well known, there is still debate about the collective work of the motors: is there 29 cooperativity between myosin motors while binding to the actin filaments? In this study, we 30 used high-speed AFM to evaluate this issue. We observed that the initial binding of small 31 arrays of myosin heads to the non-regulated actin filaments did not affect the cooperative 32 probability of subsequent bindings to neighboring sites and did not lead to an increase in the 33 fractional occupancy of the actin binding sites. These results suggest that myosin motors are 34 independent force generators when connected in small arrays, and that the binding of one 35 myosin does not alter the kinetics of other myosins. In contrast, the probability of binding of 36 myosin heads to regulated thin filaments under activating conditions (at high Ca2+ 37 concentration and with 2 μM ATP) was increased with the initial binding of one myosin, 38 leading to a larger occupancy of neighboring available binding sites. The result suggests that 39 myosin cooperativity is defined by the activation status of the thin filaments. 40 41 2 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint eLife digest 42 43 Muscle contraction is the result of large ensembles of the molecular motor myosin II working 44 in coordination while attached to actin. Myosin II produces the power stroke, responsible for 45 force generation. In this paper, we used High-Speed Atomic Force Microscopy (HS-AFM) to 46 determine the potential cooperativity between myosin motors bound to non-regulated and 47 regulated thin filaments. Abstract 26 The actin-attached fraction of the ATP turnover time, the duty 71 ratio, is ~ 5% (Howard, 1997), which enables high speeds of shortening (Pertici et al., 2018; 72 Cheng et al., 2020). Although most studies looking to the mechanics of isolated myosin II 73 have been performed with single molecules, assemblies of myosin II have been investigated 74 in arrays developed with a small number of motors adsorbed to silica beads (Debold et al., 75 2005), optical fiber surfaces (Pertici et al., 2018) or with the native thick filaments (Cheng et 76 al., 2020). These small ensemble studies show a load dependence and force-velocity relation 77 that is similar to that observed in myofibrillar (Lowey et al., 2018) and cellular preparations 78 (Edman & Hwang, 1977). Furthermore, these force-velocity relationships can be modelled 79 using single molecule properties (Mansson et al., 2018; Mansson, 2019), and experimental 80 Abstract 26 Based on the direct visualization of myosin-actin interaction, 48 probability of myosin binding, and the myosin fractional occupancy of binding sites along non- 49 regulated and regulated actin filaments, our results show no cooperative effects over ~100 50 nm of the actin filament length. In contrast, there is myosin cooperativity within the activated 51 thin filament, that induces a high affinity of myosin heads to the filaments. Our results support 52 the independent behaviour of myosin heads while attached to actin filaments, but a 53 cooperative behavior when attached to regulated thin filaments. 54 56 3 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Introduction 57 Myosin II is a non-processive molecular motor that binds to actin filaments to produce 58 mechanical work, using the chemical free energy of adenosine triphosphate (ATP). After an 59 initial attachment to actin, the myosin motor domain undergoes conformational changes 60 associated with release of the ATP hydrolysis products inorganic phosphate (Pi) and ADP 61 from the active site of myosin. In this process, a force-generating power stroke, with swing of 62 the myosin lever arm, is generated and there is a transition of myosin from the weak to the 63 strong actin-binding states (Rayment et al., 1993; Fisher et al., 1995; Mansson et al., 2018; 64 Robert-Paganin et al., 2020). 65 66 Myosin II molecules form bipolar filaments in skeletal, cardiac and smooth muscles and this 67 filamentous form of myosin II allows the motors to collectively produce high forces during 68 muscle contraction despite a low duty ratio (Finer et al., 1994; Ishijima et al., 1994; Yanagida 69 & Ishijima, 1995; Kaya & Higuchi, 2010; Kaya et al., 2017; Pertici et al., 2018; Cheng et al., 70 2019; Cheng et al., 2020). Introduction 57 CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint data from single molecules (Kaya & Higuchi, 2010; Capitanio et al., 2012; Sung et al., 2015) 81 suggesting that myosin II motors are independent force generators, as postulated decades 82 ago (Huxley, 1957), even when they are attached to a common thick filament. 83 data from single molecules (Kaya & Higuchi, 2010; Capitanio et al., 2012; Sung et al., 2015) 81 suggesting that myosin II motors are independent force generators, as postulated decades 82 ago (Huxley, 1957), even when they are attached to a common thick filament. 83 84 However, there are also suggestions that myosin molecules work cooperatively, and the work 85 produced by motor assemblies is different from individual motors (Kaya et al., 2017). 86 Accordingly, the attachment of one motor would interfere with the kinetics and attachment 87 mechanics of other motors when working in arrays. The result casts doubt on the concept of 88 independent force generators in motor assemblies. Cooperativity could also arise in double- 89 headed molecules (Huxley & Tideswell, 1997; Brunello et al., 2007) or myosin motors that 90 bind to adjacent actin sites (Caremani et al., 2013; Rahman et al., 2018). X-ray diffraction 91 studies using muscle fibre preparations provide evidence that the coordinated movements of 92 myosin heads may indeed regulate force generation (Irving et al., 1992; Linari et al., 2015). 93 Finally, this form of cooperativity may arise from allosteric changes of the actin filament itself 94 so that binding of one myosin molecule modifies the kinetics of myosin binding to nearby 95 sites (Orlova et al., 1993; Tokuraku et al., 2009; Prochniewicz et al., 2010). 96 97 However, there are also suggestions that myosin molecules work cooperatively, and the work 85 produced by motor assemblies is different from individual motors (Kaya et al., 2017). 86 Accordingly, the attachment of one motor would interfere with the kinetics and attachment 87 mechanics of other motors when working in arrays. The result casts doubt on the concept of 88 independent force generators in motor assemblies. Introduction 57 Myosin II is a non-processive molecular motor that binds to actin filaments to produce 58 mechanical work, using the chemical free energy of adenosine triphosphate (ATP). After an 59 initial attachment to actin, the myosin motor domain undergoes conformational changes 60 associated with release of the ATP hydrolysis products inorganic phosphate (Pi) and ADP 61 from the active site of myosin. In this process, a force-generating power stroke, with swing of 62 the myosin lever arm, is generated and there is a transition of myosin from the weak to the 63 strong actin-binding states (Rayment et al., 1993; Fisher et al., 1995; Mansson et al., 2018; 64 Robert-Paganin et al., 2020). 65 Myosin II molecules form bipolar filaments in skeletal, cardiac and smooth muscles and this 67 filamentous form of myosin II allows the motors to collectively produce high forces during 68 muscle contraction despite a low duty ratio (Finer et al., 1994; Ishijima et al., 1994; Yanagida 69 & Ishijima, 1995; Kaya & Higuchi, 2010; Kaya et al., 2017; Pertici et al., 2018; Cheng et al., 70 2019; Cheng et al., 2020). The actin-attached fraction of the ATP turnover time, the duty 71 ratio, is ~ 5% (Howard, 1997), which enables high speeds of shortening (Pertici et al., 2018; 72 Cheng et al., 2020). Although most studies looking to the mechanics of isolated myosin II 73 have been performed with single molecules, assemblies of myosin II have been investigated 74 in arrays developed with a small number of motors adsorbed to silica beads (Debold et al., 75 2005), optical fiber surfaces (Pertici et al., 2018) or with the native thick filaments (Cheng et 76 al., 2020). These small ensemble studies show a load dependence and force-velocity relation 77 that is similar to that observed in myofibrillar (Lowey et al., 2018) and cellular preparations 78 (Edman & Hwang, 1977). Furthermore, these force-velocity relationships can be modelled 79 using single molecule properties (Mansson et al., 2018; Mansson, 2019), and experimental 80 4 4 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . Introduction 57 Cooperativity could also arise in double- 89 headed molecules (Huxley & Tideswell, 1997; Brunello et al., 2007) or myosin motors that 90 bind to adjacent actin sites (Caremani et al., 2013; Rahman et al., 2018). X-ray diffraction 91 studies using muscle fibre preparations provide evidence that the coordinated movements of 92 myosin heads may indeed regulate force generation (Irving et al., 1992; Linari et al., 2015). 93 Finally, this form of cooperativity may arise from allosteric changes of the actin filament itself 94 so that binding of one myosin molecule modifies the kinetics of myosin binding to nearby 95 sites (Orlova et al., 1993; Tokuraku et al., 2009; Prochniewicz et al., 2010). 96 Other forms of cooperativity between myosin motors involve activation of the thin filament 98 where several cooperative phenomena have been described (Gordon et al., 2000). In skeletal 99 muscle sarcomeres, actin–myosin interactions are regulated by Ca2+ through the regulatory 100 proteins troponin (Tn) and tropomyosin (Tm), that form the thin filament complex with actin. 101 Each of the Tm molecules contact seven actin monomers and is associated with the three Tn 102 subunits: Tn-T, Tn-I and Tn-C . Upon Ca2+ binding to Tn-C, conformational changes are 103 triggered in the Tn–Tm complex resulting in a displacement of Tm that allows for myosin 104 5 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint binding to actin (Galinska-Rakoczy et al., 2008; Lehman et al., 2009). We previously have 105 shown that under relaxing conditions, thin filaments presented a combination of activated and 106 non-activated segments along their lengths, and were not blocked from myosin; the 107 equilibrium between blocked and closed states was defined by Ca2+-induced Tn-Tm 108 conformational changes (Matusovsky et al., 2019). Introduction 57 In addition, myosin binding to actin is also 109 required for full activation, or to induce the open state of activation of the thin filament 110 (McKillop and Geeves, 1993; Smith & Geeves, 2003; Desai et al., 2015). When myosin binds 111 to actin, it may directly affect the regulatory system by changing the conformation of Tm, such 112 that other myosin heads can attach to thin filaments (Geeves & Holmes; 1999; Gordon et al., 113 2000). Furthermore, the question remains if one or two myosin heads in a molecule are 114 required for the full activation of the thin filament. 115 Therefore, cooperativity during myosin II-actin interactions can conceptually arise from at 117 least two sources: cooperativity among myosin molecules within the thick filaments due to 118 structural changes in the actin filament or cooperativity through activation of the regulated, 119 thin filaments. Each cooperativity source may present different mechanisms. In this study, we 120 used High-Speed Atomic Force Microscopy (HS-AFM) to evaluate the potential cooperativity 121 of double-headed heavy meromyosin fragments (HMM) of myosin II that were connected 122 through the S2 tail regions, while attaching between non-regulated actin filaments, or 123 regulated thin filaments. Because HS-AFM allows the investigation of protein dynamics with 124 nanometer spatial and millisecond temporal resolutions (Kodera et al., 2021; Heath & 125 Scheuring, 2018; Matusovsky et al., 2021) our experimental approach allows us to 126 investigate important aspects of myosin cooperativity, with a better resolution than previous 127 fluorescence microscopy studies (spatial resolution limitation of >100 nm) (Desai et al., 128 6 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 2015) . Introduction 57 Specifically for this study, we developed a method in which HMM motors, attached by 129 their S2 regions to form a structure with up to 8-10 individual myosin heads (4-5 HMM 130 molecules) bound to nearby sites along two actin filaments or two thin filaments (Fig. 1 and 131 Fig. S1). The benefit of this approach is the ability to monitor the behavior of each of the 132 HMM heads over the time of an experiment to evaluate the potential cooperative binding of 133 HMM heads with either actin or thin filaments during the ATPase cycle. The approach also 134 allows investigation of aspects of inter-head cooperativity as well as the potential to 135 investigate cooperative changes along actin or thin filaments at spatial resolution similar to 136 the inter-monomer distance along the filaments. 137 Experimental design to study myosin cooperativity by HS-AFM 140 In order to track the cooperativity behavior of the myosin heads within a sequence of 141 successive HS-AFM images, we used an experimental approach in which pairs of HMM 142 heads are attached to two actin filaments (Fig. 1, Fig. S1), as explained in details in a 143 previous study from our laboratory (Matusovsky et al., 2021). Briefly, we aimed for an 144 experimental situation in which two non-regulated actin filaments (F-actin) or two regulated 145 cardiac thin filaments (cTFs) were bound to an underlying mica-supported lipid bilayer (SLB) 146 surface, in parallel to each other, and with enough space for binding of double-headed HMM 147 molecules between them. A cross-section analysis showed that the distance between two 148 actin filaments during the experiments was 40.56 ± 9.65 nm, and the distance between cTFs 149 was 67.69 ± 15.92 nm (Fig. S2). The observed difference in distance (27.1 ± 6.1 nm) between 150 non-regulated F-actin and regulated cTFs did not affect the HMM binding and displacement 151 analysis (Figs 1f, 2c). 152 7 7 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 153 153 Once the filaments were found in a parallel orientation, HMM fragments were added into the 154 HS-AFM chamber filled with an experimental solution or placed on the top of the mica-SLB 155 surface, in a solution containing 0.2-10 μM of NPE-caged or non-caged ATP. We then 156 searched for events where each of the two HMM heads would interact with two parallel 157 filaments. Immediately after both HMM heads were bound between parallel actin filaments, 158 we activated the NPE-caged ATP in the solution by photolysis using a UV laser (340 nm) 159 installed into the HS-AFM system (see Materials and Methods). Experimental design to study myosin cooperativity by HS-AFM 140 The HS-AFM snapshots of 160 two parallel non-regulated F-actins or regulated cTFs showed regularly bound HMM 161 molecules between them in the absence or in the presence of ATP (Figs S1, S2). The 162 globular upper and lower heads of each HMM molecule were bound in ~30-37 nm proximity 163 from each other, along the actin half-helical pitch structure (Fig. 1, Fig. S2). HMM heads were 164 not bound to all the available actin-binding sites along the filaments at various experimental 165 conditions, including rigor or in the presence of ADP in similarity to electron microscopy 166 studies (Orlova et al., 1993). This observation may be related to the immobilization of S2 167 regions of each HMM molecule to the underlying lipid bilayer, allowing it to reach a maximum 168 of ~2-3 binding spots between neighboring actin monomers, i.e., 11 nm or 2 × 5.5 nm 169 (Fig.1b). The ~37 nm arrangement of myosin heads in our HS-AFM experiments is similar to 170 the preferable binding sites of myosin heads along actin filaments (Steffen et al., 2001) and 171 relate to the ~37 nm hotspots for myosin head bindings along the thin filaments in the A-band 172 of the sarcomere (Wang et al 2021) 173 Once the filaments were found in a parallel orientation, HMM fragments were added into the 154 HS-AFM chamber filled with an experimental solution or placed on the top of the mica-SLB 155 surface, in a solution containing 0.2-10 μM of NPE-caged or non-caged ATP. We then 156 searched for events where each of the two HMM heads would interact with two parallel 157 filaments. Immediately after both HMM heads were bound between parallel actin filaments, 158 we activated the NPE-caged ATP in the solution by photolysis using a UV laser (340 nm) 159 installed into the HS-AFM system (see Materials and Methods). The HS-AFM snapshots of 160 two parallel non-regulated F-actins or regulated cTFs showed regularly bound HMM 161 molecules between them in the absence or in the presence of ATP (Figs S1, S2). The 162 globular upper and lower heads of each HMM molecule were bound in ~30-37 nm proximity 163 from each other, along the actin half-helical pitch structure (Fig. 1, Fig. S2). Experimental design to study myosin cooperativity by HS-AFM 140 HMM heads were 164 not bound to all the available actin-binding sites along the filaments at various experimental 165 conditions, including rigor or in the presence of ADP in similarity to electron microscopy 166 studies (Orlova et al., 1993). This observation may be related to the immobilization of S2 167 regions of each HMM molecule to the underlying lipid bilayer, allowing it to reach a maximum 168 of ~2-3 binding spots between neighboring actin monomers, i.e., 11 nm or 2 × 5.5 nm 169 (Fig.1b). The ~37 nm arrangement of myosin heads in our HS-AFM experiments is similar to 170 the preferable binding sites of myosin heads along actin filaments (Steffen et al., 2001) and 171 relate to the ~37 nm hotspots for myosin head bindings along the thin filaments in the A-band 172 of the sarcomere (Wang et al., 2021). 173 174 8 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 175 Figure 1. The kinetics of double-headed myosin motors bound to non-regulated actin 176 filaments. (a) Diagram illustrating the approach used to study cooperative behavior of HMM 177 molecules bound between two parallel filaments. (b) Structural model of actin-myosin complex for the 178 experimental design used in the study; actin-myosin complex in rigor conditions (PDB:1M8Q) with 179 upper and lower heads bound between two parallel actin filaments and attached by their S2 fragments 180 (PDB: 2FXO). (c) Kinetics model of actin-myosin interaction and approach used to calculate the 181 backward and forward myosin displacements; the green dots indicate the center of mass of the head. 182 KT and K-AD are the constants of the ATP binding and ADP release; Kw and Kws are the constants of 183 the weak biding and weak-to-strong transition, respectively. Experimental design to study myosin cooperativity by HS-AFM 140 (d-e) Representative time course of 184 binomial binding (d) and head displacements (e) calculated for the individual HMM molecules (M1- 185 M4) at the given time for upper and lower heads of each HMM molecule in the presence of 2 μM ATP. 186 (f) The backward and forward myosin displacements in the F-actin-HMM complex (n=6, 589 events, 187 ~43 HMM molecules); data sets were fitted by sum of two Gaussians (r2=0.97 and r2=0.99, 188 respectively). The two peaks for backward displacement: 1.8±0.2 nm and 3.7±2.0 nm; the two peaks 189 for forward displacement: 2.7±0.5 nm and 5.1±2.1 nm. (g) Simulated HS-AFM image of the structural 190 model shown in (b) performed in Bio-AFM viewer software (Amyot & Flechsig, 2020). (h) 191 Representative HS-AFM snapshots of HMM molecules bound between two actin filaments at the 192 indicated times (M1-M4 shown in color code duplicated in the d-e and across all of the figures), scale 193 bar 30 nm. Related to Movies S1-S3. 194 195 Figure 1. The kinetics of double-headed myosin motors bound to non-regulated actin 176 filaments. (a) Diagram illustrating the approach used to study cooperative behavior of HMM 177 molecules bound between two parallel filaments. (b) Structural model of actin-myosin complex for the 178 experimental design used in the study; actin-myosin complex in rigor conditions (PDB:1M8Q) with 179 upper and lower heads bound between two parallel actin filaments and attached by their S2 fragments 180 (PDB: 2FXO). (c) Kinetics model of actin-myosin interaction and approach used to calculate the 181 backward and forward myosin displacements; the green dots indicate the center of mass of the head. 182 KT and K-AD are the constants of the ATP binding and ADP release; Kw and Kws are the constants of 183 the weak biding and weak-to-strong transition, respectively. (d-e) Representative time course of 184 binomial binding (d) and head displacements (e) calculated for the individual HMM molecules (M1- 185 M4) at the given time for upper and lower heads of each HMM molecule in the presence of 2 μM ATP. 186 (f) The backward and forward myosin displacements in the F-actin-HMM complex (n=6, 589 events, 187 ~43 HMM molecules); data sets were fitted by sum of two Gaussians (r2=0.97 and r2=0.99, 188 respectively). The two peaks for backward displacement: 1.8±0.2 nm and 3.7±2.0 nm; the two peaks 189 for forward displacement: 2.7±0.5 nm and 5.1±2.1 nm. Experimental design to study myosin cooperativity by HS-AFM 140 (g) Simulated HS-AFM image of the structural 190 model shown in (b) performed in Bio-AFM viewer software (Amyot & Flechsig, 2020). (h) 191 Representative HS-AFM snapshots of HMM molecules bound between two actin filaments at the 192 indicated times (M1-M4 shown in color code duplicated in the d-e and across all of the figures), scale 193 bar 30 nm. Related to Movies S1-S3. 194 196 Kinetics of actin-myosin interaction in the non-regulated and regulated systems 197 215 In the presence of Mg.ATP, myosin heads detach from the filament and re-attach to the same 216 or a new binding site, allowing us to determine the displacement of the myosin head by the 217 change in COM. The calculated displacement in our study is slightly larger than the working 218 stroke size of 5 nm reported for S1 (Capitanio et al., 2006) and slightly smaller than the 219 values obtained from structural studies with single-headed myosin (~10-12 nm) (Geeves et 220 al., 2005). It is comparable with studies performed with myosin molecules evaluated with 221 laser tweezers (Finer et al., 1994; Tyska et al., 1999) and single fiber mechanics (Piazzesi et 222 al 2002) 223 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made py g p p p y , ; p g p p displacements (d) of myosin heads in the presence of ATP (and high Ca2+ = pCa 4.5 in the 200 case of cTFs). The HMM displacements calculated as a change in the center of mass (COM) 201 of the myosin head at the given time during the experiment (Fig. 1c, see also Materials and 202 Methods) were in the range of 6-8 nm. The backward (towards minus end of the filament) and 203 forward (towards plus end of the filament) HMM displacements were calculated. The size 204 distribution of HMM displacements revealed two distinct peaks in F-actin-HMM and cTFs- 205 HMM complexes that most likely represent the events occurring through ADP (1-3 nm 206 displacements) and Pi releases (over 3 nm displacements as previously described 207 (Matusovsky et al., 2021). The sum of two peaks for backwards and forward displacements of 208 HMM molecules were 5.5 ± 1.68 nm and 7.8 ± 1.96 nm on the non-regulated actin filaments, 209 and 7.4 ± 1.73 nm and 7.6 ± 1.95 nm on the regulated cTFs (Figs. 1f and 2c). 210 The evaluation of displacement and working stroke of myosin heads in the HS-AFM was 212 described in details in the Materials and Methods. Briefly, the working stroke is viewed as a 213 transition from the weak to the strong-binding states evaluated by the changes in the lever 214 arm movement. 196 Kinetics of actin-myosin interaction in the non-regulated and regulated systems 197 We characterized functional parameters of the myosin heads bound between two parallel 198 non-regulated F-actins or regulated cTFs, including the average backward and forward 199 non-regulated F-actins or regulated cTFs, including the average backward and forward 199 9 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint displacements (d) of myosin heads in the presence of ATP (and high Ca2+ = pCa 4.5 in the 200 case of cTFs). The HMM displacements calculated as a change in the center of mass (COM) 201 of the myosin head at the given time during the experiment (Fig. 1c, see also Materials and 202 Methods) were in the range of 6-8 nm. The backward (towards minus end of the filament) and 203 forward (towards plus end of the filament) HMM displacements were calculated. The size 204 distribution of HMM displacements revealed two distinct peaks in F-actin-HMM and cTFs- 205 HMM complexes that most likely represent the events occurring through ADP (1-3 nm 206 displacements) and Pi releases (over 3 nm displacements as previously described 207 (Matusovsky et al., 2021). The sum of two peaks for backwards and forward displacements of 208 HMM molecules were 5.5 ± 1.68 nm and 7.8 ± 1.96 nm on the non-regulated actin filaments, 209 and 7.4 ± 1.73 nm and 7.6 ± 1.95 nm on the regulated cTFs (Figs. 1f and 2c). 210 211 The evaluation of displacement and working stroke of myosin heads in the HS-AFM was 212 described in details in the Materials and Methods. Briefly, the working stroke is viewed as a 213 transition from the weak to the strong-binding states evaluated by the changes in the lever 214 arm movement. The change in the lever arm considers a defined polarity of the actin filament. 196 Kinetics of actin-myosin interaction in the non-regulated and regulated systems 197 241 242 243 Probability of HMM binding to the non-regulated and regulated actin filaments 244 To monitor the probability of binding events between individual myosin heads we applied a 245 probability analysis based on a binary combination: HMM bound to the filaments equals to “1” 246 and HMM detached from the filaments equals to “0”. To use this analysis, we need to 247 evaluate if there is any directional bias in the myosin bindings along the F-actin and cTFs, 248 either towards the barbed plus end or the pointed minus end of the filaments. Therefore, the 249 polarity of F-actin and cTFs complexed with HMM was determined by using the morphology 250 of the myosin heads bound to the filaments (Ngo et al., 2015). The bound myosin heads 251 observed in the presence of Mg.ATP, Mg.ADP or in the rigor state allowed us to determine 252 the polarity of the filaments (see Figs S1, S2-S5). According to our observations the most 253 frequent myosin head orientation in the weak binding state (presence of Mg.ATP-γ-S) or 254 strong binding state (rigor state) is the one where the heads of HMM molecules are 255 positioned toward the minus end of the filament (Fig. 1h, Fig. 3a-b, Fig. 4b-d, Fig. S5). 256 Therefore, binding events that occurred towards to the plus end of the filament (Mn  Mn+1) 257 for individual upper and lower myosin heads at ATP concentrations ranging from 0.2-10 μM 258 were used in the analysis. 259 Figure 2. The kinetics of double-headed myosin motors bound to regulated cTFs. (a-b) 236 Representative time course of binomial binding (a) and heads displacements (b) calculated for the 237 individual HMM molecules (M1-M4) at the given time for upper and lower heads of each HMM 238 molecule in the presence of 0.5 μM ATP and high Ca2+ concentrations. (c) The backward and forward 239 myosin displacements in the cTFs-HMM complex (n=5, 911 events, ~35 HMM molecules); data sets 240 were fitted by sum of two Gaussians (r2=0.99 and r2=0.99, respectively). 241 196 Kinetics of actin-myosin interaction in the non-regulated and regulated systems 197 It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Figure 2. The kinetics of double-headed myosin motors 236 Representative time course of binomial binding (a) and heads d 237 individual HMM molecules (M1-M4) at the given time for uppe 238 molecule in the presence of 0.5 μM ATP and high Ca2+ concentra 239 myosin displacements in the cTFs-HMM complex (n=5, 911 even 240 were fitted by sum of two Gaussians (r2=0.99 and r2=0.99, respect 241 242 243 Probability of HMM binding to the non-regulated and reg 244 To monitor the probability of binding events between indivi 245 probability analysis based on a binary combination: HMM bo 246 and HMM detached from the filaments equals to “0”. To 247 evaluate if there is any directional bias in the myosin bindi 248 either towards the barbed plus end or the pointed minus en 249 polarity of F-actin and cTFs complexed with HMM was dete 250 of the myosin heads bound to the filaments (Ngo et al., 2 251 observed in the presence of Mg.ATP, Mg.ADP or in the rig 252 the polarity of the filaments (see Figs S1, S2-S5). Accordi 253 frequent myosin head orientation in the weak binding sta 254 strong binding state (rigor state) is the one where the 255 positioned toward the minus end of the filament (Fig. 1h 256 Therefore, binding events that occurred towards to the plus 257 for individual upper and lower myosin heads at ATP conce 258 were used in the analysis. 259 260 Figure 2. The kinetics of double-headed myosin motors bound to regulated cTFs. (a-b) 236 Representative time course of binomial binding (a) and heads displacements (b) calculated for the 237 individual HMM molecules (M1-M4) at the given time for upper and lower heads of each HMM 238 molecule in the presence of 0.5 μM ATP and high Ca2+ concentrations. (c) The backward and forward 239 myosin displacements in the cTFs-HMM complex (n=5, 911 events, ~35 HMM molecules); data sets 240 were fitted by sum of two Gaussians (r2=0.99 and r2=0.99, respectively). 196 Kinetics of actin-myosin interaction in the non-regulated and regulated systems 197 The change in the lever arm considers a defined polarity of the actin filament. 215 In the presence of Mg.ATP, myosin heads detach from the filament and re-attach to the same 216 or a new binding site, allowing us to determine the displacement of the myosin head by the 217 change in COM. The calculated displacement in our study is slightly larger than the working 218 stroke size of 5 nm reported for S1 (Capitanio et al., 2006) and slightly smaller than the 219 values obtained from structural studies with single-headed myosin (~10-12 nm) (Geeves et 220 al., 2005). It is comparable with studies performed with myosin molecules evaluated with 221 laser tweezers (Finer et al., 1994; Tyska et al., 1999) and single fiber mechanics (Piazzesi et 222 al., 2002). 223 10 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 224 224 The representative binomial binding traces of the individual upper and lower myosin heads in 225 the F-actin-HMM (Fig. 1d, Movies S1-S3) and in the cTFs-HMM (Fig. 2a, Movie S4) revealed 226 that binding of one HMM head is not necessarily accompanied by the binding of the second 227 HMM head for the given HMM molecule (M1-M4, Figs 1d and 2a). To specifically investigate 228 the coordination between two heads in a molecule we analyzed the binding events at the 229 different ATP concentrations. Tellingly, the binding events of two heads of given HMM 230 molecule bound between two filaments was higher at lower ATP concentrations. At the higher 231 ATP concentrations, the binding of either one head or two heads was approximately equally 232 distributed (Fig. S3). 233 234 234 11 5 235 11 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Probability of HMM binding to the non-regulated and regulated actin filaments 244 CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint high fractional occupancies: ~95% for the non-regulated F-actin and ~79% for the native 264 cTFs. The latter observation is consistent with the idea that the binding sites on cTFs in the 265 absence of Ca2+ and ATP are present in an equilibrium between the blocked, closed and 266 open states (Movie S4) (Matusovsky et al., 2019; Risi et al., 2017; Risi et al., 2021). 267 269 Figure 3. Probability of myosin heads binding to the non-regulated actin filaments. (a-b) 270 Successive HS-AFM images of F-actin-HMM complexes, where HMM heads bind between two actin 271 filaments in the presence of 0.2 μM ATP (a) or 2 μM ATP (b). Dashed color boxes indicate upper and 272 / or lower HMM heads bound between two actin filaments. Numbers on each frame show the time in 273 seconds. Related to Movies S1-S3. Scale bars: 30 nm. (c-e) Probabilities of the individual HMM head 274 binding to the 4 or 5 binding sites on the non-regulated actin filaments in the presence of 0.2 μM ATP 275 (c, 1235 events), 2 μM ATP (d, 567 events) and 10 μM ATP (e, 934 events). The average fractional 276 occupancies by HMM heads for all of the sites for each of the ATP conditions showed a decrease in 277 the occupancy with time (right panels in c-e, n=4). (f) Frequency distributions for the number of 278 myosin heads attached to 4 neighboring binding sites along a given actin filament or thin filament over 279 time (n=8 experiments, ~35 HMM molecules, 898 events). Mean number (± 95% CI) of attached 280 myosin heads relative to the 4 available binding sites given in the inset text. Full lines and dashed 281 lines in the right panels in (c-f) represent the regression lines and 95% confidence intervals, 282 suggesting a decline in the number of myosin molecules in time. 283 Figure 3. Probability of myosin heads binding to the non-regulated a Figure 3. Probability of myosin heads binding to the non-regulated actin filaments. Probability of HMM binding to the non-regulated and regulated actin filaments 244 (a-b) 270 Successive HS-AFM images of F-actin-HMM complexes, where HMM heads bind between two actin 271 filaments in the presence of 0.2 μM ATP (a) or 2 μM ATP (b). Dashed color boxes indicate upper and 272 / or lower HMM heads bound between two actin filaments. Numbers on each frame show the time in 273 seconds. Related to Movies S1-S3. Scale bars: 30 nm. (c-e) Probabilities of the individual HMM head 274 binding to the 4 or 5 binding sites on the non-regulated actin filaments in the presence of 0.2 μM ATP 275 (c, 1235 events), 2 μM ATP (d, 567 events) and 10 μM ATP (e, 934 events). The average fractional 276 occupancies by HMM heads for all of the sites for each of the ATP conditions showed a decrease in 277 the occupancy with time (right panels in c-e, n=4). (f) Frequency distributions for the number of 278 myosin heads attached to 4 neighboring binding sites along a given actin filament or thin filament over 279 time (n=8 experiments, ~35 HMM molecules, 898 events). Mean number (± 95% CI) of attached 280 myosin heads relative to the 4 available binding sites given in the inset text. Full lines and dashed 281 lines in the right panels in (c-f) represent the regression lines and 95% confidence intervals, 282 suggesting a decline in the number of myosin molecules in time. 283 284 Figure 3. Probability of myosin heads binding to the non-regulated actin filaments. (a-b) 270 Successive HS-AFM images of F-actin-HMM complexes, where HMM heads bind between two actin 271 filaments in the presence of 0.2 μM ATP (a) or 2 μM ATP (b). Dashed color boxes indicate upper and 272 / or lower HMM heads bound between two actin filaments. Numbers on each frame show the time in 273 seconds. Related to Movies S1-S3. Scale bars: 30 nm. (c-e) Probabilities of the individual HMM head 274 binding to the 4 or 5 binding sites on the non-regulated actin filaments in the presence of 0.2 μM ATP 275 (c, 1235 events), 2 μM ATP (d, 567 events) and 10 μM ATP (e, 934 events). The average fractional 276 occupancies by HMM heads for all of the sites for each of the ATP conditions showed a decrease in 277 the occupancy with time (right panels in c-e, n=4). Probability of HMM binding to the non-regulated and regulated actin filaments 244 To monitor the probability of binding events between individual myosin heads we applied a 245 probability analysis based on a binary combination: HMM bound to the filaments equals to “1” 246 and HMM detached from the filaments equals to “0”. To use this analysis, we need to 247 evaluate if there is any directional bias in the myosin bindings along the F-actin and cTFs, 248 either towards the barbed plus end or the pointed minus end of the filaments. Therefore, the 249 polarity of F-actin and cTFs complexed with HMM was determined by using the morphology 250 of the myosin heads bound to the filaments (Ngo et al., 2015). The bound myosin heads 251 observed in the presence of Mg.ATP, Mg.ADP or in the rigor state allowed us to determine 252 the polarity of the filaments (see Figs S1, S2-S5). According to our observations the most 253 frequent myosin head orientation in the weak binding state (presence of Mg.ATP-γ-S) or 254 strong binding state (rigor state) is the one where the heads of HMM molecules are 255 positioned toward the minus end of the filament (Fig. 1h, Fig. 3a-b, Fig. 4b-d, Fig. S5). 256 Therefore, binding events that occurred towards to the plus end of the filament (Mn  Mn+1) 257 for individual upper and lower myosin heads at ATP concentrations ranging from 0.2-10 μM 258 were used in the analysis. 259 260 Initially, we tested binding of HMM between two filaments in rigor conditions, i.e., in the 261 absence of ATP and Ca2+, or in the presence of ATP-γ-S, a slowly hydrolyzed analog of ATP 262 (Fig. S5). At these conditions the HMM heads were tightly bound between two filaments with 263 260 Initially, we tested binding of HMM between two filaments in rigor conditions, i.e., in the 261 absence of ATP and Ca2+, or in the presence of ATP-γ-S, a slowly hydrolyzed analog of ATP 262 (Fig. S5). At these conditions the HMM heads were tightly bound between two filaments with 263 12 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . Probability of HMM binding to the non-regulated and regulated actin filaments 244 (f) Frequency distributions for the number of 278 myosin heads attached to 4 neighboring binding sites along a given actin filament or thin filament over 279 time (n=8 experiments, ~35 HMM molecules, 898 events). Mean number (± 95% CI) of attached 280 myosin heads relative to the 4 available binding sites given in the inset text. Full lines and dashed 281 lines in the right panels in (c-f) represent the regression lines and 95% confidence intervals, 282 suggesting a decline in the number of myosin molecules in time. 283 284 13 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint The probability analyses revealed that binding of Mn myosin head to the non-regulated actin 286 filaments did not affect the subsequent bindings of the next Mn+1 molecule (towards the plus 287 end of the filament) in the presence of different ATP concentrations (Fig. 3, Movies S1-S3). 288 While we can observe some random increase in the binding probabilities with 0.2 μM ATP or 289 2 μM ATP concentrations (Fig.3c-d) towards the plus end of the filament, the average 290 fractional occupancy indicates a constant decrease in the occupancy of the binding sites with 291 time, in all ATP concentrations used in this study (Fig. 3c-e, right panels). These data are 292 consistent with the results pooled from 8 different experiments, suggesting that in the F-actin- 293 HMM complex the most frequently observed events represent occupation of one binding site 294 or no binding with the average number of occupied sites calculated as 1.27 ± 0.07 (Fig. 3f). 295 These results suggest that HMM molecules are frequently detached from actin in the 296 presence of ATP due to a lower affinity of myosin to actin in comparison to the affinity to the 297 thin filaments (Fig. S4). Probability of HMM binding to the non-regulated and regulated actin filaments 244 This idea is consistent with the different time evolutions of the 298 number of HMM molecules with actin and thin filaments (Figs. 3f and 4h) in the presence of 299 ATP. It is also consistent with findings that the fractional occupancy of actin-binding sites in 300 the absence of ATP (rigor) or presence of slowly hydrolyzed ATP-γ-S did not change with 301 time (Fig. S5) when HMM is bound to both actin and the thin filaments all the time, 302 suggesting that the myosin heads did not detach from the filaments over the time of the 303 experiment due to interaction with the scanning cantilever tip. 304 305 At the non-activating conditions with thin filaments in the blocked state (0.5 µM ATP, pCa 306 9.0), myosin heads revealed a similar decrease in the binding probability and fractional 307 occupancy (Fig. 4e) compared to the bare F-actin-HMM complex. In contrast, the probability 308 of myosin heads binding to cTFs in the closed state under activating conditions (0.5 µM ATP, 309 14 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint pCa 4.5) was increased (Fig. 4f-g, Movie S4) compared to myosin heads binding to actin 310 filaments (Fig. 3c-d). This feature is reflected in the increased average number of attached 311 myosin heads in the cTFs-HMM complex under activating conditions, almost doubling the 312 average number of bound heads (Fig. 4h) compared to the situation with the F-actin-HMM 313 complex (Fig. 3f). The increase in probability of binding of the myosin heads to the cTFs is 314 also matched in the kymograph images of cTFs-HMM complex at the high [Ca2+] and different 315 ATP concentrations, when compared to the kymograph images obtained from F-actin-HMM 316 complex at the various ATP concentrations (Fig. S4). Probability of HMM binding to the non-regulated and regulated actin filaments 244 The random presence of activated and 317 non-activated sites across cTFs at the relaxing (pCa > 8) or activating (pCa 4.0-4.5) 318 conditions (Risi et al., 2017; Matusovsky et al., 2019) complicated the analysis and can 319 explain the pattern of varied mean probabilities between HMM molecules (Fig. 4f-g). 320 321 321 321 322 15 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Figure 4. Probability of myosin heads binding to the regulated cTFs. (a, c) Fitting the simulated 323 HS-AFM images (Amyot & Flechsig, 2020) of the cTFs at Ca2+- free (a) and Ca2+-bound (c) states into 324 the cTFs molecular structures, scale bar: 2.5 nm. PDB: 6KN7 and 6KN8 for Ca2+ free and high Ca2+ 325 states, respectively. (b, d) Representative HS-AFM snapshots of HMM molecules bound between two 326 cTFs in the presence of 0.5 μM ATP and either low Ca2+ (b) or high Ca2+ (d) concentrations at the 327 indicated times. Scale bars: 30 nm. Note, that the number of bound heads increase with time in (d). 328 (e-g) Probabilities of the HMM heads binding to the regulated cTFs in the presence of 0.5 μM ATP 329 and low Ca2+ (492 events) (e) or 0.5 μM ATP and high Ca2+ concentrations (2830 events) (f) or 2 μM 330 ATP and high Ca2+ concentration (1839 events) (g) with the corresponding average fractional 331 occupancy vs time shown in the right panels. (h) Frequency distributions for the number of myosin 332 heads attached to 4 neighboring binding sites along all studied thin filaments over time in the 333 activating conditions (n=7 experiments, ~35 HMM molecules, 932 events). Mean number (± 95% CI) 334 of attached myosin heads relative to the 4 available binding sites given in the inset text. Probability of HMM binding to the non-regulated and regulated actin filaments 244 Full lines and 335 dashed lines in the right panels in (e-h) represent regression lines and 95% confidence intervals. The 336 analysis suggested a decline in the number of myosin molecules with time, but to a lower degree 337 compared to the F-actin-HMM complex. (i) Cooperative probability of binding of myosin heads in F- 338 actin-HMM and cTFs-HMM complexes (n=7: cTFs-HMM n=8: F-actin-HMM, ~40-50 HMM molecules 339 in each data sets were analyzed; the data points shown as mean value ± 95% CI). 340 Cooperativity in the non-regulated and regulated actin-myosin systems 343 To quantify the observed probability binding pattern in the binary system, we applied the 344 following equation 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘, where C denotes cooperative probability of binding 345 between neighboring myosin heads, n = number of total events or subsequent frames of the 346 experiment; k = number of binding events in the subsequent frames, p = probability of 347 binding, i.e. the ratio between binding events and total events, and (𝑛 𝑘) represents the 348 combination of total and binding events expressed as 𝑛! 𝑘!(𝑛−𝑘)! (see Supplementary Table 1). 349 350 Following this analysis, we found no change in the probability of cooperative binding of the 351 HMM heads to the non-regulated actin filaments. It suggests that the interaction in the F- 352 actin-HMM complex is largely random. The linear regression slope of the cooperative 353 probability binding between neighboring myosin heads in the F-actin-HMM complex showed 354 no significant deviation from zero (p = 0.295) with the Pearson’s r = 0.766 and r2 = 0.59. In 355 contrast, the linear regression slope of the cooperative probability binding between 356 Cooperativity in the non-regulated and regulated actin-myosin systems 343 363 364 365 Discussion 366 In this study, we used a binomial probability analysis to evaluate the potential cooperat 367 between myosin motors while attached to actin or regulated thin filaments. Our experime 368 approach – using myosin motors that can attach between two filaments positioned in par 369 on the surface - is particularly well-suited for this analysis, and we could visualize sev 370 different motors at the same time. Despite this approach has geometrical features that 371 distinct from those of the actin-myosin arrays operating within a muscle sarcomere, it 372 been recently shown that each of the double myosin heads can acquire different lever 373 confirmations and bound two different thin filaments in rigor (Wang et al., 2021). This is 374 true when myosin attaches to actin and thin filaments in the presence of ATP, when e 375 head may be in a different state, at a given time of the ATPase cycle (Matusovsky et 376 2021). This feature may enable myosin double heads to interact with two different 377 filaments within the sarcomere, potentially maximizing muscle power and efficiency. 378 particular interest, is the fact that myosin motors can be assumed as the independent fo 379 neighboring myosin heads in the cTFs-HMM complex showed significant deviation from zero 357 (p = 0.022) with the Pearson’s r = 0.953 and r2 = 0.91 (Fig. 4i). In accordance with these 358 results, the individual fitting for each experiment demonstrated an increase in the cooperative 359 probability of binding of the HMM heads to the regulated cTFs in comparison to that of in F- 360 actin-HMM complex (Figs S6 and S7). This is broadly consistent with cooperativity, although 361 the degree of cooperative binding in the cTFs-HMM was variable between experiments as 362 can be noticed from confidence intervals (Fig. 4i). 363 Cooperativity in the non-regulated and regulated actin-myosin systems 343 To quantify the observed probability binding pattern in the binary system, we applied the 344 following equation 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘, where C denotes cooperative probability of binding 345 between neighboring myosin heads, n = number of total events or subsequent frames of the 346 experiment; k = number of binding events in the subsequent frames, p = probability of 347 binding, i.e. the ratio between binding events and total events, and (𝑛 𝑘) represents the 348 combination of total and binding events expressed as 𝑛! 𝑘!(𝑛−𝑘)! (see Supplementary Table 1). 349 Following this analysis, we found no change in the probability of cooperative binding of the 351 HMM heads to the non-regulated actin filaments. It suggests that the interaction in the F- 352 actin-HMM complex is largely random. The linear regression slope of the cooperative 353 probability binding between neighboring myosin heads in the F-actin-HMM complex showed 354 no significant deviation from zero (p = 0.295) with the Pearson’s r = 0.766 and r2 = 0.59. In 355 contrast, the linear regression slope of the cooperative probability binding between 356 16 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint neighboring myosin heads in the cTFs-HMM complex showed significant deviation from z 357 (p = 0.022) with the Pearson’s r = 0.953 and r2 = 0.91 (Fig. 4i). In accordance with th 358 results, the individual fitting for each experiment demonstrated an increase in the coopera 359 probability of binding of the HMM heads to the regulated cTFs in comparison to that of i 360 actin-HMM complex (Figs S6 and S7). This is broadly consistent with cooperativity, altho 361 the degree of cooperative binding in the cTFs-HMM was variable between experiments 362 can be noticed from confidence intervals (Fig. 4i). Discussion 366 In this study, we used a binomial probability analysis to evaluate the potential cooperativity 367 between myosin motors while attached to actin or regulated thin filaments. Our experimental 368 approach – using myosin motors that can attach between two filaments positioned in parallel 369 on the surface - is particularly well-suited for this analysis, and we could visualize several 370 different motors at the same time. Despite this approach has geometrical features that are 371 distinct from those of the actin-myosin arrays operating within a muscle sarcomere, it has 372 been recently shown that each of the double myosin heads can acquire different lever arm 373 confirmations and bound two different thin filaments in rigor (Wang et al., 2021). This is also 374 true when myosin attaches to actin and thin filaments in the presence of ATP, when each 375 head may be in a different state, at a given time of the ATPase cycle (Matusovsky et al., 376 2021). This feature may enable myosin double heads to interact with two different thin 377 filaments within the sarcomere, potentially maximizing muscle power and efficiency. Of 378 particular interest, is the fact that myosin motors can be assumed as the independent force 379 generators even when connected in small assemblies. 380 17 17 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint The displacements produced by each individual myosin head within a given HMM molecule in 381 our experimental conditions were in the range of ~6-8 nm, slightly larger than the working 382 stroke size of 5 nm reported for myosin S1 (Capitanio et al., 2006) and slightly smaller than 383 the values obtained from structural studies with single-headed myosin (~10-12 nm) (Geeves 384 et al., 2005). Discussion 366 It is comparable with studies performed with myosin molecules evaluated with 385 laser tweezers and single fiber mechanics (Finer et al., 1994; Tyska et al., 1999; Piazzesi et 386 al., 2002). Therefore, our results are consistent with studies utilizing different techniques. 387 388 Cooperativity between myosin motors 389 However, because we have only 408 performed our studies under a limited number of specific conditions, we cannot completely 409 exclude that such allosteric effects occur under certain conditions, e.g. at submicromolar 410 concentrations of Mg.ATP as in some of the experiments of Tokuraku et al 2009. In contrast 411 to the results with pure F-actin we found strong evidence for cooperativity between 412 neighboring thin filament target zones as further considered in detail below. 413 414 Another form of cooperativity has been suggested by X-ray diffraction studies using muscle 415 fibre preparations, indicating that the two heads of a given myosin molecule may bind 416 sequentially to resist stretch of the active muscle (Brunello et al., 2007). Such sequential 417 actions of the two heads have also been suggested (Edman et al., 1997, Huxley & Tideswell, 418 1997; Conibear & Geeves, 1998) to occur during shortening to account for rapid repriming of 419 the myosin power-stroke after a quick release, high power output during shortening and other 420 phenomena. To the best of our knowledge interhead cooperativity has, however, not 421 previously been observed experimentally under dynamic conditions in the presence of ATP. 422 Our demonstration that binding of one myosin head increases the probability for binding of 423 the second head is thus unique by demonstrating the potential for inter-head cooperativity 424 where binding of one head increases the probability of binding of the second head to another. 425 The demonstration for this potential is of interest despite the fact that the distance between 426 neighboring, roughly parallel actin filaments in our study is appreciably larger than in the 427 muscle sarcomere. On the other hand, the inter-filament distance in our experiments is not 428 dynamic conditions using fluorescence microscopy (e.g. Desai et al, 2015). This result does 405 not seem to be consistent with previous findings suggesting that binding of a myosin head 406 allosterically affects the properties of the entire actin filament with potential changes of 407 myosin affinity at other sites (Tokuraku et al., 2009). However, because we have only 408 performed our studies under a limited number of specific conditions, we cannot completely 409 exclude that such allosteric effects occur under certain conditions, e.g. at submicromolar 410 concentrations of Mg.ATP as in some of the experiments of Tokuraku et al 2009. Cooperativity between myosin motors 389 There are several studies suggesting that myosin molecules work cooperatively (Vilfan & 390 Duke, 2003; Hilbert et al., 2013; Kaya et al., 2017; Hwang et al., 2021), and that the work 391 produced by assemblies of motors is different from individual motors. For instance, a study 392 using synthetic myosin filaments measured 4 nm stepwise actin displacements at a high load 393 (>30 pN). Due to the fact that the mechanical work of 4 × 30 pN nm = 120 pN nm ≈ 30 kBT 394 (kB: Bolzmann constant; T absolute temperature) is greater than the free energy of Mg.ATP 395 turnover (25 kBT), the authors concluded that the steps they observed could not be produced 396 by single motors but potentially due to coordinated force generation by several myosin 397 motors (Kaya et al., 2017, Hwang et al., 2021). Despite the fact that theoretical analysis 398 (Duke, 1999; Månsson, 2020) suggests that this finding is consistent with previous models of 399 independent force generators as proposed previously (Huxley, 1957; Huxley, 1988; Hill, 400 1974), it casted some doubt on this concept when motors work in arrays. In this regard, the 401 present study is consistent with fully independent force-generators along the actin filaments. 402 Importantly, the HS-AFM approach allows us to demonstrate that this applies for neighboring 403 actin target zones separated by ~37 nm, appreciably shorter than possible to resolve under 404 18 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint dynamic conditions using fluorescence microscopy (e.g. Desai et al, 2015). This result does 405 not seem to be consistent with previous findings suggesting that binding of a myosin head 406 allosterically affects the properties of the entire actin filament with potential changes of 407 myosin affinity at other sites (Tokuraku et al., 2009). Cooperativity between myosin motors 389 In contrast 411 to the results with pure F-actin we found strong evidence for cooperativity between 412 neighboring thin filament target zones as further considered in detail below. 413 414 Another form of cooperativity has been suggested by X-ray diffraction studies using muscle 415 fibre preparations, indicating that the two heads of a given myosin molecule may bind 416 sequentially to resist stretch of the active muscle (Brunello et al., 2007). Such sequential 417 actions of the two heads have also been suggested (Edman et al., 1997, Huxley & Tideswell, 418 1997; Conibear & Geeves, 1998) to occur during shortening to account for rapid repriming of 419 the myosin power-stroke after a quick release, high power output during shortening and other 420 phenomena. To the best of our knowledge interhead cooperativity has, however, not 421 previously been observed experimentally under dynamic conditions in the presence of ATP. 422 Our demonstration that binding of one myosin head increases the probability for binding of 423 the second head is thus unique by demonstrating the potential for inter-head cooperativity 424 where binding of one head increases the probability of binding of the second head to another. 425 The demonstration for this potential is of interest despite the fact that the distance between 426 neighboring, roughly parallel actin filaments in our study is appreciably larger than in the 427 muscle sarcomere. On the other hand, the inter-filament distance in our experiments is not 428 19 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint very different from the next-neighbour inter-filament distance between actin filaments in the 429 hexagonal arrangement of thin filaments that surround each thick filament in the sarcomere. Cooperativity between myosin motors 389 430 In contrast to the inter-head cooperativity involving binding each of the HMM heads between 431 two filaments, we did not study cooperativity of the double-head HMM binding to a filament 432 (similar to in vitro motility or laser-trapping), due to the uncertainty to recognize the binding of 433 specific HMM molecules in subsequent HS-AFM frames (Matusovsky et al., 2021). 434 435 Cooperativity between myosin motors that involves activation of the thin filament 436 Studies have shown that myosin binding to actin is required for full activation of the thin 437 filament (McKillop and Geeves 1993, Smith & Geeves, 2003; Desai et al., 2015). When 438 myosin binds to actin, it may directly affect the regulatory system by changing the 439 conformation of Tm, such that other myosin heads can attach to thin filaments (Geeves & 440 Holmes, 1999; Gordon, 2000; Smith et al., 2003). According to this model, with the transition 441 from weak to strong actin–myosin binding, the myosin heads transfer Tm to an open state, 442 making neighboring myosin binding sites on actin available for myosin binding. Our data are 443 consistent with this model, as we observed that the binding of one motor to the activated thin 444 filament (pCa 4.5) has changes the attachment kinetics of neighboring motors compared to 445 non-activated thin filaments (pCa 9.0) in the blocked state or the bare actin filaments. Most 446 specifically, when one motor is bound to the activated thin filament at the pCa 4.5, it moves 447 the thin filament from the closed to the open state, which allows for further motor binding at 448 nearby sites. 449 In a previous study, we showed that the interaction of HMM with cTFs caused a change in 451 the thin filament conformation, both in the absence and presence of Ca2+, and in the absence 452 20 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint and presence of different concentrations of ATP (Matusovsky et al., 2019).Our new data 453 strengthen those findings and corroborate the idea that cooperativity of myosin heads in 454 striated muscles is defined by thin filaments and their state of activation. We evaluated 455 whether one head in a HMM molecule could activate the thin filament in the presence of ATP 456 at low or high Ca2+ concentrations. Cooperativity between myosin motors that involves activation of the thin filament 436 Under non-activating conditions (presence of ATP, pCa 457 9.0) when cTFs were in the blocked state, myosin heads were able to bind to cTFs but not 458 able to switch the filaments from the blocked to the closed state, showing a similar decrease 459 in the binding probability and fractional occupancy (Fig. 4e) when compared to the F-actin- 460 HMM complex (Fig.3). Thus, binding of the two myosin heads are required for the transition 461 of a thin filament from the blocked to the close state (Fig. 4). However, the situation is 462 changed if myosin heads bind to cTFs under activation conditions (presence of ATP and pCa 463 4.5), showing an increase in the probability of binding and the relative number of motors 464 attached to thin filaments, as a result of a first myosin binding (Figs 3 and 4g, Movies S3, S4). 465 These results suggest that one head (upper or lower heads of a given HMM molecule bound 466 between two filaments) is able to switch a thin filament from the closed to the open state. 467 468 strengthen those findings and corroborate the idea that cooperativity of myosin heads in 454 striated muscles is defined by thin filaments and their state of activation. We evaluated 455 whether one head in a HMM molecule could activate the thin filament in the presence of ATP 456 at low or high Ca2+ concentrations. Under non-activating conditions (presence of ATP, pCa 457 9.0) when cTFs were in the blocked state, myosin heads were able to bind to cTFs but not 458 able to switch the filaments from the blocked to the closed state, showing a similar decrease 459 in the binding probability and fractional occupancy (Fig. 4e) when compared to the F-actin- 460 HMM complex (Fig.3). Thus, binding of the two myosin heads are required for the transition 461 of a thin filament from the blocked to the close state (Fig. 4). However, the situation is 462 changed if myosin heads bind to cTFs under activation conditions (presence of ATP and pCa 463 4.5), showing an increase in the probability of binding and the relative number of motors 464 attached to thin filaments, as a result of a first myosin binding (Figs 3 and 4g, Movies S3, S4). Cooperativity between myosin motors that involves activation of the thin filament 436 ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 477 To summarize, our data suggest that cooperativity between neighboring myosin molecule 478 along a filament is primarily defined by the state of thin filament activation. In contrast, we find 479 no evidence for cooperative effects attributed to allosteric changes along pure actin filaments. 480 481 482 Materials and Methods 483 Proteins 484 Native thin filaments were purified from rabbit right and left ventricular cardiac muscle that 485 had been glycerinated and actin was purified from acetone powder of rabbit skeletal muscle 486 (Sigma-Aldrich, USA), following a protocol previously used in our laboratory (Matusovsky et 487 al., 2019). The double-headed skeletal myosin II was purified from rabbit psoas muscle and 488 HMM fragments were prepared by proteolysis of the myosin with α-chymotrypsin (Oakville, 489 Ontario, Canada) as previously described (Cheng et al., 2019). Prior of the HS-AFM 490 experiments, HMM, F-actin and thin filaments were tested for their functionality using in-vitro 491 motility and Mg2+-ATPase activity assays, as previously described (Matusovsky et al., 2019). 492 493 The lipid bilayer template surface and experimental design 494 The lipid composition for HS-AFM imaging contained 1,2-Dipalmitoyl-sn-glycero-3- 495 phosphocholine (DPPC, Avanti Polar Lipids), 1,2-Dipalmitoyl-3-trimethylammonium-propane 496 (DPTAP, Avanti Polar Lipids) and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(cap 497 biotinyl) (biotin-cap-DPPE, Avanti Polar Lipids). DPPC: DPTAP: biotin-cap-DPPE were mixed 498 477 Native thin filaments were purified from rabbit right and left ventricular cardiac muscle that 485 had been glycerinated and actin was purified from acetone powder of rabbit skeletal muscle 486 (Sigma-Aldrich, USA), following a protocol previously used in our laboratory (Matusovsky et 487 al., 2019). The double-headed skeletal myosin II was purified from rabbit psoas muscle and 488 HMM fragments were prepared by proteolysis of the myosin with α-chymotrypsin (Oakville, 489 Ontario, Canada) as previously described (Cheng et al., 2019). Prior of the HS-AFM 490 experiments, HMM, F-actin and thin filaments were tested for their functionality using in-vitro 491 motility and Mg2+-ATPase activity assays, as previously described (Matusovsky et al., 2019). 492 493 Cooperativity between myosin motors that involves activation of the thin filament 436 465 These results suggest that one head (upper or lower heads of a given HMM molecule bound 466 between two filaments) is able to switch a thin filament from the closed to the open state. 467 468 In addition to the cooperative phenomena considered above, our results also demonstrate 469 higher affinity of myosin heads to the thin filaments in comparison to the actin filaments. This 470 follows from the higher average number of the HMM heads bound (2.32 ± 0.06) to cTFs (Fig. 471 4h) than to the non-regulated actin filaments (1.27 ± 0.07; Fig. 3f) and the slower decline in 472 the total number of available heads in the former case. These findings are broadly consistent 473 with previous observations that both tension and the average number of attached cross- 474 bridges was increased in actin-reconstituted skinned muscle fibres after furhter reconstitution 475 with thin filament regulatory proteins (Fujita et al, 2002). 476 In addition to the cooperative phenomena considered above, our results also demonstrate 469 higher affinity of myosin heads to the thin filaments in comparison to the actin filaments. This 470 follows from the higher average number of the HMM heads bound (2.32 ± 0.06) to cTFs (Fig. 471 4h) than to the non-regulated actin filaments (1.27 ± 0.07; Fig. 3f) and the slower decline in 472 the total number of available heads in the former case. These findings are broadly consistent 473 with previous observations that both tension and the average number of attached cross- 474 bridges was increased in actin-reconstituted skinned muscle fibres after furhter reconstitution 475 with thin filament regulatory proteins (Fujita et al, 2002). 476 21 21 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. The lipid bilayer template surface and experimental design 494 The lipid composition for HS-AFM imaging contained 1,2-Dipalmitoyl-sn-glycero-3- 495 phosphocholine (DPPC, Avanti Polar Lipids), 1,2-Dipalmitoyl-3-trimethylammonium-propane 496 (DPTAP, Avanti Polar Lipids) and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(cap 497 biotinyl) (biotin-cap-DPPE, Avanti Polar Lipids). DPPC: DPTAP: biotin-cap-DPPE were mixed 498 in a weight ratio of 89:10:1. The preparation of lipid vesicles and deposition on mica substrate 499 22 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint to form a mica-supported lipid bilayer surface (mica-SLB) has been previously described 500 (Matusovsky et al., 2021). 501 The mica-SLB surface was rinsed with the buffer A, containing 25 mM KCl, 2 mM MgCl2, 502 0.25 mM EGTA, 1.25 mM Imidazole-HCl, 0.5 mM DTT, (pH 7.0). Subsequently, 2.8 µl of 503 either 7 µM non-regulated actin filaments or 1.0 µM regulated cTFs diluted in the buffer A 504 were deposited on the mica-SLB surface and incubated for 10 minutes in a wet chamber. At 505 these conditions, many filaments were attached to the surface in close proximity to each 506 other. The distance distributions between two parallel non-regulated actin filaments and 507 regulated cTFs are shown in Fig.S2. The observed distances were enough for binding the 508 HMM heads between two parallel filaments, allowing counting of the exact number of HMM 509 molecules at the given time of the experiment. We studied cooperativity of binding of the 510 myosin heads in F-actin-HMM or cTF-HMM complexes using the following experimental 511 conditions: i) nucleotide-free (NF) state; ii) presence of ATP analogs (ATP- γ -S); iii) presence 512 of ATP and Ca2+. 513 HS-AFM imaging of cTFs-HMM complex 531 The procedure for imaging the cTFs-HMM complex was similar to that explained above for 532 non-regulated actin filaments. Imaging of the cTFs-HMM complex at low Ca2+ (pCa 9.0) or 533 high Ca2+ (pCa 4.5) concentrations, using skeletal muscle HMM was performed in the 534 following way: 2-20 μL of TFs (1 μM) in the buffer A (relaxing conditions, absence of Ca2+) 535 were placed on a mica-SLB surface for 10 min in the wet chamber and unbound cTFs were 536 removed by exchanging for the buffer B, containing low or high Ca2+ concentrations. Then, 537 3.0 μL of skeletal muscle HMM (8 nM) in buffer A was placed on top of the mica-SLB surface 538 with bound cTFs for 10 min in the wet chamber. Unbound HMM was washed out by buffer A 539 followed by washing several times with appropriate buffer B with low or high Ca2+ 540 concentrations containing 0.5-2 μM of caged or non-caged ATP as desired. 541 HS-AFM imaging of F-actin-HMM complex 515 After rinsing unbound actin filaments with buffer A, 3.0 µl of 8 nM HMM diluted in buffer A 516 was placed on top of non-regulated actin filaments on the mica-SLB surface, and incubated 517 for an additional 3 minutes. The F-actin-HMM complex in nucleotide-free (NF) conditions was 518 rinsed by 10 μl of buffer A, containing either NPE-caged ATP, non-caged ATP (0.5, 2 or 10 519 µM) or 0.5 µM non-hydrolyzable ATP-γ-S. NPE-caged ATP (adenosine 5’-triphosphate, P3- 520 (1-(2-nitrophenyl ethyl) ester) (Invitrogen) dissolved in attachment buffer was photolyzed in 521 the AFM chamber using an UV light source at 340 nm. A delay of ~5-10 seconds was found 522 after activation of caged ATP, likely because caged ATP molecules were in solution and 523 23 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint required this time to attach to and get hydrolyzed in the motor domain of HMM. To ensure 524 nucleotide free conditions 1 U/µl of apyrase was added to the solution. Further, 1 U/ml of 525 hexokinase and 10 mM glucose were added to the ADP solutions to remove contaminating 526 ATP. The F-actin-HMM complex was formed on the mica-SLB surface in the buffer A with 527 low (pCa 9.0) or high (pCa 4.5) Ca2+ concentrations to ensure similar experimental conditions 528 as for cTFs-HMM complex. 529 Probability of binding, fractional occupancies and cooperativity analysis 543 Probability of the HMM heads binding to the non-regulated or regulated actin filaments was 544 calculated by binomial distribution evaluated in HS-AFM experiments. This assumes that the 545 binding situation in each frame is treated as an independent event because each myosin 546 head is assumed to undergo independent cycling (possibly several times per frame). The 547 24 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint bound and unbound events (0 – no binding, 1 – binding) were visualized directly to compute 548 probabilities for the binding-unbinding process as a ratio of the binding events to the total 549 number of events in each independent frame. Our experimental design allows to visualize 4-6 550 HMM molecules (or 8-12 individual heads) bound between two filaments. Two typical 551 scanning views and rates were used: 150 × 75 nm2 (80 × 40 pixels2) at the 6.7-10 frame per 552 second (fps) and 200 × 200 nm2 (120 × 120 pixels2) at the 2 fps. 553 554 Fractional Occupancy (θ) is the ratio of the actin-binding sites occupied by HMM heads to the 555 total number of the actin binding sites experimentally observed in the given time of the 556 experiment and calculated from: 557 558 θ = [bound sites] [bound sites]+[unbound sites] 559 Cooperative probability which is related to the probability of binding was calculated from: 560 561 bound and unbound events (0 – no binding, 1 – binding) were visualized directly to compute 548 probabilities for the binding-unbinding process as a ratio of the binding events to the total 549 number of events in each independent frame. Our experimental design allows to visualize 4-6 550 HMM molecules (or 8-12 individual heads) bound between two filaments. included both binding events (head was bound to actin filament) and unbinding events (head 570 was unbind from actin filament). 571 included both binding events (head was bound to actin filament) and unbinding events (head 570 was unbind from actin filament). 571 572 Analysis of the myosin displacements 573 To analyze the HMM displacement, each of the HMM heads bound between two parallel non- 574 regulated or regulated actin filaments were tracked individually in successive HS-AFM 575 frames. The tracked parameters included the height of the HMM head used for subsequent 576 determination of the center of mass (COM) in each myosin head. The height of the HMM 577 heads was determined in semi-automatic mode using the x, y, and z data of the HS-AFM 578 frames in Kodec software (v. 4.4.7.39) (Ngo et al., 2015). The x and y data correspond to the 579 lateral coordinates, while the observed z values correspond to the highest point in the center 580 of the HMM heads. To obtain the z values for the highest point in HMM head(s) the image 581 was automatically searched within a 5 × 5 pixels area. Next, the obtained height values and 582 x, y positions within the 5 × 5 pixels area were used to automatically calculate the COM. To 583 obtain the accurate COM values, the height of the surface outside of the actin-HMM position 584 was subtracted from the average COM of the HMM heads. The displacement size was 585 calculated as a difference in the COM position of HMM head in the reference frame and the 586 next frame, in successive HS-AFM frames. The forward and backward displacements were 587 calculated for each myosin head, plotted and fitted by sum of two Gaussians. The 588 Probability of binding, fractional occupancies and cooperativity analysis 543 Two typical 551 scanning views and rates were used: 150 × 75 nm2 (80 × 40 pixels2) at the 6.7-10 frame per 552 second (fps) and 200 × 200 nm2 (120 × 120 pixels2) at the 2 fps. 553 Fractional Occupancy (θ) is the ratio of the actin-binding sites occupied by HMM heads to the 555 total number of the actin binding sites experimentally observed in the given time of the 556 experiment and calculated from: 557 Fractional Occupancy (θ) is the ratio of the actin-binding sites occupied by HMM heads to the 555 total number of the actin binding sites experimentally observed in the given time of the 556 experiment and calculated from: 557 Cooperative probability which is related to the probability of binding was calculated from: 560 Cooperative probability which is related to the probability of binding was calculated from: 560 Cooperative probability which is related to the probability of binding was calculated from: 560 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘 562 where C denotes cooperative probability of binding, n = number of total events or subsequent 563 frames of the experiment; k = number of binding events in the experiment, p = probability of 564 binding, 1-p = probability of unbinding and (𝑛 𝑘) represents the combination of the total and 565 binding events expressed as: 𝑛! 𝑘!(𝑛−𝑘)!. 566 where C denotes cooperative probability of binding, n number of total events or subsequent 563 frames of the experiment; k = number of binding events in the experiment, p = probability of 564 binding, 1-p = probability of unbinding and (𝑛 𝑘) represents the combination of the total and 565 ! binding events expressed as: 𝑛! 𝑘!(𝑛−𝑘)!. 566 binding events expressed as: 𝑛! 𝑘!(𝑛−𝑘)!. 566 The events for each myosin head calculated from the reference frame, i.e. a moment when 568 the head was bound to the filament until the end of image acquisition. The total events 569 The events for each myosin head calculated from the reference frame, i.e. a moment when 568 the head was bound to the filament until the end of image acquisition. The total events 569 25 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. Probability of binding, fractional occupancies and cooperativity analysis 543 ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint included both binding events (head was bound to actin filament) and unbinding events (head 570 was unbind from actin filament). 571 HS-AFM system and cantilevers 599 The experiments were performed on a tapping-mode HS-AFM system (RIBM) (Ando et al., 600 2001), equipped with an additional UV laser. Olympus cantilevers BL-AC10DS-A2 with the 601 following parameters were used: spring constant 0.08-0.15 N/m; quality factor in water ~1.4- 602 1.6; resonance frequency in water 0.6-1.2 MHz. The additional carbon probe tip was 603 fabricated on the tip of a cantilever by electron-beam deposition and sharped by plasma 604 etcher, giving a ~4 nm tip apex. The tip–sample loading force can be modulated and 605 decreased by the free oscillation peak-to-peak amplitude (A0) of the cantilever set to ~2.0 nm 606 and the amplitude set point adjusted to more than 0.9 A0. 607 Analysis of the myosin displacements 573 It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint data with the sampling rate of 589 events for the non-regulated actin filaments a 593 events for the regulated cTFs. 594 595 Cross-sectional analysis 596 Cross-sectional analysis was performed by Kodec software (Ngo et al., 2015) to calcu 597 distance between two filaments (Fig. S2). 598 HS-AFM system and cantilevers 599 The experiments were performed on a tapping-mode HS-AFM system (RIBM) (Ando 600 2001), equipped with an additional UV laser. Olympus cantilevers BL-AC10DS-A2 w 601 following parameters were used: spring constant 0.08-0.15 N/m; quality factor in wate 602 1.6; resonance frequency in water 0.6-1.2 MHz. The additional carbon probe t 603 fabricated on the tip of a cantilever by electron-beam deposition and sharped by 604 etcher, giving a ~4 nm tip apex. The tip–sample loading force can be modulate 605 decreased by the free oscillation peak-to-peak amplitude (A0) of the cantilever set to ~ 606 and the amplitude set point adjusted to more than 0.9 A0. 607 608 Data analysis and processing of HS-AFM images 609 To remove spike noise and to make the xy-plane flat, the HS-AFM images were pro 610 with low-pass filtering using Kodec software (4.4.7.39). The COM and cross-cor 611 analyses were performed in Kodec software. Fittings of equations to the observed dat 612 data with the sampling rate of 589 events for the non-regulated actin filaments and 911 593 events for the regulated cTFs. 594 Cross-sectional analysis was performed by Kodec software (Ngo et al., 2015) to calculate the 597 distance between two filaments (Fig. S2). 598 Analysis of the myosin displacements 573 To analyze the HMM displacement, each of the HMM heads bound between two parallel non- 574 regulated or regulated actin filaments were tracked individually in successive HS-AFM 575 frames. The tracked parameters included the height of the HMM head used for subsequent 576 determination of the center of mass (COM) in each myosin head. The height of the HMM 577 heads was determined in semi-automatic mode using the x, y, and z data of the HS-AFM 578 frames in Kodec software (v. 4.4.7.39) (Ngo et al., 2015). The x and y data correspond to the 579 lateral coordinates, while the observed z values correspond to the highest point in the center 580 of the HMM heads. To obtain the z values for the highest point in HMM head(s) the image 581 was automatically searched within a 5 × 5 pixels area. Next, the obtained height values and 582 x, y positions within the 5 × 5 pixels area were used to automatically calculate the COM. To 583 obtain the accurate COM values, the height of the surface outside of the actin-HMM position 584 was subtracted from the average COM of the HMM heads. The displacement size was 585 calculated as a difference in the COM position of HMM head in the reference frame and the 586 next frame, in successive HS-AFM frames. The forward and backward displacements were 587 calculated for each myosin head, plotted and fitted by sum of two Gaussians. The 588 displacement size of the upper and lower HMM heads did not differ between each other, 589 although the frequency and binding events were not correlated between two heads within one 590 HMM molecule. The displacement size was also not affected by the range of the ATP 591 concentrations used in our experiments (0.2 μM, 0.5 μM, 2 μM, 10 μM), thus we averaged the 592 26 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Data analysis and processing of HS-AFM images 609 To remove spike noise and to make the xy-plane flat, the HS-AFM images were processed 610 with low-pass filtering using Kodec software (4.4.7.39). The COM and cross-correlation 611 analyses were performed in Kodec software. Fittings of equations to the observed data were 612 performed in GraphPad Prism software (v.9.3.0). Values are reported as mean ± Standard 613 Deviation or 95% Confidential Intervals throughout the paper as indicated. Number of n 614 equals to independent experiments. A level of significance of p < 0.05 was used for all 615 analyses. 616 27 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 617 Data availability 618 All data required for evaluation of the conclusions in the paper are present in the main body 619 of the paper and/or in the Supporting Information. 620 621 Acknowledgments 622 This work was supported by the Natural Science and Engineering Research Council of 623 Canada (to D.E.R). A.M. was supported by the Swedish Research Council (grant number 624 2019-03456). D.E.R. is a Canada Research Chair in Muscle Biophysics. We thank Dr. Y.-S. 625 Cheng for the HMM preparation. 626 627 Author contributions 628 O.S.M. and D.E.R. designed research; O.S.M. performed HS-AFM experiments and all 629 authors were involved in analysis and interpretation of the data. O.S.M., A.M., D.E.R. wrote 630 the paper and all authors approved the final version of the manuscript. 631 632 Ethics declarations 633 Competing interests 634 The authors declare no competing interests. 635 617 Data availability 618 All data required for evaluation of the conclusions in the paper are present in the main body 619 of the paper and/or in the Supporting Information. 620 All data required for evaluation of the conclusions in the paper are present in the main body 619 of the paper and/or in the Supporting Information. 620 Acknowledgments 622 This work was supported by the Natural Science and Engineering Research Council of 623 Canada (to D.E.R). A.M. was supported by the Swedish Research Council (grant number 624 2019-03456). D.E.R. is a Canada Research Chair in Muscle Biophysics. We thank Dr. Y.-S. 625 Cheng for the HMM preparation. 626 Cheng for the HMM preparation. 626 627 Author contributions 628 O.S.M. and D.E.R. designed research; O.S.M. performed HS-AFM experiments and all 629 authors were involved in analysis and interpretation of the data. O.S.M., A.M., D.E.R. wrote 630 the paper and all authors approved the final version of the manuscript. 631 632 Ethics declarations 633 Competing interests 634 The authors declare no competing interests. 635 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint References 636 Rayment, I. et al. Three-dimensional structure of myosin subfragment-1: a molecular motor. 637 Science 261, 50-58 (1993). 638 Rayment, I. et al. Three-dimensional structure of myosin subfragment-1: a molecular motor. 637 Science 261, 50-58 (1993). 638 Fisher, A. J. et al. Structural studies of myosin:nucleotide complexes: a revised model for the 639 molecular basis of muscle contraction. Biophys.J. 68, 19S-26S (1995). 640 Fisher, A. J. et al. Structural studies of myosin:nucleotide complexes: a revised model for the 639 molecular basis of muscle contraction. Biophys.J. 68, 19S-26S (1995). 640 Mansson, A., Usaj, M., Moretto, L. & Rassier, D. E. Do actomyosin single-molecule 641 mechanics data predict mechanics of contracting muscle? Int.J.Mol.Sci. 19, 1863 (2018). 642 Mansson, A., Usaj, M., Moretto, L. & Rassier, D. E. Do actomyosin single-molecule 641 mechanics data predict mechanics of contracting muscle? Int.J.Mol.Sci. 19, 1863 (2018). 642 Robert-Paganin, J., Pylypenko, O., Kikuti, C., Sweeney, H. L. & Houdusse, A. Force 643 generation by myosin motors: A structural perspective. Chem.Rev. 120, 5-35 (2020). 644 Robert-Paganin, J., Pylypenko, O., Kikuti, C., Sweeney, H. L. & Houdusse, A. Force 643 generation by myosin motors: A structural perspective. Chem.Rev. 120, 5-35 (2020). 644 Finer, J. T., Simmons, R. M. & Spudich, J. A. Single myosin molecule mechanics: piconewton 645 forces and nanometre steps. Nature 368, 113-119 (1994). 646 Finer, J. T., Simmons, R. M. & Spudich, J. A. Single myosin molecule mechanics: piconewton 645 forces and nanometre steps. Nature 368, 113-119 (1994). 646 Ishijima, A. et al. Single-molecule analysis of the actomyosin motor using nano-manipulation. 647 Biochem.Biophys.Res.Commun. 199, 1057-1063 (1994). 648 Ishijima, A. et al. Single-molecule analysis of the actomyosin motor using nano-manipulation. 647 Biochem.Biophys.Res.Commun. 199, 1057-1063 (1994). 648 Yanagida, T. & Ishijima, A. Forces and steps generated by single myosin molecules. 649 Biophys.J. 68, 312S-318S (1995). 650 Yanagida, T. & Ishijima, A. Forces and steps generated by single myosin molecules. 649 Biophys.J. 68, 312S-318S (1995). 650 Kaya, M. & Higuchi, H. Nonlinear elasticity and an 8-nm working stroke of single myosin 651 molecules in myofilaments. Science 329, 686-689 (2010). 652 Kaya, M. & Higuchi, H. Nonlinear elasticity and an 8-nm working stroke of single myosin 651 molecules in myofilaments. Science 329, 686-689 (2010). 652 Kaya, M., Tani, Y., Washio, T., Hisada, T. & Higuchi, H. Coordinated force generation of 653 skeletal myosins in myofilaments through motor coupling. Nat.Commun. 8, 16036 (2017). 654 Pertici, I. et al. 627 Author contributions 628 O.S.M. and D.E.R. designed research; O.S.M. performed HS-AFM experiments and all 29 authors were involved in analysis and interpretation of the data. O.S.M., A.M., D.E.R. wrote 30 the paper and all authors approved the final version of the manuscript. 31 28 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint References 636 A myosin II nanomachine mimicking the striated muscle. Nat.Commun. 9, 655 3532 (2018). 656 Cheng, Y. S., Matusovskiy, O. S. & Rassier, D. E. Cleavage of loops 1 and 2 in skeletal 657 muscle heavy meromyosin (HMM) leads to a decreased function. Arch.Biochem.Biophys. 658 661, 168-177 (2019). 659 Cheng, Y. S., de Souza Leite, F. & Rassier, D. E. The load dependence and the force- 660 velocity relation in intact myosin filaments from skeletal and smooth muscles. 661 Am.J.Physiol.Cell Physiol. 318, C103-C110 (2020). 662 Howard, J. Molecular motors: structural adaptations to cellular functions. Nature 389, 561- 663 567 (1997). 664 Debold, E. P., Patlak, J. B. & Warshaw, D. M. Slip sliding away: load-dependence of velocity 665 generated by skeletal muscle myosin molecules in the laser trap. Biophys.J. 89, L34-L36 666 (2005). 667 29 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Lowey, S. et al. Hypertrophic cardiomyopathy R403Q mutation in rabbit beta-myosin reduces 668 contractile function at the molecular and myofibrillar levels. Proc.Natl.Acad.Sci.USA 115, 669 11238-11243 (2018). 670 Lowey, S. et al. Hypertrophic cardiomyopathy R403Q mutation in rabbit beta-myosin reduces 668 contractile function at the molecular and myofibrillar levels. Proc.Natl.Acad.Sci.USA 115, 669 11238-11243 (2018). 670 Edman, K. A. & Hwang, J. C. The force-velocity relationship in vertebrate muscle fibres at 671 varied tonicity of the extracellular medium. J.Physiol. 269, 255-272 (1977). 672 Edman, K. A. & Hwang, J. C. The force-velocity relationship in vertebrate muscle fibres at 671 varied tonicity of the extracellular medium. J.Physiol. 269, 255-272 (1977). 672 Mansson, A. Comparing models with one versus multiple myosin-binding sites per actin 673 target zone: The power of simplicity. J.Gen.Physiol. 151, 578-592 (2019). 674 Mansson, A. Comparing models with one versus multiple myosin-binding sites per actin 673 target zone: The power of simplicity. J.Gen.Physiol. 151, 578-592 (2019). 674 Capitanio, M. et al. References 636 Ultrafast force-clamp spectroscopy of single molecules reveals load 675 dependence of myosin working stroke. Nat.Methods 9, 1013-1019 (2012). 676 Capitanio, M. et al. Ultrafast force-clamp spectroscopy of single molecules reveals load 675 dependence of myosin working stroke. Nat.Methods 9, 1013-1019 (2012). 676 Sung, J. et al. Harmonic force spectroscopy measures load-dependent kinetics of individual 677 human beta-cardiac myosin molecules. Nat.Commun. 6, 7931 (2015). 678 Huxley, A. F. Muscle structure and theories of contraction. Prog.Biophys.Biophys.Chem. 7, 679 255-318 (1957). 680 Huxley, A. F. & Tideswell, S. Rapid regeneration of power stroke in contracting muscle by 681 attachment of second myosin head. J Muscle Res Cell Motil. 18, 111-114 (1997). 682 Brunello, E. et al. Skeletal muscle resists stretch by rapid binding of the second motor domain 683 of myosin to actin. Proc.Natl.Acad.Sci.USA 104, 20114-9 (2007). 684 Caremani, M., Melli, L., Dolfi, M., Lombardi, V., Linari, M. The working stroke of the myosin II 685 motor in muscle is not tightly coupled to release of orthophosphate from its active site. J 686 Physiol. 591, 5187-5205 (2013). 687 Rahman, M. A., Usaj, M., Rassier, D. E. & Mansson, A. Blebbistatin effects expose hidden 688 secrets in the force-generating cycle of actin and myosin. Biophys.J. 115, 386-397 (2018). 689 Irving, M., Lombardi, V., Piazzesi, G. & Ferenczi, M. A. Myosin head movements are 690 synchronous with the elementary force-generating process in muscle. Nature 357, 156-158 691 (1992). 692 Linari, M. et al. Force generation by skeletal muscle is controlled by mechanosensing in 693 myosin filaments. Nature 528, 276-279 (2015). 694 Orlova, A. & Egelman, E. H. A conformational change in the actin subunit can change the 695 flexibility of the actin filament. J.Mol.Biol. 232, 334-341 (1993). 696 Orlova, A. & Egelman, E. H. A conformational change in the actin subunit can change the 695 flexibility of the actin filament. J.Mol.Biol. 232, 334-341 (1993). 696 Tokuraku, K., Kurogi, R., Toya, R. & Uyeda, T. Q. Novel mode of cooperative binding 697 between myosin and Mg2+ -actin filaments in the presence of low concentrations of ATP. 698 J.Mol.Biol. 386, 149-162 (2009). 699 Tokuraku, K., Kurogi, R., Toya, R. & Uyeda, T. Q. Novel mode of cooperative binding 697 between myosin and Mg2+ -actin filaments in the presence of low concentrations of ATP. 698 J.Mol.Biol. 386, 149-162 (2009). 699 Tokuraku, K., Kurogi, R., Toya, R. & Uyeda, T. Q. References 636 Novel mode of cooperative binding 697 between myosin and Mg2+ -actin filaments in the presence of low concentrations of ATP. 698 J.Mol.Biol. 386, 149-162 (2009). 699 30 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Prochniewicz, E. et al. Myosin isoform determines the conformational dynamics and 700 cooperativity of actin filaments in the strongly bound actomyosin complex. J.Mol.Biol. 396, 701 501-509 (2010). 702 Prochniewicz, E. et al. Myosin isoform determines the conformational dynamics and 700 cooperativity of actin filaments in the strongly bound actomyosin complex. J.Mol.Biol. 396, 701 501-509 (2010). 702 Prochniewicz, E. et al. Myosin isoform determines the conformational dynamics and 700 cooperativity of actin filaments in the strongly bound actomyosin complex. J.Mol.Biol. 396, 701 501-509 (2010). 702 Gordon, A. M., Homsher, E. & Regnier, M. Regulation of contraction in striated muscle. 703 Physiol.Rev. 80, 853-924 (2000). 704 Galinska-Rakoczy, A. et al. Structural basis for the regulation of muscle contraction by 705 troponin and tropomyosin. J.Mol.Biol. 379, 929-935 (2008). 706 Lehman, W., Galinska-Rakoczy, A., Hatch, V., Tobacman, L. S. & Craig, R. Structural basis 707 for the activation of muscle contraction by troponin and tropomyosin. J.Mol.Biol. 388, 673-681 708 (2009). 709 Matusovsky, O. S., Mansson, A., Persson, M., Cheng, Y. S. & Rassier, D. E. High-speed 710 AFM reveals subsecond dynamics of cardiac thin filaments upon Ca2+ activation and heavy 711 meromyosin binding. Proc.Natl.Acad.Sci.USA 116, 16384-16393 (2019). 712 McKillop, D. F. & Geeves, M. Regulation of the interaction between actin and myosin 713 subfragment 1: evidence for three states of the thin filament. Biophys.J. 65, 693-701 (1993). 714 Smith, D. A. & Geeves, M. A. Cooperative regulation of myosin-actin interactions by a 715 continuous flexible chain II: actin-tropomyosin-troponin and regulation by calcium. Biophys.J. 716 84, 3168-3180 (2003). 717 Desai, R., Geeves, M. A. & Kad, N. M. References 636 PLoS Comput Biol. 16, e1008444 (2020). 733 Amyot, R. & Flechsig, H. BioAFMviewer: An interactive interface for simulated AFM scanning 732 of biomolecular structures and dynamics. PLoS Comput Biol. 16, e1008444 (2020). 733 Capitanio, M. et al. Two independent mechanical events in the interaction cycle of skeletal 734 muscle myosin with actin. Proc.Natl.Acad.Sci.USA 103, 87-92 (2006). 735 Capitanio, M. et al. Two independent mechanical events in the interaction cycle of skeletal 734 muscle myosin with actin. Proc.Natl.Acad.Sci.USA 103, 87-92 (2006). 735 Geeves, M. A., Fedorov, R. & Manstein, D. J. Molecular mechanism of actomyosin-based 736 motility. Cell Mol.Life Sci. 62, 1462-1477 (2005). 737 Geeves, M. A., Fedorov, R. & Manstein, D. J. Molecular mechanism of actomyosin-based 736 motility. Cell Mol.Life Sci. 62, 1462-1477 (2005). 737 Tyska, M. J. et al. Two heads of myosin are better than one for generating force and motion. 738 Proc.Natl.Acad.Sci.USA 96, 4402-4407 (1999). 739 Tyska, M. J. et al. Two heads of myosin are better than one for generating force and motion. 738 Proc.Natl.Acad.Sci.USA 96, 4402-4407 (1999). 739 Piazzesi, G., Lucii, L. & Lombardi, V. The size and the speed of the working stroke of muscle 740 myosin and its dependence on the force. J.Physiol. 545, 145-151 (2002). 741 Ngo, K. X., Kodera, N., Katayama, E., Ando, T., Uyeda, T. Q. Cofilin-induced unidirectional 742 cooperative conformational changes in actin filaments revealed by high-speed atomic force 743 microscopy. eLife 4, e04806 (2015). 744 Ngo, K. X., Kodera, N., Katayama, E., Ando, T., Uyeda, T. Q. Cofilin-induced unidirectional 742 cooperative conformational changes in actin filaments revealed by high-speed atomic force 743 microscopy. eLife 4, e04806 (2015). 744 Risi, C.M. et al. Ca2+-induced movement of tropomyosin on native cardiac thin filaments 745 evealed by cryoelectron microscopy. Proc.Natl.Acad.Sci.USA 114, 6782–6787 (2017). 746 Risi, C. M. et al. The structure of the native cardiac thin filament at systolic Ca2+ levels. 747 Proc.Natl.Acad.Sci.USA 118, e2024288118 (2021). 748 Vilfan, A. & Duke, T. Instabilities in the transient response of muscle. Biophys.J. 85, 818-827 749 (2003). 750 Hilbert, L., Cumarasamy, S., Zitouni, N. B., Mackey, M. C., Lauzon, A.-M. The kinetics of 751 mechanically coupled myosins exhibit group size-dependent regimes. Biophys.J. 105, 1466- 752 1474 (2013). 753 Hwang, Y., Washiob, T., Hisada, T., Higuchia, Kaya, M. A reverse stroke characterizes the 754 force generation of cardiac myofilaments, leading to an understanding of heart function. 755 Proc.Natl.Acad.Sci.USA 118, e2011659118 (2021). 756 Duke, T. A. References 636 Using fluorescent myosin to directly visualize 718 cooperative activation of thin filaments. J.Biol.Chem. 290, 1915-1925 (2015). 719 Geeves, M. A. & Holmes, K. C. Structural mechanism of muscle contraction. 720 Annu.Rev.Biochem. 68, 687-728 (1999). 721 Kodera, N. et al. Structural and dynamics analysis of intrinsically disordered proteins by high- 722 speed atomic force microscopy Nat Nanotechnol. 16, 181-189 (2021). 723 Kodera, N. et al. Structural and dynamics analysis of intrinsically disordered proteins by high- 722 speed atomic force microscopy Nat Nanotechnol. 16, 181-189 (2021). 723 speed atomic force microscopy Nat Nanotechnol. 16, 181-189 (2021 723 Heath, G. & Scheuring, S. High-speed AFM height spectroscopy reveals µs-dynamics of 724 unlabeled biomolecules. Nat Commun. 9, 4983 (2018). 725 Heath, G. & Scheuring, S. High-speed AFM height spectroscopy reveals µs-dynamics of 724 unlabeled biomolecules. Nat Commun. 9, 4983 (2018). 725 Matusovsky, O. S. et al. Millisecond conformational dynamics of skeletal myosin II power 726 stroke studied by high-speed atomic force microscopy. ACS Nano 15, 2229-2239 (2021). 727 Matusovsky, O. S. et al. Millisecond conformational dynamics of skeletal myosin II power 726 stroke studied by high-speed atomic force microscopy. ACS Nano 15, 2229-2239 (2021). 727 Steffen, W., Smith, D., Simmons, R., Sleep, J. Mapping the actin filament with myosin. 728 Proc Natl Acad Sci USA 96 14949-14954 (2001) 729 Steffen, W., Smith, D., Simmons, R., Sleep, J. Mapping the actin filament with myosin. 728 Proc.Natl.Acad.Sci.USA 96, 14949-14954 (2001). 729 Steffen, W., Smith, D., Simmons, R., Sleep, J. Mapping the actin filament with myosin. 728 Proc.Natl.Acad.Sci.USA 96, 14949-14954 (2001). 729 Wang, Z. et al. The molecular basis for sarcomere organization in vertebrate skeletal muscle. 730 Cell 184, 2135-2150 (2021). 731 Cell 184, 2135-2150 (2021). 731 31 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Amyot, R. & Flechsig, H. BioAFMviewer: An interactive interface for simulated AFM scanning 732 of biomolecular structures and dynamics. References 636 Molecular model of muscle contraction. Proc.Natl.Acad.Sci.USA 96, 2770-2775 757 (1999). 758 Mansson, A. Hypothesis: single actomyosin properties account for ensemble behavior in 759 active muscle shortening and isometric contraction. Int.J.Mol.Sci. 21, 8399 (2020). 760 Huxley, A. Muscular contraction. Ann.Rev.Physiol. 50, 1-16 (1988). 761 Hill, T. L. Theoretical formalism for the sliding filament model of contraction of striated 762 muscle. Part I. Prog.Biophys.Mol.Biol. 28, 267-340 (1974). 763 32 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Edman, K. A., Mansson, A., Caputo, C. The biphasic force-velocity relationship in frog muscle 764 fibres and its evaluation in terms of cross-bridge function. J.Physiol. 503, 141-156 (1997). 765 Edman, K. A., Mansson, A., Caputo, C. The biphasic force-velocity relationship in frog muscle 764 fibres and its evaluation in terms of cross-bridge function. J.Physiol. 503, 141-156 (1997). 765 Conibear, P. B., Geeves, M. Cooperativity between the two heads of rabbit skeletal muscle 766 heavy meromyosin in binding to actin. Biophys. J. 75, 926-937 (1998) 767 773 33 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 4 5 Oleg S. Matusovsky1 6 Alf Mansson2 7 Dilson E. References 636 Rassier1 8 9 1 Department of Kinesiology and Physical Education, McGill University, Canada 10 11 2 Department of Chemistry and Biomedical Sciences, Linnaeus University, 12 Kalmar, Sweden. 13 14 15 16 17 18 Supplementary file includes: 19 Figures S1 to S7 20 Supplementary Table S1 21 Captions for Movies S1 to S4 22 23 Other Supplementary Materials for this manuscript include the following: 24 Movies S1 to S4 25 26 27 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 0 Figure S2. Arrangement of myosin heads between two parallel filaments. (a-b) Cross- 1 sections of myosin upper and lower heads bound between two parallel cTFs (left, red profiles) 2 and F-actin (right, blue profiles), scale bars: 30 nm. (c) The measured distance between two 3 parallel cTFs (left) and F-actin (right) averaged in (d). The difference in distance between non- 4 regulated F-actin and regulated cTFs (p=0.001, unpaired t-test) did not affect the HMM binding 5 and displacement analysis (Figs 1f and 2c). (e-f) Two types of HMM binding between parallel 6 cTFs or actin filaments were observed: the upper and lower heads bound towards the similar 7 direction (right HS-AFM images in e-f) or the upper and lower heads bound towards the 8 opposite directions (left HS-AFM images in e-f). The captured snapshots in left e and left f 9 panels indicate the rare situation occurred due to head displacement on the filament in the 0 presence of ATP. HMM heads between two actin filaments highlighted by red colors, S2 region 1 of the HMM molecules highlighted by green color The corresponding polarity of the filaments is 2 40 Figure S2. Arrangement of myosin heads between two parallel filaments. (a-b) Cross- 41 sections of myosin upper and lower heads bound between two parallel cTFs (left, red profiles) 42 and F-actin (right, blue profiles), scale bars: 30 nm. (c) The measured distance between two 43 parallel cTFs (left) and F-actin (right) averaged in (d). The difference in distance between non- 44 regulated F-actin and regulated cTFs (p=0.001, unpaired t-test) did not affect the HMM binding 45 and displacement analysis (Figs 1f and 2c). (e-f) Two types of HMM binding between parallel 46 cTFs or actin filaments were observed: the upper and lower heads bound towards the similar 47 direction (right HS-AFM images in e-f) or the upper and lower heads bound towards the 48 opposite directions (left HS-AFM images in e-f). The captured snapshots in left e and left f 49 panels indicate the rare situation occurred due to head displacement on the filament in the 50 presence of ATP. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 1 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Supplementary Figures Supplementary Figures 28 29 Figure S1. Successive HS-AFM images of F-actin-HMM complex in the presence of 0.2 μM ATP. (a) Double-headed heavy meromyosin (HMM) motors bound between two actin filaments in ~37 nm distance to form a stable structure with up to eight individual myosin heads (four HMM molecules), attached by their S2 regions. (b) HMM heads between two actin filaments highlighted by red colors, S2 region of the HMM molecules highlighted by green color. (c) high contrast images from (a) panel to highlight the connected S2 regions between HMM molecules. Numbers indicate the time in seconds, the scan rate: 14.4 fps, scale bar: 30 nm. Figure S1. Successive HS-AFM images of F-actin-HMM complex in the presence of 0.2 μM ATP. (a) Double-headed heavy meromyosin (HMM) motors bound between two actin filaments in ~37 nm distance to form a stable structure with up to eight individual myosin heads (four HMM molecules), attached by their S2 regions. (b) HMM heads between two actin filaments highlighted by red colors, S2 region of the HMM molecules highlighted by green color. (c) high contrast images from (a) panel to highlight the connected S2 regions between HMM molecules Numbers indicate the time in seconds the scan rate: 14 4 fps 30 30 31 32 33 34 35 36 37 38 Figure S1. Successive HS-AFM images of F-actin-HMM complex in the presence of 0.2 μM ATP. (a) Double-headed heavy meromyosin (HMM) motors bound between two actin filaments in ~37 nm distance to form a stable structure with up to eight individual myosin heads (four HMM molecules), attached by their S2 regions. (b) HMM heads between two actin filaments highlighted by red colors, S2 region of the HMM molecules highlighted by green color. (c) high contrast images from (a) panel to highlight the connected S2 regions between HMM molecules. Numbers indicate the time in seconds, the scan rate: 14.4 fps, scale bar: 30 nm. 39 2 2 . Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 HMM heads between two actin filaments highlighted by red colors, S2 region 51 of the HMM molecules highlighted by green color. The corresponding polarity of the filaments is 52 shown. Scale bars: 30 nm. 53 3 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Figure S3. Frequency distribution of one HMM head or two HMM heads bound between two filaments in the presence of ATP. (a-c) Distribution of one (1) and two (2) HMM heads in Fa-HMM complex in the presence of 0.2 μM ATP (319 binding events), 2.0 μM ATP (488 binding events) or 10 μM ATP (1072 binding events); (d-e) Distribution of one (1) and two (2) HMM heads in cTFs-HMM complex in the presence of 0.5 μM ATP, high Ca2+ concentration (545 binding events) and 2.0 μM ATP high Ca2+ concentration (440 binding events). The zero indicates the temporarily no bindings of HMM heads along the length of the filaments in analyzed datasets. HMM heads are indicated by white arrows (one head bound) and red arrows (two heads bound). Scale bars: 30 nm; z-scales indicated for each HS-AFM image. 54 Figure S3. Frequency distribution of one HMM head or two HMM heads bound between 55 two filaments in the presence of ATP. (a-c) Distribution of one (1) and two (2) HMM heads in 56 Fa-HMM complex in the presence of 0.2 μM ATP (319 binding events), 2.0 μM ATP (488 57 binding events) or 10 μM ATP (1072 binding events); (d-e) Distribution of one (1) and two (2) 58 HMM heads in cTFs-HMM complex in the presence of 0.5 μM ATP, high Ca2+ concentration 59 (545 binding events) and 2.0 μM ATP high Ca2+ concentration (440 binding events). The zero 60 indicates the temporarily no bindings of HMM heads along the length of the filaments in 61 analyzed datasets. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 Scan 72 area and scan rates of the top HS-AFM images for cTFs-HMM complex at the activating 73 conditions: (e)120 x 120 nm2, 200 × 200 pixels2, 6.7 fps; (f) 120 × 120 nm2, 200 × 200 pixels2, 2 74 fps; (g) 150 × 75 nm2, 80 × 40 pixels2, 6.7 fps; (h) 200 × 200 nm2, 120 × 120 pixels2, 2 fps. The 75 horizontal scale bars: 30 nm; the vertical scale bars in kymographs: 50 ms (e-f) and 100 ms (g- 76 h). The dashed yellow line indicated the initial position to create a kymograph image. (i) Average 77 dwell-time of myosin heads in F-actin-HMM complex in the presence of 0.5 μM ATP (n=3, 101 78 events, ~10 HMM heads) and 2 μM ATP (n=4, 592 events, ~20 HMM heads). (j). Average 79 dwell-time of myosin heads in cTFs-HMM complex in the presence of 0.5 μM ATP (n=3, 825 80 events, ~10 HMM heads) and 2 μM ATP (n=3, 498 events, ~10 HMM heads). 81 Figure S4. Kymograph images of the F-actin-HMM (a-d) and cTFs-HMM (e-h) complexes 68 at the different ATP concentrations. Scan area and scan rates of the top HS-AFM images for 69 F-actin-HMM complex: (a)150 × 75 nm2, 80 × 40 pixels2, 10 fps; (b) 150 x 90 nm2, 80 × 48 70 pixels2, 20 fps; (c) 200 × 120 nm2, 80 × 40 pixels2, 10 fps; (d) 200 × 120 nm2, 80 × 40 pixels2, 71 12.5 fps. The horizontal scale bars: 30 nm; the vertical scale bars in kymographs: 20 ms. Scan 72 area and scan rates of the top HS-AFM images for cTFs-HMM complex at the activating 73 conditions: (e)120 x 120 nm2, 200 × 200 pixels2, 6.7 fps; (f) 120 × 120 nm2, 200 × 200 pixels2, 2 74 fps; (g) 150 × 75 nm2, 80 × 40 pixels2, 6.7 fps; (h) 200 × 200 nm2, 120 × 120 pixels2, 2 fps. The 75 horizontal scale bars: 30 nm; the vertical scale bars in kymographs: 50 ms (e-f) and 100 ms (g- 76 h). The dashed yellow line indicated the initial position to create a kymograph image. (i) Average 77 dwell-time of myosin heads in F-actin-HMM complex in the presence of 0.5 μM ATP (n=3, 101 78 events, ~10 HMM heads) and 2 μM ATP (n=4, 592 events, ~20 HMM heads). (j). Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 HMM heads are indicated by white arrows (one head bound) and red arrows 62 (two heads bound). Scale bars: 30 nm; z-scales indicated for each HS-AFM image. 63 Figure S3. Frequency distribution of one HMM head or two HMM heads bound between 55 two filaments in the presence of ATP. (a-c) Distribution of one (1) and two (2) HMM heads in 56 Fa-HMM complex in the presence of 0.2 μM ATP (319 binding events), 2.0 μM ATP (488 57 binding events) or 10 μM ATP (1072 binding events); (d-e) Distribution of one (1) and two (2) 58 HMM heads in cTFs-HMM complex in the presence of 0.5 μM ATP, high Ca2+ concentration 59 (545 binding events) and 2.0 μM ATP high Ca2+ concentration (440 binding events). The zero 60 indicates the temporarily no bindings of HMM heads along the length of the filaments in 61 analyzed datasets. HMM heads are indicated by white arrows (one head bound) and red arrows 62 (two heads bound). Scale bars: 30 nm; z-scales indicated for each HS-AFM image. 63 64 4 4 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 65 66 67 Figure S4. Kymograph images of the F-actin-HMM (a-d) and cTFs-HMM (e-h) complexes 68 at the different ATP concentrations. Scan area and scan rates of the top HS-AFM images for 69 F-actin-HMM complex: (a)150 × 75 nm2, 80 × 40 pixels2, 10 fps; (b) 150 x 90 nm2, 80 × 48 70 pixels2, 20 fps; (c) 200 × 120 nm2, 80 × 40 pixels2, 10 fps; (d) 200 × 120 nm2, 80 × 40 pixels2, 71 12.5 fps. The horizontal scale bars: 30 nm; the vertical scale bars in kymographs: 20 ms. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 Average 79 dwell-time of myosin heads in cTFs-HMM complex in the presence of 0.5 μM ATP (n=3, 825 80 events, ~10 HMM heads) and 2 μM ATP (n=3, 498 events, ~10 HMM heads). 81 82 5 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 83 Figure S5. Probability of binding of myosin heads to the non-regulated or regulated 84 filaments in rigor conditions or in the presence of ATP--S. (a-c) HS-AFM images of HMM 85 molecules bound between two actin filaments in the presence of ATP--S (a), HMM molecules 86 bound between two actin filaments in the absence of Ca2+ and ATP (b) and HMM molecules 87 bound between two cardiac thin filaments in the absence of Ca2+ and ATP (c). Scale bars: 30 88 nm (d-f) Probabilities of binding of the HMM heads in the F-actin-HMM complex in the presence 89 of ATP--S (d, 1145 events), in the absence of Ca2+ and ATP (e, 1408 events) and in the cTFs- 90 HMM complex in the absence of Ca2+ and ATP (f, number of events 663). Individual data points 91 are shown for each HMM molecules (M1-M4) as mean values ± 95% CI, n=3. (g-i) 92 Corresponded fractional occupancies of the binding sites by HMM heads for the conditions 93 shown in d-f. 94 83 Figure S5. Probability of binding of myosin heads to the non-regulated or regulated 84 filaments in rigor conditions or in the presence of ATP--S. (a-c) HS-AFM images of HMM 85 molecules bound between two actin filaments in the presence of ATP--S (a), HMM molecules 86 bound between two actin filaments in the absence of Ca2+ and ATP (b) and HMM molecules 87 bound between two cardiac thin filaments in the absence of Ca2+ and ATP (c). Scale bars: 30 88 nm (d-f) Probabilities of binding of the HMM heads in the F-actin-HMM complex in the presence 89 of ATP--S (d, 1145 events), in the absence of Ca2+ and ATP (e, 1408 events) and in the cTFs- 90 HMM complex in the absence of Ca2+ and ATP (f, number of events 663). Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 Individual data points 91 are shown for each HMM molecules (M1-M4) as mean values ± 95% CI, n=3. (g-i) 92 Corresponded fractional occupancies of the binding sites by HMM heads for the conditions 93 shown in d-f. 94 95 6 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 96 Figure S6. Cooperative probability of the myosin heads bound to the non-regulatory F- 97 actin. The independent experiments showing different patterns of the cooperative probabilities 98 between the one HMM molecule Mn and the next molecule Mn+1 in the F-actin-HMM complex. 99 The linear regression fittings showed in the solid lines and the dashed curves represented 95% 100 CI. 101 Figure S6. Cooperative probability of the myosin heads bound to the non-regulatory F- 97 actin. The independent experiments showing different patterns of the cooperative probabilities 98 between the one HMM molecule Mn and the next molecule Mn+1 in the F-actin-HMM complex. 99 The linear regression fittings showed in the solid lines and the dashed curves represented 95% 100 CI. 101 102 102 7 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint 103 104 Figure S7. Cooperative probability of the myosin heads bound to the regulatory cardiac 105 TFs. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 The independent experiments showing different patterns of the cooperative probabilities 106 between the one HMM molecule Mn and the next molecule Mn+1 in the cTFs-HMM complex. The 107 observed differences, most likely related to the diverse population of activated / non-activated 108 segments across the thin filament length that led to different degree of positive cooperativity of 109 myosin bindings. The linear regression fittings showed in the solid lines and the dashed curves 110 represented 95% CI. 111 103 104 Figure S7. Cooperative probability of the myosin heads bound to the regulatory cardiac 105 TFs. The independent experiments showing different patterns of the cooperative probabilities 106 between the one HMM molecule Mn and the next molecule Mn+1 in the cTFs-HMM complex. The 107 observed differences, most likely related to the diverse population of activated / non-activated 108 segments across the thin filament length that led to different degree of positive cooperativity of 109 myosin bindings. The linear regression fittings showed in the solid lines and the dashed curves 110 represented 95% CI. 111 112 113 114 115 116 Figure S7. Cooperative probability of the myosin heads bound to the regulatory cardiac 105 TFs. The independent experiments showing different patterns of the cooperative probabilities 106 between the one HMM molecule Mn and the next molecule Mn+1 in the cTFs-HMM complex. The 107 observed differences, most likely related to the diverse population of activated / non-activated 108 segments across the thin filament length that led to different degree of positive cooperativity of 109 myosin bindings. The linear regression fittings showed in the solid lines and the dashed curves 110 represented 95% CI. 111 116 8 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint To calculate the cooperative probability from observed probability patterns, we applied 117 the following equation 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘 118 where C denotes cooperative probability of binding, n = number of total events or 119 subsequent frames of the experiment; k = number of binding events in the experiment, p 120 = probability of binding, i.e. the ratio between binding events and total events and (𝑛 𝑘) 121 represents the combination of the total and binding events expressed as: 𝑛! 𝑘!(𝑛−𝑘)!. 122 123 Table S1. A concrete example of calculation of the cooperativity of binding for 4 124 neighboring HMM molecules bound to actin filaments in the presence of 10 μM ATP. 125 126 HMM molecules M1 M2 M3 M4 Number of total events (frames) for each HMM (n) 176 176 176 176 Binding events (k) 171 61 37 21 Binding probability (p) 0.97 0.35 0.21 0.12 1-p 0.03 0.65 0.79 0.88 p^k 7.24E-03 8.49E-29 8.70E-26 4.08E-20 n-k 5.00 115.00 139.00 155.00 (1-p)^(n-k) 1.85E-08 5.58E-22 5.66E-15 2.80E-09 (𝑛 𝑘) = 𝑛! 𝑘! (𝑛−𝑘)! 1.33E+09 1.33E+48 1.50E+38 8.09E+26 Cooperative probability (C) 0.178 0.063 0.074 0.092 127 To calculate the cooperative probability from observed probability patterns, we applied 117 the following equation 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘 118 To calculate the cooperative probability from observed probability patterns, we applied 117 the following equation 𝐶= (𝑛 𝑘)𝑝𝑘(1 −𝑝)𝑛−𝑘 118 where C denotes cooperative probability of binding, n = number of total events or 119 subsequent frames of the experiment; k = number of binding events in the experiment, p 120 = probability of binding, i.e. the ratio between binding events and total events and (𝑛 𝑘) 121 represents the combination of the total and binding events expressed as: 𝑛! 𝑘!(𝑛−𝑘)!. 122 123 Table S1. A concrete example of calculation of the cooperativity of binding for 4 124 neighboring HMM molecules bound to actin filaments in the presence of 10 μM ATP. 125 126 HMM molecules M1 M2 M3 M4 Number of total events (frames) for each HMM (n) 176 176 176 176 Binding events (k) 171 61 37 21 Binding probability (p) 0.97 0.35 0.21 0.12 1-p 0.03 0.65 0.79 0.88 p^k 7.24E-03 8.49E-29 8.70E-26 4.08E-20 n-k 5.00 115.00 139.00 155.00 (1-p)^(n-k) 1.85E-08 5.58E-22 5.66E-15 2.80E-09 (𝑛 𝑘) = 𝑛! 𝑘! (𝑛−𝑘)! 1.33E+09 1.33E+48 1.50E+38 8.09E+26 Table S1. Supplementary 1 Cooperativity of myosin II motors in the non-regulated and regulated thin 2 filaments investigated with high-speed AFM 3 A concrete example of calculation of the cooperativity of binding for 4 124 neighboring HMM molecules bound to actin filaments in the presence of 10 μM ATP. 125 HMM molecules M1 M2 M3 M4 Number of total events (frames) for each HMM (n) 176 176 176 176 Binding events (k) 171 61 37 21 Binding probability (p) 0.97 0.35 0.21 0.12 1-p 0.03 0.65 0.79 0.88 p^k 7.24E-03 8.49E-29 8.70E-26 4.08E-20 n-k 5.00 115.00 139.00 155.00 (1-p)^(n-k) 1.85E-08 5.58E-22 5.66E-15 2.80E-09 (𝑛 𝑘) = 𝑛! 𝑘! (𝑛−𝑘)! 1.33E+09 1.33E+48 1.50E+38 8.09E+26 Cooperative probability (C) 0.178 0.063 0.074 0.092 7 9 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Supplementary Movies 133 Scan area: 150 x 75 nm2, 80 158 x 40 pixels2; recording rate 3.3 fps, playing rate 10 fps; scale bar is 30 nm (left movie); 159 100 x 60 nm2, 80 x 48 pixels2; recording rate 6.7 fps, playing rate 10 fps, scale bar is 20 160 nm (central movie); 100 x 60 nm2, 80 x 40 pixels2; recording rate 6.7 fps, playing rate 10 161 fps, scale bar is 20 nm (right movie). 162 Movie S1 134 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 135 bound between two actin filaments in the presence of 0.2 μM Mg.ATP (top panel). The 136 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 137 color code corresponded to the position of the HMM molecule on the actin filaments – 138 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 139 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 140 used in the mean probability binding of the HMM molecules to the actin filaments at the 141 0.2 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 142 binding sites in the presence of 0.2 μM Mg.ATP (bottom right graph). Scan area: 150 x 143 90 nm2, 80 x 48 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 200 x 120 144 nm2, 80 x 48 pixels2, recording rate 12.5 fps, playing rate 10 fps (central movie); 200 x 145 120 nm2, 80 x 48 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The 146 scale bars are 30 nm. 147 148 Movie S2 149 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 150 bound between two actin filaments in the presence of 2 μM Mg.ATP (top panel). The 151 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 152 color code corresponded to the position of the HMM molecule on the actin filaments – 153 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 154 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 155 used in the mean probability binding of the HMM molecules to the actin filaments at the 156 2 μM Mg.ATP (bottom left graph). Supplementary Movies 133 The average fractional occupation of the actin binding 157 sites in the presence of 2 μM Mg.ATP (bottom right graph). Scan area: 150 x 75 nm2, 80 158 x 40 pixels2; recording rate 3.3 fps, playing rate 10 fps; scale bar is 30 nm (left movie); 159 100 x 60 nm2, 80 x 48 pixels2; recording rate 6.7 fps, playing rate 10 fps, scale bar is 20 160 nm (central movie); 100 x 60 nm2, 80 x 40 pixels2; recording rate 6.7 fps, playing rate 10 161 fps, scale bar is 20 nm (right movie). 162 Supplementary Movies 133 Movie S1 134 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 135 bound between two actin filaments in the presence of 0.2 μM Mg.ATP (top panel). The 136 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 137 color code corresponded to the position of the HMM molecule on the actin filaments – 138 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 139 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 140 used in the mean probability binding of the HMM molecules to the actin filaments at the 141 0.2 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 142 binding sites in the presence of 0.2 μM Mg.ATP (bottom right graph). Scan area: 150 x 143 90 nm2, 80 x 48 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 200 x 120 144 nm2, 80 x 48 pixels2, recording rate 12.5 fps, playing rate 10 fps (central movie); 200 x 145 120 nm2, 80 x 48 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The 146 scale bars are 30 nm. 147 148 Movie S2 149 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 150 bound between two actin filaments in the presence of 2 μM Mg.ATP (top panel). The 151 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 152 color code corresponded to the position of the HMM molecule on the actin filaments – 153 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 154 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 155 used in the mean probability binding of the HMM molecules to the actin filaments at the 156 2 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin binding 157 sites in the presence of 2 μM Mg.ATP (bottom right graph). Movie S1 134 The 166 color code corresponded to the position of the HMM molecule on the actin filaments – 167 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 168 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 169 used in the mean probability binding of the HMM molecules to the actin filaments at the 170 10 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 171 binding sites in the presence of 10 μM Mg.ATP (bottom right graph). Scan area: 150 x 172 75 nm2, 80 x 40 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 150 x 75 173 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (central movie); 150 x 75 174 nm2, 80 x 40 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The scale 175 bars are 30 nm. 176 177 Movie S4 178 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 179 bound between two cTFs in the presence of 0.5 or 2 μM Mg.ATP (top panel). The 180 colored boxes indicated upper and / or lower HMM heads bound to cTFs. The color 181 code corresponded to the position of the HMM molecule on the thin filaments – 1st 182 molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 4th 183 molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 184 used in the mean probability binding of the HMM molecules to the cTFs at the 0.5 μM or 185 2 μM Mg.ATP (middle panel). The average fractional occupations of the binding sites in 186 the presence of 0.5 μM or 2 μM Mg.ATP (bottom panel). Scan area: 200 x 200 nm2, 120 187 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (left movie); Scan area: 200 x 200 188 nm2, 120 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (central movie); 150 x 75 189 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (top right movie); 150 x 190 75 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (bottom right movie). 191 The scale bars are 30 nm. 192 Movie S1 134 Movie S1 134 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 135 bound between two actin filaments in the presence of 0.2 μM Mg.ATP (top panel). The 136 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 137 color code corresponded to the position of the HMM molecule on the actin filaments – 138 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 139 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 140 used in the mean probability binding of the HMM molecules to the actin filaments at the 141 0.2 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 142 binding sites in the presence of 0.2 μM Mg.ATP (bottom right graph). Scan area: 150 x 143 90 nm2, 80 x 48 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 200 x 120 144 nm2, 80 x 48 pixels2, recording rate 12.5 fps, playing rate 10 fps (central movie); 200 x 145 120 nm2, 80 x 48 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The 146 scale bars are 30 nm. 147 10 . CC-BY 4.0 International license available under a which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted February 24, 2022. ; https://doi.org/10.1101/2022.02.24.481751 doi: bioRxiv preprint Movie S3 163 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 164 bound between two actin filaments in the presence of 10 μM Mg.ATP (top panel). The 165 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 166 color code corresponded to the position of the HMM molecule on the actin filaments – 167 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 168 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The scale bars are 30 nm. 192 The scale bars are 30 nm. Movie S1 134 The same color code was 169 used in the mean probability binding of the HMM molecules to the actin filaments at the 170 10 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 171 binding sites in the presence of 10 μM Mg.ATP (bottom right graph). Scan area: 150 x 172 75 nm2, 80 x 40 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 150 x 75 173 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (central movie); 150 x 75 174 nm2, 80 x 40 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The scale 175 bars are 30 nm. 176 177 Movie S4 178 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 179 bound between two cTFs in the presence of 0.5 or 2 μM Mg.ATP (top panel). The 180 colored boxes indicated upper and / or lower HMM heads bound to cTFs. The color 181 code corresponded to the position of the HMM molecule on the thin filaments – 1st 182 molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 4th 183 molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 184 used in the mean probability binding of the HMM molecules to the cTFs at the 0.5 μM or 185 2 μM Mg.ATP (middle panel). The average fractional occupations of the binding sites in 186 the presence of 0.5 μM or 2 μM Mg.ATP (bottom panel). Scan area: 200 x 200 nm2, 120 187 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (left movie); Scan area: 200 x 200 188 nm2, 120 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (central movie); 150 x 75 189 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (top right movie); 150 x 190 75 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (bottom right movie). 191 The scale bars are 30 nm. 192 Movie S3 163 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 164 bound between two actin filaments in the presence of 10 μM Mg.ATP (top panel). The 165 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. Movie S3 163 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 164 bound between two actin filaments in the presence of 10 μM Mg.ATP (top panel). The 165 colored boxes indicated upper and / or lower HMM heads bound to actin filaments. The 166 color code corresponded to the position of the HMM molecule on the actin filaments – 167 1st molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 168 4th molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 169 used in the mean probability binding of the HMM molecules to the actin filaments at the 170 10 μM Mg.ATP (bottom left graph). The average fractional occupation of the actin 171 binding sites in the presence of 10 μM Mg.ATP (bottom right graph). Scan area: 150 x 172 75 nm2, 80 x 40 pixels2; recording rate 10 fps, playing rate 10 fps (left movie); 150 x 75 173 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (central movie); 150 x 75 174 nm2, 80 x 40 pixels2, recording rate 10 fps, playing rate 10 fps (right movie). The scale 175 bars are 30 nm. 176 Movie S4 178 Movie S4 178 Representative HS-AFM movies of the transient binding of skeletal HMM molecules 179 bound between two cTFs in the presence of 0.5 or 2 μM Mg.ATP (top panel). The 180 colored boxes indicated upper and / or lower HMM heads bound to cTFs. The color 181 code corresponded to the position of the HMM molecule on the thin filaments – 1st 182 molecule (M1): red box; 2nd molecule (M2): green box; 3rd molecule (M3): blue box; 4th 183 molecule (M4): magenta box; 5th molecule (M5): cyan box. The same color code was 184 used in the mean probability binding of the HMM molecules to the cTFs at the 0.5 μM or 185 2 μM Mg.ATP (middle panel). The average fractional occupations of the binding sites in 186 the presence of 0.5 μM or 2 μM Mg.ATP (bottom panel). Scan area: 200 x 200 nm2, 120 187 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (left movie); Scan area: 200 x 200 188 nm2, 120 x 120 pixels2; recording rate 2 fps, playing rate 5 fps (central movie); 150 x 75 189 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (top right movie); 150 x 190 75 nm2, 80 x 40 pixels2, recording rate 6.7 fps, playing rate 10 fps (bottom right movie). 191 11
https://openalex.org/W3125849373	https://www.federalreserve.gov/pubs/feds/1997/199752/199752pap.pdf	English	null	Market Definition and the Analysis of Antitrust in Banking	Finance and economics discussion series	1,997	public-domain	9,414	Forthcoming: The Antitrust Bulletin (Winter 1997) Forthcoming: The Antitrust Bulletin (Winter 1997) Market Definition and the Analysis of Antitrust in Banking Market Definition and the Analysis of Antitrust in Banking John D. Wolken* See Stephen A. Rhoades, Bank Mergers and Industrywide Structure, 1 1980-94 (Staff Study No. 169, Board of Governors of the Federal Reserve System, January 1996). The above figures exclude mergers of failed banks. Beginning in the mid-1980s, and continuing through 1992, over 1,300 bank failures also contributed significantly to the consolidation of U.S. banking. October 1997 *Division of Research and Statistics, Board of Governors of the Federal Reserve System, Washington, D.C. 20551. The views expressed are those of the authors and do not necessarily reflect those of the Board of Governors or its staff. The authors thank Dean Amel, Tim Hannan, Wayne Passmore, Robin Prager, and Steve Rhoades for comments on a earlier draft. Any errors are the responsibility of the authors. BY MYRON L. KWAST, MARTHA STARR-MCCLUER, and JOHN D. WOLKEN BY MYRON L. KWAST, MARTHA STARR-MCCLUER, and JOHN D. WOLKEN I. Introduction The U.S. banking industry is currently experiencing the most significant consolidation in its history. The years 1981-1995 have averaged some 434 mergers among healthy banks per year, and have cumulated a stunning 1.6 trillion (1995) dollars in acquired bank assets. Several of these mergers, and especially some in the 1990s, have been the largest 1 in U.S. history -- including the mergers of BankAmerica and Security Pacific, Chase Manhattan and Chemical, and Wells Fargo and First Interstate. Passage of interstate banking and branching legislation at the national level, the Riegle-Neal Act of 1994, can be expected to further encourage these trends, since it will allow the development of a truly national banking structure. While the on-going consolidation of the U.S. banking system is in many ways long overdue, it also raises a number of public policy concerns. One of these concerns is the potential impact of consolidation on the competitiveness of banking markets. Under the Bank Holding Company Act, the Bank Merger Act, and other statutes, the U.S. While the on-going consolidation of the U.S. banking system is in many ways long overdue, it also raises a number of public policy concerns. One of these concerns is the potential impact of consolidation on the competitiveness of banking markets. Under the Bank Holding Company Act, the Bank Merger Act, and other statutes, the U.S. Department of Justice and the federal bank regulatory agencies are charged with enforcing the antitrust laws in banking. Thus, over the years, each of these agencies has developed procedures for analyzing the potential competitive impact of a proposed merger. Department of Justice and the federal bank regulatory agencies are charged with enforcing the antitrust laws in banking. Thus, over the years, each of these agencies has developed procedures for analyzing the potential competitive impact of a proposed merger. Current procedures for examining the potential competitive effects of bank mergers, while varying somewhat across agencies, are based on three fundamental 2 concepts. In particular: the (1) “cluster” of bank products is the relevant product line for 2 competition analysis; this cluster is normally viewed as being consumed in (2) geographically local banking markets; and (3) market structure is a key determinant of the degree of competition. The precise definition of the cluster is vague, but is meant to encompass the set of products and services that are purchased primarily from banks. It is important to note that in recent years the Department of Justice has emphasized 3 small business lending, and other small business customer relationships, as the most relevant local product, while the Federal Reserve has continued to rely on the (more general) concept of the cluster. These basic concepts were first confirmed by the U.S. Supreme Court in 1963 (the 2 Philadelphia National Bank decision), and reconfirmed by lower courts as recently as 1987. A recent forum for the expression of such criticisms was a conference on bank 4 antitrust issues sponsored by the Office of the Comptroller of the Currency in November 1995. I. Introduction Products and services for which markets are viewed as being national or regional in scope are virtually never considered to have competitive problems; the focus is almost entirely on “locally limited products.” Thus, in practice, competition analysis is almost exclusively centered on products consumed by households and small businesses, since these are the agents that are normally viewed as being the most likely to be locally limited in their purchases of at least some financial services. 3 concepts. In particular: the (1) “cluster” of bank products is the relevant product line for 2 competition analysis; this cluster is normally viewed as being consumed in (2) The current approach has received widespread, and often strong, criticism from bankers, academics, policy makers, and other market participants. Some argue that the 4 concept of a cluster of banking services, while perhaps relevant in the 1960s, is hopelessly outdated in the modern financial world where virtually any banking product is offered not only by depository institutions (including thrifts), but also by many (often specialized) nonbanking firms. At a minimum, this argument says, the cluster should be broken up into those individual products and services where banks have potential market power. Others go further and argue that changing technology has made the concept of the local market obsolete. Today, such critics say, households and small businesses can purchase virtually any financial service virtually anywhere in the country from either an insured A recent forum for the expression of such criticisms was a conference on bank 4 antitrust issues sponsored by the Office of the Comptroller of the Currency in November 1995. 3 3 depository or an uninsured nonbank institution. Electronic banking, it is said, will only accelerate this trend. These critics view existing competition as sufficient to allay any concerns regarding the competitive effects of a bank merger. In addition, some argue that even if existing competition may sometimes be inadequate, potential competition is sufficient to assuage any competitive concerns. Such observers maintain that barriers to entry in banking are low, and that recent legislation allowing fuller interstate banking and branching has reduced entry barriers even more. This paper attempts to contribute to this debate by utilizing two data sources that are particularly well-suited to examining some of the key hypotheses. Indeed, both data sets were in part designed precisely to allow such an analysis. Gregory E. Elliehausen and John D. Wolken, Banking Markets and the Use of 5 Financial Services by Small and Medium-Sized Businesses, 76 Federal Reserve Bulletin (October 1990). I. Introduction The paper uses the 1992 Survey of Consumer Finances (SCF) and the 1993 National Survey of Small Business Finances (NSSBF) to examine the extent to which households and small businesses (1) tend to focus their purchases of financial services at insured depository, as opposed to nondepository, institutions, (2) purchase their financial services locally, and (3) tend to cluster their purchases of financial services at a single “primary” financial institution. A strong effort is made to define variables comparably across surveys so that households and small businesses can be easily compared. Indeed, a unique contribution of this paper is its simultaneous consideration of the two classes of agents most often thought of as potentially subject to market power in banking. In addition, previous studies of these issues are somewhat dated, and thus provide little evidence to address the major contemporary criticisms. The next section reviews the literature, and Section III briefly describes the two surveys used in this study. Section IV considers the extent to which households and small businesses use depository and nondepository institutions for their purchases of financial services. Section V looks at the extent to which such services are purchased in geographically local markets, the next section examines the issue of clustering, and section VII addresses some difficulties in interpreting our data. Overall, the results are supportive of the current methodology used to analyze the potential competitive effects of bank mergers. Key household asset services, such as checking accounts, and important small business asset and credit services appear to a substantial degree to be obtained, and 4 preferred to be obtained, at local depository institutions. Section VIII concludes and suggests some directions for future research. preferred to be obtained, at local depository institutions. Section VIII concludes and suggests some directions for future research. II. Literature Review Recent empirical evidence on the relevancy of the current approach to competitive analysis for bank mergers is limited, and increasingly dated. Elliehausen and Wolken 5 analyzed data from the 1987 National Survey of Small Business Finances. At that 6, 7 time, local commercial banks were the most widely used source for virtually every financial service used by small businesses. The study found a significant portion of small businesses obtaining services from nonbank institutions, but generally these sources were used for single products. Multiple service use was most often associated with the firms primary institution, which was almost always a local commercial bank. Gregory E. Elliehausen and John D. Wolken, Banking Markets and the Use of 5 Financial Services by Small and Medium-Sized Businesses, 76 Federal Reserve Bulletin (October 1990). Following the Supreme Courts 1963 Philadelphia National Bank decision, a number 6 of case studies of individual geographic areas (e.g., Nassau county, St. Louis) surveyed households and sometimes businesses to determine where and from what types of institutions these groups of bank customers obtained their financial services. Most of these studies were conducted in the late 1960s and throughout the 1970s. Generally these studies provided data confirming the dependence on local commercial banks in the geographic areas surveyed. See John D. Wolken, Geographic Market Delineation: A Review of the Literature (Staff Study, No. 140, Board of Governors of the Federal Reserve System, October 1984) for a detailed review of these studies. For a recent study examining banking markets for “middle market firms,” see Robert 7 Tannanwald, The Geographic Boundaries of New England's Middle-Lending Markets, Federal Reserve Bank of Boston, 15 New England Economic Review (July-August 1994). 5 Using the same data set, Elliehausen and Wolken explored the issue of clustering for small businesses. This study found that a large segment of the small business 8 population clustered purchases of some financial services (particularly checking, lines of credit and some other services such as credit card receipt processing and cash management services). However, the results suggested that it was unlikely that the entire set of products offered by commercial banks belonged to ªthe cluster." Rather, some products likely belonged to separate markets. The most recent study of household banking markets was conducted using data from the 1989 Survey of Consumer Finances. II. Literature Review Elliehausen and Wolken found that local depository institutions, especially local commercial banks, were the main suppliers for most of the financial services used by households. Credit products were predominantly 9 local, but not as local as checking and savings accounts. Nondepository institutions were generally local as well, but to a lesser extent than depository institutions. Multiple service usage was most often associated with the checking account, and the institution at which clustering occurred was almost always a local depository institution. The typical service bundle consisted of a checking account and another savings account or credit service. As was true for small businesses, some credit products, such as mortgages and vehicle loans, were often purchased separately. III. The Surveys of Consumer and Small Business Finances Gregory E. Elliehausen and John D. Wolken, Banking Markets and the Use of 9 Financial Services by Households, 78 Federal Reserve Bulletin (March 1992). Gregory E. Elliehausen and John D. Wolken, Small Business Clustering of Financial 8 Services and the Definition of Banking Markets for Antitrust Analysis, 37 The Antitrust Bulletin (Fall 1992). The 1993 National Survey of Small Business Finances was co-sponsored by the 0 The 1993 National Survey of Small Business Finances was co-sponsored by the 10 Small Business Administration. For general descriptions of the surveys, see Arthur B. Kennickell and Martha Starr-McCluer, Changes in Family Finances, 1989-1992: Evidence from the Survey of Consumer Finances, Federal Reserve Bulletin, 80 (October 1994) and Rebel A. Cole and John D. Wolken, Financial Services Used by Small Businesses: Evidence from the 1993 National Survey of Small Business Finances, Federal Reserve Bulletin, 81 (July 1995). All data presented in this paper use the weights developed for each survey to adjust for sample design and nonresponse. Small businesses are defined as firms with fewer than 500 employees. 11 Leasing companies are not reported separately for households. In the SCF, the main 12 institutions in the "other" category are life insurance and mortgage companies. III. The Surveys of Consumer and Small Business Finances Both the 1992 SCF and the 1993 NSSBF were sponsored by the Federal Reserve Board. The 1992 SCF interviewed 3,906 households, asking detailed questions about 10 Gregory E. Elliehausen and John D. Wolken, Small Business Clustering of Financial 8 Services and the Definition of Banking Markets for Antitrust Analysis, 37 The Antitrust Bulletin (Fall 1992). Gregory E. Elliehausen and John D. Wolken, Banking Markets and the Use of 9 Financial Services by Households, 78 Federal Reserve Bulletin (March 1992). Gregory E. Elliehausen and John D. Wolken, Banking Markets and the Use of 9 Financial Services by Households, 78 Federal Reserve Bulletin (March 1992). 6 their assets, liabilities, demographic characteristics, and use of financial services. The 1993 NSSBF interviewed 5,356 small businesses about their use of credit and other financial services. The data sets are uniquely well-suited to examining markets for 11 financial services. In both surveys, respondents provide information about the institutions at which they have accounts or loans, including the institution type and its distance from the main office of the business (for the NSSBF) or from the home or workplace of the person who uses it most (for the SCF). These data permit an examination of the roles of institution type and location in the use of financial services. For purposes of this study, institutions are divided into depository and nondepository institutions. Depositories include commercial banks, savings and loans (including savings banks), and credit unions. Nondepositories include finance companies, brokerage firms and mutual funds, leasing companies, and other institutions. For the 12 NSSBF, nonfinancial institutions and unclassified sources are also shown separately. In both surveys, respondents identify their primary institution. Because our interest is in markets for financial services typically supplied by banks, the analysis focuses on certain financial services. For households, the services covered are checking accounts, savings accounts, money market accounts (both money market deposit accounts and money market mutual funds), certificates of deposit, IRA and Keogh accounts, brokerage accounts, trusts and other managed assets, mortgages and Leasing companies are not reported separately for households. In the SCF, the main 12 institutions in the "other" category are life insurance and mortgage companies. 7 home equity loans, motor vehicle loans, home equity and other lines of credit, and other loans. "Other" loans include consumer loans, home improvement loans, and education 13 loans. The analysis excludes services not traditionally provided by banks, like life insurance and non money market mutual funds. The one service of potential interest that we are not able to cover is credit cards. The SCF collects limited information on institutions used for credit cards only. III. The Surveys of Consumer and Small Business Finances For small businesses, the services covered are checking accounts, savings 13 accounts, lines of credit, leases, mortgages, equipment loans, motor vehicle loans, other loans, and five financial management services (transaction, cash management, credit- related, pension, and brokerage). A paper by Rebel A. Cole, John D. Wolken, and R. Louise Woodburn, Bank and 14 Nonbank Competition for Small Business Credit: Evidence from the 1987 and 1993 National Surveys of Small Business Finances, 82 Federal Reserve Bulletin (November 1996), compares bank and nonbank credit use by small businesses using the 1987 and 1993 NSSBFs. This study examines both the incidence and the dollar shares of small business loans provided by various types of service providers. Interestingly, while the share of businesses using nonbanks remained constant, the share using banks declined from 1987 to 1993. However, the dollar shares of small business credit outstanding provided by banks and nonbanks were essentially unchanged across the survey dates. IV. Depository vs. Nondepository Institutions Table 1 shows the percentage of households (panel A) and the percentage of small businesses (panel B) using various types of depository and nondepository institutions. Looking at column 1 in both panels, almost 99 percent of both households and small businesses using financial services have at least one account at a depository institution. In contrast, only 35.7 percent of households, and 34.9 percent of small businesses use a nondepository institution. Nonetheless, the percentage of both groups using nondepositories is substantial, and consistent with the view that nondepositories should be taken seriously in competitive analysis. The data in the rest of column 1 provide a deeper look at this segment of the market for financial services. The data show that commercial banks are by far the financial institution used most frequently by both households and small businesses. More importantly, some 96.5 percent of households, and 93.5 percent of small businesses say that a depository institution is their primary financial institution (column 2); only about four percent of both groups call a nondepository their primary institution. To an overwhelming extent, commercial banks are also the institution most frequently designated as the primary institution. Thus, the data suggest that, overall, depositories are 8 substantially more important for both households and small business than are nondepositories, although nondepositories are clearly relevant. In addition, commercial banks are, by a wide margin, the most important depository.14 substantially more important for both households and small business than are nondepositories, although nondepositories are clearly relevant. In addition, commercial banks are, by a wide margin, the most important depository.14 For households, local institutions are defined as institutions (including ATMs) 15 located within 30 miles of the home or workplace of the household member who uses it the most. Distances are coded up to a maximum of 50 miles, with greater distances recorded as “more than 50 miles.” In the NSSBF, an institution is local if it is located within 30 miles of the headquarters office of the small business. Respondents are asked to report the number of miles between the firm and institution when the institution is located in the same city or metropolitan area. If the institution is located elsewhere, miles are calculated using institution and firm zipcode latitude and longitude. V. Local vs. Nonlocal Institutions Columns 4 and 5 of table 1 and tables 2 and 3 use a variety of measures to examine the importance of local financial institutions. With respect to depositories, 15 table 1 shows that 97.5 percent of households, and 92.4 percent of small businesses, use a local depository institution. In stark contrast, only 20.2 percent of households, and not quite eight percent of small businesses use nonlocal depositories. The dominance of local institutions persists across commercial banks, savings institutions, and credit unions. However, the dominance of local institutions disappears for both households and small businesses with respect to their use of nondepositories. Indeed, the percentages of local and nonlocal use are very similar (compare columns 4 and 5) for all nondepository institutions. In short, the dominance of local institutions appears to be a characteristic of For households, local institutions are defined as institutions (including ATMs) 15 located within 30 miles of the home or workplace of the household member who uses it the most. Distances are coded up to a maximum of 50 miles, with greater distances recorded as “more than 50 miles.” In the NSSBF, an institution is local if it is located within 30 miles of the headquarters office of the small business. Respondents are asked to report the number of miles between the firm and institution when the institution is located in the same city or metropolitan area. If the institution is located elsewhere, miles are calculated using institution and firm zipcode latitude and longitude. For households, local institutions are defined as institutions (including ATMs) 15 located within 30 miles of the home or workplace of the household member who uses it the most. Distances are coded up to a maximum of 50 miles, with greater distances recorded as “more than 50 miles.” In the NSSBF, an institution is local if it is located within 30 miles of the headquarters office of the small business. Respondents are asked to report the number of miles between the firm and institution when the institution is located in the same city or metropolitan area. If the institution is located elsewhere, miles are calculated using institution and firm zipcode latitude and longitude. 9 depositories, not nondepository institutions. depositories, not nondepository institutions. It is also interesting to note that the data in column 1 of table 2 reinforce the results 16 of section IV. On average, depository institutions, and especially commercial banks, provide substantially more services to both households and small businesses than do nondepositories. See footnote 12 for the precise definitions used. 17 An interesting implication of this result is that it suggests that geographically remote 18 alternatives, such as money market funds, are generally not perfect substitutes for local transaction deposits. This supposition is supported by other data in the 1992 SCF which indicate that less than six percent of households had a money market fund. Even more telling, 98 percent of households with money market funds also had a checking or money market deposit account. This conclusion is supported by data (not shown) from the SCF regarding 19 households’ use of credit cards. Because the SCF does not collect distance information for institutions used for credit cards only, credit cards are not included in the analysis. However, the information that is available suggests that credit cards are among the less locally oriented financial services. Almost 58 percent of all general-purpose credit cards (e.g., Visa and MasterCard) held by households are issued by institutions where the household obtains no other financial service. It seems reasonable to presume that many of these institutions are not local firms. V. Local vs. Nonlocal Institutions The role of local institutions is examined using the average number of financial services consumed at different types of institutions by households (panel A) and small businesses (panel B) in table 2. Focusing on column 4, it is clear that at depository institutions local institutions dominate by this measure as well. Overall, 90.7 percent of services consumed at depositories by households, and 94.6 percent of services consumed by small business are provided by local depositories. This result is robust across all three classes of depository institutions. In contrast, local nondepositories are important, but not nearly as critical as local depositories. For example, local institutions provide only 52.1 percent of services used at nondepositories by households, and 55.1 percent used by small businesses. Interestingly, within the class of nondepositories, local brokerage firms stand out as being particularly important for both households and small businesses. 16 Our final measure of the role of local institutions considers the number of miles between users and their institutions. Specifically, for each institution where a household or small business purchased services, the distance between a household (small business) and the given institution is examined. Thus, a single household, or small business, has 17 an observation for each institution where it reports that it conducts business. Table 3 displays the distribution of the distance in miles of these user-institution pairs across types of financial services. For households, financial services are divided into types of asset and credit services, and for small businesses services are separated into categories of asset, credit, and financial management services. Using the “checking” row of panel 3A as an example, the data indicate that households get 25 percent of their checking services at institutions within one mile of their home or work. The results in table 3 generally reinforce the importance of local institutions, but also suggest a fairly complex story. For example, households obtain 75 percent of their It is also interesting to note that the data in column 1 of table 2 reinforce the results 16 of section IV. On average, depository institutions, and especially commercial banks, provide substantially more services to both households and small businesses than do nondepositories. 10 checking, savings, money market, certificate of deposit, and line of credit services at a financial institution within 10 miles of their home or workplace. The “local” importance of checking and savings services is reinforced by the high 21 degree of localness exhibited by transaction and cash management services, both of which are closely related to deposit account activity. V. Local vs. Nonlocal Institutions 11 savings services, stand out as dominated by local providers, while leases are clearly the service least likely to be local.21 savings services, stand out as dominated by local providers, while leases are clearly the service least likely to be local.21 V. Local vs. Nonlocal Institutions Checking services stand out as dominated by local institutions; households purchase 90 percent of their checking services at an institution no more than 12 miles from their home or workplace. In 18 addition, while local institutions are certainly important with respect to credit services for households, asset services generally appear to be more locally oriented than credit services. 19 checking, savings, money market, certificate of deposit, and line of credit services at a financial institution within 10 miles of their home or workplace. Checking services stand out as dominated by local institutions; households purchase 90 percent of their checking services at an institution no more than 12 miles from their home or workplace. In 18 addition, while local institutions are certainly important with respect to credit services for households, asset services generally appear to be more locally oriented than credit services. 19 The differing degree of “localness” between asset and credit services for households is not as clear for small businesses. Indeed, small businesses seem more closely tied to local institutions across a wide range of services. For example, small businesses obtain 75 percent of their checking, savings, lines of credit, mortgage, other loans, transaction, cash management, and credit related services at a financial institution within 15 miles of their headquarters office. Still, asset services, here checking and 20 This conclusion is supported by data (not shown) from the SCF regarding 19 households’ use of credit cards. Because the SCF does not collect distance information for institutions used for credit cards only, credit cards are not included in the analysis. However, the information that is available suggests that credit cards are among the less locally oriented financial services. Almost 58 percent of all general-purpose credit cards (e.g., Visa and MasterCard) held by households are issued by institutions where the household obtains no other financial service. It seems reasonable to presume that many of these institutions are not local firms. Transaction services include the provision of currency and coin, the processing of 20 credit card receipts, the collection of night deposits, and wire transfers. Cash management services include the provision of sweep accounts, zero-balance accounts, and lockbox services. Credit-related services include bankers' acceptances, letters of credit, and factoring. VI. Clustering of Financial Services Of the issues examined in this study, the degree of clustering of financial services by households and small businesses is the most difficult to address empirically. The approach adopted here is to focus on the use of financial services at the institution identified by a survey respondent as their primary financial institution. Thus, our underlying assumption is that households and small businesses would be most likely to cluster at the institution where they conduct the majority, or at least the most important aspect, of their business. Conversely, if respondents do not cluster at their primary institution, then it would seem unlikely that they would cluster at any other institution. A second aspect of our approach is also to consider the relative importance of a service to households and small businesses. Services consumed by relatively few agents are not considered likely candidates for the cluster. Tables 4 and 5 present data on the average number of financial services used per household (panel A) and per small business (panel B) at primary and nonprimary institutions. In table 4, the data are arrayed by the same types of services used in table 3; in table 5 the rows of the table are arranged by the type of primary institution. Consider column 4 of panel 4A, which gives the percent of each service purchased at the primary institution, and column 5, which gives the percent of households using the service. These data suggest that the services most frequently clustered at the primary institution may include checking, savings and money market accounts, lines of credit, and possibly certificates of deposit. Interestingly, these are precisely the same services most likely to be purchased locally (table 3). In addition, it is important to recall that primary The “local” importance of checking and savings services is reinforced by the high 21 degree of localness exhibited by transaction and cash management services, both of which are closely related to deposit account activity. 12 institutions are overwhelmingly depository institutions. Indeed, data not shown indicate that for 93 percent of households the primary institution is a local depository institution. In combination, these facts suggest that households often tend to cluster their purchases of certain financial services, and particularly asset services, at a local depository institution. institutions are overwhelmingly depository institutions. Indeed, data not shown indicate that for 93 percent of households the primary institution is a local depository institution. However, recall that only one percent of households claim that a brokerage firm is 22 their primary institution (table 1, panel 1A, column 2). In contrast, almost 97 percent of households say a depository institution is their primary institution. VI. Clustering of Financial Services In combination, these facts suggest that households often tend to cluster their purchases of certain financial services, and particularly asset services, at a local depository institution. Comparable analysis of the data in columns 4 and 5 of panel 4B suggests that the services most frequently clustered at the primary institution by small businesses may be checking, savings, lines of credit, mortgages, transaction, cash management, and credit related services. As was the case for households, these are exactly the same services that are most likely to be purchased locally, and a small business’ primary institution is almost certain to be a depository institution. Data not shown indicate that for 88 percent of small businesses the primary institution is a local depository institution. Thus, a strong circumstantial case can be made that small businesses, as well as households, frequently tend to cluster their purchases of certain financial services at a local depository institution. Unlike households, the cluster for small businesses appears to include not only asset services, but also important credit and financial management services. Table 5 focuses on households’ and small businesses’ purchases of any services at their primary institutions. Concentrating first on columns 4 and 5 of panel 5A, it can be seen that when a household’s primary institution is a depository institution, it is likely that at least 50 percent of all the financial services purchased by the household come from the primary institution. But such is not the case when a household’s primary institution is a nondepository institution. For nondepositories, only about a third of the household’s financial services tend to be purchased from the primary institution, and even then the percent of households purchasing more than one service from the primary institution tends to be 30 percent or lower (with the notable exception of purchases from brokerage firms). Taken together, these data support the view that there is some tendency for 22 households to cluster their use of financial services at their local primary depository However, recall that only one percent of households claim that a brokerage firm is 22 their primary institution (table 1, panel 1A, column 2). In contrast, almost 97 percent of households say a depository institution is their primary institution. 13 institution, and to be considerably less likely to cluster on those relatively uncommon occasions when their primary institution is a nondepository. Similar conclusions hold for small businesses. As was also true of households, recall that only one percent of small businesses 23 claimed a brokerage firm as their primary institution (table 1, panel 1B, column 2), while almost 94 percent say a depository is their primary institution. VI. Clustering of Financial Services Indeed, small businesses appear somewhat more tied to depositories than are households. When a small business’ primary institution is a depository, the small business purchases over 60 percent of its financial services at the primary institution. When the primary institution is not a depository, only a third or less of the small business’ financial services are purchased at the primary nondepository. In addition, the percent of small businesses purchasing more than one service from the nondepository primary institution tends to be well below 30 percent (again, with the notable exception of purchases from brokerage firms).23 While this discussion provides some support for the notion of clustering by both households and small businesses, it is important not to overestimate the extent of this support. For example, the data in column 2 of panel 5A indicate that households purchase, on average, only slightly over two financial services at their primary depository institution. While this equals the average number of services purchased at all nonprimary institutions, it still suggests that there are only a small number of items in the average household’s “cluster.” Similar patterns are observed in the small business data (column 2, panel 5B), except that the mean number of services purchased at a small business’ primary depository institution, while still only about two, is consistently greater than the mean number of services purchased at all nonprimary institutions. This is consistent with the view that clustering may be a more important concept for small businesses than for households. VII. A Comment on Interpretation A recent study by Booz-Allen and Hamilton, Inc., Consumer Demand 24 for Internet Banking, Financial Services Group, New York (July 1996) cites an October 1995 American Bankers Association - Gallup poll regarding the most important reason that a consumer maintains an account with a financial institution. The single most important reason by far was “convenience,” which was cited by almost 40 percent of respondents. “Friendly/Good Service” came in second at almost 20 percent; “long-standing relationship” was third, with almost 15 percent; and “checking held there” and “good interest rates” were virtually tied for fourth and fifth with about 11 percent each. Other important reasons cited for the choice of primary mortgage institution were 25 “recommended” (25 percent), “easy to qualify” (13 percent), and “assumption by or financed through developer” (12 percent). VII. A Comment on Interpretation A problem with interpreting these data is that, while they represent equilibrium As was also true of households, recall that only one percent of small businesses 23 claimed a brokerage firm as their primary institution (table 1, panel 1B, column 2), while almost 94 percent say a depository is their primary institution. 14 conditions given the current market environment, they do not necessarily indicate how households and small businesses would behave if, say, local market prices increased substantially. The evidence presented here suggests that households and small businesses may well resist moving to a nonlocal institution for some services (but not others) because of increases in explicit prices alone. That is, the data are consistent with the view that, at least for some financial services, and at current prices, the transactions and information costs of using nonlocal providers dominate the benefits of going outside the local market area. The survey data provide some direct evidence on the role of geographic proximity in the use of financial services. With respect to households, when asked to give their most important reason for choosing their main checking account institution, 44 percent of households listed “location,” and 32 percent cited “many services in one place, already do business there.” For checking services, only 14 percent of households gave “favorable interest rate and/or low fees” as the most important reason for choosing their main checking institution. In contrast, when asked the same question regarding their choice of 24 primary mortgage institution, 27 percent of households indicated “favorable interest rate and/or low fees.” Only 4 percent of households gave “location,” and 17 percent cited “many services in one place, already do business there.”25 Results are similar for small businesses. When asked about factors which influenced the firm to use its primary institution, 37 percent of respondents indicated Other important reasons cited for the choice of primary mortgage institution were 25 “recommended” (25 percent), “easy to qualify” (13 percent), and “assumption by or financed through developer” (12 percent). 15 location, 27 percent cited “many services in one place or already do business there,” but only 6 percent identified “favorable interest rates or prices.” Lastly, the importance of “many services in one place, already do business there” for both households’ and small businesses’ choice of their main checking account or primary institution suggests the clustering of some financial services by both groups. VII. A Comment on Interpretation location, 27 percent cited “many services in one place or already do business there,” but only 6 percent identified “favorable interest rates or prices.” Lastly, the importance of “many services in one place, already do business there” for both households’ and small businesses’ choice of their main checking account or primary institution suggests the clustering of some financial services by both groups. VIII. Conclusion This paper has investigated three of the most important issues that arise when analyzing the potential competitive effects of bank mergers on households and small businesses: (1) the relative importance of depository and nondepository institutions, (2) the extent to which markets for some financial services are geographically local, and (3) whether some financial services are clustered at an agent’s primary financial institution. With respect to the first issue, our results suggest that, while nondepositories are clearly relevant, depository institutions are substantially more important for both households and small businesses. In addition, commercial banks are by a wide margin the most important depository. For several important financial services, local firms are by far the dominant provider. For households, checking services stand out as overwhelmingly supplied by local firms. But other common products -- savings and money market deposits, CDs, and line of credit services -- are also in large part locally limited. In general, households appear to be more closely tied to local providers of asset services than they are to local providers of credit services, and the importance of local institutions is more a characteristic of depository institutions than it is of nondepositories. Important financial services consumed by small businesses also have a strong tendency to be provided by local institutions. Indeed, small businesses may be more closely tied to local firms than are households. Checking and savings services stand out as dominated by local providers. Other important locally supplied services include lines of credit, mortgages, other loans, transaction, cash management, and credit related services. Overall, key asset and credit services consumed by small businesses are 16 provided primarily by local institutions, and, as with households, the importance of local institutions is more a characteristic of depositories than of nondepositories. provided primarily by local institutions, and, as with households, the importance of local institutions is more a characteristic of depositories than of nondepositories. This study provides some support for the argument that substantial proportions of households and small businesses cluster their purchases of a few key financial services at their local primary depository institution. For households, these services appear to be the same asset services that are locally limited. For small businesses, clustered services appear to be the same asset and credit services that are locally limited. There is a small, and hopefully growing literature in this area. See, for example, 26 Rhoades, op. cit. and Arthur B. Kennickell and Myron L. Kwast, Who Uses Electronic Banking? Preliminary Results from the 1995 Survey of Consumer Finances, presented at the Annual Meetings of the Financial Management Association (October 1996), Board of Governors of the Federal Reserve System, Washington, D.C. VIII. Conclusion Panel 1B: Percent of Small Businesses Using Various Types of Institution By type of financial institution and selected institution characteristic (1) (2) (3) (4) (5) Institution characteristic Type of institution All Primary Nonprimary Local (≤ 30 miles) Nonlocal (> 30 miles) All 100 98.5 54.0 93.3 16.9 Depository 98.7 93.5 35.5 92.4 7.6 Commercial bank 91.4 83.9 30.2 84.9 6.4 Savings 12.1 7.7 5.1 10.9 .8 Credit union 4.4 1.9 2.5 3.6 .4 Nondepository 34.9 4.3 32.8 13.5 10.8 Finance company 13.4 1.4 12.4 3.8 4.1 Brokerage firm 9.9 1.0 9.1 5.9 2.3 Leasing 8.0 .5 7.6 2.4 3.0 Other ND financial 3.7 .2 3.5 1.1 .8 Nonfinancial 8.7 1.1 8.0 1.4 1.4 Note: Excludes small businesses that do not use financial services provided by an institutional source. Source: 1993 National Survey of Small Business Finances. Table 1 Table 1 Table 1 Panel 1A: Percent of Households Using Various Types of Institution By type of institution and selected institution characteristic (1) (2) (3) (4) (5) Institution characteristic Type of institution All Primary Nonprimary Local (≤ 30 miles) Nonlocal (> 30 miles) All 100 100 72.0 98.3 35.7 Depository 98.9 96.5 62.9 97.5 20.2 Commercial bank 83.3 71.3 40.9 80.6 11.3 Savings 29.5 13.4 19.7 26.5 4.3 Credit union 31.8 11.8 22.2 26.8 6.6 Nondepository 35.7 3.5 33.3 20.3 20.7 Finance company 13.1 1.7 11.8 6.2 7.8 Brokerage firm 16.3 1.0 15.7 10.7 7.3 Other 12.2 .8 11.4 4.6 7.8 Note: Excludes households that do not use financial services provided by an institutional source. Source: 1992 Survey of Consumer Finances. 1993 National Survey of Small Business Finances. VIII. Conclusion Finally, the data suggest that, for both households and small businesses, clustering is mostly a characteristic of the purchase of services at their local primary depository institution, and considerably less likely to occur on those relatively uncommon occasions when their primary institution is a nondepository. These results are, on balance, consistent with previous studies and supportive of the methodology currently used at the Federal Reserve and the Department of Justice for analyzing the potential competitive effects of bank mergers. Overall, they suggest that, for perhaps the vast majority of households and small businesses, use of financial services has not changed as much in recent years as some observers have argued. However, it is also clear that further research would be very useful. One interesting extension of the current work would be to investigate households’ and small businesses’ use of financial services as a function of agent and market characteristics. The authors are currently pursuing such an effort. Another important topic is the dynamics of household and small business responses to changes in the explicit and implicit prices of financial services. Our final suggestion concerns the impact of electronic banking on the importance of potential competition in banking. Studies of the nature of household and small business use and 26 acceptance of various electronic media for financial services would be quite useful, 17 especially in better understanding the possible future course of defining banking markets. 17 especially in better understanding the possible future course of defining banking markets. 18 18 18 Table 1 Panel 1A: Percent of Households Using Various Types of Institution By type of institution and selected institution characteristic (1) (2) (3) (4) (5) Institution characteristic Type of institution All Primary Nonprimary Local (≤ 30 miles) Nonlocal (> 30 miles) All 100 100 72.0 98.3 35.7 Depository 98.9 96.5 62.9 97.5 20.2 Commercial bank 83.3 71.3 40.9 80.6 11.3 Savings 29.5 13.4 19.7 26.5 4.3 Credit union 31.8 11.8 22.2 26.8 6.6 Nondepository 35.7 3.5 33.3 20.3 20.7 Finance company 13.1 1.7 11.8 6.2 7.8 Brokerage firm 16.3 1.0 15.7 10.7 7.3 Other 12.2 .8 11.4 4.6 7.8 Note: Excludes households that do not use financial services provided by an institutional source. Source: 1992 Survey of Consumer Finances. Note: Excludes small businesses that do not use financial services provided by an institutional source. Survey of Small Business Finances. small businesses that do not use financial services provided by an institutional source. VIII. Conclusion Panel 1A: Percent of Households Using Various Types of Institution By type of institution and selected institution characteristic Panel 1B: Percent of Small Businesses Using Various Types of Institution By type of financial institution and selected institution characteristic (1) (2) (3) (4) (5) Institution characteristic Type of institution All Primary Nonprimary Local (≤ 30 miles) Nonlocal (> 30 miles) All 100 98.5 54.0 93.3 16.9 Depository 98.7 93.5 35.5 92.4 7.6 Commercial bank 91.4 83.9 30.2 84.9 6.4 Savings 12.1 7.7 5.1 10.9 .8 Credit union 4.4 1.9 2.5 3.6 .4 Nondepository 34.9 4.3 32.8 13.5 10.8 Finance company 13.4 1.4 12.4 3.8 4.1 Brokerage firm 9.9 1.0 9.1 5.9 2.3 Leasing 8.0 .5 7.6 2.4 3.0 Other ND financial 3.7 .2 3.5 1.1 .8 Nonfinancial 8.7 1.1 8.0 1.4 1.4 Note: Excludes small businesses that do not use financial services provided by an institutional source. Source: 1993 National Survey of Small Business Finances. Panel 1B: Percent of Small Businesses Using Various Types of Institution By type of financial institution and selected institution characteristic 19 Panel 2A: Average Number of Financial Services Used Per Household at Local and Nonlocal Institutions By type of institution (1) (2) (3) (4) Type of institution All Local (≤ 30 miles) Nonlocal (> 30 miles) Memo: Services at local institutions as a percent of all services All 4.24 3.58 .66 84.4 Depository 3.53 3.20 .32 90.7 Commercial bank 2.23 2.07 .16 92.8 Savings .63 .57 .06 90.5 Credit union .67 .57 .10 85.1 Nondepository .71 .37 .34 52.1 Finance company .18 .08 .10 44.4 Brokerage firm .37 .23 .14 62.2 Other .16 .06 .10 37.5 Source: 1992 Survey of Consumer Finances. : 1993 National Survey of Small Business Finances. Survey of Small Business Finances. The data for the local/nonlocal distribution of nondepository sources should be regarded as preliminary. Source: 1993 National Survey of Small Business Finances. VIII. Conclusion mber of Financial Services Used Per Household at Local and Nonlocal Institutions Panel 2A: Average Number of Financial Services Used Per Household at Local and Nonlocal Institutions By type of institution Panel 2B: Average Number of Financial Services Used Per Small Business at Local and Nonlocal Institutions By type of institution (1) (2) (3) (4) Type of institution All Local (≤ 30 miles) Nonlocal (> 30 miles) Memo: Services at local institutions as a percent of all services All 3.11 2.78 .32 89.4 Depository 2.45 2.32 .13 94.6 Commercial bank 2.19 2.07 .12 94.7 Savings .20 .19 .01 95.2 Credit union .07 .06 .01 90.3 Nondepository .61 .34 .24 55.1 ND financial .48 .27 .19 56.1 Finance company .18 .08 .09 46.0 Brokerage firm .15 .11 .04 73.8 Leasing .11 .04 .06 39.0 Other .04 .02 .02 55.0 Nonfinancial .13 .06 .04 47.0 Source unknown .04 .04 .00 88.9 Note: The data for the local/nonlocal distribution of nondepository sources should be regarded as preliminary. Source: 1993 National Survey of Small Business Finances. Panel 2B: Average Number of Financial Services Used Per Small Business at Local and Nonlocal Institutions Source: 1993 National Survey of Small Business Finances. 20 Table 3 Panel 3A: Number of Miles Between Households and Their Institutions By type of financial service Distance in miles by percentile Type of service 25th Median 75th 90th Asset <1 3 8 50 Checking <1 2 5 12 Savings <1 3 8 50 Money market <1 3 10 50 Certificate of deposit <1 3 6 20 IRA orKeogh 2 5 25 50 Brokerage 4 10 50 50 Trust 3 15 50 50 Credit 2 7 50 50 Mortgage 3 8 50 50 Motor vehicle 2 7 22 50 Line of credit <1 3 10 50 Other loan 2 10 50 50 Note: Distances are recorded up to a maximum of 50 miles. Source: 1992 Survey of Consumer Finances. VIII. Conclusion Panel 3B: Number of Miles Between Small Businesses and Their Institutions By type of financial service Distance in miles by percentile Type of ervice 25th Median 75th 90th Asset 1 2 5 15 Checking 1 2 5 13 Savings 1 2 6 17 Credit 1 5 25 264 Lines of credit 1 4 11 45 Leases 5 39 431 1,214 Mortgage loan 1 4 12 35 Equipment loan 1 5 35 494 Motor vehicle loan 2 6 25 137 Other loan 1 4 12 139 Financial management services 1 3 10 55 Transaction 1 2 5 20 Cash management 1 3 10 30 Credit related 1 3 14 74 Pension 3 7 21 250 Brokerage 3 7 24 216 Source: 1993 National Survey of Small Business Finances. Table 3 Table 3 Panel 3A: Number of Miles Between Households and Their Institutions By type of financial service Distance in miles by percentile Type of service 25th Median 75th 90th Asset <1 3 8 50 Checking <1 2 5 12 Savings <1 3 8 50 Money market <1 3 10 50 Certificate of deposit <1 3 6 20 IRA orKeogh 2 5 25 50 Brokerage 4 10 50 50 Trust 3 15 50 50 Credit 2 7 50 50 Mortgage 3 8 50 50 Motor vehicle 2 7 22 50 Line of credit <1 3 10 50 Other loan 2 10 50 50 Note: Distances are recorded up to a maximum of 50 miles. Source: 1992 Survey of Consumer Finances. Panel 3A: Number of Miles Between Households and Their Institutions Panel 3B: Number of Miles Between Small Businesses and Their Institutions By type of financial service Distance in miles by percentile Type of ervice 25th Median 75th 90th Asset 1 2 5 15 Checking 1 2 5 13 Savings 1 2 6 17 Credit 1 5 25 264 Lines of credit 1 4 11 45 Leases 5 39 431 1,214 Mortgage loan 1 4 12 35 Equipment loan 1 5 35 494 Motor vehicle loan 2 6 25 137 Other loan 1 4 12 139 Financial management services 1 3 10 55 Transaction 1 2 5 20 Cash management 1 3 10 30 Credit related 1 3 14 74 Pension 3 7 21 250 Brokerage 3 7 24 216 Source: 1993 National Survey of Small Business Finances. VIII. Conclusion Panel 4B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions 22 Table 5 Panel 5A: Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions By type of primary institution (1) (2) (3) (4) (5) Memo Type of primary institution Total Primary institution All nonprimary institutions Services at primary institutions as a percent of all services Percent of households using more than one service at primary institution All 4.24 2.11 2.13 49.8 56.4 Depository 4.24 2.13 2.11 50.2 57.3 Commercial bank 4.18 2.07 2.11 49.5 54.7 Savings 4.15 2.15 2.00 51.8 57.4 Credit union 4.69 2.49 2.20 53.1 72.9 Nondepository 4.21 1.51 2.70 35.9 30.4 Finance company 3.50 1.38 2.12 39.4 17.3 Brokerage firm 6.30 2.06 4.24 32.7 68.2 Other 3.11 1.11 2.00 35.7 11.1 Source: 1992 Survey of Consumer Finances. Panel 5B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions By type of primary financial institution (1) (2) (3) (4) (5) Memo Type of primary institution Total Primary institution All nonprimary institutions Services at primary institutions as a percent of all services Percent of businesses using more than one service at primary institution All 3.10 1.86 1.24 59.9 53.2 Depository 3.07 1.92 1.16 62.4 54.6 Commercial bank. 3.10 1.94 1.16 62.5 55.4 Savings 2.73 1.68 1.05 61.6 47.0 Credit union 3.05 1.78 1.27 58.4 51.9 Nondepository 4.29 1.36 2.93 31.8 26.1 ND financial 4.35 1.44 2.91 33.2 30.6 Finance company 4.66 1.32 3.34 28.3 22.5 Brokerage 4.27 1.87 2.40 43.8 57.0 Leasing 4.11 1.08 3.03 26.2 7.8 Other 3.16 1.19 1.96 37.8 19.4 Nonfinancial 4.13 1.14 2.98 27.7 13.7 Source unknown 4.11 1.08 3.03 26.2 26.3 Source: 1993 National Survey of Small Business Finances. VIII. Conclusion Panel 3B: Number of Miles Between Small Businesses and Their Institutions 21 Table 4 Panel 4A: Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions Panel 4A: Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions By type of financial service (1) (2) (3) (4) (5) Memo Type of service All institutions Primary institution All nonprimary institutions Services at primary institutions as percent of all services Percent of households using the service All services 4.24 2.11 2.13 49.8 100 Asset 3.13 1.72 1.41 55.0 96.9 Checking 1.16 .86 .30 74.1 85.7 Savings .80 .42 .38 52.5 48.8 Money market .34 .17 .16 51.3 23.2 Certificate of deposit .25 .11 .13 46.3 18.4 IRA or Keogh .43 .13 .29 31.1 25.6 Brokerage .12 .01 .11 11.6 10.1 Trust .04 .01 .03 16.2 3.2 Credit 1.12 .39 .73 34.6 59.7 Mortgage .49 .16 .33 32.6 39.5 Motor vehicle .29 .10 .19 34.4 25.2 Line of credit .15 .07 .07 49.5 12.2 Other loan .19 .06 .14 29.1 13.7 Source: 1992 Survey of Consumer Finances. : Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions Panel 4B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions By type of financial service (1) (2) (3) (4) (5) Memo Type of service All institutions Primary institution All nonprimary institutions Services at primary institutions as percent of all services Percent of small businesses using financial service All services 3.10 1.86 1.24 59.9 100 Asset 1.46 1.10 .36 75.2 99.2 Checking 1.16 .91 .25 78.8 98.7 Savings .30 .18 .11 61.4 25.0 Credit 1.07 .46 .61 42.9 55.1 Lines of credit .34 .22 .12 64.2 26.3 Leases .13 .01 .11 11.3 9.4 Mortgage loan .07 .03 .03 51.4 6.1 Equipment loan .17 .07 .10 41.3 13.9 Motor vehicle loan .29 .09 .20 30.1 24.5 Other loan .08 .04 .04 46.4 6.7 Financial management services .58 .31 .27 52.8 36.4 Transaction .29 .19 .10 65.1 23.9 Cash management .06 .05 .01 77.5 5.4 Credit related .06 .04 .02 61.4 4.7 Pension .11 .03 .09 22.4 4.4 Brokerage .05 .01 .05 13.4 10.1 Source: 1993 National Survey of Small Business Finances. VIII. Conclusion 22 Table 5 Panel 5A: Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions By type of primary institution (1) (2) (3) (4) (5) Memo Type of primary institution Total Primary institution All nonprimary institutions Services at primary institutions as a percent of all services Percent of households using more than one service at primary institution All 4.24 2.11 2.13 49.8 56.4 Depository 4.24 2.13 2.11 50.2 57.3 Commercial bank 4.18 2.07 2.11 49.5 54.7 Savings 4.15 2.15 2.00 51.8 57.4 Credit union 4.69 2.49 2.20 53.1 72.9 Nondepository 4.21 1.51 2.70 35.9 30.4 Finance company 3.50 1.38 2.12 39.4 17.3 Brokerage firm 6.30 2.06 4.24 32.7 68.2 Other 3.11 1.11 2.00 35.7 11.1 Source: 1992 Survey of Consumer Finances. : Average Number of Financial Services Used Per Household at Primary and Nonprimary Institutions Panel 5B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions By type of primary financial institution (1) (2) (3) (4) (5) Memo Type of primary institution Total Primary institution All nonprimary institutions Services at primary institutions as a percent of all services Percent of businesses using more than one service at primary institution All 3.10 1.86 1.24 59.9 53.2 Depository 3.07 1.92 1.16 62.4 54.6 Commercial bank. 3.10 1.94 1.16 62.5 55.4 Savings 2.73 1.68 1.05 61.6 47.0 Credit union 3.05 1.78 1.27 58.4 51.9 Nondepository 4.29 1.36 2.93 31.8 26.1 ND financial 4.35 1.44 2.91 33.2 30.6 Finance company 4.66 1.32 3.34 28.3 22.5 Brokerage 4.27 1.87 2.40 43.8 57.0 Leasing 4.11 1.08 3.03 26.2 7.8 Other 3.16 1.19 1.96 37.8 19.4 Nonfinancial 4.13 1.14 2.98 27.7 13.7 Source unknown 4.11 1.08 3.03 26.2 26.3 Source: 1993 National Survey of Small Business Finances. inancial Services Used Per Small Business at Primary and Nonprimary Institutions Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions Panel 5B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary Institutions anel 5B: Average Number of Financial Services Used Per Small Business at Primary and Nonprimary I
https://openalex.org/W4241473314	https://asianpubs.org/index.php/ajchem/article/download/25_13_123/7776	English	null	Microstructural Changes of Fuel Claddings for Pressurized Water Reactors Fuel after Irradiation	Asian Journal of Chemistry/Asian journal of chemistry	2,013	cc-by	3,000	http://dx.doi.org/10.14233/ajchem.2013.05 http://dx.doi.org/10.14233/ajchem.2013.05 Microstructural Changes of Fuel Claddings for Pressurized Water Reactors Fuel after Irradiation† YANG-SOON PARK, JONG-GOO KIM, HYOUNG-MUN KWON, YEONG-KEONG HA* and KYUSEOK SONG Nuclear Chemistry Research Division, Korea Atomic Energy Research Institute, Daedeokdaero 989-111, Yuseong-gu, Daejeon, 305-353, Republic of Korea Nuclear Chemistry Research Division, Korea Atomic Energy Research Institute, Daedeokdaero 989-111, Yuseong-gu, Daejeon, 305-353, Republic of Korea *Corresponding author: Fax: +82 42 8688148; E-mail: nykha@kaeri.re.kr AJC-13627 Microstructural changes of spent fuel claddings discharged from pressurized water reactors were investigated by a radiation-shielded micro-X-ray diffraction system (µ-XRD). A zircaloy-4 cladding of UO2 fuel and a zirlo cladding of burnable poison fuel (5.98 wt % gadolinia-urania) with 41 GWd/tU burn-up and a zircaloy-4 cladding of 58 GWd/tU UO2 fuel were chosen as sample specimens to investigate the effects of burn-up on the irradiation behaviour of cladding. At 41 GWd/tU burn-up, the chemical interaction between a fuel pellet and cladding was not observed for both samples. The formation of ZrO2 and ZrH2 phases was observed in the outer surface of the zircaloy-4 cladding, while those phases were not observed for the zirlo cladding. It can be concluded that zirlo is more stable than zircaloy-4 at this burn-up. At 58 GWd/tU burn-up, the formation of a zirconium oxide (ZrOx) layer was observed at the inner surface of the zircaloy-4 cladding, which indicates that the cladding interacts with the fuel. On the outer surface of zircaloy-4 cladding, ZrO2 and ZrH2 phases were also observed. Key Words: Spent fuel cladding, Zircaloy-4, Zirlo, Chemical interaction, Oxidation. Key Words: Spent fuel cladding, Zircaloy-4, Zirlo, Chemical interaction, Oxidation. †Presented to the 4th International Symposium on Applications of Chemical and Analytical Technologies in Nuclear Industries, Daejeon, Korea Asian Journal of Chemistry; Vol. 25, No. 12 (2013), 7009-7012 International Symposium on Applications of Chemical and Analytical Technologies in Nuclear Industries, Daejeon, Korea Presented to the 4th International Symposium on Applications of Chemical and Analytical Technologies in Nuclear Industri RESULTS AND DISCUSSION composition of the corrosion region may be possible. It has been reported that a corrosion layer of a cladding under the primary circuits of pressurized water reactors is produced by the diffusion of oxygen or hydrogen from the cooling water into the cladding during the operation of a nuclear reactor9,10. Therefore, ZrO2 and ZrH2 phases in the outer layer of the cladding can be formed by the diffusion of oxygen and hydrogen from the cooling water. However, nothing was detected except zirconium in the outer surface layer of a zirlo cladding contacted by Gd2O3-UO2 fuel. It can be estimated that the corrosion resistance of the zirlo cladding is better than the zry-4 cladding. Irradiation behaviour of the zircaloy-4 and zirlo claddings at 41 GWd/tU: During irradiation, a fuel pellet can swell and contact the cladding, leading to a chemical reaction between the fuel and cladding. The characteristics of the chemical reaction in this interface region should be under- stood because they affect the integrity of the cladding. For an identification of the reaction product in this interface region, the XRD spectrum was measured for the reaction area between a fuel and its cladding. The XRD spectra of a zry-4 cladding contacted by 41 GWd/tU UO2 (UO2/zry-4) and zirlo cladding contacted by 41 GWd/tU poisoned UO2 (Gd2O3-UO2/zirlo) are shown in Figs. 1 and 2, respectively. Diffraction spectra were obtained from the inside to the outside of the cladding at 50-100-µm inter- vals. At the inner surface, zirconium oxide (ZrOx) was not observed in either sample. The observed UO2 phase resulted from the fuel due to the beam width at the interface. It appears that the chemical interaction between the fuel and cladding did not occur for both zry-4 and zirlo claddings at 41 GWd/tU fuel burn-up. Irradiation behaviour of zircaloy-4 cladding at 58 GWd/tU fuel burn-up: For the UO2/zry-4 couple of 58 GWd/ tU fuel burn-up, zirconium oxide (ZrOx) was observed at the inner surface as shown in Fig. 3. However, the expected phase of zirconium oxide cannot be easily distinguished from those of UO2 because the XRD patterns of those phases are nearly identical, as shown in Fig. 4. Thus, the elemental analysis was performed by EDS and WDS (Fig. 5) to clarify the composition of the reaction layer. A line scan analysis for the content of Zr, U and O was performed at 2.5-µm intervals around the interface. EXPERIMENTAL Preparation of fuel cladding specimens: Three sample specimens were prepared from spent fuel rods discharged from a pressurized water reactor. Two specimens were couples of the UO2 fuel/zircaloy-4 cladding, which were discharged from the NPP(Y-2). The burn-up of two fuels was 41 GWd/tU and 58 GWd/tU, as measured by the Nd-148 method (ASTM, E321-96)4,5. The third fuel was a couple of burnable poison fuel (5.98 wt % Gd2O3-UO2)/zirlo cladding, which was discharged from the NPP(K-3). The average burn-up of the poisoned fuel specimen was 41 GWd/tU. At high burn-up conditions, fuel can come into contact with cladding due to fuel swelling, which results in a chemical interaction between the fuel and cladding (FCCI). There have been many studies on FCCI to investigate fuel rod behaviour1-3. Because Zr is thermodynamically reactive with oxygen, this chemical interaction leads to the oxidation of the inner surface of the cladding. This in turn leads to a thinning of the cladding, which affects the safety of fuel pins. The extent of the reaction layer depends on the contact between UO2 and its cladding. Because oxygen transportation from UO2 to cladding is difficult to achieve, the oxidation of Zr cannot be observed without direct contact. In this study, to access the structural changes of the cladding by chemical interaction with UO2, the diffraction spectrum of a cladding's inner surface was measured for the different cladding materials and different fuel burn-ups. The corrosion of the fuel cladding under the primary circuits of pressurized water reactors and their dry storage is an important issue. Therefore, a cladding outer surface was also measured to investigate the oxidation behaviour of the cladding because this also leads to thinning of the cladding. Axial slices of the fuel rods were cut at a 3 mm height. The slices of the spent fuels were embedded in epoxy resin and then polished at the post irradiation examination facility. A radiation-shielded µ-XRD system (D8 ADVANCED, BRUKER-AXS) was used to obtain the diffraction spectra6,7. For an accurate and precise positioning of the samples, two linear stages (travel distance: 25 mm, minimum movement: 0.1 µm) and a rotation stage (rotation angle: 360º, minimum movement: 0.004º) were equipped with a microscope (magni- fication at 120 mm distance: 13x). Because the background signal of γ-radiation from a spent fuel specimen interrupts the diffracted signal from a sample and results in poor detection, 7010 Park et al. Asian J. Chem. EXPERIMENTAL 24 26 28 30 32 34 36 38 600 500 400 300 200 100 Diffraction angle 2 (º) θ Distance ( m) µ 0 Zr UO2 Zr Measuring time / step : 20 sec. / 0.04 ¡Æ(2?) UO 2 Outer Surface Pellet/Cladding Interface Zirlo UO2-5.98wt%Gd2O3 (Spent Fuel ) Fig. 2. XRD spectra from the inner surface to the outer surface of a zirlo cladding contacted with 41 GWd/tU poisoned fuel (UO2-5.98 wt % Gd2O3) discharged from NPP(K-3) the sample holder8 and detector were shielded with 5- to 10- mm thick tungsten sheets. The measurements were conducted with a scanning step of 0.04º (2θ) for 20 s per each count and a detector slit width of 0.6 mm. An X-ray diffraction spectrum was obtained by measuring one position of a specimen for approximately 2 h. The CuKα line filtered through an Ni foil and NaI (Tl) scintillation counter was used at 40 mA beam current and 40 kV beam generation power. The visual inspec- tions of the fuel pellet involving the cladding were conducted using a radiation-shielded SEM (XL30, Philips) and the chemical compositions of the surface were analyzed by shielded EDS for uranium and zirconium content and a WDS system for oxygen content. Diffraction angle 2 (º) θ Fig. 2. XRD spectra from the inner surface to the outer surface of a zirlo cladding contacted with 41 GWd/tU poisoned fuel (UO2-5.98 wt % Gd2O3) discharged from NPP(K-3) RESULTS AND DISCUSSION The uranium content decreased sharply at the interface between the rim region and reaction layer, which indicates that the uranium diffusion into the interaction layer was not significant. The O content decreased slightly in the reaction layer and then decreased sharply in the cladding. It is thus concluded that a ZrOx (x ≤ 2) was generated by an oxygen uptake of the cladding from the UO2 pellet. of the fuel (Fig. 5-c). The uranium content decreased sharply at the interface between the rim region and reaction layer, which indicates that the uranium diffusion into the interaction layer was not significant. The O content decreased slightly in the reaction layer and then decreased sharply in the cladding. It is thus concluded that a ZrOx (x ≤ 2) was generated by an oxygen uptake of the cladding from the UO2 pellet. 400 600 800 1000 1200 1400 1600 24 26 28 30 32 34 36 38 UO2 UO2 Zr Zr ZrO2 Intensity (counts) 2 (º) θ Fig. 4. XRD spectrum measured at the interaction layer between UO2 fuel and zry-4 cladding. Dotted line: index of ZrO2 phase from database Zry-4 Corrosion layer Epoxy resin 20 § Zry-4 Corrosion layer Epoxy resin 20 § Zry-4 Corrosion layer Epoxy resin 20 § Zry-4 Corrosion layer Epoxy resin 20 § Distance (mm) 0 10 20 30 40 50 14000 12000 10000 8000 6000 4000 2000 0 Intensity (counts) Zr O Zr O Fig. 6. Variations of Zr and O content from the outer corrosion layer to the inner zry-4 analyzed by EDS and WDS, respectively 400 600 800 1000 1200 1400 1600 24 26 28 30 32 34 36 38 UO2 UO2 Zr Zr ZrO2 Intensity (counts) 2 (º) θ Distance (mm) 0 10 20 30 40 50 14000 12000 10000 8000 6000 4000 2000 0 Intensity (counts) Zr O Zr O Fig. 4. XRD spectrum measured at the interaction layer between UO2 fuel and zry-4 cladding. Dotted line: index of ZrO2 phase from database (a) (b) (c) Zry-4 Interaction Layer Rim Distance (mm) (a) (b) (c) Zry-4 Interaction Layer Rim Zr Zr 30000 25000 20000 15000 10000 5000 0 U O (a) (b) U O (a) (b) (c) (c) 20 25 30 35 40 45 50 Intensity (counts) Fig. 5. RESULTS AND DISCUSSION Interaction between fuel and cladding: (top) SEM image, (bottom) the content of Zr and U measured by EDS and that of O measured by WDS (a) (b) (c) Zry-4 Interaction Layer Rim Distance (mm) (a) (b) (c) Zry-4 Interaction Layer Rim Zr Zr 30000 25000 20000 15000 10000 5000 0 U O (a) (b) U O (a) (b) (c) (c) 20 25 30 35 40 45 50 Intensity (counts) Intensity (counts) Distance (mm) Zr Zr 30000 25000 20000 15000 10000 5000 0 U O (a) (b) U O (a) (b) (c) (c) 20 25 30 35 40 45 50 Intensity (counts) Intensity (counts) Fig. 6. Variations of Zr and O content from the outer corrosion layer to the inner zry-4 analyzed by EDS and WDS, respectively RESULTS AND DISCUSSION This measured region can be divided by three parts as follows: (a) cladding, (b) reaction layer and (c) the rim region of the fuel. Zirconium is the major element in the cladding (Fig. 5-a) and its content decreased slightly in the reaction layer (Fig. 5-b) and then decreased considerably in the rim region At the outer surface, the UO2/zry-4 couple showed ZrO2 and ZrH2 phases, although these diffraction patterns were not clearly identified due to weak intensity of the peaks and peak overlap. The very thin corroded layer and limited spatial resolu- tion (ca. 50 µm) of our µ-XRD system appear to be responsible for the weak signals. Nevertheless, some estimation on the Diffraction angle, 2 (º) θ Fig. 1. XRD spectra of the zry-4 cladding contact with 41 GWd/tU UO2 fuel discharged from NPP(Y-2) Zr Zr ZrO2 ZrH2 Zr Diffraction angle, 2 (º) θ Zr Zr ZrO2 ZrH2 Zr Interface Spent Fuel (UO2) Cladding (Zr) 24 26 28 30 32 34 36 38 600 500 400 300 200 100 0 ZrO2 ZrH2 Zr Zr Zr UO UO2 Zr Zr Diffraction angle, 2 (º) θ Distance from inner surface of cladding ( m) µ Fig. 3. XRD spectra measured from the inner surface to the outer surface of the zry-4 cladding contacted with 58 GWd/tU UO2 fuel Interface Spent Fuel (UO2) Cladding (Zr) 24 26 28 30 32 34 36 38 600 500 400 300 200 100 0 ZrO2 ZrH2 Zr Zr Zr UO UO2 Zr Zr Diffraction angle, 2 (º) θ Distance from inner surface of cladding ( m) µ Diffraction angle, 2 (º) θ Fig. 1. XRD spectra of the zry-4 cladding contact with 41 GWd/tU UO2 fuel discharged from NPP(Y-2) Fig. 3. XRD spectra measured from the inner surface to the outer surface of the zry-4 cladding contacted with 58 GWd/tU UO2 fuel Microstructural Changes of Fuel Claddings for Pressurized Water Reactors Fuel after Irradiation 7011 Vol. 25, No. 12 (2013) 7011 EDS and WDS (Fig. 6) to clarify the composition of the corroded layer. The outer layer was shown to be zirconium oxide (Fig. 6-bottom). In the SEM image of the cladding (Fig. 6-top), the dark grey region represents the corrosion layer of the cladding caused by the oxygen uptake from cooling water. The thickness of the corroded layer was approximately 16 µm. of the fuel (Fig. 5-c). 5. Y.K. Ha, J.S. Kim, Y.S. Jeon, S.H. Han, H.S. Seo and K. Song, Nucl. Eng. Technol., 42, 79 (2010). 6. Y.S. Park, S.H. Han, J.G. Kim and K.Y. Jee, Design and Fabrication of Micro X-ray Diffraction System, KAERI/TR-3256/2006, Korea Atomic Energy Research Institute (2006). 9. M. Tupin, M. Pijolat, F. Valdivieso, M. Soustelle, A. Frichet and P. Barberis, J. Nucl. Mater., 317, 130 (2003). 10. J.H. Baek, K.B. Park and Y.H. Jeong, J. Nucl. Mater., 335, 443 (2004). gy 7. Y.S. Park, S.H. Han, Y.K. Ha and K.Y. Jee, Design and Fabrication of Radiation Shielded Micro X-ray Diffraction System. KAERI/TR-3295/ 2006, Korea Atomic Energy Research Institute (2006). ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS This work was supported by the National Research Foun- dation of Korea (NRF) grant funded by the Korea Government. 8. Y.S. Park, J.G. Kim, Y.K. Ha and K. Song, Sample Holder Device for X-ray Diffraction Analysis of a High Radioactive Sample, Korean Patent 1039749 (2011). 5. Y.K. Ha, J.S. Kim, Y.S. Jeon, S.H. Han, H.S. Seo and K. Song, Nucl. Eng. Technol., 42, 79 (2010). 6. Y.S. Park, S.H. Han, J.G. Kim and K.Y. Jee, Design and Fabrication of Micro X-ray Diffraction System, KAERI/TR-3256/2006, Korea Atomic Energy Research Institute (2006). 7. Y.S. Park, S.H. Han, Y.K. Ha and K.Y. Jee, Design and Fabrication of Radiation Shielded Micro X-ray Diffraction System. KAERI/TR-3295/ 2006, Korea Atomic Energy Research Institute (2006). 8. Y.S. Park, J.G. Kim, Y.K. Ha and K. Song, Sample Holder Device for X-ray Diffraction Analysis of a High Radioactive Sample, Korean Patent 1039749 (2011). 9. M. Tupin, M. Pijolat, F. Valdivieso, M. Soustelle, A. Frichet and P. Barberis, J. Nucl. Mater., 317, 130 (2003). 10. J.H. Baek, K.B. Park and Y.H. Jeong, J. Nucl. Mater., 335, 443 (2004). 3. P. Hofmann, J. Nucl. Mater., 270, 194 (1999). Conclusion The changes of claddings after irradiation were measured by µ-XRD. At lower burn-up (41 GWd/tU), the fuel and its cladding did not interact with each other, regardless of the cladding material (zry-4 or zirlo). However, the ZrO2 and ZrH2 phases in the outer layer of the zry-4 cladding were observed, but nothing was detected except zirconium in the outer layer of a zirlo cladding. We concluded that the corrosion resistance of zirlo cladding is superior to that of zry-4 cladding. Fig. 5. Interaction between fuel and cladding: (top) SEM image, (bottom) the content of Zr and U measured by EDS and that of O measured by WDS At higher burn-up (58 GWd/tU), the zirconium oxide (ZrOx) layer was observed at the interface between the UO2 fuel and zry-4 cladding, which suggests that oxygen diffused from the fuel into the cladding. In the outer surface of the zry-4 cladding, the formation of ZrO2 and ZrH2 phases by the oxygen uptake from the cooling water was observed. The thickness of the corroded layer was approximately 16 µm. Continued research on the fuels with burn-up over 58 GWd/ tU is in progress and the results may provide a clearer picture regarding FCCI and structural change. In the outer surface of the cladding, the formation of ZrO2 and ZrH2 phases by the diffusion of oxygen and hydrogen from the cooling water was observed. However, the diffraction patterns of the ZrO2 and ZrH2 phases were not clearly identi- fied due to the weak intensity of the peaks and peak overlap. The reasons for the weak signal are the very thin corroded layer and limited spatial resolution (ca. 50 µm) of our µ-XRD system. Therefore, the elemental analysis was performed by 7012 Park et al. Asian J. Chem. ACKNOWLEDGEMENTS 4. Atom Percent Fission in Uranium and Plutonium Fuel (Neodymium- 148 Method). American Society for Testing and Materials Standard Method, E321-96, 1 (1996). REFERENCES 8. Y.S. Park, J.G. Kim, Y.K. Ha and K. Song, Sample Holder Device for X-ray Diffraction Analysis of a High Radioactive Sample, Korean Patent 1039749 (2011). 1. P. Hofmann and D. Kerwin-Peck, J. Nucl. Mater., 124, 80 (1984). 2. D.R. Olander, J. Nucl. Mater., 115, 271 (1983). 3. P. Hofmann, J. Nucl. Mater., 270, 194 (1999). 3. P. Hofmann, J. Nucl. Mater., 270, 194 (1999). 9. M. Tupin, M. Pijolat, F. Valdivieso, M. Soustelle, A. Frichet and P. Barberis, J. Nucl. Mater., 317, 130 (2003). 4. Atom Percent Fission in Uranium and Plutonium Fuel (Neodymium- 148 Method). American Society for Testing and Materials Standard Method, E321-96, 1 (1996). 10. J.H. Baek, K.B. Park and Y.H. Jeong, J. Nucl. Mater., 335, 443 (2004).
https://openalex.org/W2140904449	https://www.epj-conferences.org/articles/epjconf/pdf/2010/05/epjconf_ICEM14_05005.pdf	English	null	Characterization of the mechanical behavior of ultrafinegrained metals using digital image correlation	EPJ web of conferences	2,010	cc-by	4,367	Characterization of the mechanical behavior of ultrafine- grained metals using digital image correlation M. Hokka1,2,a, J. Seidt2, T. Matrka2, A. Gilat2, V.-T. Kuokkala1, J. Kokkonen1 and S. Müller3 1Tampere University of Technology, Department of Materials Science, POB 589, Tampere, Finland 2The Ohio State University, Department of Mechanical Engineering, 201 West 19th Avenue, Columbus OH, 43210 USA 3Technische Universität Berlin, Extrusion Research and Development Center, Gustav-Meyer-Allee 25, D-13355 Berlin, Germany Abstract. Nanocrystalline and ultrafine-grained materials show many very interesting properties compared to the conventional coarse-grained materials. For example, their strength values are usually much higher but, on the other hand, their ductility and deformability can be significantly lowered by the grain size refinement processes. One of the methods used to refine the grain size of metals is severe plastic deformation (SPD), which usually produces highly oriented structures that are strong but may deform non-homogeneously. In this paper, the deformation behavior of severe plastic deformation processed AZ31 magnesium alloy and commercially pure 1070 aluminum were characterized in compression and torsion using mechanical testing and high speed photography with digital image correlation. The results show that the AZ31 magnesium alloy deforms fairly uniformly in compression until fracture and the strains calculated from the movements of the ends of the anvils and the Lagrange strain on the surface of the specimen describe the overall deformation quite well. For the 1070 UFG aluminum, the shear deformation localizes very rapidly and the average strains measured from the surface or calculated from the displacements of the bars underestimate the true strain in the shear bands significantly. Also, the deformation in the torsion tests was not uniform shear but substantial strains were also measured in the axial direction. These strains were most likely due to small experimental errors, such as slight misalignment of the bars and the specimen in the test. a e-mail : mikko.hokka@tut.fi © Owned by the authors, published by EDP Sciences, 2010 DOI:10.1051/epjconf/20100605005 EPJ Web of Conferences 6, 05005 (2010) 6 © Owned by the authors, published by EDP Sciences, 2010 DOI:10.1051/epjconf/20100605005 EPJ Web of Conferences 6, 05005 (2010) 6 1 Introduction The mechanical properties of most materials are strongly affected by strain rate. Usually the strength of material increases with increasing strain rate, especially at strain rates around 103 s-1, where a rapid increase in strength is often observed. In practice, high strain rates in the range of 103 s-1 to 104 s-1 occur in many industrial applications and high speed material forming methods, such as dynamic loading of structures and cold heading, as well as in many events of practical importance such as automotive crashes and various defense and aerospace applications. At quasi-static and intermediate This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial License 3.0, which permits unrestricted use, distribution, and reproduction in any noncommercial medium, provided the original work is properly cited. rticle available at http://www.epj-conferences.org or http://dx.doi.org/10.1051/epjconf/20100605005 EPJ Web of Conferences strain rates (<~500 s-1), the deformation behavior of crystalline materials such as coarse-grained metals is readily explained by the thermally activated dislocation motion. At higher strain rates (>~1000 s-1), the dislocation drag mechanisms start to control the deformation behavior, and usually the strength of the material increases much faster than in the thermally activated region due to the change in the deformation and transient hardening mechanisms. However, when the crystal sizes are decreased to nanometer scale, the movement of dislocations becomes more and more difficult and new phenomena may take place. At the moment, the strain rate dependent deformation and strain hardening mechanisms of nanocrystalline metals are not yet properly understood, despite the quite extensive research that has been done on these materials over the past decade. Bulk nanocrystalline (nc) metals have significantly improved strength properties and usually lower deformability and ductility compared to the coarse-grained metals. These materials can be produced using several techniques, but the two main competing technologies are compaction and sintering of nanopowders and severe plastic deformation (SPD) of bulk metals. However, only SPD techniques enable production of impurity and porosity free bulk metals from basically any metal, which cannot be done with conventional powder metallurgical techniques. In addition, severe plastic deformation does not require atomization of the metal as the nanocrystallization is performed in bulk form, which is environmentally more friendly and less hazardous since there is no need to handle the nanoparticle powders. In principle, any coarse-grained bulk metal may be plastically deformed in such a way that its grain size is reduced to submicroscopic level, but so far the most commonly studied nanocrystalline bulk metals are copper, aluminum and its alloys, titanium and its alloys, and tantalum. Ultrafine-grained (ufg) materials have similar properties as the nanocrystalline metals, i.e., very high strength, generally low ductility due to tensile or shear instability, low strain hardening capability, high fatigue strength, and altered strain rate sensitivity compared to coarse-grained materials [1]. Usually materials with grain size below 100 nm are referred to as nanocrystalline and those with grain size between 100 nm and 1 µm as ultrafine-grained. The potential applicability of both nanocrystalline and ultrafine-grained materials is wide, including aerospace applications, transportation, health care equipment and components, defense applications, and even implant technology due to the high purity of the materials. 1.1 Digital image correlation Digital image correlation (DIC) can be used to measure 3D surface displacement fields of the specimen during dynamic or static loading. Typically two digital cameras viewing the specimen at different angles are used to acquire pairs of images of the specimen. The displacements are then calculated by dividing the images into smaller subsets or facets and by matching the facets in each image pairs to track their movement in subsequent images. Theoretically, the accuracy of the image subset matching can be even as high as 1/200 of a pixel [7]. The strains on the surface of the specimen can therefore be determined with a very high spatial resolution even with relatively poor pixel resolution of the image. The digital image correlation has been found very useful in the characterization of strain distributions in compression and tension experiments [2-4]. In high strain rate tension experiments, in particular, the strains calculated based on the relative motions of the ends of the specimen can give erroneous results due to the localization of strain after necking, but also because of the yielding of the rounded fillet regions of the specimens and the uncertainties in the true gage length determination of the specimen. Due to the large plastic strains involved in the SPD processing, many ultrafine-grained metals have a very strong texture. Also, some SPD techniques such as high pressure torsion or reciprocating extrusion produce gradient structures, where some regions have much higher strains while other regions are deformed significantly less. This of course leads to different microstructures and grain sizes throughout the cross section of the material, and most likely also the mechanical properties vary along the cross section. Therefore, due to these gradient structures accurate determination of mechanical properties and analyzing the mechanical behavior of these materials can be very difficult 2 Materials and experiments The materials studied in this work were ultrafine-grained AA1070 (99.7 wt-% Al) commercial purity aluminum and near-UFG AZ31 magnesium alloy. Torsion tests at a wide range of strain rates were performed on the UFG aluminum, whereas compression tests were performed on the near-UFG AZ31 magnesium alloy. Some torsion tests were also performed on the standard AA7075 aluminum alloy for comparison. The grain size of the 1070 aluminum was refined by Equal Channel Angular Pressing by passing the billet four times through a 2-turn 90° channel die at room temperature, leading to a total accumulated shear strain of about 9.4. The average grain size after the ECA –pressing was estimated to be roughly 750 nm based on transmission electron micrographs [5]. The aluminum was produced at the Institute of Materials Processing Technology at Warsaw University of Technology. The technique and equipment used for the processing are described in more details in refs. [9-10] q q p p g The original size of the billets was 2626120 mm3, from which the torsion specimens were manufactured. First the billets were dissected into four rectangular blocks of 1212120 mm3 and then machined into thin walled tube specimens. The gage length of the specimens was 5.08 mm (0.2 in) or 2.54 mm (0.1 in), the mean gage diameter 9.27 mm (0.365 in), and the wall thickness 0.38 mm (0.015 in) and 0.76 mm (0.03 in) for the 7075 aluminum alloy and ultrafine grained 1070 aluminum, respectively. The AZ31 magnesium alloy was processed by reciprocating extrusion at 300 °C. A schematic picture of the reciprocating extrusion tool is shown in Figure 1. A more detailed description of the processing apparatus can be found in ref. [6]. The 30 mm diameter billet was first extruded back and forth through the kneeding die with a 26 mm diameter collar in the middle of the extrusion channel. After five passes through the die, the billet was extruded into a round bar with a 10 mm diameter and then cooled in air. The material was processed both in the as-cast and as-cast and extruded conditions, and the compression behavior of both alloys was studied in a wide range of strain rates. The microstructures for the cast and reciprocating extrusion processed AZ31 was very close to that of the cast, extruded and reciprocating extrusion processed AZ31. 05005-p.2 14th International Conference on Experimental Mechanics using conventional material testing methods. Also, while conducting mechanical testing at a very wide range of strain rates, the average strain in the specimen is usually measured using several different techniques that may lead to problems when comparing the properties in different strain rate regimes. The digital image correlation can be used virtually at any strain rate provided that the frame rates of the cameras are sufficient. In this paper, the mechanical properties of a reciprocating extrusion processed AZ31 magnesium alloy were tested in compression and ECAP processed 1070 commercially pure (99.7-wt%) aluminum in torsion. The tests were monitored with two digital cameras and the strains were calculated using the digital image correlation technique. Special emphasis was placed on how the digital image correlation can be effectively used to analyze the development of strains during the test. 2 Materials and experiments The average grain sizes for both grades were around 6-8 micrometers, and their microstructures consisted of very small grains surrounding larger diameter less-deformed grains. The small grains were equiaxed for the AZ31 extruded prior to the grain size refinement. More elongated and deformed grains were found in the AZ31 without prior extrusion. The magnesium alloys were prepared at the Technische Universität Berlin. 05005-p.3 EPJ Web of Conferences a) b) Fig. 1. Schematics of the grain size refinement procedures used for the tested materials, a) the reciprocating extrusion used for the AZ31 magnesium alloy [after 6], and b) the ECAP die used for the 1070 aluminum. EPJ Web of Conferences a) b) Fig. 1. Schematics of the grain size refinement procedures used for the tested materials, a) the reciprocating extrusion used for the AZ31 magnesium alloy [after 6], and b) the ECAP die used for the 1070 aluminum. Mechanical testing at low strain rates was performed using a servohydraulic materials testing machine equipped with both axial and angular actuators and load cells, whereas the high strain rate testing was done using Split Hopkinson Pressure Bar devices. The compression Split Hopkinson Pressure Bar consisted of two 12.7 mm (0.5 in) titanium bars, between which the specimen was sandwiched. The SHPB test is performed by impacting a striker bar on the free end of the incident bar, producing a compressive stress pulse that propagates in the incident bar towards the specimen. As the pulse reaches the bar-specimen interface, part of the pulse is reflected back as a wave of tension while part of the incident wave is transmitted through the specimen into the transmitted bar. The three stress pulses, incident, reflected, and transmitted, are measured using resistive strain gages bonded on the surfaces of the bars, amplified and finally recorded on a digital oscilloscope. The stress, strain, and strain rate in the specimen can then be easily calculated from the stress pulses using Matlab based calculation routines. The torsion Split Hopkinson Pressure Bar device is somewhat more complicated than the compression equipment. The torsional stress is first stored in the incident bar by a hydraulic pulley and a clamp-release system. The torsion stress pulse is generated simply by releasing the stored torque by breaking the clamp pin by a second hydraulic press. 2 Materials and experiments The bar stresses are measured by three strain gage stations and the pulses are calculated from the measured signals by applying the appropriate time shifts [8]. 2.1 Strain Measurements Measuring strain at a wide range of strain rates usually requires several different measurement techniques to cover each strain rate regime. At low strain rates, the strain is usually measured directly from the specimen using extensometers, or calculated from the motion of the anvils from the LVDT transducer signals. At high strain rates, on the other hand, the strains are usually calculated from the time resolved strain signals measured from the stress bars with strain gages. With digital image correlation, the strains can be measured using the same method at a very wide range of strain rates, and the maximum strain rate is only limited by the maximum frame rate, resolution, and optical properties of the digital cameras. In this study, two Grasshopper (Point Gray Research) 2 megapixel cameras were used in the low strain rate tests at the maximum frame rate of 19 s-1. Two Photron Fastcam SA1.1 high speed cameras were, in turn, used at higher strain rates operating at frame rates between 1000 and 120.000 s-1. In the compression tests, the Lagrange strains were calculated from the images recorded directly from the surface of the specimen. The spatial Lagrange strains were then converted into average engineering strains using Equation 1, where E11 is the Lagrange strain in the axial direction of the specimen. 14th International Conference on Experimental Mechanics 14th International Conference on Experimental Mechanics N E N i i E ∑ = − + = 1 11 1 2 1 ε (1) N E N i i E ∑ = − + = 1 11 1 2 1 ε (1) N E N i i E ∑ = − + = 1 11 1 2 1 ε (1) N In the torsion tests, the strain measurements were complicated by the very high degrees of strain and the problems in maintaining a good contrast pattern on the surface of the specimen. First, a black base coat with hand painted white coarse dots were used for maximum contrast, but the paint chipped off from the surface after 30-40% of strain and the DIC signal was lost. A white Krylon base coat and spray painted black contrast patterns were found to have better adhesion to the surface, but the lower contrast lead to slightly lower spatial accuracy of the measurements. To overcome the problem of losing the contrast pattern at high strains, the ends of the specimen outside the gage section were also painted and the displacements of the ends of the gage section were calculated from the images. From these displacements, the rotation or the twist angle of the specimen was calculated, and finally the average strains in the specimen were calculated using Equation 2, where rs and ls are the average radius and length of the gage section of the specimen, θ is the angle of twist, and a and b are the vectors from the centerline of the specimen to the surface at the opposite ends of the gage section. )\|\| \|\| * \|\| \|\| arccos( b a b a l r l r s s s s • = = θ γ (2) (2) 3 Results and discussion In Figure 2a, compressive stress-strain curves for the cast, extruded and SPD processed near-UFG AZ31 alloy measured at the strain rate of 10-3 s-1 are plotted using both the strains measured directly from the specimen and the strains calculated from the displacements of the anvils of the materials testing machine. The difference between the curves is quite small at low strains, but at higher strains there is an increasing shift between the curves, indicating that the strains measured directly from the specimen are much smaller than those calculated from the displacements of the anvils. This is typical for a compression test, and the difference is mainly due to the barreling of the specimen during compression. The different compressive strains measured from the anvil displacements with DIC or directly from the surface of the specimen are well comparable for this type of a material, which deforms fairly uniformly and fractures at a fairly low strain. If the strains were much higher, the strains would most likely differ more due to the increasing barreling at higher strains. y g g g In the dynamic torsion test, the shear strain can be calculated either from the stress pulses in the bars, from the Lagrange shear strain (γ=2εxy) measured directly from the surface of the specimen, or from the displacement of the ends of the specimen using Eq. 2. Figure 2b shows the comparison of the shear stress-shear strain curves obtained with strains determined with the three above mentioned methods for the UFG 1070 aluminum at the strain rate of 1000 s-1. The Lagrange strain could be calculated only to about 50% strain, where the paint on the surface of the gage section chips off. As seen in the figure, at higher strains the shear strain calculated from the stress waves is about 8% higher than the strain calculated from the displacements of the ends of the specimen. 05005-p.5 EPJ Web of Conferences a) b) Fig. 2. Stress-strain curves obtained with different strain determination methods for a) cast, extruded and SPD processed AZ31 in compression at the strain rate of 10-3 s-1, and b) UFG 1070 aluminum in torsion at the strain rate of 1000 s-1. a) b) b) a) Fig. 2. 3 Results and discussion Stress-strain curves obtained with different strain determination methods for a) cast, extruded and SPD processed AZ31 in compression at the strain rate of 10-3 s-1, and b) UFG 1070 aluminum in torsion at the strain rate of 1000 s-1. The strains measured and calculated from the torsion specimens are better shown in Figure 3, which shows the transverse and axial strain components, εxx, εyy, and the three different shear strains measured with digital image correlation and using the stress waves in the bar. Figure 3a shows the strains for the 7075-T6 aluminum alloy at the strain rate of 480 s-1, and Figure 3b for the UFG 1070 aluminum at the strain rate of 1000 s-1. For both materials it is clear that the deformation is not uniform shear, as assumed by the one dimensional wave propagation that is used in calculating the specimen strain from the stress pulses measured from the pressure bars. The specimen undergoes deformation also in other directions, especially in the direction of the axis of the specimen (εyy), whereas the transverse strain component εxx is very small. The shear strain calculated from the stress pulses increases quite linearly, whereas the strain calculated from the displacements of the ends of the bars shows somewhat less linear behavior. The average shear strain measured directly from the surface of the specimen over the gage section is clearly higher than the average shear strain calculated from the rotations of the bars. The strains εxx and εyy are most likely caused by a combination of several small experimental errors, including slight misalignment of the bars, bending of the bars near the specimen, eccentric positioning of the specimen, and eccentric machining of the tubular specimen. a) b) Fig. 3. Lagrange strains measured from the surface of the specimen (εxx, εyy) and the shear strains calculated from the Lagrange strain, displacement of the ends of the specimen, and from the stress pulses in the bars for a) the 7075-T6 aluminum alloy and b) the UFG 1070 aluminum. Figure 4 shows the waterfall plots (strain as a function of position and time) for both a compression test of AZ31 (Fig.4a.) and a torsion test of UFG AA1070 (Fig.4b.). In the compression test of AZ31, b) a) b) a) Fig. 3. 3 Results and discussion Lagrange strains measured from the surface of the specimen (εxx, εyy) and the shear strains calculated from the Lagrange strain, displacement of the ends of the specimen, and from the stress pulses in the bars for a) the 7075-T6 aluminum alloy and b) the UFG 1070 aluminum. Figure 4 shows the waterfall plots (strain as a function of position and time) for both a compression test of AZ31 (Fig.4a.) and a torsion test of UFG AA1070 (Fig.4b.). In the compression test of AZ31, Figure 4 shows the waterfall plots (strain as a function of position and time) for both a compression test of AZ31 (Fig.4a.) and a torsion test of UFG AA1070 (Fig.4b.). In the compression test of AZ31, 05005-p.6 05005-p.6 14th International Conference on Experimental Mechanics the strain is somewhat asymmetric with respect to position but increases quite uniformly with time. In different tests, however, the degree of asymmetry varied quite much for reasons that are not completely understood. The average strain (dashed lines in Figure 4a) represent fairly well the overall deformation of the specimen, especially at low strains, which is typical for compression tests [2]. In the compression tests of AZ31, the strain localized and shear bands were formed just before the final fracture. In the torsion tests of UFG aluminum, strain localizes very rapidly after yielding and most of the deformation occurs in a very narrow region roughly in the middle of the gage section. A fully developed shear band can be seen in Fig 4c. Therefore, in a torsion test the average strain over the whole gage section is significantly lower than the local strains in the shear band. As seen in Figure 4b, the maximum strain at the ends of the specimen is around 5% at one end and 25% at the other end, whereas in the shear band it is around 45%. The average shear strain calculated over the whole gage section of the specimen from the Lagrange strain (γ=2εxy) is significantly higher than the other average shear strains, i.e., shear strain calculated from the stress pulses and from the displacements of the ends of the specimen, but it still is much lower than the shear strains in the shear band. Therefore, the “true” shear strain should be calculated only from the localized region instead of using an average over the whole gage section of the specimen. a) b) c) Fig. 4. 3 Summary Mechanical testing was performed on reciprocating extrusion processed near-ultrafine-grained AZ31 magnesium alloy, ECAP processed ultrafine-grained commercially pure AA 1070 aluminum, and standard 7075-T6 aluminum alloy. Compression tests were performed on the AZ31 alloy and torsion tests on the UFG aluminum and 7075 aluminum alloy. The tests were monitored using digital cameras, and the strain distributions on the specimen surfaces were calculated using the digital image correlation technique. In the compression tests of AZ31, the strains in the specimen were relatively homogeneously distributed and uniform, and no localization of strain was observed until the final 45º shear fracture of the specimen that was preceded by the formation of a shear band with very high local strains. The behavior of the ultrafine-grained aluminum in the torsion tests was more complicated, and the strains localized very rapidly already in the early stages of the test. In general, the shear band that is formed during the deformation has much higher strains than the surrounding areas, and the average strain measured from the displacements of the ends of the bars do not describe the material’s true behavior correctly. Also, the deformation in the torsion tests was found not to be uniform shear, but also significant axial strains were observed during the tests. These axial strains were obviously due to slight experimental errors such as the slight misalignment of the bars and the specimen. Strain measured from the rotational displacements of the ends of the specimen seem to be more accurate than the strain obtained from the signals of the strain gages glued in the stress bars, which do not take into account the possible eccentric motions of the ends of the bars. EPJ Web of Conferences EPJ Web of Conferences Acknowledgements Dr. A. Rosochowski and Dr. L. Oljenik are thanked for providing the ultrafine-grained 1070 aluminum material. This work was partly funded by the Finnish Funding Agency for Technology and Innovation (MIVA project) and the Academy of Finland under the grant No. 130780. The support from the Finnish Foundation for Technology Promotion is also acknowledged. 3 Results and discussion Waterfall plots for a) AZ31 in compression and b) ultrafine-grained 1070 aluminum in torsion, and c) a fully developed shear band in a torsion test of ultrafine-grained 1070 at the strain rate of 1.6 10-4 s-1. b) a) b) a) c) c) Fig. 4. Waterfall plots for a) AZ31 in compression and b) ultrafine-grained 1070 aluminum in torsion, and c) a fully developed shear band in a torsion test of ultrafine-grained 1070 at the strain rate of 1.6 10-4 s-1. 05005-p.7 05005-p.7 References 1. R. Valiev, T. Langdon, Progress in Materials Science. 51, (2006) 2. A. Gilat, T. Schmidt, A. Walker, Experimental Mechanics, 49, (2009) 3. A. Gilat, T. Schmidt, J. Tyson, Society for Experimental Mechanics, (2006) 4. A. Gilat, A. Walker, J. Seidt, Society for Experimental Mechanics, (2008) 5. J. Kokkonen, V.-T. Kuokkala, L. Olejnik, A. Rosochowski, Society for Experimental Mechanics, (2008) ( ) 6. K. Müller, S. Müller, Journal of Materials Processing Technology, 178-188, (2007) 7. M. Sutton, J.-J. Orteu, H. Schreier, Image Correlation for Shape, Motion and Deformation Measurements (Springer Science, New York, NY, USA, 2009) 8. A. Gilat, ASM Handbook, Volume 8, Mechanical Testing and Evaluation (ASM International, Materials Park, OH, USA, 2000) ) 9. A Rosochowski, L. Olejnik, Journal of Materials Processing Technology, 125-126, (2002 10. L. Olejnik, A. Rosochowski, Bulletin of the Polish Academy of Sciences, 53, (2005 05005-p.8
W2993657663.txt	https://link.springer.com/content/pdf/10.1007/978-94-024-1720-3_6.pdf	en		Engaging Small and Medium-Sized Enterprises in Responsible Innovation	SpringerBriefs in research and innovation governance	2,019	cc-by	5,577	Chapter 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation Catherine Flick, Malcolm Fisk, and George Ogoh Abstract A significant part of responsible innovation is engagement with diverse groups of stakeholders; this remains true for projects investigating responsible innovation practices. This chapter discusses strategies for engaging small and medium- sized enterprises (SMEs) in co-creating visions of and plans for implementing responsible innovation, drawing on the example of engagement with United Kingdom cyber security companies. The key aspect of the engagement was building trust between the responsible innovation researchers and the companies. Trust was built by a movement away from traditional recruitment procedures for research projects, towards proactive engagement with the culture and traditions of the sector – participating in company sponsored talks and conferences, finding ways to communicate effectively, and ensuring a tailored message that fit the expectations and requirements of the sector. This chapter reviews the context in which the recruitment took place, the assumptions made prior to recruitment, the approaches taken, the revisions made to these approaches, and ultimately offers some general recommendations for industry engagement in responsible innovation activities. Keywords Cyber security · Responsible innovation · Engagement · Small- medium enterprises · Trust 6.1 Introduction Some of the most significant challenges for responsible innovation in industry include raising awareness of the concept, showing businesses its value, and capturing businesses’ interest in implementing responsible innovation in their own research and development practice. This chapter looks at the engagement of cyber security SMEs (small and medium-sized enterprises) in responsible innovation, by C. Flick (*) · M. Fisk · G. Ogoh School of Computer Science and Informatics, De Montfort University, Leicester, UK e-mail: cflick@dmu.ac.uk; malcolm.fisk@dmu.ac.uk; george.ogoh@dmu.ac.uk © The Author(s) 2020 K. Jarmai (ed.), Responsible Innovation, SpringerBriefs in Research and Innovation Governance, https://doi.org/10.1007/978-94-024-1720-3_6 71 72 C. Flick et al. investigating the techniques that were used to recruit companies for a series of online and face-to-face peer co-creative workshops on implementing responsible innovation within the United Kingdom cyber security sector. The analysis of these engagement methods culminates in a set of general requirements and recommendations for engaging primarily with cyber security companies, but which also have general relevance to other industry sectors. Responsible innovation, as has been seen in previous chapters, is a set of practices by which researchers and innovators engage with society to identify social and ethical impacts and issues of the technologies they are developing. Largely referred to in the academic world as the more cumbersome “responsible research and innovation” (RRI), definitions of responsible innovation are many and varied, but the general idea is that innovation should include society, deliberate on ethical and social issues, and align with societal needs (European Commission and Directorate- General for Research and Innovation 2013; Owen et al. 2013; Von Schomberg 2013). However, the concept of RRI does not have much penetration into industry (Stahl et al. 2017), and industry players are more likely to know and recognise terms such as corporate social responsibility (CSR) (European Commission 2011) or more simply, business ethics. To reflect this finding, and for reasons we discuss below, we henceforth refer to RRI as “responsible innovation” (RI). In order to engage effectively with cyber security companies on topics surrounding RI and engage them in the planned workshops, a communications strategy needed to be devised. The approaches that were successful focused on the opportunities available to SMEs, were individually tailored to their spaces and requirements, and helped to ensured that the SMEs were comfortable in discussing confidential information. These experiences found that a desire for the development of trust with the general public, consumers of companies’ products and services, and/or other businesses was a major driving factor in their engagement with RI. This chapter reviews the context in which the recruitment of the companies took place, the assumptions made prior to recruitment, the recruitment approaches taken, the revisions made to these approaches, and offers some general recommendations for industry engagement. It argues that one of the most effective strategies for recruitment and engagement of SMEs is to become involved in the existing communication spaces of the sector, rather than expecting companies to respond to calls for interest. 6.2 Responsible Innovation for Cyber Security Companies Previous chapters have explored the potential benefits of following a value-based approach to corporate innovation. However, the value propositions need to be well- defined and to generally align with existing goals within the company if they are to be considered useful. For example, for cyber security, trust seems to be a significant factor in interest in RI. The value that public and customer trust has for each cyber security company is significant, although this might not initially have been seen by 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 73 cyber security SMEs in monetary terms and business sustainability. However, when RI activities were explained to cyber security SMEs in the context of ethics, responsibility, privacy, and trust, and with only a passing mention of ‘responsible innovation’ (instead of attempting to define RI explicitly), companies could see the alignment with their existing value statements, medium-long term goals, and discussions that had already taken place internally (especially regarding ethics). In fact, for cyber security companies, ethics and trust are regular topics of industry discussion, with philosophical differences arising between different camps on particular ethical dilemmas, such as disclosure of vulnerabilities (responsible disclosure vs. full disclosure), and bug bounties (rewards offered by companies for finding exploitable bugs in their software) (Hughes 2015; Lefkowitz 2017). The emphasis on security is growing in a more uncertain and technologically- dependent world. Cyber security is therefore a natural growth area for industry, and a good example subsector of the more general IT industry, much of which grapples with uncertainty. It is a loosely defined sector encompassing many different types of security-related products and services. Much of the cyber security market is business-to-business, offering reputation protection, security of data, forensics and fraud detection, and server security. However, cyber security companies are also responsible for products and services that consumers use, such as security cameras, identity management apps, encryption of devices, and educational materials. The nature of cyber security’s past can be suggestive of a somewhat ‘cowboy’ culture, with its frontiers of technological crime prevention often seen as a ‘grey area’ - including ‘white-hat‘ (those who operate within legal and ethical norms), ‘grey- hat‘ (those who operate mainly in a legal sphere, but occasionally exploit opportunities of policy vacuums, usually within ethical norms), and ‘black-hat’ hackers (those who break legal and ethical norms) operating on both sides of the law to meet their goals. Coupled with the complexity of the topic and issues, as well as poor representation of the field in movies and TV shows, there is a significant lack of understanding of what cyber security is, what its goals are, and how it works. This can translate into a lack of trust between end-users and security companies and their products, or to a view of cyber security products and services as ‘grudge purchases’ made by companies who view the sector much as they see insurance. Thus, the value of RI to cyber security companies is in helping them to develop these trust relationships with their clients, whether they are individual end-users or companies. In this way, a company can show its trustworthiness to users who may not understand the technicalities, theoretical aspects, or even the user interfaces for cyber security. And in helping the cyber security sector to engage in openness, transparency, ethics and responsibility, along with other RI practices, clients who do not understand the inner workings of the technologies involved can develop a stronger trust relationship with the company. We found in our work that companies are eager to engage with the concept of trust. The strategies detailed in the following section point to ways of harnessing companies’ interest. 74 6.3 C. Flick et al. Recruitment Aims and Strategies This section looks briefly at the aims of the RI research to provide context, then in more depth at the strategies which were chosen to approach the cyber security community to participate in the project, and describes the most effective engagement processes. It also discusses complications that arose after companies had made a commitment to the process. The resulting approach was effective in engaging companies for the RI workshops and helped build a significant rapport with the companies that, in turn, improved the outcomes of the workshops. The aim of these interventions was to engage companies in a series of on- and off-line workshops. It was envisaged that the companies would work together to develop a ‘responsible innovation roadmap’ co-creatively with their peers, facilitated by the workshop leaders. Initially, there were to be three webinar-style online workshops, and two face-to-face workshops, where the companies would come together to co-create the shared roadmap using foresight and backcasting methodologies.1 The cyber security companies were to be from the UK and considered as SMEs (up to 250 employees). SMEs were targeted as approximately 50% of SMEs in the UK are engaged in innovation activities (Department for Business, Energy, and Industrial Strategy 2017), but unlike large companies, they often do not have significant corporate social responsibility (or similar) arms. Prior to the strategy being developed by which SMEs would be approached, however, a concern arose that cyber security companies might be more difficult to engage than other sectors due to their more secretive nature, particularly if this was to be in a peer-led co-creative exercise such as the planned workshops. This concern was based on discussions with cyber security experts within academia about company involvement with their research, but, as this article will show, the concerns were relatively unfounded, as the topics of ethics, trust, and other technical philosophical discussions were seen by the companies individually to be interesting and relevant. However, it took some time to realise this specific entry point for engaging with companies, as is explained below. The peer-led co-creative exercise however, was correctly identified to be a problematic approach for this sector, regardless of interest in the topics. The evolution of the planned co-creative exercises is also detailed below.2 Firstly, a generic, academic-style call for participation was developed. This was sent to a number of contacts identified by members of the research project. Some effort was made to circulate this call through established cyber security fora, for example, the UK Cyber Security Forum, as well as more personal networks, such as university cyber security partners. This was based on the assumption that companies would be most likely to respond to personal contacts and through advertisements on an industry website. More information on the workshop methodology and approaches can be found in D2.5 at https:// innovation-compass.eu/deliverables-2/ 2 Examples of the drafts discussed below are available from the authors by request; due to space limitations we have included only the final, successful, recruitment letter in Appendix. 1 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 75 After poor engagement with this method (i.e. none), discussions with several cyber security experts were undertaken (university researchers with industry contacts; cyber security experts from other countries). Advice was taken on the nature of the ‘sales pitch’ (i.e. the description of the activities and the benefits to the companies in taking part) to make it more focused on the benefits that companies might gain from participating, as well as to avoid the implication that this would be largely an academic activity that might berate companies for unethical behaviour. The ‘RRI’ terminology was removed at this stage as it was considered by our advisors to be jargon, and could result in restricted discussion to the constituent parts, such as ethics. A more conversational tone was adopted, addressing some of the companies’ potential concerns; avoiding what might be seen as any moralising attitude or the pursuit of impractical theoretical outputs from academics; and included clear reference to links with established business organisations that were partners in the project. With this new pitch greater interest in the project was generated, but no companies confirmed any commitment to engagement. It seemed there was still some confusion as to what the benefits of participating in the research were and what was required of the companies, especially in terms of the time commitment. Significant discussion at a project meeting came up with the idea of pitching the workshops as free ‘innovation consulting’ to see if that would impact the involvement of companies. In this rather lengthy pitch, it was possible to demonstrate knowledge of the issues cyber security companies faced. Unfortunately, this new pitch did not work very well either, perhaps because of its length (six paragraphs and some bullet points), or perhaps because it seemed a bit too good to be true (in fact, one of the participant companies regularly checked to make sure they didn’t have to pay for anything). Also, it seemed that the relatively lengthy time commitments envisaged (“less than a day and a half spread over a couple of months”) were considered particularly onerous, and the collaborative working was seen as too complicated, in part due to the intellectual property that could be compromised if collaborative activities were undertaken. Further discussions within the project offered a revised and final (Appendix) research protocol, with two 2–2.5 h face-to-face workshops in which the researchers came to the companies. A revised sales pitch concentrated on the potential benefits for the companies from engaging in the activities, focusing on topics such as trust-building and ethics. Another change in strategy was to become engaged in activities that the companies were running themselves. In this way, rather than asking companies to come into what they might perceive as an academic world somewhat detached from commerce; the academics would be working in the world of industry. This was complemented by engaging in talks and networking events (De Montfort University Cyber Forum, IOActive’s HACK::SOHO, Malvern and South Wales Cyber Security clusters seminar sessions and workshops, a company launch) and speaking at industry venues. The ability to engage with the audience on the topics of ethics, r esponsibility and trust helped to validate the expertise of the researchers and the development of trust relationships with company representatives. For some companies, knowing that others had already taken up the offer also helped establish this trust relationship 76 C. Flick et al. with project engagement and workshops taking place, sometimes reinforced through recommendations from their advisory boards. Personal connections made through face-to-face discussions at networking events or talks also made a significant difference to the uptake of our subsequent workshops, compared with email introductions, and even more than cold-emailing. Frequently, the companies pointed to a specific set of issues they wished to be discussed in workshops, either problems they had encountered that we might give tailored advice on, or asking us to help them consider different options available to them as they moved from being a very small company of only a few employees to a more structured and larger company. Once again, the focus was around how the companies would benefit. They did not want to generally contribute to research without a well-thought-out set of benefits that they would receive in the process. Additionally (again reinforcing the importance of the interpersonal relationships) being able to show expertise in the specific area of cyber security (i.e. being able to ‘talk shop’) had a definite advantage in terms of showing trustworthiness and the relevance of the RI activities the companies were being asked to participate in. Once the companies had taken up the offer, and the initial workshops were set up, some interesting issues around informed consent forms emerged. Discussing confidential business information is relatively taboo in cyber security as these companies are by nature generally quite secretive. It was necessary, therefore, to reinforce the initial trust that had been established through e.g. the use of appropriate consent forms, signing non-disclosure agreements, and other mechanisms. The informed consent procedures followed a fairly standard approach that is typical for university-led research – ensuring that participants understand what the research is about, what information will be taken, how the information can be used, and how they can withdraw from the study. For the workshops, the written work the participants developed and the discussions that were recorded (video or audio) were the main pieces of information taken from the experience. Usually, for this sort of research, these procedures are easy to gain ethical approval for. This project was no different, and ethical approval was gained from the De Montfort University Ethics Review Board for the Faculty of Technology. However, the companies participating in the workshops, often with their legal advisors present, had difficulty with the (UK academic standard) consent documents. One company had issues with the representativeness of the discussion – with the employees in question being subject to non-disclosure agreements about company procedures and otherwise not speak for the company. Related issues were: How could they engage in this sort of research where they are being asked to discuss company approaches to responsible innovation? Were they speaking personally, or representing the company? After the CEO reassured the employees that they would not be breaking their contracts to discuss anything he or she was open to, the workshop continued. Another company asked that the researchers should also sign non- disclosure agreements about the specific company processes and procedures that might be discussed although all of these conditions were covered by the informed 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 77 consent form and research ethics approval underpinning the research. Clearly the companies felt they needed an added level of security for their intellectual property. There was, furthermore, a seeming parallel between the lack of understanding of how university research projects function and the relationships between cyber security companies and end-users or clients (as previously discussed) who often don’t understand how the cyber security technologies work. With trust in the research process having been reaffirmed, the companies were prepared to trust the researchers with significant amounts of useful information to further understand the opportunities, challenges, costs and barriers to implementing RI practices in their businesses. This allowed unparalleled access to their processes and gave emphasis to the need for trust in the research process. Having succeeded in establishing such trust with four cyber security SMEs, a total of eight workshops took place. 6.4 Discussion The lessons from the approaches discussed above are important in the context of recruiting and engaging with companies for academic research around RI. These may be generalisable and any recruitment strategy could adapt these lessons to their own specific industry sector. The lessons are illustrated in Table 6.1 and discussed below. Table 6.1 Summary of findings Coming down from the ivory tower Standard ethical approaches may not be recognised The need for expertise Tailoring Failure of standard academic approaches One-to-one instead of one-tomany 78 6.4.1 C. Flick et al. The Importance of Coming Down from the Ivory Tower One of the key lessons was that small companies in particular do not often have the resources to engage with research if it involves them coming to the researchers. More importantly, in going to the activities that the companies themselves initiated, a signal was sent that the researchers a) understood both their space, and that they had these activities in the first place; and b) were happy to engage on their terms (including accommodating and facilitating discussion on topics of particular concern to them). This helped to establish the element of trust whereby the companies would ‘host’ (food and refreshment and meeting venue) as well as engage with the researchers on a reciprocal basis. 6.4.2 Standard Ethical Approaches May Not Be Recognised One of the more surprising lessons was the pointer to how much academic researchers may trust in ethical procedures and research ethics committee approvals granted for these sorts of activities. The fact that some of the companies required additional layers of protection for their intellectual property and procedural approaches was particularly interesting considering that they were, in fact, covered by the ethical approval processes. Is this a sign that there is little trust propensity for scientific research ethics processes outside of academia? Or is it more indicative of the particularly secretive natures of cyber security companies? No other sector companies engaged in our project had issues with the consent documentation, but perhaps this is because those other sectors addressed in the project (biomedicine and nanotechnology) are more closely aligned with traditional academic scientific research, where there is familiarity with and trust in these procedures. It is important that this issue is considered by researchers when engaging with companies, and certainly those in the cyber security sector. It follows that the ability of companies to sign non-disclosure agreements that cover the same conditions as more standard academic consent procedures should be discussed with university legal services and ethics review committees, and legal teams within companies given time to investigate them. Additionally, fall-back options should be considered. For one company, for instance, workshops were only recorded audio, as video recording was considered too invasive. 6.4.3 The Need for Expertise in the Target Area Throughout this whole procedure, the need for the researchers to ‘prove themselves’ as experts with reasonable knowledge in the specific sector area, and not just in applied ethics/responsible innovation was clearly important. A significant 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 79 u nderstanding of technical issues was definitely advantageous when working with the companies. Being able to tailor questions to help each company delve into the ethical questions surrounding their specific lines of work was very helpful to get detailed, in depth, responses. Cyber security is a widely varied sector, and with expertise of many of the different areas it is clearly easier for the researcher to establish trustworthiness, and more likely that the company will have a trust propensity for the researchers. Indeed, the company’s understanding must be that the researchers will understand some of the complexities of the sector and their business and, therefore, be able to use the research outcomes effectively. Similarly, the use of “known experts” as part of the pitch, particularly those in cyber security companies’ areas of interest, including the in-house expertise of cyber security researchers at the university, the local police, business support organisations, and others, improved the credentials of the research team, showing that we were engaged with other organisations and businesses outside of the university. 6.4.4 Tailoring Is Advantageous Expertise in the subject area can also help to fulfil another requirement, that of tailoring the discussions to the specific company. The cyber security workshops were characterised by co-creation activity by peers and were conducted with several members of the same company. This allowed for tailoring of the information provided to the company, rather than a more generic approach. Such tailoring requires more understanding of the company involved, and expertise on the part of the researchers to be able to analyse and report back on the results. By following this approach, the results from the cyber security workshops allowed a richer set of outcomes than those which arose from the ‘collective’ approach to workshops that were undertaken for the biomedicine and nanotechnology sectors elsewhere in the project. 6.4.5 The Failure of Standard Academic Approaches Standard academic approaches for research recruitment generally include calls for participation via email lists, or newsletters, or other methods that are often picked up by multipliers. These kinds of ‘passive consumption’ requests for engagement were largely unsuccessful in this study. Unlike with academic calls for papers or similar, these kinds of activities are not part of the day-to-day business of cyber security companies, which may explain why such calls were regularly ignored. Other standard academic approaches to potential participants, such as offering to pay for travel and accommodation, food, etc., also did not work. This may be explained by the fact that many of the SMEs engaged with were time-poor, with several potential participants dropping-out of the process due to lack of time or the 80 C. Flick et al. inability to agree a mutually convenient time. Clearly to contribute a day or two of their time is overly burdensome for many SMEs, even with financial compensation. The fact that the researchers were willing to travel to the companies was well- received by the companies involved, as was the reduction of the time investment required. 6.4.6 One-to-One Instead of One-to-Many Finally, the advantages of sending personalised, follow-up emails after a personal introduction or meeting at a networking or talk event are significant. As has been noted, the original approaches of sending information to potentially interested parties via multipliers (e.g. the university’s cyber security network, the UK cyber security forum, and larger multipliers such as more general business networks) were largely unsuccessful. Large-scale advertising allows for relative anonymity and, it is suggested, can lead to a lack of response. Ignoring personal emails after initial connections are made is much less socially acceptable and, even when invitations are declined, these refusals can offer useful insights into the reasons (e.g. time constraints, concerns about confidentiality). Additionally, recommendations from boards of trustees/advisory boards for their companies to participate, as well as their having knowledge that other well-respected companies are participating, helps increase the predisposition to take part. The trust companies have in advice from these boards also contributes to the overall trust propensity of the cyber security practitioners in the researchers themselves. 6.5 Conclusion and Recommendations These experiences describe how hard it sometimes is to recruit companies to work with RI research projects. Often there are conflicting ideas of roles, benefits, what is required, and what outputs are created. In moving from an academic sphere to a business sphere, going into their world and becoming involved in their events, approaches, and ultimately understanding their positions, it was possible to recruit companies who not only initially engaged, but over time became longer-term partners with the project, offering to go above and beyond the minimal engagement requirements. These interactions point to a high level of trust between the researchers and the companies: not just that the researchers were trusted, but that they were trustworthy. This reflects, it is considered, the desire that the companies have to, themselves, be seen as trustworthy beyond the cyber security sector: pointing to such “trust” being a key reason for engaging with the activities during the work- 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 81 shops. This validates the usefulness of locating a key value that the industry is likely to engage with, and that aligns with RI principles and practices, in order to use it as a method for engagement. Overall, the following approaches worked best for engaging with cyber security companies about RI: • Removing the academic terminology of “responsible research and innovation” as a concept in its own right, and talking about its constituent parts using industry language. Hence, the use of the simpler term “responsible innovation”, acknowledging the overlap with the more familiar concept of corporate social responsibility. • Being clear about the benefits of ethical approaches in commerce. • Making a positive effort to understand the commercial context within which SMEs operate (i.e. through engaging in or speaking at their events), rather than expecting them to come into the academic world. • Engaging with external advisory organisations to boost credentials and trustworthiness. • Extending academic knowledge around responsible innovation and ethics in order to understand the key technical and commercial dilemmas, challenges and opportunities that confront companies in the sector in question. • Minimising the requirements for companies to participate (e.g. time, travel, etc.). • Being positioned to assist with any particular ethical dilemmas or issues faced by the companies. • Engaging in personalised and often face-to-face discussions with key members of the companies in order to demonstrate understanding, and to establish a rapport conducive to outcomes within workshops. Some of these approaches may be more specific to cyber security companies, but there are wider lessons for other sectors. Perhaps most notable (and generalisable) is the importance of understanding the sector in question and its commercial context in order to engage with the staff, often at a senior level, of SMEs. This positions the researcher more clearly as an equal in the search for insights and truths that the workshops can reveal. Linked with this is the need not to offer RI as a model or blueprint, but rather to demonstrate knowledge of the sector; personalise and tailor information to the specific company; and to focus on those components of RI which are already recognised by the company. Image Credits Tower by iconcheese from the Noun Project contract by Templet from the Noun Project consulting by Vectors Market from the Noun Project Tailor by Pham Duy Phuong Hung from the Noun Project Recruitment by Massupa Kaewgahya from the Noun Project Conversation by Olivia from the Noun Project 82 C. Flick et al. Appendix Dear ______, As a security company, you’re probably very concerned about ethics, and ensuring your business acts as responsibly as possible. What we want to do is to help your company be even more ethical in your business practices. We want to be pragmatic, useful, and responsive to your company’s needs and goals. Much like the security sector sells an idea – that security needs to be built in from the beginning – we will convince you that if you build in responsible and ethical practice from the beginning, you’ll benefit from it in the medium-long term through: • • • • • better relationships with clients; broader and more sensitive outreach and sales approaches; higher levels of client trust in your company; a more embedded community presence; and an agility for future challenges and opportunities. We will work directly and confidentially with you and your company, identifying your areas of good practice and injecting good practice identified by interviews with practitioners, CEOs, and developers of other tech companies. We have successfully done this with the health technology sector in the past, and now we want to open up our methods to the security sector. We want your company to be prepared for what the future might bring – 2, 5, even 10 years down the line, and help you to put good practice in place to be able to deal with these challenges and opportunities. You’ll also learn how to use our techniques to help potential clients think about their own futures – and how security can benefit them. We’ll need around 5 h of your time total, spread over 2 face-to-face meetings where we come to you, and a couple of short follow-up phone calls/emails after each meeting. In between, we will integrate expert opinion from our research for the COMPASS project, the East Midlands Police, academic security researchers, business support organisations such as B Labs and EBN Innovation Network, and professional organisations to help you look above and beyond your everyday practice. You’ll get a tailored, future-looking roadmap to practically implement responsible and ethical practice in your company, so you can benefit from being more trustworthy, learn from our methods, and end up with a more agile, future-looking company that can be relied on by customers and the public to behave ethically and responsibly. For more information please contact … Sincerely, 6 Engaging Small and Medium-Sized Enterprises in Responsible Innovation 83 References Department for Business, Energy & Industrial Strategy. (2017). UK innovation survey 2017: headline findings [WWW Document]. GOV.UK. https://www.gov.uk/government/statistics/ uk-innovation-survey-2017-headline-findings. Accessed 11.8.18. European Commission. (2011). A renewed EU strategy 2011–14 for corporate social responsibility’, Communication from the commission to the European Parliament, the council, the European economic and social committee and the Committee of the Regions COM/2011/0681 final. European Commission, Directorate-General for Research and Innovation. (2013). Options for strengthening responsible research and innovation. Luxembourg: EUR-OP. Hughes, M. (2015). Full or responsible disclosure: How security vulnerabilities are disclosed [WWW document]. MakeUseOf. http://www.makeuseof.com/tag/responsible-disclosuresecurity-vulnerabilities/. Accessed 13.12.17. Lefkowitz, J. (2017). Responsible disclosure – Critical for security. Critical for Intelligence \| SecurityWeek.Com [WWW Document]. URL http://www.securityweek.com/responsible-disclosure-critical-security-critical-intelligence. Accessed 13.12.17. Owen, R., Stilgoe, J., Macnaghten, P., Gorman, M., Fisher, E., & Guston, D. (2013). A framework for responsible innovation. In R. Owen, J. Bessant, & M. Heintz (Eds.), Responsible innovation (pp. 27–50). Chichester: Wiley. https://doi.org/10.1002/9781118551424.ch2. Stahl, B., Flick, C., Mantovani, E., Borsella, E., Porcari, A., Barnett, S. J., Yaghil, A., Ladikas, M., Hahn, J., Obach, M., Garzo, A., Schroeder, D., Chatfield, K., Antoniou, J., Paspallis, N., Brem, A., Yaghmaei, E., Brey, P., Søraker, J. H., Gauttier, S., Gurzawska, A., Ikonen, V., Leikas, J., & Mäkinen, M. (2017). Benefits of responsible research and innovation in ICT for an ageing society. Responsible-Industry Project. https://doi.org/10.5281/zenodo.1050357. Von Schomberg, R. (2013). A vision of responsible research and innovation. In R. Owen, M. Heintz, & J. Bessant (Eds.), Responsible innovation. Managing the responsible emergence of science and innovation in society (pp. 51–74). Wiley. Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
https://openalex.org/W1968749867	https://hal.inrae.fr/hal-02674834/document	French	null	Enhanced intestinal motor response to cholecystokinin in post‐<i>Nippostrongylus brasiliensis</i>‐infected rats: modulation by CCK receptors and the vagus nerve	Neurogastroenterology & motility/Neurogastroenterology and motility	2,001	cc-by-sa	143	To cite this version: Jérôme Gay-Quéheillard, Jean Fioramonti, Rafael Garcia Villar, Lionel Bueno. Enhanced intestinal motor response to cholecystokinin in post-Nippostrongylus brasiliensis infected rats : modulation by CCK receptors and the vagus nerve. Journal of Neurogastroenterology and Motility, 2001, 13, pp.155- 162. hal-02674834 HAL Id: hal-02674834 https://hal.inrae.fr/hal-02674834v1 Submitted on 31 May 2020 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution - ShareAlike 4.0 International License
https://openalex.org/W4249770551	https://hess.copernicus.org/preprints/hess-2021-466/hess-2021-466.pdf	English	null	Comment on hess-2021-466	null	2,021	cc-by	19,809	Flood frequency analysis using mean daily flows vs. instantaneous peak flows Anne Bartens1, Uwe Haberlandt1 1Institute of Hydrology and Water Resources Management, Leibniz University of Hannover, Germany Correspondence to: Anne Bartens (fangmann@iww.uni-hannover.de) 5 Abstract. In many cases flood frequency analysis needs to be carried out on mean daily flow (MDF) series without any available information on the instantaneous peak flow (IPF). We analyze the error of using MDFs instead of IPFs for flood quantile estimation on a German dataset and assess spatial patterns and factors that influence the deviation of MDF floods from their IPF counterparts. The main dependence could be found for catchment area but also gauge elevation appeared to have some influence. Based on the findings we propose simple linear models to correct both MDF flood peaks of individual 10 flood events and overall MDF flood statistics. Key predictor in the models is the event-based ratio of flood peak and flood volume obtained directly from the daily flow records. This correction approach requires a minimum of data input, is easily applied, valid for the entire study area and successfully estimates IPF peaks and flood statistics. The models perform particularly well in smaller catchments, where other IPF estimation methods fall short. Still, the limit of the approach is reached for catchment sizes below 100 km², where the hydrograph information from the daily series is no longer capable of 15 approximating instantaneous flood dynamics. have some influence. Based on the findings we propose simple linear models to correct both MDF flood peaks of individual 10 flood events and overall MDF flood statistics. Key predictor in the models is the event-based ratio of flood peak and flood volume obtained directly from the daily flow records. This correction approach requires a minimum of data input, is easily applied, valid for the entire study area and successfully estimates IPF peaks and flood statistics. The models perform particularly well in smaller catchments, where other IPF estimation methods fall short. Still, the limit of the approach is reached for catchment sizes below 100 km², where the hydrograph information from the daily series is no longer capable of 15 approximating instantaneous flood dynamics. for catchment sizes below 100 km², where the hydrograph information from the daily series is no longer capable of 15 approximating instantaneous flood dynamics. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. 1 Introduction Common flood frequency analysis (FFA) is based on samples of maximum flows. The dimensions and variability of these maxima pose the baseline for the choice of probability distribution, the estimation of its parameters and eventually the deduction of flood quantiles as design criteria. For FFA to be as accurate as possible, it is important to have a large number of 20 peak flows measured with high precision, so that flood magnitude and dynamics are well assessable. 20 peak flows measured with high precision, so that flood magnitude and dynamics are well assessable. However, embracing the true dimension of a peak requires continuous measurement of the flow on a high temporal resolution. Such data is rarely available and oftentimes FFA needs to be carried out on average daily flow records instead. The daily averaging naturally flattens the flood peak and the true maximum becomes unknowable. The degree of this smoothing, i.e. the difference between the true instantaneous peak flow (IPF) and the maximum mean daily flow (MDF) depends on the response 25 time of a system, which is controlled by a multitude of factors. The average relationship between MDF and IPF peaks at a site depends greatly on its basin area (Fuller, 1914) and characteristics related to topography, like altitude, relief and channel slope (Canuti and Moisello 1982) The internal variability of the MDF-IPF ratio within a site’s flow record is largely determined by However, embracing the true dimension of a peak requires continuous measurement of the flow on a high temporal resolution. Such data is rarely available and oftentimes FFA needs to be carried out on average daily flow records instead. The daily averaging naturally flattens the flood peak and the true maximum becomes unknowable. The degree of this smoothing, i.e. the difference between the true instantaneous peak flow (IPF) and the maximum mean daily flow (MDF) depends on the response 25 time of a system, which is controlled by a multitude of factors. The average relationship between MDF and IPF peaks at a site depends greatly on its basin area (Fuller, 1914) and characteristics related to topography, like altitude, relief and channel slope (Canuti and Moisello, 1982). The internal variability of the MDF-IPF ratio within a site’s flow record is largely determined by the type of meteorological input causing the individual flood events (Viglione and Blöschl, 2009; Gaál et al. 2013). 1 Introduction They also 35 propose their own method based on the rising and falling slopes of the event hydrograph, estimated from the three consecutive days around the peak. They found that their slope-based method and Fill and Steiner’s method perform well and are probably applicable under a wide range of climates. However, both methods’ performances decrease with decreasing catchment size and work best for areas larger than 500 km². day. Chen et al. (2017) compare two of these methods, namely those of Sangal (1983) and Fill and Steiner (2003). They also 35 propose their own method based on the rising and falling slopes of the event hydrograph, estimated from the three consecutive days around the peak. They found that their slope-based method and Fill and Steiner’s method perform well and are probably applicable under a wide range of climates. However, both methods’ performances decrease with decreasing catchment size and work best for areas larger than 500 km². There naturally exist more complex means to correct the divergence between MDFs and IPFs. This includes disaggregation of 40 the daily flow series to a finer scale, as done by e.g. Stedinger and Vogel (1984), Tarboton et al. (1998), Kumar et al. (2000), Tan et al. (2007) and Acharya and Ryu (2014). Also, hydrological modelling may be applied for IPF estimation, e.g. in combination with high-resolution disaggregated rainfall (Ding et al., 2016) and regionalized model parameters (Ding and Haberlandt, 2017). These methods, however, require a variety of computational steps and/or data sources. There naturally exist more complex means to correct the divergence between MDFs and IPFs. This includes disaggregation of 40 the daily flow series to a finer scale, as done by e.g. Stedinger and Vogel (1984), Tarboton et al. (1998), Kumar et al. (2000), Tan et al. (2007) and Acharya and Ryu (2014). Also, hydrological modelling may be applied for IPF estimation, e.g. in combination with high-resolution disaggregated rainfall (Ding et al., 2016) and regionalized model parameters (Ding and Haberlandt, 2017). These methods, however, require a variety of computational steps and/or data sources. This study aims at analyzing the differences between IPF and MDF with focus on flood frequency. The errors in mean 45 maximum flows, distribution parameters and flood quantiles are assessed and analyzed for spatial patterns. 1 Introduction The p/V of individual events can describe the internal variability at a site by reflecting different types of floods caused by different rainfall and/or snowmelt inputs. At the same time the p/V accounts for the variability between sites caused by local flood generating processes governed by general physiographic 55 conditions. Accordingly, the proposed method is tested for IPF estimation for individual events, which are then used for FFA, and for direct correction of site-specific distribution parameters and flood quantiles. shape of a flood event, which in turn gives an idea about the expected instantaneous peak: the larger the daily peak and the 50 smaller the event volume, the larger the expected difference between IPF and MDF and vice versa. We assume that the peak- volume ratio (p/V) holds important information on the general behavior of flood events (Tan et al., 2006; Gaál et al. 2015; Fischer, 2018), and thus the expected magnitude of the IPF. The p/V of individual events can describe the internal variability at a site by reflecting different types of floods caused by different rainfall and/or snowmelt inputs. At the same time the p/V accounts for the variability between sites caused by local flood generating processes governed by general physiographic 55 conditions. Accordingly, the proposed method is tested for IPF estimation for individual events, which are then used for FFA, and for direct correction of site-specific distribution parameters and flood quantiles. In a first step, the general differences between IPF and MDF statistics are analyzed. An error is computed as percentage 60 deviation of the MDF statistic MDFstat from the IPF statistic IPFstat 1 Introduction A variety difference between the true instantaneous peak flow (IPF) and the maximum mean daily flow (MDF) depends on the response 25 time of a system, which is controlled by a multitude of factors. The average relationship between MDF and IPF peaks at a site depends greatly on its basin area (Fuller, 1914) and characteristics related to topography, like altitude, relief and channel slope (Canuti and Moisello, 1982). The internal variability of the MDF-IPF ratio within a site’s flow record is largely determined by the type of meteorological input causing the individual flood events (Viglione and Blöschl, 2009; Gaál et al. 2013). A variety 1 of studies make use of the dependencies named above in order to estimate IPFs from MDFs including Taguas et al (2008) https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. of studies make use of the dependencies named above in order to estimate IPFs from MDFs, including Taguas et al. (2008), 30 Muñoz et al. (2012) and Ding et al. (2015). Other IPF estimation methods aim at using the bare minimum of available data, i.e. solely the available daily flow record. In these cases, usually the shape of hydrographs are used to estimate the instantaneous peaks of events (e.g. Langbein, 1944; Ellis and Gray, 1966). Several approaches make use of the maximum daily flow and the discharge of the previous and successive these cases, usually the shape of hydrographs are used to estimate the instantaneous peaks of events (e.g. Langbein, 1944; Ellis and Gray, 1966). Several approaches make use of the maximum daily flow and the discharge of the previous and successive day. Chen et al. (2017) compare two of these methods, namely those of Sangal (1983) and Fill and Steiner (2003). They also 35 propose their own method based on the rising and falling slopes of the event hydrograph, estimated from the three consecutive days around the peak. They found that their slope-based method and Fill and Steiner’s method perform well and are probably applicable under a wide range of climates. However, both methods’ performances decrease with decreasing catchment size and work best for areas larger than 500 km². day. Chen et al. (2017) compare two of these methods, namely those of Sangal (1983) and Fill and Steiner (2003). 1 Introduction Based on the findings, a method is proposed that facilitates IPF estimation using a combination of daily event hydrographs and functional dependencies with geomorphic catchment descriptors, while keeping the data input to a minimum. Key predictor in this approach is the ratio of direct event peak runoff and direct event volume. This ratio is expected to effectually describe the This study aims at analyzing the differences between IPF and MDF with focus on flood frequency. The errors in mean 45 maximum flows, distribution parameters and flood quantiles are assessed and analyzed for spatial patterns. Based on the findings, a method is proposed that facilitates IPF estimation using a combination of daily event hydrographs and functional dependencies with geomorphic catchment descriptors, while keeping the data input to a minimum. Key predictor in this approach is the ratio of direct event peak runoff and direct event volume. This ratio is expected to effectually describe the shape of a flood event, which in turn gives an idea about the expected instantaneous peak: the larger the daily peak and the 50 smaller the event volume, the larger the expected difference between IPF and MDF and vice versa. We assume that the peak- volume ratio (p/V) holds important information on the general behavior of flood events (Tan et al., 2006; Gaál et al. 2015; Fischer, 2018), and thus the expected magnitude of the IPF. The p/V of individual events can describe the internal variability at a site by reflecting different types of floods caused by different rainfall and/or snowmelt inputs. At the same time the p/V accounts for the variability between sites caused by local flood generating processes governed by general physiographic 55 conditions. Accordingly, the proposed method is tested for IPF estimation for individual events, which are then used for FFA, and for direct correction of site-specific distribution parameters and flood quantiles. shape of a flood event, which in turn gives an idea about the expected instantaneous peak: the larger the daily peak and the 50 smaller the event volume, the larger the expected difference between IPF and MDF and vice versa. We assume that the peak- volume ratio (p/V) holds important information on the general behavior of flood events (Tan et al., 2006; Gaál et al. 2015; Fischer, 2018), and thus the expected magnitude of the IPF. 2.1 Analysis and estimation of IPF peaks The combination of hydrograph shape and catchment characteristics as predictors is expected to better reproduce both the at-site and between-site variability in the 75 IPF MDF l ti hi d i ld i l d l where CD denotes additional catchment descriptors that may be included in the models. The combination of hydrograph shape where CD denotes additional catchment descriptors that may be included in the models. The combination of hydrograph shape and catchment characteristics as predictors is expected to better reproduce both the at-site and between-site variability in the 75 IPF-MDF relationship and yield a more universal model. The event correction method will be compared with the slope correction method developed by Chen et al. (2017). This method and catchment characteristics as predictors is expected to better reproduce both the at-site and between-site variability in the 75 IPF-MDF relationship and yield a more universal model. The event correction method will be compared with the slope correction method developed by Chen et al. (2017). This method estimates an instantaneous event peak flow based on the slopes of the daily peak Qpeak to its preceding and following daily flows Qpre and Qsuc. The IPF is thus estimated as The event correction method will be compared with the slope correction method developed by Chen et al. (2017). This method estimates an instantaneous event peak flow based on the slopes of the daily peak Qpeak to its preceding and following daily flows Qpre and Qsuc. The IPF is thus estimated as IPFevent = Qpeak + (Qpeak−Qpre)∗(Qpeak−Qsuc) 2∗Qpeak−Qpre−Qsuc . (4) 80 Qpeak−Qpre)∗(Qpeak−Qsuc) 2∗Qpeak−Qpre−Qsuc . (4) IPFevent = Qpeak + (Qpeak−Qpre)∗(Qpeak−Qsuc) 2∗Qpeak−Qpre−Qsuc . 80 (4) For validation, both methods are applied to estimate IPFs for all separated events in the daily flow series. The maximum flows needed for FFA are then taken from this corrected event series rather than from corrected maximum daily peaks. This procedure is assumed to be more accurate, since maxima in IPF and MDF do not necessarily overlap. More precisely, events with maximum instantaneous peaks can have rather inconsiderable daily peaks in some instances. Correcting only the maximum For validation, both methods are applied to estimate IPFs for all separated events in the daily flow series. The maximum flows needed for FFA are then taken from this corrected event series rather than from corrected maximum daily peaks. 2.1 Analysis and estimation of IPF peaks In a first step, the general differences between IPF and MDF statistics are analyzed. An error is computed as percentage 60 deviation of the MDF statistic MDFstat from the IPF statistic IPFstat 2 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. (1) This error is computed at each station for any desired quantity stat, like the mean annual maximum flow (MHQ), L-moments, distribution parameters and flood quantiles. 65 This error is computed at each station for any desired quantity stat, like the mean annual maximum flow (MHQ), L-moments, distribution parameters and flood quantiles. This error is computed at each station for any desired quantity stat, like the mean annual maximum flow (MHQ), L-moments, distribution parameters and flood quantiles. 65 65 In order to improve the IPF estimation by MDF, several correction methods are applied, which make use of the peak-volume ratio. This ratio is computed for events in the average daily time series after baseflow separation using the direct peak flow Qdir and the direct flood volume Voldir p/V [ 1 𝑑] = Qdir [m3d−1] Voldir [m³] . (2) p/V [ 1 𝑑] = Qdir [m3d−1] Voldir [m³] . (2) The first IPF estimation method aims at correcting individual events. For calibration, all events are identified that contain a 70 monthly maximum instantaneous peak. For these events the daily and instantaneous peaks, as well as the daily p/Vs are computed. Then a linear regression model of the following form is fitted IPFevent = MDFevent ∗(𝑎+ 𝑏1 ∗p/Vevent + 𝑏2 ∗CD1 + ⋯+ 𝑏𝑛+1 ∗CD𝑛), (3) The first IPF estimation method aims at correcting individual events. For calibration, all events are identified that contain a 70 monthly maximum instantaneous peak. For these events the daily and instantaneous peaks, as well as the daily p/Vs are computed. Then a linear regression model of the following form is fitted ent = MDFevent ∗(𝑎+ 𝑏1 ∗p/Vevent + 𝑏2 ∗CD1 + ⋯+ 𝑏𝑛+1 ∗CD𝑛), ∗(𝑎+ 𝑏1 ∗p/Vevent + 𝑏2 ∗CD1 + ⋯+ 𝑏𝑛+1 ∗CD𝑛), (3) (3) where CD denotes additional catchment descriptors that may be included in the models. where CD denotes additional catchment descriptors that may be included in the models. The combination of hydrograph shape where CD denotes additional catchment descriptors that may be included in the models. 2.2 Event separation 100 For separation of the flood events, the initial steps of the procedure used by Tarasova et al. (2018) are carried out, which has proven effective and convenient for their German dataset. For the initial step of baseflow separation they selected the simple nonparametric algorithm by the Institute of Hydrology (1980). This method is applied here with the same settings, i.e. 5-day non-overlapping blocks are used to find minima that are identified as turning points if they are more than 1.1 times smaller than their neighboring minima. The baseflow is then derived by linear interpolation between the turning points. Discharge 105 peaks are subsequently determined from the flow series and for every peak the start and end of the belonging flow event is defined by the nearest surrounding turning points. In order to prevent false identification of events due to natural variability, events are discarded if their peak discharge is not at least 10% larger than the baseflow. The final step of re-defining events with multiple peaks is not carried out as it requires rainfall and snowmelt information than their neighboring minima. The baseflow is then derived by linear interpolation between the turning points. Discharge 105 peaks are subsequently determined from the flow series and for every peak the start and end of the belonging flow event is defined by the nearest surrounding turning points. In order to prevent false identification of events due to natural variability, events are discarded if their peak discharge is not at least 10% larger than the baseflow. The final step of re-defining events with multiple peaks is not carried out, as it requires rainfall and snowmelt information, which are not available in our case. It is assumed that the majority of events, especially the larger ones relevant for FFA, are 110 separated correctly. which are not available in our case. It is assumed that the majority of events, especially the larger ones relevant for FFA, are 110 separated correctly. which are not available in our case. It is assumed that the majority of events, especially the larger ones relevant for FFA, are 110 separated correctly. 2.1 Analysis and estimation of IPF peaks The p/Vmean in itself is expected to be a good predictor that reflects local conditions like spatial scale, climate, geology and other external factors that control flow variability obtainable from daily flow records. The additional inclusion of catchment descriptors is tested case by case and may contribute to the reproduction of spatial variability. The model is expected to represent the average conditions that determine the average deviation of MDF from IPF estimates. The p/Vmean in itself is expected to be a good predictor that reflects local conditions like spatial scale, climate, geology and other external factors that control flow variability obtainable from daily flow records. The additional inclusion of catchment descriptors is tested case by case and may contribute to the reproduction of spatial variability. 2.1 Analysis and estimation of IPF peaks This procedure is assumed to be more accurate, since maxima in IPF and MDF do not necessarily overlap. More precisely, events with maximum instantaneous peaks can have rather inconsiderable daily peaks in some instances. Correcting only the maximum MDFs would lead to underestimation of the IPFs in these cases 85 3 MDFs would lead to underestimation of the IPFs in these cases. 85 Since the p/V method in equation (3) can only be calibrated on events for which monthly maxima exist but is eventually applied to all events, including very small and potentially improperly separated ones, unrealistic IPF estimates may be created that adversely affect the subsequent FFA. In order to avoid this problem and to be able to estimate IPF statistics directly from 3 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. daily records, a second type of IPF estimation methods is analyzed. These involve the estimation of flood statistics, i.e. mean annual and seasonal maximum flows, sample L-moments, estimated distribution parameters and derived flood quantiles based 90 on averaged p/Vs. These averages are obtained from all annual maximum MDF events at each station. These major events are expected to be properly separated by the algorithm and to be most influential in FFA. Although the maximum MDF events may not necessarily be identical to the maximum IPF events, as discussed before, this approach appears to be the most sensible here. The model set up is analogous to the event correction approach IPFstat = MDFstat ∗(𝑎+ 𝑏1 ∗p/Vmean + 𝑏2 ∗CD1 + ⋯+ 𝑏𝑛+1 ∗CD𝑛). (5) 95 IPFstat = MDFstat ∗(𝑎+ 𝑏1 ∗p/Vmean + 𝑏2 ∗CD1 + ⋯+ 𝑏𝑛+1 ∗CD𝑛). (5) 95 (5) The model is expected to represent the average conditions that determine the average deviation of MDF from IPF estimates. The p/Vmean in itself is expected to be a good predictor that reflects local conditions like spatial scale, climate, geology and other external factors that control flow variability obtainable from daily flow records. The additional inclusion of catchment descriptors is tested case by case and may contribute to the reproduction of spatial variability. The model is expected to represent the average conditions that determine the average deviation of MDF from IPF estimates. 2.3 Distribution fitting and flood quantile estimation (7) This approach allows the combined estimation of flood quantiles from multiple underly This approach allows the combined estimation of flood quantiles from multiple underlying distributions and thus the assessment of errors in seasonal FFA. The approach is described in detail in Fischer et al. 2016. Other than in their study, we 125 This approach allows the combined estimation of flood quantiles from multiple underlying distributions and thus the This approach allows the combined estimation of flood quantiles from multiple underlying distributions and thus the assessment of errors in seasonal FFA. The approach is described in detail in Fischer et al. 2016. Other than in their study, we 125 do not censor our data with thresholds, i.e. for matters of simplicity we assume that every seasonal maximum is indeed a flood event. assessment of errors in seasonal FFA. The approach is described in detail in Fischer et al. 2016. Other than in their study, we 125 do not censor our data with thresholds, i.e. for matters of simplicity we assume that every seasonal maximum is indeed a flood event. 2.4 Uncertainty Since both distribution fitting and IPF estimation via linear models are approximations an Since both distribution fitting and IPF estimation via linear models are approximations and not fully accurate, we eventually assess the overall level of uncertainty in the final IPF flood quantile estimates. This is done using simple bootstrapping 130 procedures. In a first step, the series of annual maxima from both daily and monthly maximum data are analogously resampled 1000 times with replacement. For each resampling the desired flood quantiles are estimated using L-moments. The range of these estimates provides the baseline level of uncertainty due to distribution fitting. Since both distribution fitting and IPF estimation via linear models are approximations and not fully accurate, we eventually assess the overall level of uncertainty in the final IPF flood quantile estimates. This is done using simple bootstrapping 130 procedures. In a first step, the series of annual maxima from both daily and monthly maximum data are analogously resampled 1000 times with replacement. For each resampling the desired flood quantiles are estimated using L-moments. The range of these estimates provides the baseline level of uncertainty due to distribution fitting. In a second step, linear regression models are fitted to each pairing of estimated IPF and MDF flood quantiles over all stations in the study area. In order to assess the uncertainty of the fitted models, another resampling is carried out, this time shuffling 135 the set of considered stations, again 1000 times with replacement. For each station this procedure yields 1000 estimates of paired flood quantiles from both the IPF and MDF series (IPF-bs and MDF-bs), 1000 full-model quantile estimates resulting from the original p/V model fitted to each permutation (p/V-full), and 10001000 quantile estimates resulting from permutation of the p/V model for all IPF and MDF transpositions (p/V-bs-bs). In order to assess the overall level of uncertainty, several indices will be assessed at the individual stations. The first one is the 140 relative width of the 95% confidence intervals (CI) calculated for all aforementioned bootstrap sample estimates of the desired flood quantile In order to assess the overall level of uncertainty, several indices will be assessed at the individual stations. where xbs;0.025 and xbs;0.975 are the 2.5% and 97.5% quantile and xbs;0.5 the median of the respective 2.3 Distribution fitting and flood quantile estimation For extrapolation of the time series and estimation of floods with specific return periods, distributions were fitted to the annual and seasonal samples of both IPF and MDF. This enables the direct comparison of both the higher flood quantiles and of the estimated distribution parameters. Here, the General Extreme Value distribution (GEV) of the following form was used for all 115 samples For extrapolation of the time series and estimation of floods with specific return periods, distributions were fitted to the annual and seasonal samples of both IPF and MDF. This enables the direct comparison of both the higher flood quantiles and of the estimated distribution parameters. Here, the General Extreme Value distribution (GEV) of the following form was used for all 115 For extrapolation of the time series and estimation of floods with specific return periods, distributions were fitted to the annual and seasonal samples of both IPF and MDF. This enables the direct comparison of both the higher flood quantiles and of the estimated distribution parameters. Here, the General Extreme Value distribution (GEV) of the following form was used for all 115 samples 115 𝐹(𝑥) = 𝑒−exp (1 𝑘∗ log(1−𝑘∗𝑥−𝜉 𝛼)) (6) 4 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. with location parameter ξ, scale parameter α and shape parameter k. The parameters were estimated using sample L- moments. The goodness of fit of the distributions was determined with the Cramer-von-Mises test. with location parameter ξ, scale parameter α and shape parameter k. The parameters were estimated using sample L- moments. The goodness of fit of the distributions was determined with the Cramer-von-Mises test. Additionally, for seasonal considerations, mixed models were applied, which combine two or more GEV distributions fitted 120 to different subsamples of the data, like summer and winter floods. A simple maximum mixing approach is used to combine the individual distributions: Additionally, for seasonal considerations, mixed models were applied, which combine two or more GEV distributions fitted 120 to different subsamples of the data, like summer and winter floods. A simple maximum mixing approach is used to combine the individual distributions: 120 𝐹mix(𝑥) = ∏ 𝐹𝑖(𝑥) 𝑛 𝑖=1 . 2.4 Uncertainty The first one is the 140 relative width of the 95% confidence intervals (CI) calculated for all aforementioned bootstrap sample estimates of the desired flood quantile CIb = 𝑥bs;0.975−𝑥bs;0.025 (8) (8) where xbs;0.025 and xbs;0.975 are the 2.5% and 97.5% quantile and xbs;0.5 the median of the respective sample. The second one is the deviation of the individual MDF and p/V-model bootstrap samples from the IPF sample, which allows the assessment of 145 error distributions the deviation of the individual MDF and p/V-model bootstrap samples from the IPF sample, which allows the assessment of 145 error distributions the deviation of the individual MDF and p/V-model bootstrap samples from the IPF sample, which allows the assessment of 145 error distributions 5 (9) (9) From the resulting error vector, a variety of statistics can be computed for comparison. Finally, the agreement of the 95% confidence intervals of the MDF and p/V-model samples with the IPF confidence bands are determined as percentage overlap: overlap = min(𝑥bs;0.975,IPFbs;0.975)−max (𝑥bs;0.025,IPFbs;0.025) max(𝑥bs;0.975,IPFbs;0.975)−min (𝑥bs;0.025,IPFbs;0.025) ∗100%. (10) 150 overlap = min(𝑥bs;0.975,IPFbs;0.975)−max (𝑥bs;0.025,IPFbs;0.025) max(𝑥bs;0.975,IPFbs;0.975)−min (𝑥bs;0.025,IPFbs;0.025) ∗100%. (10) 50 150 (10) 3 Study area and data This study uses data from 653 discharge gauges distributed over Germany. For the analyses, average daily flow and maximum monthly flow are required. The selected stations represent the datasets of the federal agencies, who provide online access to both parameters (Lower Saxony, Saxony-Anhalt, Saxony, Bavaria and Baden-Württemberg). Germany poses a transition zone from an oceanic climate in the northwest to a humid continental climate in the southeast. The 155 northwestern parts are influenced by wet air and have mild winters, while the more southeastern parts are drier and exhibit larger temperature ranges. The average temperature for the entire country is 8.9 °C, the monthly averages ranging between 0.4°C in January and 18°C in July (reference period 1981-2010; DWD). The average precipitation is 819 mm, where amounts generally decrease in west-east direction and in strong dependence on topography. Annual rainfall sums are generally highest Germany poses a transition zone from an oceanic climate in the northwest to a humid continental climate in the southeast. The 155 northwestern parts are influenced by wet air and have mild winters, while the more southeastern parts are drier and exhibit larger temperature ranges. The average temperature for the entire country is 8.9 °C, the monthly averages ranging between 0.4°C in January and 18°C in July (reference period 1981-2010; DWD). The average precipitation is 819 mm, where amounts generally decrease in west-east direction and in strong dependence on topography. Annual rainfall sums are generally highest over the Alps at the very Southern border and the various secondary mountain ranges. The flat continental east is driest. 160 Temporally, the summer months are wettest with rainfall often occurring in convective events. Snowfall occurs between October and April, where amount and depth of snow cover increase with decreasing oceanic influence and increasing altitude. Even though not the entire area of Germany is covered by the available data, the selected gauges provide a cross section through the climatically and topographically distinct regions, from the flat oceanic northwest to the mountainous continental over the Alps at the very Southern border and the various secondary mountain ranges. The flat continental east is driest. 160 Temporally, the summer months are wettest with rainfall often occurring in convective events. Snowfall occurs between October and April, where amount and depth of snow cover increase with decreasing oceanic influence and increasing altitude. 3 Study area and data Even though not the entire area of Germany is covered by the available data, the selected gauges provide a cross section through the climatically and topographically distinct regions, from the flat oceanic northwest to the mountainous continental 6 Figure 1: Location of the 653 gauges used for analysis. The 103 stations used for model calibration are marked in blue. Digital elevation data by Jarvis et al. (2008). https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Figure 1: Location of the 653 gauges used for analysis. The 103 stations used for model calibration are marked in blue. Digital elevation data by Jarvis et al. (2008). The lengths of the discharge records vary substantially from 11 to 183 years with a mean assessment of differences in IPF and MDF floods and final model validation, all 653 stations with their variable record lengths 170 are considered. For assessment of flood frequency criteria only those stations with at least 30 years of observations were used (490). Model fitting was carried out on a subset of 103 gauges, whose discharge series were thoroughly checked. Also, their records were cropped to a common period from 1979 to 2012, in order to eliminate potential non-stationary effects. For the 103 stations used for calibration a catalogue of catchment descriptors is available. For the remaining stations only 170 For the 103 stations used for calibration a catalogue of catchment descriptors is available. For the remaining stations only rudimentary information was obtained, i.e. catchment size, geographical position and altitude of the gauges. 175 For the 103 stations used for calibration a catalogue of catchment descriptors is available. For the remaining stations only rudimentary information was obtained, i.e. catchment size, geographical position and altitude of the gauges. 175 rudimentary information was obtained, i.e. catchment size, geographical position and altitude of the gauges. 175 4.1 Comparison of MDF and IPF peaks In theory, the relative deviation between MDF and IPF peaks depends greatly on catchment size. Small catchments without appreciable buffering capacity react fast to even small rainfall, leading to short and steep flood waves that are hardly reproduced on coarsely averaged time scales. Factors like steep slopes, impermeable underground and short but intense rainfall 180 contribute to the “burstiness” of storm events and make these even less representable through daily flow records. In theory, the relative deviation between MDF and IPF peaks depends greatly on catchment size. Small catchments without appreciable buffering capacity react fast to even small rainfall, leading to short and steep flood waves that are hardly reproduced on coarsely averaged time scales. Factors like steep slopes, impermeable underground and short but intense rainfall 180 contribute to the “burstiness” of storm events and make these even less representable through daily flow records. The effect of the catchment size is clearly visible in the data set. Fig. 2 demonstrates by means of the mean annual maximum flow that the larger the area, the smaller the deviation between MDF and IPF. Also, errors appear to be especially large in In theory, the relative deviation between MDF and IPF peaks depends greatly on catchment size. Small catchments without appreciable buffering capacity react fast to even small rainfall, leading to short and steep flood waves that are hardly appreciable buffering capacity react fast to even small rainfall, leading to short and steep flood waves that are hardly reproduced on coarsely averaged time scales. Factors like steep slopes, impermeable underground and short but intense rainfall 180 contribute to the “burstiness” of storm events and make these even less representable through daily flow records. The effect of the catchment size is clearly visible in the data set. Fig. 2 demonstrates by means of the mean annual maximum flow that the larger the area, the smaller the deviation between MDF and IPF. Also, errors appear to be especially large in reproduced on coarsely averaged time scales. Factors like steep slopes, impermeable underground and short but intense rainfall 180 contribute to the “burstiness” of storm events and make these even less representable through daily flow records. The effect of the catchment size is clearly visible in the data set. Fig. 2 demonstrates by means of the mean annual maximum flow that the larger the area, the smaller the deviation between MDF and IPF. 4.1 Comparison of MDF and IPF peaks Figure 3: Error in the MHQ in relation to catchment size and elevation for the entire year (a) summer (b) and winter (c 95 Error in the MHQ in relation to catchment size and elevation for the entire year (a), summer (b) and w Figure 3: Error in the MHQ in relation to catchment size and elevation for the entire year (a), summer (b) and winter (c). 195 In Fig. 3 the error in the MHQ is shown for the entire year and the summer and winter half year. The relationship with catchment area is clearly visible in all three cases. Also, the effect of the elevation becomes obvious, namely in the lowest elevations (yellow points, below 100 m) showing very small errors, even for small catchment sizes down to approximately 100 km². This is the clearest stratification in the error due to elevation; the errors at higher altitudes appear less distinguishable. Figure 3: Error in the MHQ in relation to catchment size and elevation for the entire year (a), summer (b) and winter (c). 195 In Fig. 3 the error in the MHQ is shown for the entire year and the summer and winter half year. The relationship with catchment area is clearly visible in all three cases. Also, the effect of the elevation becomes obvious, namely in the lowest elevations (yellow points, below 100 m) showing very small errors, even for small catchment sizes down to approximately 100 km². This is the clearest stratification in the error due to elevation; the errors at higher altitudes appear less distinguishable. There is, however, a clear distinction between summer and winter. As expected, the error is overall smaller in the winter 200 months, where snowmelt and stratiform events prevail, while the convective events in summer are poorly reproduced by MDF. The error in the annual peaks is a mixture of the two seasons. Which season contributes mainly to the annual peaks depends on the individual flood regimes. At 69.1% of the considered gauges the winter floods exceed their summer counterparts on average, the remaining 30.9% are dominated by summer floods. These seasonality statistics are established on basis of the IPF. When considering MDF instead, only 23.2% of the gauges are identified as having maximum peaks in summer. 4.1 Comparison of MDF and IPF peaks Also, errors appear to be especially large in 180 The effect of the catchment size is clearly visible in the data set. Fig. 2 demonstrates by means of the mean annual maximum flow that the larger the area, the smaller the deviation between MDF and IPF. Also, errors appear to be especially large in 7 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. higher altitudes. Generally, the error seems to increase in north-south direction, which could be a secondary effect of both increasing altitude and decreasing catchment size. 185 185 Figure 2: Spatial distribution of the MDF error in the MHQ. Figure 2: Spatial distribution of the MDF error in the MHQ. Figure 2: Spatial distribution of the MDF error in the MHQ. When assessing the differences between average daily and instantaneous peaks, it is also meaningful to take a closer look at different types of floods. For our German dataset the two most opposite types are a) flood events induced by short intense rainfall, especially convective events, and b) extended flood events with significant volume, as caused by snowmelt and/or 90 stratiform rain. Presumably, the latter flood type is much better represented by average daily flow than the former. In order to roughly distinguish between the two types, the flow records are divided into summer (May - October) and winter (November - April) half years. rainfall, especially convective events, and b) extended flood events with significant volume, as caused by snowmelt and/or 190 stratiform rain. Presumably, the latter flood type is much better represented by average daily flow than the former. In order to roughly distinguish between the two types, the flow records are divided into summer (May - October) and winter (November 190 - April) half years. 8 Figure 3: Error in the MHQ in relation to catchment size and elevation for the entire year (a), summer (b) and winter (c). 95 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. 4.1 Comparison of MDF and IPF peaks This indicates 205 that the average daily flow smooths significant peaks to a point where they are no longer relevant for the overall flood behavior. Fig. 4 a shows the percentage of annual maxima at each gauge that are attributed to the wrong season. Negative values are falsely attributed to summer, positive values to winter. It is obvious that with decreasing catchment size an increasing number of annual maxima are falsely identified in the winter half year, while the actual instantaneous maxima occur in summer. Apart When considering MDF instead, only 23.2% of the gauges are identified as having maximum peaks in summer. This indicates 205 that the average daily flow smooths significant peaks to a point where they are no longer relevant for the overall flood behavior. Fig. 4 a shows the percentage of annual maxima at each gauge that are attributed to the wrong season. Negative values are falsely attributed to summer, positive values to winter. It is obvious that with decreasing catchment size an increasing number of annual maxima are falsely identified in the winter half year, while the actual instantaneous maxima occur in summer. Apart When considering MDF instead, only 23.2% of the gauges are identified as having maximum peaks in summer. This indicates 205 that the average daily flow smooths significant peaks to a point where they are no longer relevant for the overall flood behavior. Fig. 4 a shows the percentage of annual maxima at each gauge that are attributed to the wrong season. Negative values are falsely attributed to summer, positive values to winter. It is obvious that with decreasing catchment size an increasing number of annual maxima are falsely identified in the winter half year, while the actual instantaneous maxima occur in summer. Apart from not being able to properly identify flood magnitudes when using daily flow series, this is a serious issue for classification 210 of flood regimes, identification of dominating flood types and application of heterogeneous flood frequency analysis when daily data is the only available option. Another general issue highlighted by this analysis, independent of seasonality, is the asynchronous occurrence of IPFs and MDFs. Instantaneous maxima are not always identifiable in the daily flow series, i.e. 4.1 Comparison of MDF and IPF peaks the maxima obtained from the daily series from not being able to properly identify flood magnitudes when using daily flow series, this is a serious issue for classification 210 of flood regimes, identification of dominating flood types and application of heterogeneous flood frequency analysis when daily data is the only available option. Another general issue highlighted by this analysis, independent of seasonality, is the asynchronous occurrence of IPFs and MDFs. Instantaneous maxima are not always identifiable in the daily flow series, i.e. the maxima obtained from the daily series Another general issue highlighted by this analysis, independent of seasonality, is the asynchronous occurrence of IPFs and MDFs. Instantaneous maxima are not always identifiable in the daily flow series, i.e. the maxima obtained from the daily series are inevitably found in other places. In general, the smaller the catchment, the smaller the temporal overlap between 215 instantaneous and daily peaks, as seen in Fig. 4 b. This problem needs to be kept in mind when attempting to estimate are inevitably found in other places. In general, the smaller the catchment, the smaller the temporal overlap between 215 instantaneous and daily peaks, as seen in Fig. 4 b. This problem needs to be kept in mind when attempting to estimate are inevitably found in other places. In general, the smaller the catchment, the smaller the temporal overlap between 215 instantaneous and daily peaks, as seen in Fig. 4 b. This problem needs to be kept in mind when attempting to estimate 9 9 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. instantaneous peaks from daily peaks, since the two may belong to significantly different events and thus to different populations. Figure 4: Percentage of peaks falsely attributed by MDF to the winter or summer half-year (a) and percentage of peaks in MDF and 20 IPF overlapping in time (with a 5-day buffer; b). Figure 4: Percentage of peaks falsely attributed by MDF to the winter or summer half-year (a) and percentage of peaks in MDF and 220 IPF overlapping in time (with a 5-day buffer; b). 4.2 Estimation of mean annual IPF For both correction of the individual events and of the MHQ, linear models appeared appropriate. The fitted models are listed in Table 1. For all models, the p/V is able to explain the majority of the variance in the IPF-MDF relationship and they may well be applied with the p/V as sole regressor. However, including the gauge elevation in some models led to minor 225 improvements of the performances, especially for the two seasons. Gauge location and catchment size, on the other hand, did not prove relevant. It turned out that when correcting the peaks of individual events, the performance could not be significantly improved by considering the seasons separately. For correction of the MHQ, however, individual seasonal models did prove meaningful. It turned out that when correcting the peaks of individual events, the performance could not be significantly improved by considering the seasons separately. For correction of the MHQ, however, individual seasonal models did prove meaningful. This observation indicates that the relationship between average IPF, MDF and p/V differs between seasons. This is attributed 230 to the inability of the daily series to identify instantaneous events, which is more severe in summer than in winter. The individually fitted models tend to correct for this deficiency. Table 1: Linear models fitted for correction of individual events and the MHQ. Type Model Events MDF (0.94 – 0.49 * p/Vevent - 0.000078 * elevation) MHQ Year MHQMDF * (1.11 - 1.27 * p/Vmean) Summer MHQMDF * (1.34 - 1.54 * p/Vmean - 0.00011 * elevation) Winter MHQMDF * (1.34 - 1.63 * p/Vmean - 0.00019 * elevation) 235 Fig. 5 shows the change in mean absolute error in the annual MHQ after correction with the different methods in relation to catchment size and elevation. The slope method (a) applied to the individual events yields a rather constant reduction of the Table 1: Linear models fitted for correction of individual events and the MHQ. Type Model Events MDF * (0.94 – 0.49 * p/Vevent - 0.000078 * elevation) MHQ Year MHQMDF * (1.11 - 1.27 * p/Vmean) Summer MHQMDF * (1.34 - 1.54 * p/Vmean - 0.00011 * elevation) Winter MHQMDF * (1.34 - 1.63 * p/Vmean - 0.00019 * elevation) 5 Table 1: Linear models fitted for correction of individual events and the MHQ. 235 Fig. 4.2 Estimation of mean annual IPF This deterioration appears to affect those stations that have been highlighted earlier, namely the ones with the lowest elevations in the data set. Here the error is significantly lower than the p/Vs suggest. Again, the average p/V method appears more robust than the event-correction and leads to larger improvements in smaller catchments. 245 It should be noted that working with large data and automatic event separation without manual post-correction leads to problems that could potentially be avoided when considering individual time series more carefully. Several events are identified as too long or too short (or not at all), so their volumes are over- or understated, respectively. This results in false p/Vs and in some cases to severe over- or underestimation of the peak. The weight of such events is assumed to be significantly leads to larger improvements in smaller catchments. 245 It should be noted that working with large data and automatic event separation without manual post-correction leads to problems that could potentially be avoided when considering individual time series more carefully. Several events are identified as too long or too short (or not at all), so their volumes are over- or understated, respectively. This results in false p/Vs and in some cases to severe over- or underestimation of the peak. The weight of such events is assumed to be significantly It should be noted that working with large data and automatic event separation without manual post-correction leads to problems that could potentially be avoided when considering individual time series more carefully. Several events are identified as too long or too short (or not at all), so their volumes are over- or understated, respectively. This results in false p/Vs and in some cases to severe over- or underestimation of the peak. The weight of such events is assumed to be significantly lower when correcting flood statistics based on average p/Vs, which could be a particular reason for the latter method being 250 favored here. In addition, the overall performance can only be assessed for events that contain the monthly maximum flow, i.e. primarily larger events. How the event correction performs for minor events cannot be analyzed here. lower when correcting flood statistics based on average p/Vs, which could be a particular reason for the latter method being 250 favored here. In addition, the overall performance can only be assessed for events that contain the monthly maximum flow, i.e. 4.2 Estimation of mean annual IPF 5 shows the change in mean absolute error in the annual MHQ after correction with the different methods in relation to catchment size and elevation. The slope method (a) applied to the individual events yields a rather constant reduction of the 10 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. error independent of catchment size. However, there are several outliers produced by the method, which can be attributed to improper separation of smaller events. Applying the slope method only to the annual maximum MDF events, as done in Fig. 5 b shows a much smoother and more constant error reduction. The two methods using the p/V ratio (Fig. 5 c and d) yield a 240 much larger improvement for the smaller catchments where the error is generally larger than in the bigger catchments. However, both methods simultaneously lead to an increase of the error in several cases. This deterioration appears to affect those stations that have been highlighted earlier, namely the ones with the lowest elevations in the data set. Here the error is significantly lower than the p/Vs suggest. Again, the average p/V method appears more robust than the event-correction and leads to larger improvements in smaller catchments 245 5 b shows a much smoother and more constant error reduction. The two methods using the p/V ratio (Fig. 5 c and d) yield a 240 much larger improvement for the smaller catchments where the error is generally larger than in the bigger catchments. However, both methods simultaneously lead to an increase of the error in several cases. This deterioration appears to affect those stations that have been highlighted earlier, namely the ones with the lowest elevations in the data set. Here the error is significantly lower than the p/Vs suggest. Again, the average p/V method appears more robust than the event-correction and 5 b shows a much smoother and more constant error reduction. The two methods using the p/V ratio (Fig. 5 c and d) yield a 240 much larger improvement for the smaller catchments where the error is generally larger than in the bigger catchments. However, both methods simultaneously lead to an increase of the error in several cases. 4.2 Estimation of mean annual IPF primarily larger events. How the event correction performs for minor events cannot be analyzed here. lower when correcting flood statistics based on average p/Vs, which could be a particular reason for the latter method being 250 favored here. In addition, the overall performance can only be assessed for events that contain the monthly maximum flow, i.e. primarily larger events. How the event correction performs for minor events cannot be analyzed here. Figure 5: Error reduction (negative values) / increase (positive values) in the mean maximum flow for different IPF estimation methods when compared to MDF. 255 Figure 5: Error reduction (negative values) / increase (positive values) in the mean maximum flow for different IPF estimation methods when compared to MDF. 255 11 11 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Fig. 6 summarizes the overall model performances for the mean annual/seasonal maximum flow at all 653 stations and compares the individual methods to the error in using MDF directly. It is obvious that all methods give significantly better IPF estimates than the mere MDFs. The slope correction has quite a large bias, which is, as seen above, not only disadvantageous. Still, the overall error is smaller for the p/V methods, with fewer outliers produced by the average p/V approach. 260 Figure 6: Comparison of performances of different IPF estimation methods for the entire year (a), summer (b) and winter(c). 0 Figure 6: Comparison of performances of different IPF estimation methods for the entire year (a), summer (b) and winter(c). Figure 6: Comparison of performances of different IPF estimation methods for the entire ye Table 2 summarizes the normalized root mean square error (NRMSE) and the percentage bias (PBIAS) for all model variants. The values in parentheses indicate the performance criteria for gauges with catchment areas below 500 km². Here, the advantage of both p/V-approaches over the slope method become apparent, as their errors are significantly smaller. 265 Table 2 summarizes the normalized root mean square error (NRMSE) and the percentage bias (PBIAS) for all model variants. The values in parentheses indicate the performance criteria for gauges with catchment areas below 500 km². Here, the advantage of both p/V-approaches over the slope method become apparent, as their errors are significantly smaller. 4.2 Estimation of mean annual IPF 265 Table 2: Performances of different IPF estimation methods in terms of NRMSE and percentage bias. The values in parentheses show the performances for catchment sizes below 500 km². Year Summer Winter NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] MDF 17.0 (48.0) -18.0 (-32.5) 18.1 (49.1) -20.6 (-38.2) 14.9 (44.3) -16.4 (-28.8) Slope-events 9.2 (32.4) -8.6 (-20.5) 8.9 (34.6) -11.6 (-26.7) 10.3 (30.3) -7.8 (-17.3) p/V-events 12.1 (24.6) 2.0 (-12.3) 11.5 (26.7) -0.1 (-18.0) 13.5 (23.1) 2.9 ( -9.0) p/V-MHQ 7.6 (21.3) 0.7 ( -8.1) 8.2 (23.4) -0.4 (-12.3) 8.2 (22.5) 1.1 ( -5.7) Table 2: Performances of different IPF estimation methods in terms of NRMSE and percentage bias. The values in parentheses show the performances for catchment sizes below 500 km². Year Summer Winter NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] MDF 17.0 (48.0) -18.0 (-32.5) 18.1 (49.1) -20.6 (-38.2) 14.9 (44.3) -16.4 (-28.8) Slope-events 9.2 (32.4) -8.6 (-20.5) 8.9 (34.6) -11.6 (-26.7) 10.3 (30.3) -7.8 (-17.3) p/V-events 12.1 (24.6) 2.0 (-12.3) 11.5 (26.7) -0.1 (-18.0) 13.5 (23.1) 2.9 ( -9.0) p/V-MHQ 7.6 (21.3) 0.7 ( -8.1) 8.2 (23.4) -0.4 (-12.3) 8.2 (22.5) 1.1 ( -5.7) Table 2: Performances of different IPF estimation methods in terms of NRMSE and percentage bias. T the performances for catchment sizes below 500 km². 4.3 Comparison of IPF and MDF distributions The GEV distribution appeared to be a generally suitable distribution for the stations in the dataset, as shown in Fig. 7. The 270 Cramer-von-Mises test certifies a good fit for both the IPF and MDF samples, as well as for the slope and p/V corrected samples. Only a few stations lie close to the 5% line which would suggest a rejection of the null hypothesis. The GEV distribution appeared to be a generally suitable distribution for the stations in the dataset, as shown in Fig. 7. The 270 Cramer-von-Mises test certifies a good fit for both the IPF and MDF samples, as well as for the slope and p/V corrected samples. Only a few stations lie close to the 5% line which would suggest a rejection of the null hypothesis. 12 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Figure 7: Cramer-von-Mises p-values for the GEV distribution fitted to various samples of annual and seasonal maxima. p p c⃝Author(s) 2021. CC BY 4.0 License. Figure 7: Cramer von Mises p values for the GEV distribution fitted to various samples of annual and seasonal maxima Figure 7: Cramer-von-Mises p-values for the GEV distribution fitted to various samples of annual von-Mises p-values for the GEV distribution fitted to various samples of annual and seasonal maxima. A comparison between the estimated parameters for the IPF and MDF samples for the year and the seasons are shown in Fig. 275 8. As expected, the location parameters are consistently underestimated by the MDF series, with the largest errors in summer. This naturally leads to an overall downward shift of the “true” distribution when estimated from MDF values. The scales, here normalized by the location, appear to be primarily overestimated in summer, leading to distributions that are steeper for MDF than for IPF samples. For the year and winter, the error in in the scale parameter appears to be balanced in its direction. The shape parameters differ quite substantially between the seasons. In summer the vast majority of estimated parameter values 280 is negative, both in IPF and MDF. This indicates a heavy tail behavior for the summer floods. 4.3 Comparison of IPF and MDF distributions The fact that these negative values are in most cases significantly smaller in the MDF than in the IPF sample, suggests that the tails are overstated in the former case. This in combination with the underestimation of the location parameter leads to an overall underestimation of the lower and an overestimation of the higher flood quantiles by the MDF sample. For the year and winter, again, no clear trend is visible. 285 Some distinct patterns emerge from analysis of the estimated parameters: a) higher altitude catchments have small scale and shapes close to zero. This “mild” behavior is quite consistent over all seasons and the year. b) Lowest altitude catchments generally have a large scale and, especially in summer, negative shape. These catchments thus exhibit the strongest seasonal alternation with heaviest tails in summer. At the same time these catchments show overall largest deviations between IPF and is visible. 285 Some distinct patterns emerge from analysis of the estimated parameters: a) higher altitude catchments have small scale and shapes close to zero. This “mild” behavior is quite consistent over all seasons and the year. b) Lowest altitude catchments generally have a large scale and, especially in summer, negative shape. These catchments thus exhibit the strongest seasonal alternation with heaviest tails in summer. At the same time these catchments show overall largest deviations between IPF and is visible. 285 Some distinct patterns emerge from analysis of the estimated parameters: a) higher altitude catchments have small scale and shapes close to zero. This “mild” behavior is quite consistent over all seasons and the year. b) Lowest altitude catchments generally have a large scale and, especially in summer, negative shape. These catchments thus exhibit the strongest seasonal alternation with heaviest tails in summer. At the same time these catchments show overall largest deviations between IPF and MDF, especially in the summer half year. This suggests once more that IPF estimation for these catchments is particularly 290 difficult. MDF, especially in the summer half year. This suggests once more that IPF estimation for these catchments is particularly 290 difficult. 13 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Figure 8: Estimated GEV parameters from the IPF vs. MDF samples for the year and the two seasons. G ll h h il f h di ib i i h fl il i i l h fl d b https://doi.org/10.5194/hess-2021-466 Preprint. 4.3 Comparison of IPF and MDF distributions Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Figure 8: Estimated GEV parameters from the IPF vs. MDF samples for the year and the two seasons Generally, the heavy tails of the summer distributions in contrast to the flatter tails in winter let Generally, the heavy tails of the summer distributions in contrast to the flatter tails in winter let the summer floods become dominant at higher quantiles. For a return period of 100 years, the summer floods exceed the winter peaks at 61.9% of the 295 stations. For 50 and 10 years this exceedance occurs at 51.2% and 35.7% of stations, respectively. This behavior is also noticeable in the MDF but for fewer gauges, namely 53.4%, 43.2% and 21.0% for 100, 50 and 10-year return periods. dominant at higher quantiles. For a return period of 100 years, the summer floods exceed the winter peaks at 61.9% of the 295 stations. For 50 and 10 years this exceedance occurs at 51.2% and 35.7% of stations, respectively. This behavior is also noticeable in the MDF but for fewer gauges, namely 53.4%, 43.2% and 21.0% for 100, 50 and 10-year return periods. 4.4 Estimation of IPF quantiles Three approaches were tested for estimating IPF flood quantiles: a) correcting the sample L-moments required for parameter estimation (p/V-Lmoms), b) correcting the parameters of the fitted distribution (p/V-params), and c) directly correcting the 300 desired flood quantiles (p/V-quant). Method a) is convenient since a single model for each L-moment facilitates a correction of the complete distribution and hence each desired flood quantile. Estimating the L-moments has the additional advantage of not being restricted to a certain type of probability distribution. A proper distribution can be selected and fitted locally using 14 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. the corrected L-moments. Still, the other methods may prove more robust and are hence tested as well. Additionally, distributions were fitted to the annual and seasonal maxima that have been previously corrected using the slope (slope-events) 305 and p/V methods for events (p/V-events). the corrected L-moments. Still, the other methods may prove more robust and are hence tested as well. Additionally, distributions were fitted to the annual and seasonal maxima that have been previously corrected using the slope (slope-events) 305 and p/V methods for events (p/V-events). the corrected L-moments. Still, the other methods may prove more robust and are hence tested as well. Additionally, distributions were fitted to the annual and seasonal maxima that have been previously corrected using the slope (slope-events) 5 the corrected L-moments. Still, the other methods may prove more robust and are hence tested as well. Additionally, distributions were fitted to the annual and seasonal maxima that have been previously corrected using the slope (slope-events) 305 and p/V methods for events (p/V-events). Fig. 9 shows the errors in parameter estimates for the different approaches in comparison to the original uncorrected MDF error at the 490 validation stations with minimum 30-year flow records. All methods clearly improve the estimation for the location and shape parameters, where the L-Moment correction shows overall smallest error and bias. The improvement in the shape parameter for any of the methods is not as obvious. Since the shape parameter is generally difficult to model and the 310 overall error is comparably low, the direct use of the MDF shape parameter estimates is sensible. Nonetheless, the results presented here base on estimated shape parameters. the corrected L moments. Still, the other methods may prove more robust and are hence tested as well. 4.4 Estimation of IPF quantiles Additionally, distributions were fitted to the annual and seasonal maxima that have been previously corrected using the slope (slope-events) 305 and p/V methods for events (p/V-events). Fig. 9 shows the errors in parameter estimates for the different approaches in comparison to the original uncorrected MDF error at the 490 validation stations with minimum 30-year flow records. All methods clearly improve the estimation for the location and shape parameters, where the L-Moment correction shows overall smallest error and bias. The improvement in the 305 Fig. 9 shows the errors in parameter estimates for the different approaches in comparison to the original uncorrected MDF error at the 490 validation stations with minimum 30-year flow records. All methods clearly improve the estimation for the location and shape parameters, where the L-Moment correction shows overall smallest error and bias. The improvement in the shape parameter for any of the methods is not as obvious. Since the shape parameter is generally difficult to model and the 310 overall error is comparably low, the direct use of the MDF shape parameter estimates is sensible. Nonetheless, the results presented here base on estimated shape parameters. location and shape parameters, where the L-Moment correction shows overall smallest error and bias. The improvement in the shape parameter for any of the methods is not as obvious. Since the shape parameter is generally difficult to model and the 310 overall error is comparably low, the direct use of the MDF shape parameter estimates is sensible. Nonetheless, the results presented here base on estimated shape parameters. shape parameter for any of the methods is not as obvious. Since the shape parameter is generally difficult to model and the 310 overall error is comparably low, the direct use of the MDF shape parameter estimates is sensible. Nonetheless, the results presented here base on estimated shape parameters. Figure 9: Comparison of performances of various IPF-estimation methods for the GEV distribution parameters for the year and the two seasons. 15 Figure 9: Comparison of performances of various IPF-estimation methods for the GEV distribution parameters for the year and the two seasons. 315 315 15 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Fig. 10 demonstrates the quality of the different correction approaches by means of the 10-, 50- and 100-year flood at the 490 validation stations. 4.4 Estimation of IPF quantiles Other than for the mean, the differences between the individual methods are not as distinct here. It turned out that with increasing return period, the advantage of the L-Moment, parameter and quantile correction methods declines. The errors in the higher quantiles appear more random and do not relate as much to the average daily peak-volume relationship. Moreover, even if the parameter estimates of the correction methods are generally good, slightest deviations manifest 320 themselves in the tails of the distribution. This turns out to be especially valid for the low-altitude catchments. The overcorrection that was observed for the mean is even more pronounced here, which leads to an average decline in model performance. Last but not least, the general uncertainty in parameter estimation and extrapolation far beyond the time series length need to be kept in mind. Overall, even the estimation of the “true” IPF quantiles is potentially defective in itself, as will be discussed in the next section 325 Fig. 10 demonstrates the quality of the different correction approaches by means of the 10-, 50- and 100-year flood at the 490 validation stations. Other than for the mean, the differences between the individual methods are not as distinct here. It turned out that with increasing return period, the advantage of the L-Moment, parameter and quantile correction methods declines. The errors in the higher quantiles appear more random and do not relate as much to the average daily peak-volume relationship. Moreover, even if the parameter estimates of the correction methods are generally good, slightest deviations manifest 320 themselves in the tails of the distribution. This turns out to be especially valid for the low-altitude catchments. The overcorrection that was observed for the mean is even more pronounced here, which leads to an average decline in model performance. Last but not least, the general uncertainty in parameter estimation and extrapolation far beyond the time series length need to be kept in mind. Overall, even the estimation of the “true” IPF quantiles is potentially defective in itself, as will be discussed in the next section. 325 Figure 10: Comparison of performances of various IPF-estimation methods for different flood quantiles for the entire year. Figure 10: Comparison of performances of various IPF-estimation methods for different flood quantiles for the entire year. 4.4 Estimation of IPF quantiles Figure 10: Comparison of performances of various IPF-estimation methods for different flood qua Still, even for the 100-year flood, the L-moment and quantile approaches perform slightly better than the event correction methods, which becomes obvious from Table 3, where the normalized root mean square error (NRMSE) and the percentage bias (PBIAS) for all methods and return periods are summarized. It appears that all three average p/V methods perform 0 similarly well. However, due to its previously named advantages, the L-moment method is considered the superior approach in this setting. Still, even for the 100-year flood, the L-moment and quantile approaches perform slightly better than the event correctio methods, which becomes obvious from Table 3, where the normalized root mean square error (NRMSE) and the percentag bias (PBIAS) for all methods and return periods are summarized. It appears that all three average p/V methods perform 330 similarly well. However, due to its previously named advantages, the L-moment method is considered the superior approach in this setting. Between the event correction techniques, the slope method performs better than the p/V in terms of overall error but is strongly biased. The outlier problem, observed for the MHQ appears to propagate severely in the p/V-event method, leading to the high NRMSE values. When focusing on the catchments with areas below 500 km², the superiority of the p/V-methods becomes 335 once again apparent. Still, the difference in performance between the event methods and the p/Vmean methods decreases with increasing return period. For 50 and 100 years, the p/V-event model poses the best approach in terms of NRMSE and PBIAS, which suggests that the average p/V ratio is no longer able to properly explain the IPF-MDF relationship in the tails of the NRMSE values. When focusing on the catchments with areas below 500 km², the superiority of the p/V-methods becomes 335 once again apparent. Still, the difference in performance between the event methods and the p/Vmean methods decreases with increasing return period. For 50 and 100 years, the p/V-event model poses the best approach in terms of NRMSE and PBIAS, which suggests that the average p/V ratio is no longer able to properly explain the IPF-MDF relationship in the tails of the 16 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. 4.4 Estimation of IPF quantiles distributions and that the severe overestimations of the event correction occur primarily at the larger catchments with overall small deviation between MDF and IPF. 40 340 Table 3: Performances of different IPF estimation methods in terms of NRMSE and percentage bias The values in parentheses show the performances for catchment sizes below 500 km². mances of different IPF estimation methods in terms of NRMSE and percentage bias for different floo rentheses show the performances for catchment sizes below 500 km². Table 3: Performances of different IPF estimation methods in terms of NRMSE and percentage bias for different flood quantiles. The values in parentheses show the performances for catchment sizes below 500 km². T = 10 years T = 50 years T = 100 years NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] MDF 17.8 (50.0) -18.0 (-32.9) 17.8 (48.1) -18.3 (-32.4) 17.9 (47.6) -18.4 (-32.2) Slope-events 8.4 (30.9) -6.0 (-18.1) 10.8 (28.6) -4.6 (-16.4) 12.8 (29.1) -4.2 (-16.0) p/V-events 14.7 (23.6) 5.0 ( -9.7) 17.0 (21.7) 6.3 ( -7.4) 18.3 (23.0) 6.7 ( -6.7) p/V-Lmoms 8.8 (23.1) 1.1 ( -8.7) 10.0 (24.0) 0.1 ( -9.4) 10.9 (25.7) -0.4 ( -9.8) p/V-params 8.5 (24.2) 0.6 ( -9.3) 9.4 (25.8) -0.5 (-10.6) 10.1 (27.7) -1.1 (-11.4) p/V-quants 8.9 (23.0) 1.2 ( -8.5) 10.3 (23.8) 0.6 ( -9.2) 11.3 (25.5) 0.2 ( -9.7) PF estimation methods in terms of NRMSE and percentage bias for different flood quantiles. performances for catchment sizes below 500 km². Finally, the model performances of the mixed models, combining summer and winter floods, are analyzed for different flood 345 quantiles. Their behavior is generally comparable to the annual maximum series approach, as shown in Fig. 11. Even though the quantiles obtained with the mixed models may be more extreme and more parameters need to be estimated and corrected, there is no indication that the IPF correction will not function in this case. Finally, the model performances of the mixed models, combining summer and winter floods, are analyzed for different flood 345 quantiles. Their behavior is generally comparable to the annual maximum series approach, as shown in Fig. 11. Even though the quantiles obtained with the mixed models may be more extreme and more parameters need to be estimated and corrected, there is no indication that the IPF correction will not function in this case. 4.4 Estimation of IPF quantiles Where this was 360 the case, the utilized variance function is given in the table. The final models for the L-moment correction can be found in Table 5. Some models exhibited a non-constant error variance and were thus re-fitted using generalized least squares from the “nlme” R-package (Pinhero & Bates, 2018). Where this was 360 the case, the utilized variance function is given in the table. The final models for the L-moment correction can be found in Table 5. Some models exhibited a non-constant error variance and were thus re-fitted using generalized least squares from the “nlme” R-package (Pinhero & Bates, 2018). Where this was 360 the case, the utilized variance function is given in the table. Table 5: Linear models fitted for correction of L-moments with variance function used for variance stabilization in generalized least squares fitting. Model Variance Function L1 Year L1MDF * (1.10 - 1.27 * p/Vmean) - Summer L1MDF * (1.28 - 1.32 * p/Vmean - 0.00014 * elevation) Exponential Winter L1MDF * (1.29 - 1.43 * p/Vmean - 0.00021 * elevation) Exponential L2 Year L2MDF * (1.06 - 1.06 * p/Vmean) - Summer L2MDF * (1.24 - 1.09 * p/Vmean- 0.00013 * elevation) Exponential Winter L2MDF * (1.28 - 1.37 * p/Vmean- 0.00026 * elevation) Exponential T3 Year 0.95 * T3MDF - 0.00062 * p/Vmean - Summer 1.07 * T3MDF - 0.19 * p/Vmean - Winter 0.94 * T3MDF - 0.024 * p/Vmean Power 365 for correction of L-moments with variance function used for variance stabilization in generalized least Table 5: Linear models fitted for correction of L-moments with variance function used for variance stabilization in generalized least squares fitting. Linear models fitted for correction of L-moments with variance function used for variance stabilization itting Table 5: Linear models fitted for correction of L-moments with variance function used for variance sta squares fitting. 4.5 Uncertainty The results of the bootstrapping procedure used to assess uncertainty are exemplary shown in Fig. 12 for the HQ100 at a single station with a reduced number of 100 permutations. In panel a, the IPF and MDF estimates for each permutation of the annual maximum series are plotted against each other. This shows the bandwidths of both the IPF and MDF estimates as a result of uncertainty in the distribution fitting. Fig. 12 b shows the estimated IPF flood quantiles vs. the quantiles estimated using the 370 p/V models for each permutation. The dark blue points represent the full linear models using all available stations in the study area, while the light blue points represent 100 resampled model estimates. In this example, it becomes obvious that the range in flood quantile estimates due to permutation in the linear models is significantly smaller than the range in estimates due to distribution fitting. This is valid for the majority of stations. The results of the bootstrapping procedure used to assess uncertainty are exemplary shown in Fig. 12 for the HQ100 at a single station with a reduced number of 100 permutations. In panel a, the IPF and MDF estimates for each permutation of the annual maximum series are plotted against each other. This shows the bandwidths of both the IPF and MDF estimates as a result of uncertainty in the distribution fitting. Fig. 12 b shows the estimated IPF flood quantiles vs. the quantiles estimated using the 370 p/V models for each permutation. The dark blue points represent the full linear models using all available stations in the study The results of the bootstrapping procedure used to assess uncertainty are exemplary shown in Fig. 12 for the HQ100 at a single station with a reduced number of 100 permutations. In panel a, the IPF and MDF estimates for each permutation of the annual maximum series are plotted against each other. This shows the bandwidths of both the IPF and MDF estimates as a result of uncertainty in the distribution fitting. Fig. 12 b shows the estimated IPF flood quantiles vs. the quantiles estimated using the 370 p/V models for each permutation. The dark blue points represent the full linear models using all available stations in the study area, while the light blue points represent 100 resampled model estimates. 4.4 Estimation of IPF quantiles Finally, the model performances of the mixed models, combining summer and winter floods, are analyzed for different flood 345 quantiles. Their behavior is generally comparable to the annual maximum series approach, as shown in Fig. 11. Even though the quantiles obtained with the mixed models may be more extreme and more parameters need to be estimated and corrected, there is no indication that the IPF correction will not function in this case. Figure 11: Comparison of performances of various IPF-estimation methods for different seasonally mixed flood quantiles. 350 The NRMSE and PBIAS values for the mixed approach are shown in table 4. Again, for the smaller catchments, the p/V-event correction method shows the best performance. The NRMSE and PBIAS values for the mixed approach are shown in table 4. Again, for the smaller catchments, the p/V-event correction method shows the best performance. 17 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. 355 Table 4: Mixed-model performances of different IPF estimation methods in terms of NRMSE and percentage bias for different flood quantiles. The values in parentheses show the performance for catchment sizes below 500 km². Table 4: Mixed-model performances of different IPF estimation methods in terms of NRMSE and percentage bias for different flood quantiles. The values in parentheses show the performance for catchment sizes below 500 km². Table 4: Mixed-model performances of different IPF estimation methods in terms of NRMSE and percentage bias for different flood quantiles. The values in parentheses show the performance for catchment sizes below 500 km². T = 10 years T = 50 years T = 100 years NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] NRMSE [%] PBIAS [%] MDF 17.7 (50.2) -17.9 (-33.0) 17.5 (48.4) -18.0 (-32.3) 17.6 (48.0) -18.1 (-32.1) Slope-events 8.0 (31.2) -6.1 (-18.3) 9.4 (28.4) -4.4 (-16.2) 10.6 (28.4) -3.7 (-15.5) p/V-events 14.6 (23.8) 4.8 ( -9.8) 16.7 (21.8) 6.4 ( -7.2) 17.8 (23.3) 6.9 ( -6.2) p/V-Lmoms 9.1 (25.7) 0.0 (-10.0) 8.6 (26.0) -2.0 (-11.4) 9.2 (27.3) -3.1 (-12.5) p/V-params 9.3 (26.1) -0.8 (-10.1) 8.8 (26.3) -2.8 (-11.7) 9.4 (27.7) -4.0 (-12.8) The final models for the L-moment correction can be found in Table 5. Some models exhibite The final models for the L-moment correction can be found in Table 5. Some models exhibited a non-constant error variance and were thus re-fitted using generalized least squares from the “nlme” R-package (Pinhero & Bates, 2018). 4.5 Uncertainty In this example, it becomes obvious that the range in flood quantile estimates due to permutation in the linear models is significantly smaller than the range in estimates due to distribution fitting. This is valid for the majority of stations. uncertainty in the distribution fitting. Fig. 12 b shows the estimated IPF flood quantiles vs. the quantiles estimated using the 370 p/V models for each permutation. The dark blue points represent the full linear models using all available stations in the study area, while the light blue points represent 100 resampled model estimates. In this example, it becomes obvious that the range in flood quantile estimates due to permutation in the linear models is significantly smaller than the range in estimates due to distribution fitting. This is valid for the majority of stations. 18 375 Figure 12: Example of bootstrapping results at a station with 100 permutations. (a) HQ100 from IFP vs. MDF for each permutation of the time series, (b) HQ100 from IPF vs. 100 model estimates per permutation; the dark blue dots represent the full model. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. 375 Figure 12: Example of bootstrapping results at a station with 100 permutations. (a) HQ100 from IFP vs. MDF for each permutation of the time series, (b) HQ100 from IPF vs. 100 model estimates per permutation; the dark blue dots represent the full model. Fig. 13 shows the relative widths of the 95% confidence intervals for all bootstrapping samples. The average widths of the IPF-bs, MDF-bss and p/V-full seem to be similar with a larger variability in the IPF sample. The width of the average range of Fig. 13 shows the relative widths of the 95% confidence intervals for all bootstrapping samples. The average widths of the IPF-bs, MDF-bss and p/V-full seem to be similar with a larger variability in the IPF sample. The width of the average range of the individual model permutations (p/V-bs-mean) is very small at all stations and therefore contributes little to the overall level 380 of uncertainty (p/V-bs-bs). the individual model permutations (p/V-bs-mean) is very small at all stations and therefore contributes little to the overall level 380 of uncertainty (p/V-bs-bs). Fi 13 R l ti idth f i b t t l f diff t fl d til Figure 13: Relative widths of various bootstrap samples for different flood quantiles. 4.5 Uncertainty Figure 13: Relative widths of various bootstrap samples for different flood quantiles. In order to assess the full bandwidth of the errors in the linear model estimates, they are compare In order to assess the full bandwidth of the errors in the linear model estimates, they are compared to the range of errors in the MDF estimates. Fig. 14 shows the mean deviations from the perturbed IPF quantiles, as well as the lower and upper limits of 385 the 50% and 95% confidence intervals of the errors for the 10-, 50- and 100-year flood quantiles. It is obvious that the overall uncertainty gets larger with increasing return period, as can be seen by the increasing distance between lower and upper confidence limits. The p/V-model estimates appear to be slightly positively biased and positively skewed, which is especially noticeable in the 95% confidence interval for the HQ100. At many stations there is a significant overestimation of the true IPF MDF estimates. Fig. 14 shows the mean deviations from the perturbed IPF quantiles, as well as the lower and upper limits of 385 the 50% and 95% confidence intervals of the errors for the 10-, 50- and 100-year flood quantiles. It is obvious that the overall uncertainty gets larger with increasing return period, as can be seen by the increasing distance between lower and upper confidence limits. The p/V-model estimates appear to be slightly positively biased and positively skewed, which is especially noticeable in the 95% confidence interval for the HQ100. At many stations there is a significant overestimation of the true IPF 19 quantile with some of the linear model transpositions. The MDF estimates on the other hand exhibit the expected persistent https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. quantile with some of the linear model transpositions. The MDF estimates on the other hand exhibit the expected persistent 390 quantile with some of the linear model transpositions. The MDF estimates on the other hand exhibit the expected persistent 0 underestimation. quantile with some of the linear model transpositions. The MDF estimates on the other hand exhibit the expected persistent underestimation. 390 Figure 14: Error distribution of the MDF and p/V bootstrap samples for three flood quantiles. 4.5 Uncertainty Shown are the median errors (left), as well as the lower and upper limits of the 50% (center) and 95% confidence intervals (right). Figure 14: Error distribution of the MDF and p/V bootstrap samples for three flood quantiles. Shown are the median errors (left), as well as the lower and upper limits of the 50% (center) and 95% confidence intervals (right). Figure 14: Error distribution of the MDF and p/V bootstrap samples for three flood quantiles. Shown as well as the lower and upper limits of the 50% (center) and 95% confidence intervals (right). Fig. 15 summarizes the general overlap of the confidence intervals of MDF and estimated IPF with the confidence intervals of 395 the observed IPF for the three flood quantiles. It becomes obvious that the agreement between IPF and the p/V model estimates is significantly larger than with the MDF values. This observation suggests that with high probability the p/V model estimates are in the range of the “true” IPF quantiles. Figure 15: Percentage overlap between the 95% confidence intervals of all IPF bootstrap estimates and MDF and p/V model 400 bootstrap estimates for three flood quantiles. Figure 15: Percentage overlap between the 95% confidence intervals of all IPF bootstrap estimates and MDF and p/V model 400 bootstrap estimates for three flood quantiles. Figure 15: Percentage overlap between the 95% confidence intervals of all IPF bootstrap estimates and MDF and p/V model 400 bootstrap estimates for three flood quantiles. 4.6 Range of applications and limitations We therefore discourage the application of the suggested correction methods at gauges that are both situated below 100 m a.s.l. and have catchment areas larger than 200 km². A problem for IPF correction, which has been exhaustively discussed above, are gauges that exhibit little difference between 420 MDF and IPF floods, even though their p/V ratio would suggest a much larger error. For our dataset this applies to the lowest- altitude gauges in the dataset. The MDFs at these stations are overcorrected and thus exhibit severe overestimation of the true IPFs. We therefore discourage the application of the suggested correction methods at gauges that are both situated below 100 m a.s.l. and have catchment areas larger than 200 km². This observation may also suggest that other factors need to be considered for proper error estimation or that the parameters 425 of the correction models need to be adjusted for different subsets of data. This is also relevant for the question of universality of the proposed method. Our data set is limited and representative of a temperate humid climate and moderate altitude. Thus, a qualitative sensitivity analysis is carried out on the full 653-stations dataset in order to identify patterns that may be extrapolatable to other regions. The subsets are selected by combinations of geographical location, catchment size and gauge elevation. Target variable is the mean annual maximum IPF. Differences in the individual models due to different degrees of 430 freedom are natural, which is why only those subsets that lead to significant deviations from the original model are mentioned here. Two sets of stations deviate noticably from the original model. The first one includes the low-altitude gauges discussed before. Here the overall error is so small that no correction yields better results than correction by the linear model. The second group elevation. Target variable is the mean annual maximum IPF. Differences in the individual models due to different degrees of 430 freedom are natural, which is why only those subsets that lead to significant deviations from the original model are mentioned here. Two sets of stations deviate noticably from the original model. The first one includes the low-altitude gauges discussed before. Here the overall error is so small that no correction yields better results than correction by the linear model. 4.6 Range of applications and limitations Thus, the additional step of refining multiple peak events, as suggested by Tarasova et al. (2018) should be carried out, when rainfall and snowmelt information is available. Using the p/Vevent in order to correct individual events and then using the corrected series for FFA poses in theory a more necessarily coincide with maximum IPF events, which is why correcting all events first and then selecting the annual maxima 415 should yield a more appropriate IPF sample. But again, correcting individual events depends greatly on a very careful event separation, which could not be achieved in this case and which led to some unrealistic IPF estimates. Nonetheless, if a proper event separation is possible, the event correction method may have the larger potential. In such a case, a single model would be sufficient to account for all aspects of IPF estimation, including high flood quantiles. necessarily coincide with maximum IPF events, which is why correcting all events first and then selecting the annual maxima 415 should yield a more appropriate IPF sample. But again, correcting individual events depends greatly on a very careful event separation, which could not be achieved in this case and which led to some unrealistic IPF estimates. Nonetheless, if a proper event separation is possible, the event correction method may have the larger potential. In such a case, a single model would be sufficient to account for all aspects of IPF estimation, including high flood quantiles. A problem for IPF correction, which has been exhaustively discussed above, are gauges that exhibit little difference between 420 MDF and IPF floods, even though their p/V ratio would suggest a much larger error. For our dataset this applies to the lowest- altitude gauges in the dataset. The MDFs at these stations are overcorrected and thus exhibit severe overestimation of the true IPFs. We therefore discourage the application of the suggested correction methods at gauges that are both situated below 100 m a.s.l. and have catchment areas larger than 200 km². A problem for IPF correction, which has been exhaustively discussed above, are gauges that exhibit little difference between 420 MDF and IPF floods, even though their p/V ratio would suggest a much larger error. For our dataset this applies to the lowest- altitude gauges in the dataset. The MDFs at these stations are overcorrected and thus exhibit severe overestimation of the true IPFs. 4.6 Range of applications and limitations The method of correcting the error of MDF floods via p/Vs performs well and is easily applicable in our study area. However, its great simplification and mere approximation of physical flood generating processes results in some problems and limitations that will be listed and discussed here. 405 20 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. The first aspect that may influence the performance of the proposed IPF correction method is the event separation technique. The chosen technique determines how flood events and thus the required hydrograph characteristics are defined. The choice of baseflow separating algorithm can greatly affect the identification of start and end points of flood events. Strict independence criteria and thresholds for event recognition may lead to rejection of crucial flood events when considering daily time series. Lax criteria, on the other hand, may create unnaturally long multi-peak events and false inclusion of small events, both leading 0 to unrealistic hydrograph characteristics and IPF estimates. Thus, the additional step of refining multiple peak events, as suggested by Tarasova et al. (2018) should be carried out, when rainfall and snowmelt information is available. The first aspect that may influence the performance of the proposed IPF correction method is the event separation technique. The chosen technique determines how flood events and thus the required hydrograph characteristics are defined. The choice of baseflow separating algorithm can greatly affect the identification of start and end points of flood events. Strict independence criteria and thresholds for event recognition may lead to rejection of crucial flood events when considering daily time series. Lax criteria, on the other hand, may create unnaturally long multi-peak events and false inclusion of small events, both leading 410 to unrealistic hydrograph characteristics and IPF estimates. Thus, the additional step of refining multiple peak events, as suggested by Tarasova et al. (2018) should be carried out, when rainfall and snowmelt information is available. Using the p/Vevent in order to correct individual events and then using the corrected series for FFA poses in theory a more sensible approach than using the p/Vmean from the annual MDF maxima. As mentioned before, maximum MDF events do not Lax criteria, on the other hand, may create unnaturally long multi-peak events and false inclusion of small events, both leading 410 to unrealistic hydrograph characteristics and IPF estimates. 4.6 Range of applications and limitations Finally, one should note that the type of distribution for flood quantile estimation can only be selected based on daily data and may differ from the optimal IPF distribution. For our data, the GEV proved flexible enough to be a good match in both MDF and IPF but this could differ in other cases. 455 between different flood types for IPF estimation proved meaningful in our case, as it revealed different dynamics and MDF- 450 IPF relationships. This observation could be further exploited by more carefully defining and distinguishing flood types, as e.g. proposed by Fischer (2018) or Tarasova et al. (2020). Finally, one should note that the type of distribution for flood quantile estimation can only be selected based on daily data and may differ from the optimal IPF distribution. For our data, the GEV proved flexible enough to be a good match in both MDF and IPF but this could differ in other cases. 455 between different flood types for IPF estimation proved meaningful in our case, as it revealed different dynamics and MDF- 450 IPF relationships. This observation could be further exploited by more carefully defining and distinguishing flood types, as e.g. proposed by Fischer (2018) or Tarasova et al. (2020). Finally, one should note that the type of distribution for flood quantile estimation can only be selected based on daily data and Finally, one should note that the type of distribution for flood quantile estimation can only be selected based on daily data and may differ from the optimal IPF distribution. For our data, the GEV proved flexible enough to be a good match in both MDF and IPF but this could differ in other cases. 455 Finally, one should note that the type of distribution for flood quantile estimation can only be selected based on daily data and may differ from the optimal IPF distribution. For our data, the GEV proved flexible enough to be a good match in both MDF and IPF but this could differ in other cases. 455 5 Conclusions and Outlook As in other studies before, it could be shown that the IPF-MDF relationship depends primarily on catchment size. It could also be observed that other factors, in this case gauge elevation, play a role in determining the difference between MDF and IPF floods. The relationship also appeared to differ between the two types of floods considered here, namely winter and summer floods. Since summer floods are often caused by short but intense rain events and thus exhibit steep rising and falling limbs, 460 th i bd il k h l th d diffi lt t ti t f th th d d il k L l i As in other studies before, it could be shown that the IPF-MDF relationship depends primarily on catchment size. It could also be observed that other factors, in this case gauge elevation, play a role in determining the difference between MDF and IPF floods. The relationship also appeared to differ between the two types of floods considered here, namely winter and summer floods. Since summer floods are often caused by short but intense rain events and thus exhibit steep rising and falling limbs, 460 their subdaily peaks are much larger than and difficult to estimate from the smoothed average daily peaks. Long, voluminous winter floods on the other hand show a much smaller IPF-MDF ratio and are easier to model. This study has also shown that hydrograph characteristics, like the peak-volume ratio of flood events can be used to estimate As in other studies before, it could be shown that the IPF-MDF relationship depends primarily on catchment size. It could also be observed that other factors, in this case gauge elevation, play a role in determining the difference between MDF and IPF floods. The relationship also appeared to differ between the two types of floods considered here, namely winter and summer their subdaily peaks are much larger than and difficult to estimate from the smoothed average daily peaks. Long, voluminous winter floods on the other hand show a much smaller IPF-MDF ratio and are easier to model. This study has also shown that hydrograph characteristics, like the peak-volume ratio of flood events can be used to estimate instantaneous peak flows when only average daily series are available. The p/V ratio may be used to predict both IPFs of instantaneous peak flows when only average daily series are available. 4.6 Range of applications and limitations Longitude and latitude do not appear to have any effect on the model fitting. Dividing the study area into quadrants does not 445 result in any differences between the subsets, even when equalizing the other factors catchment size and elevation. Also, neither record length nor period of record appear to have an influence. Longitude and latitude do not appear to have any effect on the model fitting. Dividing the study area into quadrants does not 445 result in any differences between the subsets, even when equalizing the other factors catchment size and elevation. Also, neither record length nor period of record appear to have an influence. The distinction between summer and winter for representation of the two most opposite flood types is particularly valid for this study area and should be adjusted where flood types are otherwise distributed. In general, even the rough distinction Longitude and latitude do not appear to have any effect on the model fitting. Dividing the study area into quadrants does not 445 result in any differences between the subsets, even when equalizing the other factors catchment size and elevation. Also, neither record length nor period of record appear to have an influence. The distinction between summer and winter for representation of the two most opposite flood types is particularly valid for this study area and should be adjusted where flood types are otherwise distributed. In general, even the rough distinction Longitude and latitude do not appear to have any effect on the model fitting. Dividing the study area into quadrants does not 445 result in any differences between the subsets, even when equalizing the other factors catchment size and elevation. Also, neither record length nor period of record appear to have an influence. The distinction between summer and winter for representation of the two most opposite flood types is particularly valid for this study area and should be adjusted where flood types are otherwise distributed. In general, even the rough distinction between different flood types for IPF estimation proved meaningful in our case, as it revealed different dynamics and MDF- 450 IPF relationships. This observation could be further exploited by more carefully defining and distinguishing flood types, as e.g. proposed by Fischer (2018) or Tarasova et al. (2020). 4.6 Range of applications and limitations The second group Here the overall error is so small that no correction yields better results than correction by the linear model. The second group includes the catchments with areas below 50 km². The errors for these stations appear very scattered and randomly distributed. 435 Comparing the p/V from the daily series with the p/V obtained from instantaneous events, it becomes obvious that the difference increases with decreasing catchment size and becomes excessively large and random for catchment sizes below 100 km². The correction using mean daily p/V only functions where unknown instantaneous flood dynamics are roughly approximated by observed daily flow variability. The smaller the temporal scale of an instantaneous flood event, the poorer it includes the catchments with areas below 50 km². The errors for these stations appear very scattered and randomly distributed. 435 Comparing the p/V from the daily series with the p/V obtained from instantaneous events, it becomes obvious that the difference increases with decreasing catchment size and becomes excessively large and random for catchment sizes below 100 km². The correction using mean daily p/V only functions where unknown instantaneous flood dynamics are roughly approximated by observed daily flow variability. The smaller the temporal scale of an instantaneous flood event, the poorer it 21 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. is reproduced in the daily records. If instantaneous events manifest themselves primarily on a subdaily basis, the possibility to 440 describe their dynamics via daily flows becomes ineligible. This observation is also in accordance with the observed temporal shifts between MDF and IPF events, which is increasingly pronounced in smaller catchments. In summary, the proposed correction method founders at smaller scales below 100 km². Even though the IPF estimation leads to a general improvement at this scale, the daily flood time scale poses a poor predictor in these catchments. is reproduced in the daily records. If instantaneous events manifest themselves primarily on a subdaily basis, the possibility to 440 describe their dynamics via daily flows becomes ineligible. This observation is also in accordance with the observed temporal shifts between MDF and IPF events, which is increasingly pronounced in smaller catchments. In summary, the proposed correction method founders at smaller scales below 100 km². Even though the IPF estimation leads to a general improvement at this scale, the daily flood time scale poses a poor predictor in these catchments. 5 Conclusions and Outlook There are two limitations, where the proposed method should be handled with care: a) at stations with an elevation below 100 m and catchment areas above 200 km², since it overestimates the true difference between IPF and MDF and b) at catchments 475 smaller than 100 km², where it underestimates the error so that the full correction potential cannot be achieved. Still, in comparison to the slope method, the time-scale approach works significantly better for smaller catchment areas, especially below 500 km². For larger catchments, the two methods are more comparable in performance. For future analyses it will be meaningful to test the universality of the proposed approach in other study regions. Also, the 475 For future analyses it will be meaningful to test the universality of the proposed approach in other study regions. Also, the effect of the flood event separation on the IPF estimation performance should be analyzed in more detail, especially in order 480 to improve the event correction technique. Finally, it will be interesting to see if explicit consideration of more carefully defined flood types can improve the IFP estimation in mixed models. effect of the flood event separation on the IPF estimation performance should be analyzed in more detail, especially in order 480 to improve the event correction technique. Finally, it will be interesting to see if explicit consideration of more carefully defined flood types can improve the IFP estimation in mixed models. effect of the flood event separation on the IPF estimation performance should be analyzed in more detail, especially in order 480 to improve the event correction technique. Finally, it will be interesting to see if explicit consideration of more carefully defined flood types can improve the IFP estimation in mixed models. Data availability y The discharge data used in this study is publicly available on the websites of the respective federal agencies. 485 The discharge data used in this study is publicly available on the websites of the respective federal agencies. 485 Lower Saxony: Niedersächsischer Landesbetrieb für Wasserwirtschaft, Küsten- und Naturschutz (NLWKN) http://www.wasserdaten.niedersachsen.de/cadenza/ Saxony-Anhalt: Landesbetrieb für Hochwasserschutz und Wasserwirtschaft Sachsen-Anhalt (LHW) https://gld-sa.dhi- wasy.de/GLD-Portal/ The discharge data used in this study is publicly available on the websites of the respective federal agencies. 485 Lower Saxony: Niedersächsischer Landesbetrieb für Wasserwirtschaft, Küsten- und Naturschutz (NLWKN) http://www.wasserdaten.niedersachsen.de/cadenza/ Saxony-Anhalt: Landesbetrieb für Hochwasserschutz und Wasserwirtschaft Sachsen-Anhalt (LHW) https://gld-sa.dhi- wasy.de/GLD-Portal/ wasy.de/GLD Portal/ Saxony: Sächsisches Landesamt für Umwelt, Landwirtschaft und Geologie (LFULG) 490 https://www.umwelt.sachsen.de/umwelt/infosysteme/ida/ Bavaria: Bayerisches Landesamt für Umwelt (LfU) https://www.gkd.bayern.de/de/ Baden-Württemberg: Landesanstalt für Umwelt Baden Württemberg (LUBW) https://udo.lubw.baden- wuerttemberg.de/public/ 495 Author contribution UH formulated the research goal. The study was designed by both authors and carried out by AB. AB prepared the manuscript with contributions from UH. Competing interests 500 The authors declare that they have no conflict of interest. Saxony: Sächsisches Landesamt für Umwelt, Landwirtschaft und Geologie (LFULG) 490 https://www.umwelt.sachsen.de/umwelt/infosysteme/ida/ Bavaria: Bayerisches Landesamt für Umwelt (LfU) https://www.gkd.bayern.de/de/ Baden-Württemberg: Landesanstalt für Umwelt Baden Württemberg (LUBW) https://udo.lubw.baden- wuerttemberg.de/public/ 495 Author contribution UH formulated the research goal. The study was designed by both authors and carried out by AB. AB prepared the manuscript with contributions from UH. Competing interests 500 The authors declare that they have no conflict of interest. Acknowledgements This work is part of the research group FOR 2416 “Space-Time Dynamics of Extreme Floods (SPATE)” funded by the Germ Research Foundation (“Deutsche Forschungsgemeinschaft”, DFG). 505 5 Conclusions and Outlook The p/V ratio may be used to predict both IPFs of individual events and instantaneous flood statistics, including mean annual and seasonal maximum flows and flood quantiles. 465 Due to improper flood event separation, the event-correction method produced some outliers in our case but may work significantly better when flood events can be defined more carefully. In general, the p/V method requires a minimum of data and can be applied using mere information from the daily series itself. The performance could be marginally improved by including gauge elevation as additional predictor in the models. individual events and instantaneous flood statistics, including mean annual and seasonal maximum flows and flood quantiles. 465 Due to improper flood event separation, the event-correction method produced some outliers in our case but may work significantly better when flood events can be defined more carefully. In general, the p/V method requires a minimum of data and can be applied using mere information from the daily series itself. The performance could be marginally improved by including gauge elevation as additional predictor in the models. individual events and instantaneous flood statistics, including mean annual and seasonal maximum flows and flood quantiles. 465 Due to improper flood event separation, the event-correction method produced some outliers in our case but may work significantly better when flood events can be defined more carefully. In general, the p/V method requires a minimum of data and can be applied using mere information from the daily series itself. The performance could be marginally improved by including gauge elevation as additional predictor in the models. The general recommendation for estimating IPF flood quantiles is to use the average p/V approach for correction of L- 470 moments. This method is convenient since L-moments can be globally corrected while distributions may be locally fitted The general recommendation for estimating IPF flood quantiles is to use the average p/V approach for correction of L- 470 moments. This method is convenient since L-moments can be globally corrected while distributions may be locally fitted 22 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. afterwards. It turned out that the first two L-moments are easily estimated using p/Vmean, while higher order L-moments or L- moment ratios are more difficult to model with this approach. (DWD): https://www.dwd.de/DE/leistungen/klimadatendeutschland/vielj_mittelwerte.html, last access: 17 September, 2021. 520 Ellis, W. and Gray, M.: Interrelationships between the peak instantaneous and average daily discharges of small prairie streams, Can Agr Eng–39, 1966. Fill, H. D. and Steiner, A. A.: Estimating instantaneous peak flow from mean daily flow data, J Hydrol Eng, 8, 365-369, doi: 10.1061/(ASCE)1084-0699(2003)8:6(365), 2003. https://www.dwd.de/DE/leistungen/klimadatendeutschland/vielj_mittelwerte.html, last access: 17 September, 2021. 520 https://www.dwd.de/DE/leistungen/klimadatendeutschland/vielj_mittelwerte.html, last access: 17 September, 2021. 520 Ellis, W. and Gray, M.: Interrelationships between the peak instantaneous and average daily discharges of small prairie streams, Can Agr Eng–39, 1966. Fill H D d S i A A E i i i k fl f d il fl d J H d l E 8 365 369 d i Ellis, W. and Gray, M.: Interrelationships between the peak instantaneous and average daily discharges of small prairie streams, Can Agr Eng–39, 1966. streams, Can Agr Eng–39, 1966. Fill, H. D. and Steiner, A. A.: Estimating instantaneous peak flow from mean daily flow data, J Hydrol Eng, 8, 365-369, doi: 10.1061/(ASCE)1084-0699(2003)8:6(365), 2003. Fill, H. D. and Steiner, A. A.: Estimating instantaneous peak flow from mean daily flow data, J Hydrol Eng, 8, 365-369, doi: 10.1061/(ASCE)1084-0699(2003)8:6(365), 2003. Fischer, S.: A seasonal mixed-POT model to estimate high flood quantiles from different event types and seasons, J Appl Stat, 525 45, 2831-2847, doi:10.1080/02664763.2018.1441385, 2018. Fischer, S.: A seasonal mixed-POT model to estimate high flood quantiles from different event types and seasons, J Appl Stat, 525 45, 2831-2847, doi:10.1080/02664763.2018.1441385, 2018. Fischer, S., Schumann, A., and Schulte, M.: Characterisation of seasonal flood types according to timescales in mixed probability distributions, J Hydrol, 539, 38-56, doi:10.1016/j.jhydrol.2016.05.005, 2016. Fuller, W. E.: Flood flows, T Am Soc Civ Eng, 77, 564– 617, doi:10.1061/taceat.0002552, 1914. 45, 2831-2847, doi:10.1080/02664763.2018.1441385, 2018. Fischer, S., Schumann, A., and Schulte, M.: Characterisation of seasonal flood types according to timescales in mixed probability distributions, J Hydrol, 539, 38-56, doi:10.1016/j.jhydrol.2016.05.005, 2016. Fischer, S., Schumann, A., and Schulte, M.: Characterisation of seasonal flood types according to timescales in mixed probability distributions, J Hydrol, 539, 38-56, doi:10.1016/j.jhydrol.2016.05.005, 2016. Fuller, W. E.: Flood flows, T Am Soc Civ Eng, 77, 564– 617, doi:10.1061/taceat.0002552, 1914. Fischer, S., Schumann, A., and Schulte, M.: Characterisation of seasonal flood types according to timescales in mixed probability distributions, J Hydrol, 539, 38-56, doi:10.1016/j.jhydrol.2016.05.005, 2016. Fuller, W. E.: Flood flows, T Am Soc Civ Eng, 77, 564– 617, doi:10.1061/taceat.0002552, 1914. Acknowledgements This work is part of the research group FOR 2416 “Space-Time Dynamics of Extreme Floods (SPATE)” funded by the German Research Foundation (“Deutsche Forschungsgemeinschaft”, DFG). 505 23 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. References Acharya, A. and Ryu, J. H.: Simple Method for Streamflow Disaggregation, J Hydrol Eng, 19, 509-519, doi: 10.1061/(ASCE)HE.1943-5584.0000818, 2014. Acharya, A. and Ryu, J. H.: Simple Method for Streamflow Disaggregation, J Hydrol Eng, 19, 509-519, doi: 10.1061/(ASCE)HE.1943-5584.0000818, 2014. Canuti, P. and Moisello, U.: Relationship between the Yearly Maxima of Peak and Daily Discharge for Some Basins in Tuscany, Hydrolog Sci J, 27, 111-128, doi:10.1080/02626668209491094, 1982. 510 Chen, B., Krajewski, W. F., Liu, F., Fang, W. H., and Xu, Z. X.: Estimating instantaneous peak flow from mean daily flow, Hydrol Res, 48, 1474-1488, doi: 10.2166/nh.2017.200, 2017. Ding, J. and Haberlandt, U.: Estimation of instantaneous peak flow from maximum mean daily flow by regionalization of catchment model parameters, Hydrol Process, 31, 612-626, doi:10.1002/hyp.11053, 2017. Acharya, A. and Ryu, J. H.: Simple Method for Streamflow Disaggregation, J Hydrol Eng, 19, 509-519, doi: 10.1061/(ASCE)HE.1943-5584.0000818, 2014. Canuti, P. and Moisello, U.: Relationship between the Yearly Maxima of Peak and Daily Discharge for Some Basins in Tuscany, Hydrolog Sci J, 27, 111-128, doi:10.1080/02626668209491094, 1982. 10 Chen, B., Krajewski, W. F., Liu, F., Fang, W. H., and Xu, Z. X.: Estimating instantaneous peak flow from mean daily flow, Hydrol Res, 48, 1474-1488, doi: 10.2166/nh.2017.200, 2017. Ding, J. and Haberlandt, U.: Estimation of instantaneous peak flow from maximum mean daily flow by regionalization of catchment model parameters, Hydrol Process, 31, 612-626, doi:10.1002/hyp.11053, 2017. anuti, P. and Moisello, U.: Relationship between the Yearly Maxima of Peak and Daily Discha Canuti, P. and Moisello, U.: Relationship between the Yearly Maxima of Peak and Daily Discharge for Some Basins in Tuscany, Hydrolog Sci J, 27, 111-128, doi:10.1080/02626668209491094, 1982. Chen, B., Krajewski, W. F., Liu, F., Fang, W. H., and Xu, Z. X.: Estimating instantaneous peak flow from mean daily flow, Hydrol Res, 48, 1474-1488, doi: 10.2166/nh.2017.200, 2017. Ding, J. and Haberlandt, U.: Estimation of instantaneous peak flow from maximum mean daily flow by regionalization of catchment model parameters, Hydrol Process, 31, 612-626, doi:10.1002/hyp.11053, 2017. Ding, J., Haberlandt, U., and Dietrich, J.: Estimation of the instantaneous peak flow from maximum daily flow: a comparison 515 of three methods, Hydrol Res, 46, 671-688, doi:10.2166/nh.2014.085, 2015. Ding, J., Wallner, M., Muller, H., and Haberlandt, U.: Estimation of instantaneous peak flows from maximum mean daily flows using the HBV hydrological model, Hydrol Process, 30, 1431-1448, doi:10.1002/hyp.10725, 2016. References Deutscher Wetterdienst (DWD): Vieljährige Mittelwerte Ding, J., Haberlandt, U., and Dietrich, J.: Estimation of the instantaneous peak flow from maximum daily flow: a comparison 515 of three methods, Hydrol Res, 46, 671-688, doi:10.2166/nh.2014.085, 2015. Ding J Wallner M Muller H and Haberlandt U : Estimation of instantaneous peak flows from maximum mean daily of three methods, Hydrol Res, 46, 671 688, doi:10.2166/nh.2014.085, 2015. Ding, J., Wallner, M., Muller, H., and Haberlandt, U.: Estimation of instantaneous peak flows from maximum mean daily flows using the HBV hydrological model, Hydrol Process, 30, 1431-1448, doi:10.1002/hyp.10725, 2016. h di ( ) i lj h i i l Deutscher Wetterdienst (DWD): Vieljährige Tan, K.-S., Chiew, F. H. S. and Grayson, R. B.: A steepness index unit volume flood hydrograph approach for sub-daily flow disaggregation, Hydrological Processes, 21, 2807–2816, doi: 10.1002/hyp.6501, 2007. 555 Viglione, A. and Blöschl, G.: On the role of storm duration in the mapping of rainfall to flood return periods, Hydrol Earth Syst Sc, 13 (2), 205–216, doi:10.5194/hess-13-205–200, 2009. (DWD): Gaál, L., Szolgay, J., Kohnová, S., Hlavčová, K., Parajka, J., Viglione, A., Merz, R., and Blöschl, G.: Dependence between 530 flood peaks and volumes: a case study on climate and hydrological controls, Hydrolog Sc J, 60, 968-984, doi: 10.1080/02626667.2014.951361, 2015. Jarvis, A., Reuter, H. I., Nelson, A. and Guevara, E.: Hole-filled seamless SRTM data V4, International Centre for Tropical Agriculture (CIAT), 2008. Gaál, L., Szolgay, J., Kohnová, S., Hlavčová, K., Parajka, J., Viglione, A., Merz, R., and Blöschl, G.: Dependence between 530 flood peaks and volumes: a case study on climate and hydrological controls, Hydrolog Sc J, 60, 968-984, doi: 10.1080/02626667.2014.951361, 2015. flood peaks and volumes: a case study on climate and hydrological controls, Hydrolog Sc J, 60, 968-984, doi: 10.1080/02626667.2014.951361, 2015. Jarvis, A., Reuter, H. I., Nelson, A. and Guevara, E.: Hole-filled seamless SRTM data V4, International Centre for Tropical Agriculture (CIAT), 2008. Jarvis, A., Reuter, H. I., Nelson, A. and Guevara, E.: Hole-filled seamless SRTM data V4, International Centre for Tropical Agriculture (CIAT), 2008. Kumar, D. N., Lall, U., and Petersen, M. R.: Multisite disaggregation of monthly to daily streamflow, Water Resour Res, 36, 535 1823-1833, doi:10.1029/2000WR900049, 2000. Kumar, D. N., Lall, U., and Petersen, M. R.: Multisite disaggregation of monthly to daily streamflow, Water Resour Res, 36, 535 1823-1833, doi:10.1029/2000WR900049, 2000. Langbein, W. B.: Peak discharge from daily records, Water Resour Bull, pp. 145, 1944. 24 https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. https://doi.org/10.5194/hess-2021-466 Preprint. Discussion started: 17 September 2021 c⃝Author(s) 2021. CC BY 4.0 License. Muñoz, E., Arumí, J. L. and Vargas, J.: A design peak flow estimation method for medium-large and data-scarce watersheds with frontal rainfall, J Am Water Resour Assoc, 48 (3), 439–448, doi: 10.1111/j.1752-1688.2011.00622.x, 2012. Sangal, B. P.: Practical Method of Estimating Peak Flow, J Hydraul Eng-Asce, 109, 549-563, doi: 10.1061/(ASCE)0733- 540 9429(1983)109:4(549), 1983. Stedinger, J. R. and Vogel, R. M.: Disaggregation Procedures for Generating Serially Correlated Flow Vectors, Water Resour Res, 20, 47-56, doi: 10.1029/WR020i001p00047, 1984. Taguas, E. V., Ayuso, J. L., Pena, A., Yuan, Y., Sanchez, M. C., Giraldez, J. V., and Perez, R.: Testing the relationship between Sangal, B. P.: Practical Method of Estimating Peak Flow, J Hydraul Eng-Asce, 109, 549-563, doi: 10.1061/(ASCE)0733- 0 9429(1983)109:4(549), 1983. Stedinger, J. R. and Vogel, R. M.: Disaggregation Procedures for Generating Serially Correlated Flow Vectors, Water Resour Res, 20, 47-56, doi: 10.1029/WR020i001p00047, 1984. Taguas, E. V., Ayuso, J. (DWD): L., Pena, A., Yuan, Y., Sanchez, M. C., Giraldez, J. V., and Perez, R.: Testing the relationship between instantaneous peak flow and mean daily flow in a Mediterranean Area Southeast Spain, Catena, 75, 129-137, doi: 545 10.1016/j.catena.2008.04.015, 2008. Tarasova, L., Basso, S., Wendi, D., Viglione, A., Kumar, R., and Merz, R.: A Process-Based Framework to Characterize and Classify Runoff Events: The Event Typology of Germany, Water Resour Res, 56, doi:10.1029/2019WR026951, 2020. Tarasova, L., Basso, S., Zink, M., and Merz, R.: Exploring Controls on Rainfall-Runoff Events: 1. Time Series- Based Event Tarasova, L., Basso, S., Wendi, D., Viglione, A., Kumar, R., and Merz, R.: A Process-Based Framework to Characterize and Classify Runoff Events: The Event Typology of Germany, Water Resour Res, 56, doi:10.1029/2019WR026951, 2020. ova, L., Basso, S., Zink, M., and Merz, R.: Exploring Controls on Rainfall-Runoff Events: 1. Time Separation and Temporal Dynamics of Event Runoff Response in Germany, Water Resour Res, 54, 7711-7732, doi: 550 10.1029/2018WR022587, 2018. Tarboton, D. G., Sharma, A., and Lall, U.: Disaggregation procedures for stochastic hydrology based on nonparametric density estimation, Water Resour Res, 34, 107-119, doi: 10.1029/97WR02429, 1998. Tan, K.-S., Chiew, F. H. S. and Grayson, R. B.: A steepness index unit volume flood hydrograph approach for sub-daily flow Tarboton, D. G., Sharma, A., and Lall, U.: Disaggregation procedures for stochastic hydrology based on nonparametric density estimation, Water Resour Res, 34, 107-119, doi: 10.1029/97WR02429, 1998. disaggregation, Hydrological Processes, 21, 2807–2816, doi: 10.1002/hyp.6501, 2007. 555 Viglione, A. and Blöschl, G.: On the role of storm duration in the mapping of rainfall to flood return periods, Hydrol Earth Syst Sc, 13 (2), 205–216, doi:10.5194/hess-13-205–200, 2009. 560 25
https://openalex.org/W2955310308	http://bura.brunel.ac.uk/bitstream/2438/21589/2/FullText.pdf	English	null	Assessing Statistical Anxiety Among Online and Traditional Students	Frontiers in psychology	2,019	cc-by	6,820	ORIGINAL RESEARCH published: 04 July 2019 doi: 10.3389/fpsyg.2019.01440 Assessing Statistical Anxiety Among Online and Traditional Students Marta Frey-Clark1, Prathiba Natesan1* and Monique O’Bryant 2 Marta Frey-Clark1, Prathiba Natesan1* and Monique O’Bryant 2 1Educational Psychology, University of North Texas, Denton, TX, United States, 2Atlanta Public Schools, Atlanta, GA, United States 1Educational Psychology, University of North Texas, Denton, TX, United States, 2Atlanta Public Schools, Atlanta, GA, United States The purpose of this study was to determine whether scores on the Statistical Anxiety Scale (SAS) manifest in the same way for students in online and traditional statistics courses. Tests of measurement invariance indicated that invariance of the two-factor model of the SAS held at every level. Therefore, we compared the statistical anxiety of online and traditional students. Results indicated that online and traditional statistics students reported comparable levels of anxiety with slightly less anxiety in terms of seeking help for traditional students. We concluded that online instruction is a viable form of statistics education at least for undergraduate students enrolled in the social sciences. Keywords: statistical anxiety, online education, measurement invariance, statistics education, validity Specialty section: This article was submitted to Quantitative Psychology and Measurement, a section of the journal Frontiers in Psychology Specialty section: This article was submitted to Quantitative Psychology and Measurement, a section of the journal Frontiers in Psychology Received: 17 March 2019 Accepted: 04 June 2019 Published: 04 July 2019 Edited by: Laura Badenes-Ribera, University of Valencia, Spain Edited by: Laura Badenes-Ribera, University of Valencia, Spain Participation in online education has grown rapidly over the past 15 years and is expected to continue growing (Allen and Seaman, 2010). In fact, the New York Times declared the year 2012 as the “year of the MOOC” (massive open online courses, Pappano, 2012). In Fall 2015, 29.8% of the students were enrolled online in postsecondary institutions (NCES, 2015). The online learning consortium report further shows how in addition to education, professional development, and other related sources of knowledge have moved digitally (OLC Report, 2018). Indeed, online courses seem to offer distinct advantages, with being a more convenient and cost-effective alternative to traditional, face-to-face instruction. Researchers have worked to keep pace with the growth in online learning, comparing learning outcomes for students enrolled in online courses with those of students enrolled in traditional courses.if Reviewed by: Caterina Primi, University of Florence, Italy Thomas A. DeVaney, Southeastern Louisiana University, United States Reviewed by: Caterina Primi, University of Florence, Italy Thomas A. DeVaney, Southeastern Louisiana University, United States *Correspondence: Prathiba Natesan prathiba.natesan@unt.edu Although several meta-analyses have shown that there was no statistically significant difference between instruction employing technology and traditional instruction (Cavanaugh et al., 2004; Zhao et al., 2005; Jahng et al., 2007), other meta-analyses have found a statistically significant difference between online and traditional instruction (Shachar and Neumann, 2003; Allen and Seaman, 2004; Bernard et al., 2004; Sitzmann et al., 2006; Williams, 2006). In fact, students with low GPAs tend to withdraw more from an online course than from a traditional course and online students tend to persist less in their programs to attain a degree (Jaggars et al., 2013). Jaggars (2014) also reported that students reported having to “teach themselves” in an online class. With respect to performance although there was a statistically significant relationship between course format (online vs. traditional) and failure in the course for English and Math courses, this was not the case for Economics and Humanities courses (Griffiths et al., 2014). Thus, it seems that there is a difference in the relationship between student performance and course format by subject matter. Citation: By contrast, Scherrer (2011) found that when GPA, class format, and student major were included in a regression equation, class format was a statistically significant predictor of final grades, with traditional students outperforming online students. are identical across groups. Finally, an error variance invariance model is used to test if the factor structure, loadings, item intercepts, and item error variances are identical across groups. Factor means and variances may be compared only when all these levels of invariance are established. Lack of measurement invariance indicates that group-specific attributes unrelated to the latent constructs contaminate the way a person belonging to a group responds to an item (Meredith, 1993; Little, 1997). In other words, a lack of measurement invariance means that given the same factor score, individuals from different groups will have respond differently to a given item. Thus comparisons of factor scores, means, and variances in such a situation are invalid. Frontiers in Psychology \| www.frontiersin.org MEASURING STATISTICAL ANXIETY Upon establishing metric invariance, a scalar invariance model is used to test if the factor structure, loadings, and item intercepts DeVaney (2010) administered a statistical anxiety pretest and posttest to traditional and online graduate students, reporting that online students had higher anxiety at the beginning of the course, but there was no difference in student anxiety at the end of the course. However, DeVaney’s research operated on the assumption that measurement instrument operationalized statistical anxiety in the same way for online and traditional students. Given that previous research has identified situational antecedents to statistical anxiety (Onwuegbuzie and Wilson, 2003), it would seem that the distinct environments of traditional and online students may lead to distinct operationalization of the construct. Thus, a test of measurement invariance is a necessary foundation for future research before comparisons across traditional and online student groups can be conducted. Vigil-Colet et al. (2008) administered a Spanish version of the SAS to a sample of undergraduate students (n = 159) enrolled in statistics courses in Spain. An Exploratory Factor Analysis (EFA) verified the intended three-factor structure, with each item loading on its intended subscale. Shortly after the development and validation of the Spanish version of the SAS, Chiesi et al. (2011) administered an Italian version of the SAS to a sample of students (n = 512). A confirmatory factor analysis (CFA) confirmed the previously validated three- factor model, with the addition of correlated errors between two similarly phrased items on the Asking for Help subscale. Chiesi et al. (2011) also conducted measurement invariance tests across samples of Italian and Spanish students and reported that strict invariance of the modified three-factor model was tenable across both samples. Measurement invariance tests the equivalence of constructs across groups along four prescribed levels (see Mellenbergh, 1989; Meredith, 1993; Vandenberg and Lance, 2000). A configural invariance model is used to test if the factor structure is defined identically across groups. Once this is established, a metric or factorial invariance model tests the equivalence of factor loadings across groups in addition to identical factor structure. Upon establishing metric invariance, a scalar invariance model is used to test if the factor structure, loadings, and item intercepts Following the validation of the three-factor Spanish SAS (Vigil-Colet et al., 2008) as well as the Italian SAS (Chiesi et al., 2011), O’Bryant (2017) investigated the factor structure of the English version of the SAS. Citation: Frey-Clark M, Natesan P and O’Bryant M (2019) Assessing Statistical Anxiety Among Online and Traditional Students. Front. Psychol. 10:1440. doi: 10.3389/fpsyg.2019.01440 Given the prevalence of anxiety in statistics courses that are perceived to be challenging, several researchers have compared performance outcomes for students enrolled in online and traditional statistics courses. Some authors have reported no difference between the two class July 2019 \| Volume 10 \| Article 1440 1 Frontiers in Psychology \| www.frontiersin.org Frey-Clark et al. Statistical Anxiety formats (McLaren, 2004; Dotterweich and Rochelle, 2012), while one study found a difference favoring traditional instruction (Scherrer, 2011). McLaren (2004) found no statistically significant difference in the grades earned by online and traditional statistics students who completed their course; however, the researcher did find that online students demonstrated a greater tendency to drop the course or “vanish,” failing to take part in assignments and exams despite remaining on the roster. Similarly, Dotterweich and Rochelle (2012) found that students enrolled in online, traditional, and televised instruction statistics courses earned similar grades; however, when the researchers isolated students who were repeating the course, they found statistically significant differences in performance favoring traditional students. By contrast, Scherrer (2011) found that when GPA, class format, and student major were included in a regression equation, class format was a statistically significant predictor of final grades, with traditional students outperforming online students. Despite a growing body of literature comparing the performance of online and traditional statistics students, there remains a dearth of research comparing the statistical anxiety of online and traditional statistics students. Statistical anxiety is defined as “feelings of anxiety encountered when taking a statistics course or doing statistical analysis; that is, gathering, processing and interpreting data” (Cruise et al., 1985, p. 92). Statistical anxiety is a well-documented reality for statistics students (Onwuegbuzie et al., 2010; Chew and Dillon, 2014), and high statistical anxiety has consistently been associated with lower performance outcomes (Bell, 2001, 2003; Onwuegbuzie, 2004; Galli et al., 2008; Macher et al., 2012). In light of the mixed findings regarding the performance of traditional and online statistics students, as well as the documented relationship between statistics anxiety and statistics performance, it may be useful to examine the relationship between statistics anxiety and class format. formats (McLaren, 2004; Dotterweich and Rochelle, 2012), while one study found a difference favoring traditional instruction (Scherrer, 2011). Citation: McLaren (2004) found no statistically significant difference in the grades earned by online and traditional statistics students who completed their course; however, the researcher did find that online students demonstrated a greater tendency to drop the course or “vanish,” failing to take part in assignments and exams despite remaining on the roster. Similarly, Dotterweich and Rochelle (2012) found that students enrolled in online, traditional, and televised instruction statistics courses earned similar grades; however, when the researchers isolated students who were repeating the course, they found statistically significant differences in performance favoring traditional students. By contrast, Scherrer (2011) found that when GPA, class format, and student major were included in a regression equation, class format was a statistically significant predictor of final grades, with traditional students outperforming online students. Despite a growing body of literature comparing the performance of online and traditional statistics students, there remains a dearth of research comparing the statistical anxiety of online and traditional statistics students. Statistical anxiety is defined as “feelings of anxiety encountered when taking a statistics course or doing statistical analysis; that is, gathering, processing and interpreting data” (Cruise et al., 1985, p. 92). Statistical anxiety is a well-documented reality for statistics students (Onwuegbuzie et al., 2010; Chew and Dillon, 2014), and high statistical anxiety has consistently been associated with lower performance outcomes (Bell, 2001, 2003; Onwuegbuzie, 2004; Galli et al., 2008; Macher et al., 2012). In light of the mixed findings regarding the performance of traditional and online statistics students, as well as the documented relationship between statistics anxiety and statistics performance, it may be useful to examine the relationship between statistics anxiety and class format formats (McLaren, 2004; Dotterweich and Rochelle, 2012), while one study found a difference favoring traditional instruction (Scherrer, 2011). McLaren (2004) found no statistically significant difference in the grades earned by online and traditional statistics students who completed their course; however, the researcher did find that online students demonstrated a greater tendency to drop the course or “vanish,” failing to take part in assignments and exams despite remaining on the roster. Similarly, Dotterweich and Rochelle (2012) found that students enrolled in online, traditional, and televised instruction statistics courses earned similar grades; however, when the researchers isolated students who were repeating the course, they found statistically significant differences in performance favoring traditional students. MEASURING STATISTICAL ANXIETY In a review of literature on statistical anxiety, Chew and Dillon (2014) identified six extant scales, but the authors only recommended use of the Statistics Anxiety Rating Scale, or STARS (Cruise et al., 1985), and its abbreviated alternative, the Statistical Anxiety Scale, or SAS (Vigil-Colet et al., 2008). The STARS is the most widely used and well-known scale (Chew and Dillon, 2014). However, Vigil-Colet et al. (2008) criticized the STARS for its length and some of its content, which prompted their development of the SAS. The SAS has 24 items and is comprised of three subscales derived from the STARS anxiety subscales: Examination Anxiety (eight items), Interpretation Anxiety (eight items), and Asking for Help Anxiety (eight items). Examination Anxiety refers to anxiety experienced while taking a statistics test. Interpretation Anxiety refers to anxiety experienced while attempting to derive meaning from statistical formulas and output. Asking for Help Anxiety refers to anxiety experienced while requesting help of a peer, a tutor, or a professor. Each item of the SAS details a specific task, prompting respondents to indicate the level of anxiety associated with the task on a 5-point Likert-type scale ranging between no anxiety and very much anxiety. between statistics anxiety and class format. DeVaney (2010) administered a statistical anxiety pretest and posttest to traditional and online graduate students, reporting that online students had higher anxiety at the beginning of the course, but there was no difference in student anxiety at the end of the course. However, DeVaney’s research operated on the assumption that measurement instrument operationalized statistical anxiety in the same way for online and traditional students. Given that previous research has identified situational antecedents to statistical anxiety (Onwuegbuzie and Wilson, 2003), it would seem that the distinct environments of traditional and online students may lead to distinct operationalization of the construct. Thus, a test of measurement invariance is a necessary foundation for future research before comparisons across traditional and online student groups can be conducted. Measurement invariance tests the equivalence of constructs across groups along four prescribed levels (see Mellenbergh, 1989; Meredith, 1993; Vandenberg and Lance, 2000). A configural invariance model is used to test if the factor structure is defined identically across groups. Once this is established, a metric or factorial invariance model tests the equivalence of factor loadings across groups in addition to identical factor structure. MEASURING STATISTICAL ANXIETY After pilot-testing, she July 2019 \| Volume 10 \| Article 1440 2 Frey-Clark et al. Statistical Anxiety modified the items thus: Many revisions involved changing one word such as replacing doing to completing in items such as doing a final exam in a statistics course to completing a final exam in a statistics course. Other examples of changes included changing the word tutor to teacher to reflect the teaching system and terminology in the United States. O’Bryant administered the English version of the SAS to a sample of undergraduate students (n = 323) majoring in the humanities and enrolled in statistics courses throughout the United States. A CFA of the previously validated three-factor model indicated poor model fit ( cSB 2 = 153.46, df = 71.12, p < 0.001, RMSEA = 0.106, CFI = 0.838, SRMR = 0.073). Examination of residual correlations revealed that the residuals of the seven items on the Interpretation subscale were highly correlated with those of the items within the subscale, as well as with items on the other two subscales. Thus, O’Bryant (2017) eliminated the Interpretation subscale from the model. Eliminating the interpretation factor was not only warranted according to factor analytic output, but also seemed conceptually justifiable, given that taking an exam and asking for help are discrete tasks while interpreting numbers is not. comparison purposes. Data were collected online using qualtrics. Informed consent was obtained from participants who were all 18 years of age or above by asking them to click on a page that explained the study, the duration of the survey, and letting them know of the anonymity that would be maintained with the data. If they agreed to participate they could continue answering the questions by clicking on an appropriate button, else they could exit the survey. Participants were undergraduate students (n = 323) who were majoring in the social sciences and were enrolled in a statistics course. However, data screening revealed that 21 respondents took an online-traditional hybrid course, and seven respondents did not indicate their class format. Because we were only interested in online and traditional groups students, and the hybrid group was too small for analysis, these cases were dropped from the dataset, leaving 295 cases with online (n = 52) and traditional (n = 243) students. Reliability Internal consistency of the modified two-factor SAS was measured with Cronbach’s α for each class format. The α coefficients for the online class format were as follows: Total = 0.903, Exam Anxiety Subscale = 0.903, and Asking for Help Anxiety Subscale = 0.880. The α coefficients for the traditional class format were as follows: Total = 0.914, Exam Anxiety Subscale = 0.886, and Asking for Help Anxiety Subscale = 0.922. The entirety of the modified two-factor SAS and its subscales were deemed to have high internal consistent for each class format (Nunnally, 1978; Nunnally and Bernstein, 1994). McDonald’s (1999) omega was computed to be 0.94 for the online class format and 0.84 for traditional class format. MEASURING STATISTICAL ANXIETY Respondents in the final dataset were predominantly female (75%), predominantly white (59%), and predominantly freshman (38%), with ages ranging from 18 to 63 years (M = 20.64, SD = 5.37). p g Further examination of residual correlations revealed that one item on the Examination Anxiety subscale and one item on the Asking for Help subscale could be eliminated due to redundancy with other items. Finally, the residuals for four items (items 1, 4, 13, and 20) on the Examination Anxiety scale were allowed to correlate, given the similarity in their wording. The resulting model had two factors, Examination Anxiety and Asking for Help Anxiety, with seven items loading on each factor and correlated errors for four items on the Examination Anxiety factor. This modified two-factor model fit the data well ( cSB 2 = 49.37, df = 38.13, p = 0.105, RMSEA = 0.076, CFI = 0.959, SRMR = 0.035) and was retained. We extend O’Bryant (2017) validation study to validating the factors across the online and traditional samples using measurement invariance.h Screeningh The data were screened for outliers, assumptions of normality, and missing values prior to analysis. There were no outliers identified. Examination of frequency data on each item revealed severely peaked distributions, indicating that scores on the 5-point Likert-type scale were ordinal; thus, all subsequent analyses utilized non-parametric tests. Frequency data for missing values revealed a somewhat consistent distribution of missing data, with 0.3–4.7% missing per variable. Given the small percentage missing per variable and the spread of missingness across variables, data were assumed to be missing completely at random (MCAR) and were estimated via Mplus’ default estimation for ordinal outcomes with covariates, making use of all available data to estimate missing values. The purpose of the present study is to determine whether scores on O’Bryant (2017) modified two-factor model of statistical anxiety are operationalized in the same way for traditional and online statistics students. If measurement invariance is established, an additional purpose of the present study is to compare the latent scores on the Exam Anxiety subscale and the Asking for Help Anxiety subscale for online and traditional students. MATERIALS AND METHODS Institutional Review Board of the University of North Texas approved the study. A two-stage sampling procedure was used. First, simple random sampling without replacement was used to randomly select institutions with social science programs to participate in the study. Second, network sampling was used to ask instructors of statistics for social science courses to pass along the research opportunity to their students. The goal was to recruit participants similar to those used in previous validation studies (Vigil-Colet et al., 2008; Chiesi et al., 2011) for Frontiers in Psychology \| www.frontiersin.org Invariance Testing Finally, Model D, was used to test strict or error variance invariance by fixing all error variances to 1. This test deviated again from invariance testing with interval level data, in which strict invariance is established by constraining the error variances. Recall that the latent response indicators were scaled via theta parameterization, fixing each variance to 1 in the reference group. Thus, strict invariance was tested by fixing the latent indicator variances to 1 in both groups. Again, model fit was tenable. The scaled χ2 difference test reported a statistically significant difference in fit compared with Model C. However, Chen’s (2007) criteria for assessing differences in model fit using CFI and RMSEA did not indicate appreciably worse fit. Model D was retained, and we concluded that the SAS measures the statistical anxiety of students in online and traditional statistics classes identically. See Table 1 for overall and comparative fit indices.h Analysis began with a confirmatory factor analysis (CFA) for each group, confirming that the O’Bryant (2017) modified two-factor model adequately fit the online group and the traditional group individually. Therefore, measurement invariance was testing by first fitting Model A that is the configural invariance model by fixing the factor structure to be identical across groups. Goodness of fit indices and approximate fit indices were tenable, indicating that the factor structure was the same for each group. i The unstandardized estimates of Model D for both groups are displayed in Figure 1. We note that we report unstandardized estimates because these are comparable across groups of different sample sizes. Standardized factor loadings for the online group ranged from 0.682 to 0.856; all were statistically significant at the 0.001 level. The correlation between the exam factor and help factor for the online group was 0.554, indicating the factors were related but distinct. Standardized factor loadings for the traditional group ranged from 0.659 to 0.886; again, all loadings were statistically significant at the 0.001 level. The correlation between the exam factor and help factor was 0.591, again indicating the factors were related but distinct. Model B, that is, the metric invariance model, was fitted by retaining the factor structure of Model A and adding constraints on all factor loadings to be equal across groups. Invariance Testing We used Mplus version 7.6 with means and variance adjusted weighted least squares (WLSMV) estimation to test the July 2019 \| Volume 10 \| Article 1440 Frontiers in Psychology \| www.frontiersin.org 3 Frey-Clark et al. Statistical Anxiety by WLSMV estimation are corrected for ordinal level data. As such, the χ2 difference tests for nested models were also corrected by way of the DIFFTEST option in Mplus.i measurement invariance of the SAS for online and traditional statistics students. WLSMV is a robust weighted least squares estimator that has been recommended for ordinal level data with a sample size greater than 200 (Muthén et al., 1997, unpublished; Rhemtulla et al., 2012). Because the data were ordinal, WLSMV calculates threshold parameters for each response variable to estimate the latent, continuous response indicators that correspond with each item of the SAS. Response indicators were scaled via theta parameterization, fixing the variance of each latent indicator to 1 in the reference group.f Model C that is, the scalar invariance was fitted by retaining constraints on factor loadings and adding constraints on item thresholds. For interval level data, testing scalar invariance would involve constraining item intercepts. However, recall that scores on items from the SAS were deemed ordinal; as such, thresholds for response options determine scores on a latent response variable, which indicates the latent factor. Thus, scalar invariance requires each threshold for each indicator to be equal across groups. Fit indices for Model C were tenable, and the fit was not appreciably worse than Model B. Therefore, the scalar invariance model was retained. g When comparing nested models, we used χ2 difference tests to evaluate between-model statistical significance, with a statistically significant result indicating non-invariance across models. However, given the sensitivity of χ2 to sample size, an a priori decision was made to supplement the χ2 model testing parameters with differences in the Comparative Fit Index (CFI) and the Root Mean Square Error of Approximation (RMSEA), per Chen’s (2007) criteria. Thus, the criteria for rejecting model invariance included the joint decision rules of (1) a statistically significant χ2 difference (p < 0.05); (2) a change in RMSEA ≥ −0.005; and (3) a change in CFI ≤ 0.010. Note that Chen’s (2007) criteria for a change in Standardized Root Mean Square Residual (SRMR) were not included because Mplus does not calculate SRMR when using WLSMV estimation to evaluate a model with covariates. July 2019 \| Volume 10 \| Article 1440 DISCUSSION The purpose of the present study was to determine whether the operationalization of statistical anxiety via the modified two-factor Statistical Anxiety Scale is the same for samples of online students and traditional students. Previous research has indicated that online statistics students may represent a distinct demographic, being older, with more credit hours earned and more courses repeated than their traditional counterparts (Dotterweich and Rochelle, 2012). Previous research has also indicated online students may possess different intellectual strengths, having higher logical-mathematical intelligence than their traditional counterparts (Lopez and Patron, 2012). If the two populations differ with respect to demographic characteristics and intellectual strengths, it may seem probable that they could differ with respect to the manner in which they report statistical anxiety. However, this was not the case.i Std, Standardized; Unstd, Unstandardized. Invariance Testing Model fit was tenable and was not statistically significantly different from Model A, indicating that the Exam Anxiety factor and Asking for Help Anxiety factor were manifested in the same way across groups. That is, the relationships between these factors and the items that indicate them were identical across online and traditional statistics class formats. Note that the χ2 values produced TABLE 1 \| Values of selected fit statistics for measurement invariance hypotheses for modified two-factor model of statistics anxiety analyzed across online and traditional student samples. Model Model name c 2 SB df Model comparison RMSEA CFI Online 89.144 71 0.07 [0.0, 0.012] 0.983 Traditional 99.212 0.04 [0.018, 0.058] 0.995 Model A 185.71 142 0.046 [0.024, 0.053] 0.993 Model B Metric invariance 198.718 154 B vs. A 0.044 [0.023, 0.061] 0.993 Model C Scalar invariance 239.053 194 C vs. B 0.04 [0.019, 0.056] 0.993 Model D Error variance invariance 261.041 208 D vs. C 0.042 [0.023, 0.057] 0.992 CI, confidence interval. All results were computed in Mplus for theta parameterization. TABLE 1 \| Values of selected fit statistics for measurement invariance hypotheses for modified two-factor model of statistics anxie traditional student samples. 4 Frey-Clark et al. Statistical Anxiety FIGURE 1 \| Unstandardized estimates for traditional and online groups. FIGURE 1 \| Unstandardized estimates for traditional and online groups. FIGURE 1 \| Unstandardized estimates for traditional and online groups. TABLE 2 \| Robust weighted least squares estimates of unconstrained parameters for Model D of statistics anxiety analyzed across online and traditional student samples. and Asking for Help Anxiety (d = −0.129) that would be considered a very small effect (Cohen, 1988). Thus, we concluded that online statistics students expressed comparable levels of statistical exam anxiety, but slightly higher levels of asking for help anxiety than traditional statistics students. Online Traditional Parameter Unstd SE Std Unstd SE Std Exam factor Variance 0.87 0.25 1 0.77 0.17 1 Mean 0 – 0 0.05 0.15 0.05 Question factor Variance 1.57 0.41 1 2.15 0.35 1 Mean 0 – 0 −0.18 0.21 −0.13 Factor covariance 0.65 0.17 0.55 0.76 0.14 0.59 Std, Standardized; Unstd, Unstandardized. July 2019 \| Volume 10 \| Article 1440 Frontiers in Psychology \| www.frontiersin.org Differences in Statistical Anxieties Having established the measurement invariance of the modified two-factor SAS for online and traditional students, analysis proceeded with the primary purpose of this study: determining by how much the two groups differed in their average scores on the Exam Anxiety subscale and the Asking for Help Anxiety subscale. See Table 2 for the variances and means of each factor for each group. Note that the online group served as the reference group and its factor means were fixed to 0. As such, the factor means listed for the traditional group represent mean differences across groups. The mean difference in Exam Anxiety was 0.048, with online students indicating lower Exam Anxiety. The mean difference in Asking for Help Anxiety was 0.184, with online students indicating higher Asking for Help Anxiety. Cohen’s d effect sizes were calculated for both mean differences, revealing effect sizes for Exam Anxiety (d = 0.054) Invariance held at every level, indicating that the modified two-factor SAS measures statistical anxiety manifests in the same way for online and traditional statistics students. These findings are further strengthened by the fact that the sample for the present study was drawn via random cluster sampling of colleges and universities throughout the United States. Thus, the SAS would appear to be a versatile measure of statistical anxiety. This finding answers Chew and Dillon’s (2014) call Frontiers in Psychology \| www.frontiersin.org 5 Frey-Clark et al. Statistical Anxiety to confirm the factor structure of the SAS with diverse samples and provides a foundation for future research using the SAS with classes of varied formats.i to confirm the factor structure of the SAS with diverse samples and provides a foundation for future research using the SAS with classes of varied formats.i A major limitation of the present study is its small sample size. It is recommended that this study be repeated for larger samples so as to address the generalizability of the study. Perhaps administering a pre- and post-survey to examine statistics anxiety before and after taking traditional and online courses is another avenue for future research. Future research might seek to clarify the relationship between class format, statistical anxiety, and performance outcomes. ETHICS STATEMENT The institutional review board of the university of North Texas approved this study. Informed consent was obtained from participants before they answered the survey. Vulnerable populations were not involved. Institutions of higher learning have reported offering online courses in the interest of meeting student demand for flexible scheduling, providing college access to students who may not otherwise have access, making courses more available, and seeking to increase student enrollment (Parsad and Lewis, 2008). As a convenient class format for students, and a cost- effective class format for institutions of higher learning, capitalizing on the pragmatic advantages of online education may allow a greater number of students to access statistics education, and a greater number of institutions to offer statistics education. AUTHOR CONTRIBUTIONS MF-C conducted the data analysis and literature review. PN oversaw the project and added conclusion and introduction. MOB collected the data, came up with the instrument, and helped with literature review. Differences in Statistical Anxieties Given the established relationship between statistical anxiety and performance outcomes (e.g., Galli et al., 2008), and the conflicting findings regarding the relationship of class format to performance outcomes (e.g., Scherrer, 2011; Dotterweich and Rochelle, 2012), there exists the possibility that class format and statistical anxiety interact to influence performance outcomes. Examination of all three variables in context may serve to clarify their relationships and inform future instruction. Regardless, insofar as the present study stands, online and traditional statistics students experience similar levels of anxiety, indicating that online instruction is a viable means of delivering statistics education. Given that the modified two-factor model of the SAS is comprised of only 14 items, and scores on these items are valid for both online and traditional students, statistics instructors may consider administering this instrument to students in order to gauge anxiety and adjust instruction accordingly. Researchers have identified a number of effective interventions, including the use of humor (Pan and Tang, 2004), problem- solving games (D’Andrea and Waters, 2002), and instructor immediacy (Williams, 2006). Thus, the SAS could serve as a diagnostic tool, presenting instructors with student feedback to inform instruction. An added purpose of this study was to compare mean scores for Exam Anxiety and Asking for Help Anxiety across class formats. Effect size estimates revealed that mean differences were negligible for exam anxiety and a lower asking for help anxiety for traditional students. This is contrary to popular belief that students have lesser inhibitions in reaching out for help when they are learning within the relative privacy and social safety of online education. However, the effect size is too small to make conclusions regarding these differences.i DATA AVAILABILITY The datasets for this manuscript are not publicly available because the dataset is part of the MOB’s thesis. Covariance matrix may be provided upon request. But the data are subject to confidentiality agreement according to informed consent. Requests to access the datasets should be directed to monique_obryant@yahoo.com. f Our findings lend additional support to DeVaney’s (2010) finding that online and traditional students had comparable levels of anxiety upon completion of an introductory statistics course. Furthermore, DeVaney reported that online students had higher statistical anxiety than traditional students at the beginning of the course. Thus, if online students do not appear to carry greater statistical anxiety, as our study suggests, and if the online class format may even soothe statistical anxiety, as DeVaney’s work suggests, then online statistics education seems to present a viable alternative to traditional, face-to- face instruction. Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Struct. Equ. Model. 14, 464–504. doi: 10.1080/ 10705510701301834 REFERENCES Bernard, R., Brauer, A., Abrami, P., and Surkes, M. (2004). The development of a questionnaire for predicting online learning achievement. Distance Educ. 25, 31–47. doi: 10.1080/0158791042000212440 Allen, I. E., and Seaman, J. (2010). Class differences: Online education in the United States. (Newburyport, MA: Sloan Consortium). Cavanaugh, C., Gillan, K. J., Kromrey, J., Hess, M., and Blomeyer, R. (2004). The effects of distance education on K-12 student outcomes: A meta-analysis. Naperville, IL: Learning Point Associates/North Central Regional Educational Laboratory.i Allen, I. E., and Seaman, J. (2004). Entering the mainstream: the quality and extent of online education in the United States, 2003 and 2004. Newburyport, MA: Sloan Consortium. Bell, J. A. (2001). Length of course and levels of statistics anxiety. Education 121, 713–716. Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Struct. Equ. Model. 14, 464–504. doi: 10.1080/ 10705510701301834 Bell, J. A. (2003). Statistics anxiety: the nontraditional student. Education 124, 157–162. July 2019 \| Volume 10 \| Article 1440 Frontiers in Psychology \| www.frontiersin.org 6 Frey-Clark et al. Statistical Anxiety dissertation, University of North Texas. Available online at https://search. proquest.com/docview/2009455494 dissertation, University of North Texas. Available online at https://search. proquest.com/docview/2009455494 Chew, P. K. H., and Dillon, D. B. (2014). Statistics anxiety update: refining the construct and recommendations for a new research agenda. Perspect. Psychol. Sci. 9, 196–208. doi: 10.1177/1745691613518077 OLC Report (2018). Available at: https://olc-wordpress-assets.s3.amazonaws.com/ uploads/2019/04/OLC-2018-Annual-Report-Online.pdf (Accessed August 15, 2018). Chiesi, F., Primi, C., and Carmona, J. (2011). Measuring statistics anxiety: cross-country validity of the statistical anxiety scale (SAS). J. Psychoeduc. Assess. 29, 559–569. doi: 10.1177/0734282911404985 Onwuegbuzie, A. J. (2004). Academic procrastination and statistics anxiety. Assess. Eval. Higher Educ. 29, 3–19. doi: 10.1080/0260293042000160384 Cohen, J. (1988). Statistical power analysis for behavioral sciences. 2nd edn. (Hillsdale, NJ: Lawrence Earlbaum Associates). Onwuegbuzie, A. J., Leech, N. L., Murtonen, M., and Tähtinen, J. (2010). Utilizing mixed methods in teaching environments to reduce statistics anxiety. Int. J. Multiple Res. App. 4, 28–39. doi: 10.5172/mra.2010.4.1.028 Cruise, R. J., Cash, R. W., and Bolton, D. L. (1985). Development and validation of an instrument to measure statistical anxiety. Proceedings of the American Statistical Association, Section on Statistical Education, Las Vegas, NV. Multiple Res. App. 4, 28–39. doi: 10.5172/mra.2010.4.1.028 Onwuegbuzie, A. J., and Wilson, V. S. (2003). Statistics anxiety: nature, etiology, antecedents, effects, and treatments—a comprehensive review of the literature. Teach. High. Educ. 8, 195–209. doi: 10.1080/1356251032000052447 D’Andrea, L., and Waters, C. (2002). REFERENCES Teaching statistics using short stories: reducing anxiety and changing attitudes. In: Sixth International Conference on Teaching Statistics, Cape Town, South Africa. Pan, W., and Tang, M. (2004). Examining the effectiveness of innovative instructional methods on reducing statistics anxiety for graduate students in the social sciences. J. Instructional Psychol. 31, 149–159.h DeVaney, T. A. (2010). Anxiety and attitude of graduate students in on-campus vs. online statistics courses. J. Stat. Educ. 18. doi: 10.1080/10691898.2010.11889472 Pappano, L. (2012). The year of the MOOC. https://www.nytimes.com/2012/11/04/ education/edlife/massive-open-online-courses-are-multiplying-at-a-rapid-pace. html Dotterweich, D. P., and Rochelle, C. F. (2012). Online, instructional television, and traditional delivery: student characteristics and success factors in business statistics. Am. J. Bus. Educ. 5, 129–138. doi: 10.19030/ajbe.v5i2.6815 Parsad, B., and Lewis, L. (2008). Distance education at degree-granting postsecondary institutions: 2006–2007. First look (NCES 2009–044). (Washington, DC: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics). Galli, S., Ciancaleoni, M., Chiesi, F., and Primi, C. (2008). Who failed the introductory statistics examination? A study on a sample of psychology students. Paper presented at the 11th International Congress on Mathematical Education, Monterrey, Mexico. Rhemtulla, M., Brosseau-Liard, P. E., and Savalei, V. (2012). When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychol. Methods 17, 354–373. doi: 10.1037/a0029315 Griffiths, R., Chingos, M., Mulhern, C., and Spies, R. (2014). Interactive online learning on campus: Testing MOOCs and other hybrid formats in the University System of Maryland. New York: Ithaka S+R. y f y Jaggars, S. S. (2014). Choosing between online and face-to-face courses: community college student voices. Am. J. Dist. Educ. 28, 23–28. doi: 10.1080/ 08923647.2014.867697 Scherrer, C. R. (2011). Comparison of an introductory level undergraduate statistics course taught with traditional, hybrid, and online delivery methods. INFOMRS Trans. Educ. 11, 106–110. doi: 10.1287/ited.1110.0063f Jaggars, S. S., Edgecombe, N., and Stacey, G. W. (2013). What we know about online course outcomes. NY: Community College Research Center. Shachar, M., and Neumann, Y. (2003). Differences between traditional and distance education academic performances: a meta-analytic approach. Int. Rev. Res. Open Dist. Learn. 4, 1–20. doi: 10.19173/irrodl.v4i2.153h Jahng, N., Krug, D., and Zhang, Z. (2007). Student achievement in online distance education compared to face-to-face education. Eur. J. Open Dist. Online Learn. 10. http://www.eurodl.org/materials/contrib/2007/Jahng_Krug_ Zhang.htm Sitzmann, T., Kraiger, K., Steward, D., and Wisher, R. (2006). The comparative effectiveness of web-based and classroom instruction: a meta-analysis. Pers. Psychol. REFERENCES 59, 623–664. doi: 10.1111/j.1744-6570.2006.00049.x Little, T. D. (1997). Mean and covariance structures (MACS) analyses of cross- cultural data: practical and theoretical issues. Multivar. Behav. Res. 32, 53–76. doi: 10.1207/s15327906mbr3201_3 Vandenberg, R. J., and Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: suggestions, practices, and recommendations for organizational research. Organ. Res. Methods 3, 4–70. doi: 10.1177/ 109442810031002 Lopez, S., and Patron, H. (2012). Multiple intelligences in online, hybrid, and traditional business statistics courses. J. Edu. Online 9. doi: 10.9743/JEO.2012.2.2 Vigil-Colet, A., Lorenzo-Seva, U., and Condon, L. (2008). Development and validation of the statistical anxiety scale. Psicothema 20, 174–180. doi: 10.1037/ t62688-000 Macher, D., Paechter, M., Papousek, I., and Ruggeri, K. (2012). Statistics anxiety, trait anxiety, learning behavior, and academic performance. Eur. J. Psychol. Educ. 27, 483–498. doi: 10.1007/s10212-011-0090-5i Williams, S. L. (2006). The effectiveness of distance education in allied health science programs: a meta-analysis of outcomes. Am. J. Dist. Educ. 20, 127–141. doi: 10.1207/s15389286ajde2003_2h McDonald, R. P. (1999). Test theory: A unified treatment. (Mahwah, NJ: Erlbaum Associates). Zhao, H., Seibert, S. E., and Hills, G. E. (2005). The mediating role of self- efficacy in the development of entrepreneurial intentions. J. Appl. Psychol. 90, 1265–1272. doi: 10.1037/0021-9010.90.6.1265 McLaren, C. H. (2004). A comparison of student persistence and performance in online and classroom business statistics experiences. Decis. Sci. J. Innov. Educ. 2, 1–10. doi: 10.1111/j.0011-7315.2004.00015.x Mellenbergh, G. J. (1989). Item bias and item response theory. Int. J. Educ. Res. 13, 127–143. doi: 10.1016/0883-0355(89)90002-5 Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Meredith, W. (1993). Measurement invariance, factor analysis, and factorial invariance. Psychometrika 58, 525–543. doi: 10.1007/BF02294825 NCES. (2015). Available at: https://nces.ed.gov/fastfacts/display.asp?id=80 (Accessed August 15, 2018). Copyright © 2019 Frey-Clark, Natesan and O’Bryant. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2019 Frey-Clark, Natesan and O’Bryant. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). REFERENCES The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Nunnally, J. C. (1978). Psychometric theory. 2nd edn. (NY: McGraw-Hill). Nunnally, J. C., and Bernstein, I. H. (1994). Psychometric theory. 3rd edn. (NY: McGraw-Hill). O’Bryant, M. J. (2017). How attitudes towards statistics courses and the field of statistics predicts statistics anxiety among undergraduate social science majors: a validation of the Statistical Anxiety Scale. ProQuest LLC. Doctoral July 2019 \| Volume 10 \| Article 1440 Frontiers in Psychology \| www.frontiersin.org
https://openalex.org/W4385874889	https://transpublika.co.id/ojs/index.php/Transekonomika/article/download/473/390	Indonesian	null	PENGARUH DANA PIHAK KETIGA, LIKUIDITAS PENDANAAN DAN RISIKO KREDIT TERHADAP PENYALURAN KREDIT BANK SAAT COVID-19 DI INDONESIA	Transekonomika	2,023	cc-by	5,271	Abstrak TRANSEKONOMIKA \| VOLUME 3 NO. 4 (2023) https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 674 Abstrak Salah satu sektor yang terdampak pandemi COVID-19 adalah industri perbankan, di mana pembatasan sosial dan ekonomi yang diberlakukan untuk mengendalikan penyebaran virus telah berdampak pada berbagai aspek operasional bank, terutama penyaluran kredit. Dalam skenario ini, bank menghadapi tantangan baru dalam menjaga likuiditas, mengelola risiko kredit, dan memastikan kelangsungan penyaluran kredit kepada nasabah yang membutuhkan. Oleh karena itu, penelitian ini bertujuan untuk menganalisis sejauh mana faktor-faktor spesifik, seperti dana pihak ketiga, likuiditas pendanaan, dan risiko kredit, mempengaruhi penyaluran kredit oleh bank di Indonesia selama pandemi COVID-19. Metode penelitian yang digunakan adalah pendekatan kuantitatif dengan menggunakan analisis regresi data panel. Pendekatan ini memungkinkan peneliti untuk mengeksplorasi hubungan sebab-akibat antar variabel yang diteliti. Data yang digunakan terdiri dari data sekunder dari bank-bank yang dikategorikan sebagai BUKU III dan BUKU IV, yang mewakili ukuran dan kompleksitas bank tertentu. Data dikumpulkan selama periode tahun 2020-2021 yang penuh ketidakpastian, yang ditandai dengan dampak pandemi. Hasil analisis menunjukkan bahwa dana pihak ketiga berpengaruh positif terhadap penyaluran kredit bank selama pandemi COVID-19. Hal ini mengimplikasikan bahwa semakin besar jumlah dana pihak ketiga yang dimiliki oleh bank, semakin besar pula kemungkinan bank untuk menyalurkan kredit kepada nasabah. Namun, likuiditas pendanaan dan risiko kredit memberikan pengaruh negatif terhadap penyaluran kredit. Hal ini menunjukkan bahwa likuiditas pendanaan yang lebih rendah dan risiko kredit yang lebih tinggi membuat bank lebih sulit untuk memberikan kredit kepada nasabah selama pandemi. Kata Kunci: Dana Pihak Ketiga, Kredit Bank, Likuiditas Pendanaan, Risiko Kredit Salah satu sektor yang terdampak pandemi COVID-19 adalah industri perbankan, di mana pembatasan sosial dan ekonomi yang diberlakukan untuk mengendalikan penyebaran virus telah berdampak pada berbagai aspek operasional bank, terutama penyaluran kredit. Dalam skenario ini, bank menghadapi tantangan baru dalam menjaga likuiditas, mengelola risiko kredit, dan memastikan kelangsungan penyaluran kredit kepada nasabah yang membutuhkan. Oleh karena itu, penelitian ini bertujuan untuk menganalisis sejauh mana faktor-faktor spesifik, seperti dana pihak ketiga, likuiditas pendanaan, dan risiko kredit, mempengaruhi penyaluran kredit oleh bank di Indonesia selama pandemi COVID-19. Metode penelitian yang digunakan adalah pendekatan kuantitatif dengan menggunakan analisis regresi data panel. Pendekatan ini memungkinkan peneliti untuk mengeksplorasi hubungan sebab-akibat antar variabel yang diteliti. Data yang digunakan terdiri dari data sekunder dari bank-bank yang dikategorikan sebagai BUKU III dan BUKU IV, yang mewakili ukuran dan kompleksitas bank tertentu. Data dikumpulkan selama periode tahun 2020-2021 yang penuh ketidakpastian, yang ditandai dengan dampak pandemi. TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan Abstract One of the sectors affected by the COVID-19 pandemic is the banking industry, where social and economic restrictions imposed to control the spread of the virus have impacted various aspects of bank operations, especially lending. In this scenario, banks face new challenges in maintaining liquidity, managing credit risk, and ensuring continuous lending to customers in need. Thus, this study aims to analyze the extent to which specific factors, such as third-party funds, funding liquidity, and credit risk, influence lending by banks in Indonesia during the COVID-19 pandemic. The research method employed is a quantitative approach involving panel data regression analysis. This approach enables researchers to explore the cause-and-effect relationships among the variables under scrutiny. The data utilized consists of secondary data from banks categorized as BUKU III and BUKU IV, representing a certain size and complexity of banks. The data was collected during the uncertain period of 2020-2021, marked by the pandemic's impact. The analysis revealed that third-party funds positively influence banks' lending during the COVID-19 pandemic. This implies that the greater the amount of third-party funds held by banks, the more likely they are to extend credit to customers. However, funding liquidity and credit risk exert a negative influence on lending. This suggests that lower funding liquidity and higher credit risk make it more challenging for banks to provide credit to customers during the pandemic. d k d d k d d h d d Keywords: Bank Credit, Credit Risk, Funding Liquidity, Third Party Funds 1. PENDAHULUAN Sharia Upaya meningkatkan pertumbuhan ekonomi terus di tingkatkan pemerintah Indonesia untuk terus mendekatkan diri menjadi negara maju dan mungkin saja sedang berlomba dengan negara berkembang lain. Tidak hanya melalui kebijakan fiskal yang menstimulus pertumbuhan ekonomi tapi juga stabilitas yang terjaga melaui kebijakan moneter. Dalam hal ini bank sebagai sarana kebijakan moneter memiliki peran besar yang mempengaruhi jalanya perekonomian. Sebagai lembaga keuangan yang menghimpun dana serta menyalurkan kredit kepada masyarakat, bank memiliki pengaruh yang signifikan terhadap masyarakat (Sinaga & Masdjojo, 2022). Dalam pasal 1 Undang-Undang Perbankan No. 10 Tahun 1998 menyebutkan bahwa bank merupakan badan usaha yang menghimpun dana dari masyarakat dalam bentuk simpanan dan menyalurkannya kepada masyarakat dalam bentuk kredit dan atau bentuk- bentuk lainnya dalam rangka meningkatkan taraf hidup rakyat banyak. Peran bank sangat penting sebagai perantara antara pihak yang kelebihan dana (unit surplus) dengan pihak yang membutuhkan dana (unit defisit). Pihak yang kelebihan dana akan menyimpan dana yang dimilikinya di bank dalam bentuk tabungan, giro, maupun deposito sedangkan pihak yang membutuhkan dana akan memperoleh dana dari bank dalam bentuk kredit (Fildzah & Adnan, 2018). Disisi lain fungsi penyaluran kredit pada bank bukan hanya sebagai sumber pendapatan bagi industri perbankan melalui bunga pinjaman tapi juga untuk mempercepat arus uang, meningkatkan produktifitas, menstimulus usaha masyarakat dan menambah modal kerja masyarakat. Bank akan menyaurkan dana pada sektor bisnis maupun sektor yang membutuhkan (Solicha, 2021) Pada awal tahun 2020 telah muncul masalah kesehatan di seluruh dunia berupa pandemi global Corona Virus Disease (COVID-19). COVID-19 menyebar sangat cepat ke seluruh dunia termasuk Indonesia, mengakibatkan gangguan kesehatan dan ancaman kematian. Kondisi ini mengakibatkan pencegahan penyebaran COVID-19 dilakukan dengan langkah-langkah ekstrim dalam membatasi interaksi antar manusia. Pembatasan sosial dilakukan dalam bentuk pelarangan perjalanan (travel ban), penutupan perbatasan antarnegara (closed borders), penutupan sekolah, kantor, dan tempat ibadah bahkan isolasi suatu wilayah tertentu (lockdown). Hal ini mengakibatkan ketidakpastian yang tinggi sehingga aktivitas ekonomi mengalami penurunan drastis termasuk penyaluran kredit perbankan di Indonesia. Selain itu bagi perusahaan dan UMKM pinjaman ke bank merupakan salah satu cara untuk mendapatkan pendanaan yang cukup dalam membiayai berbagai kebijakan maupun kegiatan operasional usaha itu sendiri. Kondisi pandemi COVID-19 membuat kinerja perusahaan dan UMKM mengalami penurunan (Syahwildan & Parulian, 2023). Kondisi ini mengurangi jumlah pendanaan dari bank yang berdampak pada menurunya kegiatan usaha untuk melakukan pengembangan dan ekspansinya terhadap pasar. TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 675 Abstrak Hasil analisis menunjukkan bahwa dana pihak ketiga berpengaruh positif terhadap penyaluran kredit bank selama pandemi COVID-19. Hal ini mengimplikasikan bahwa semakin besar jumlah dana pihak ketiga yang dimiliki oleh bank, semakin besar pula kemungkinan bank untuk menyalurkan kredit kepada nasabah. Namun, likuiditas pendanaan dan risiko kredit memberikan pengaruh negatif terhadap penyaluran kredit. Hal ini menunjukkan bahwa likuiditas pendanaan yang lebih rendah dan risiko kredit yang lebih tinggi membuat bank lebih sulit untuk memberikan kredit kepada nasabah selama pandemi. Kata Kunci: Dana Pihak Ketiga, Kredit Bank, Likuiditas Pendanaan, Risiko Kredit Kata Kunci: Dana Pihak Ketiga, Kredit Bank, Likuiditas Pendanaan, Risiko Kredit TRANSEKONOMIKA \| VOLUME 3 NO. 4 (2023) https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 674 TRANSEKONOMIKA \| VOLUME 3 NO. 4 (2023) https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 TRANSEKONOMIKA \| VOLUME 3 NO. 4 (2023) https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 674 TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 675 PENGARUH DANA PIHAK KETIGA, LIKUIDITAS PENDANAAN, DAN RISIKO KREDIT TERHADAP PENYALURAN KREDIT BANK SELAMA PANDEMI COVID-19 DI INDONESIA Amril Muharyadi Etty Gurendrawati Dwi Handarini Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. Gambar 1. Data penyaluran kredit di Indonesia tahun 2015-2019 Sumber: Otoritas Jasa Keuangan : Statistik Perbankan Indonesia (diolah) 0 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000 2015 2016 2017 2018 2019 4,092,104 4,413,414 4,781,931 5,358,012 5,683,757 Penyaluran Kredit Perbankan Gambar 1. Data penyaluran kredit di Indonesia tahun 2015-2019 Sumber: Otoritas Jasa Keuangan : Statistik Perbankan Indonesia (diolah) Menurut data tersebut setiap tahun pertumbuhan kredit perbankan di Indonesia mengalami peningkatan walaupun terjadi perlambatan pertumbuhan pada tahun 2016 dan 2019. Kredit perbankan hanya tumbuh 6,08% pada tahun 2019, padahal ditahun 2018 dapat tumbuh mencapai 11,7%. Perlambatan ini dialami oleh bank-bank besar kelompok BUKU III dan BUKU IV dimana pada tahun 2019 kredit bank BUKU III hanya tumbuh 2,4% yang di tahun sebelumnya mencapai 12,3%, sementara pertumbuhan kredit bank BUKU IV melambat menjadi 7,8% dari 12,3%. Menurut Bank Indonesia perlambatan pertumbuhan kredit paling rendah pada September 2020 yaitu sebesar 0,12% yang menimbulkan kekhaatiran pada sektor perbankan (Rilka, 2020). p p Dalam penyaluran kredit, dana yang dihimpun dari masyarakat atau sering disebut dana pihak ketiga merupakan sumber dana terbesar yang dikelola oleh bank dimana ketika terjadi peningkatan dana pihak ketiga maka jumlah kredit yang disalurkan juga akan meningkat (Fitri, 2017). Dana pihak ketiga akan menjadi ukuran keberhasilan bank jika mampu membiayai kegiatan operasionalnya dari sumber dana ini. Dalam masa pandemi COVID-19 masyarakat menjadi waspada dalam pemakaian dananya dan memilih melakukan penyimpanan di bank dalam mengantisipasi kondisi yang tidak menentu. Disisi lain bank juga diharapkan mampu menciptakan kondisi yang sehat dan mampu bertahan melalui likuiditas pendanaan dalam situasi pandemi COVID-19. Pengelolaan aktivitas perbankan berdasarkan pengelolaan aset dan transaksi rekening administratif yang baik selama pandemi diharapkan dapat diasalurkan melalui penyaluran kredit dengan sumber dana yang stabil untuk meminimalisir risiko kesulitan pendanaan pada masa depan. Risiko likuiditas menjadi sebuah ancaman serius bagi lembaga keuangan dan stabilitas sistem keuangan dimana bank diharuskan untuk menjaga penyangga likuiditas untuk mengelola risiko likuiditas (Saifuddin et al., 2017). Risiko likuiditas pendanaan menjadi penting ditengah perekonomian yang lemah akibat pandemi COVID-19 karena berhubungan dengan distribusi hasil dimasa depan dengan perbedaan skala waktu yang tersirat pada satu titik waktu tertentu. 2.1. Loanabe Funds Theory Bertocco (dalam Ohlin dan Robertson, 1937) menjelaskan tentang Loanable Fund Theory bahwa permintaan dan penawaran dana pinjaman menentukan suku bunga dan ketersediaan kredit yang dapat diberikan. Menurut teori ini, tabungan dari rumah tangga dan bisnis adalah sumber utama pendanaan yang dapat dialokasikan bank kepada pihak peminjam, dimana suku bunga bertindak sebagai mekanisme penyeimbang untuk menyamakan tabungan dengan permintaan investasi. y g g p Dalam penyaluran kredit, bank akan memperhitungkan suku bunga dan risiko dari peminjaman yang dilakukan. Meskipun melakukan monitoring dan selesksi terhadap calon debitur, bank akan tetap mempunyai pengetahuan yang tidak lengkap atas kondisi debitur yang sebenarnya karena adanya ketidaksimetrisan informasi. Hal itu membuat bank akan cenderung menambahkan risiko kredit macet dalam keputusan suku bunga dan jumlah kredit yang akan ditawarkan kepada debitur (Yulian et al., 2019). TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) Risiko kredit ini tidak dapat dihindari oleh industri perbankan di tengah pandemi karena banyaknya penyaluran kredit pada sektor UMKM, dimana usahanya mengalami penurunan dari segi pendapatan. Hal ini mendorong bank untuk bisa menekan risiko kreditnya agar tidak mempengarui kondisi kesehatan bank itu sendiri. Bank dengan risiko kredit yang rendah atau kredit bermasalahnya kecil akan membutuhkan biaya pengawasan yang relatif lebih rendah sehingga efisiensi bank meningkat (Sparta, 2016). Risiko kredit menjadi ancaman bagi bank mengingat salah satu pendapatan terbesar yang diperoleh oleh bank berasal dari pinjaman kredit. Diversifikasi portofolio pinjaman yang efisien dapat memastikan bahwa risiko kredit dapat diminimalisir tetapi tetap waspada terhadap risiko kredit dalam mengelola masing-masing pinjaman individu (Taiwo, 2014). TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 Dalam hal risiko, bank juga dihadapkan pada arus masuk kas yang terhambat dari penyaluran kredit yang disebut kredit bermasalah akibat pandemi COVID-19. Akibatnya sejumlah perbankan terpaksa dan telah mengambil kebijakan hapus buku dan hapus tagih piutang (write off) untuk kredit yang sudah masuk katagori macet dalam jangka waktu lama. TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 2.3. Dana Pihak Ketiga Dana pihak ketiga merupakan sumber dana terpenting yang menjadi ukuran keberhasilan bank dalam menjalankan kegiatan operasionalnya sehingga besar kecilnya dana pihak ketiga yang dapat dicapai oleh perbankan dapat mempengaruhi besar kecilnya kredit yang akan disalurkan (Putra et al., 2014). Dana pihak ketiga menjadi dorongan besar bagi bank dalam kegiatan bisnis yang dilakukan sehingga setiap bank berusaha keras untuk mendapatkan dana pihak ketiga yang besar. Disisi lain masyarakat menjadi lebih selektif dalam memilih bank yang akan dijadikan tempat menaruh kepercayaan terhadap dananya. Bank yang menerima dana pihak ketiga yang cukup besar akan lebih mudah dalam memberikan kredit terhadap masyarakat karena arus masuk uang yang besar menciptakan peluang bagi bank menjalankan bisnis dengan memberikan kredit terhadap masayrakat. Hal ini sesuai dengan loanable fund theory yang menjelaskan bahwa tabungan masyarakat baik dari rumah tangga maupun bisnis merupakan sumber utama pendanaan yang dapat dialokasikan untuk memberikan kredit. Begitupula sebaliknya semakin sedikit dana pihak ketiga yang diterima oleh bank maka cenderung memberikan kredit dengan batasan yang lebih sedikit baik jumlah penerima maupun nominalnya. H1= Dana pihak ketiga berpengaruh positif terhadap kredit bank. Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. Setiap kredit yang diberikan memiliki jangka waktu tertentu dimana jangka waktu ini mencakup masa pengembalian kredit yang telah disepakati baik berupa jangka waktu yang pendek, menegah ataupun jangka panjang. d. Risiko Faktor risiko kerugian dapat diakibatkan karena ketidakmampuan nasabah dalam membayar angsuran sesuai dengan perjanjian. Semakin panjang jangka waktu pengembalian suatu kredit maka semakin besar juga risikonya untuk tidak tertagih begitu juga sebaliknya. e. Balas jasa Fasilitas kredit yang diberikan oleh bank tentu mengharapkan suatu keuntungan dalam jumlah tertentu. Balas jasa dalam bentuk bunga, biaya provisi dan komisi serta biaya administrasi kredit ini merupakan keuntungan utama bank. Sedangkan bagi bank yang berdasarkan prinsip syariah balas jasanya ditentukan dengan bagi hasil. 2.2. Kredit Bank Kredit adalah penyediaan uang atau tagihan berdasarkan persetujuan atau kesepakatan pinjam meminjam antara bank dengan pihak lain yang mewajibkan pihak peminjam melunasi utangnya setelah jangka waktu tertentu dengan pemberian bunga (Ginoga & Syahwani, 2022). Kredit dalam perbankan sendiri memiliki kriteria tertentu sebagai persyaratan agar risiko kehilangan uang dalam peminjman dapat diminimalkan. Menurut ( ) p y g a. Kepercayaan Kepercayaan merupakan keyakinan pemberi kredit bahwa kredit yang diberikan baik berupa uang, barang atau jasa akan diterima kembali di masa yang akan datang sesuai jangka waktu kredit. b. Kesepakatan Kesepakatan yang terjadi antara pemberi kredit dan penerima kredit dinyatakan dalam suatu perjanjian dimana masing-masing pihak menandatangani hak dan kewajibannya. Kesepakatan penyaluran kredit dituangkan dalam akad kredit oleh pihak bank dan nasabah. TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 677 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 677 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 677 2.5. Risiko Kredit Bank dalam memberikan kredit harus melakukan analisis terhadap kemampuan debitur untuk membayar kembali kewajibannya. Setelah kredit diberikan bank wajib melakukan pemantauan terhadap penggunaan kredit serta kemampuan dan kepatuhan debitur dalam memenuhi kewajibannya. Bank melakukan peninjauan, penilaian, dan pengikatan terhadap agunan untuk memperkecil risiko kredit. Pemberian kredit yang dilakukan oleh bank kepada masyarakat dapat mengandung risiko berupa tidak lancarnya pembayaran yang mampu mempengaruhi kinerja bank yang biasa disebut dengan kredit macet atau Non Performing Loan (NPL). Bank Indonesia sudah menetapkan ambang batas untuk NPL yaitu sebesar 5%. Semakin besar nilai NPL maka akan menyebabkan berkurangnya penerimaan bank dari kredit sehingga bank harus membentuk cadangan penghapusan yang lebih besar yang berpotensi menjadi kerugian bank. NPL terjadi karena kurangnya kesadaran dari debitur dalam pengembalian kredit yang sudah disalurkan (Putu & Puspita, 2018). H3= risiko kredit berpengaruh negatif terhadap kredit bank. TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) Dalam sudut pandang bank, kredit merupakan jenis penempatan dana dengan tenor jangka panjang dimana bank akan membutuhkan likuiditas untuk memenuhi kegiatan bisnis tersebut (Mayasari, 2017). Dalam hal ini bank yang memiliki likuiditas pendanaan yang tinggi maka bank akan secara aktif memberikan penyaluran kredit kepada masyarakat. Disisi lain bank akan mengurangi kegiatan penyaluran kredit jika bank rentan terhadap penarikan jalur kredit dan tidak mampu mempertahankan likuiditas pendanaan yang cukup. H2 = likuiditas pendanaan berpengaruh positif terhadap kredit bank 2 = likuiditas pendanaan berpengaruh positif terhadap kredit bank 2.4. Likuiditas Pendanaan Sebagai media antara entitas yang kelebihan dana dan entitas yang kekurangan dana, bank berperan dalam penyaluran kredit terhadap entitas yang kekurangan dana dengan tetap memperhatikan risiko yang timbul. Ketika bank memiliki sumber simpanan besar dari deposan, maka bank akan memiliki risiko likuiditas pendanaan yang lebih rendah sehingga mendorong pihak bank untuk secara agresif menurunkan suku bunga pinjaman, dengan tujuan untuk meningkatkan volume pinjaman serta meningkatkan pangsa pasar mereka sendiri (Dang, 2019). Dalam loanable fund theory bunga dikatakan sebagai mekanisme penyeimbang tabungan dengan permintaan investasi maka jumlah tabungan yang besar dari deposan dapat mempengaruhi likuiditas pendanaan bank dalam memberikan kredit dimana bunga menjadi penyeimbang antara tabungan dan kredit. 678 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 3. METODE PENELITIAN Penelitian ini merupakan penelitian dengan pendekatan kuantitatif dengan metode korelasi yang bertujuan untuk menguji hipotesis mengenai hubungan sebab akibat.Data yang digunakan dalam penelitian ini adalah data skunder berupa laporan keuangan yang diperoleh dari situs Bursa Eek Indonesia (www.idx.co.id) dan situs masing-masing bank yang akan diteliti. Populasi dan sampel pada penelitian ini adalah bank yang termasuk dalam kategori BUKU III dan BUKU IV dengan periode pengamatan tahun 2020 dan 2021. Proses pemilihan sampel untuk penelitian ini dapat dilihat pada table berikut ini: Tabel 1. Pemilihan Sampel Kriteria Tabel 1. Pemilihan Sampel No. Kriteria Jumlah 1 Bank umum konvensional yang ada di Indonesia periode 2020 – 2021 110 2 Bank yang TIDAK berada pada kategori BUKU III dan BUKU IV atau bank yang memiliki modal inti di atas 5 triliun rupiah yang terdaftar di Otoritas Jasa Keuangan selama periode 2020 – 2021 (79) Total Sample (perusahaan) 31 Tahun Amatan 2020 – 2021 (2 tahun) Total Obeservasi 2 x 31 62 Sumber : Diola penuis (2023) TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 679 Selanjutnya masing-masing variable akan diukur dengan penjeasan sebagai berikut: PENGARUH DANA PIHAK KETIGA, LIKUIDITAS PENDANAAN, DAN RISIKO KREDIT TERHADAP PENYALURAN KREDIT BANK SELAMA PANDEMI COVID-19 DI INDONESIA Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. Tabel 2. Pengukuran Variabel Variabel Pengukuran Kredit Bank (Y) Kredit bank = Ln (jumlah kredit yang disalurkan) Dana Pihak Ketiga (X1) Dana Pihak Ketiga = Ln (tabungan + giro + deposito) Likuiditas Pendanaan (X2) NSFR = ASF / RSF ASF = jumlah liabilitas + ekuitas yang stabil selama setahun RSF = jumlah aset + transaksi rekening administratif Risiko Kredit (X3) 𝑁𝑃𝐿= 𝐾𝑟𝑒𝑑𝑖𝑡 𝐵𝑒𝑟𝑚𝑎𝑠𝑎𝑙𝑎ℎ 𝑇𝑜𝑡𝑎𝑙 𝐾𝑟𝑒𝑑𝑖𝑡 𝑥 100% Sumber: Diola penulis (2023) ( ) setahun RSF = jumlah aset + transaksi rekening administratif Risiko Kredit (X3) 𝑁𝑃𝐿= 𝐾𝑟𝑒𝑑𝑖𝑡 𝐵𝑒𝑟𝑚𝑎𝑠𝑎𝑙𝑎ℎ 𝑇𝑜𝑡𝑎𝑙 𝐾𝑟𝑒𝑑𝑖𝑡 𝑥 100% Sumber: Diola penulis (2023) 4. HASIL DAN PEMBAHASAN 4.1. Hasil Penelitian 4.1.1. Statistik Deskriptif Tabel 3. Statistik Deskriptif Analisis N Mean Median Maximum Minimum Std. Deviasi X1 62 32,08 32,06 34,68 29,49 1,24 X2 62 1,395 1,371 2,537 0.977 0,306 X3 62 0,032 0,032 0,107 0,007 0,018 Y 62 31,93 31,83 34,49 29,6 1,17 Sumber: Diolah penulis dengan Eviews 12 (2023) 4.1.2. Uji Asumsi Klasik a. Uji Normalitas Gambar 2. Jargue-Bera Test Sumber : Diolah penulis dengan Eviews 12 (2023) 4. HASIL DAN PEMBAHASAN 4.1. Hasil Penelitian 4.1.1. Statistik Deskriptif 4. HASIL DAN PEMBAHASAN 4.1. Hasil Penelitian 4.1.1. b. Uji Multikolinearitas b. Uji Multikolinearitas Dari hasil uji multikolinearitas pada bagian Centered VIF dapat diketahui bahwa semua nilai VIF kurang dari 10. Hal ini dapat memberikan kesimpulan bahwa tidak terdapat multikolinearitas atau hubungan antarvariabel bebas di dalam model regresi. b. Uji Multikolinearitas Dari hasil uji multikolinearitas pada bagian Centered VIF dapat diketahui bahwa semua nilai VIF kurang dari 10. Hal ini dapat memberikan kesimpulan bahwa tidak terdapat multikolinearitas atau hubungan antarvariabel bebas di dalam model regresi. c. Uji Heteroskedastisitas Dari hasil uji heteroskedastisitas pada tabel dapat diketahui bahwa nilai ObsR-squared adalah sebesar 4,13 dengan probabilitas sebesar 0,9. Karena nilai probabilitas lebih besar dari 0,05 maka dapat disimpulkan bahwa tidak terdapat heteroskedastisitas dalam model. c. Uji Heteroskedastisitas Dari hasil uji heteroskedastisitas pada tabel dapat diketahui bahwa nilai ObsR-squared adalah sebesar 4,13 dengan probabilitas sebesar 0,9. Karena nilai probabilitas lebih besar dari 0,05 maka dapat disimpulkan bahwa tidak terdapat heteroskedastisitas dalam model. d. Uji Autokorelasi Dari hasil autokoreasi dapat diketahui bahwa nilai Berdasarkan hasil uji Durbin-Watson (DW) dapat diperoleh nilai d sebesar 1,8 sehingga masuk dalam kriteria du < d < 4-du maka dapat disimpulkan bahwa data tidak ada autokorelasi. d. Uji Autokorelasi Dari hasil autokoreasi dapat diketahui bahwa nilai Berdasarkan hasil uji Durbin-Watson (DW) dapat diperoleh nilai d sebesar 1,8 sehingga masuk dalam kriteria du < d < 4-du maka dapat disimpulkan bahwa data tidak ada autokorelasi. TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) Berdasarkan uji normalitas di atas, probabilitas sebesar 0,315 lebih besar dari 0,05 sehingga dapat disimpulkan data tersebut berdistribusi normal. 3. METODE PENELITIAN Statistik Deskriptif Tabel 3. Statistik Deskriptif Analisis N Mean Median Maximum Minimum Std. Deviasi X1 62 32,08 32,06 34,68 29,49 1,24 X2 62 1,395 1,371 2,537 0.977 0,306 X3 62 0,032 0,032 0,107 0,007 0,018 Y 62 31,93 31,83 34,49 29,6 1,17 Sumber: Diolah penulis dengan Eviews 12 (2023) 4.1.2. Uji Asumsi Klasik a. Uji Normalitas Gambar 2. Jargue-Bera Test Sumber : Diolah penulis dengan Eviews 12 (2023) g Sumber : Diolah penulis dengan Eviews 12 (2023) TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) 4.1.3. Analisis Kelayakan Model a. Analisis Regresi Linier Berganda Dari analisis memperhatikan angka yang berada pada kolom coefficient, maka dapat disusun persamaan regresi berganda sebagai berikut : a. Analisis Regresi Linier Berganda Dari analisis memperhatikan angka yang berada pada kolom coefficient, maka dapat disusun persamaan regresi berganda sebagai berikut : 4.2.1. Dana Pihak ketiga dan Penyaluran Kredit Bank Dalam situasi pandemi COVID-19 masyarakat cenderung lebih berhemat dalam mengeluarkan simpanannya dan menjadi lebih waspada akan kondisi yang tidak menentu di masa yang akan datang. Akibatnya dana pihak ketiga masyarakat yang mengalami peningkatan mempengaruhi peningkatan jumlah penyaluran kredit bank. Hal ini tentu saja menjadi kabar baik baik industri perbankan karena salah satu sumber pendapatan terbesar perbankan berasal dari penyaluran kredit. Penyaluran kredit yang ikut meningkat karena dana piak ketiga yang mengalami peningkatan tidak lepas dari sikap masyarakat yang lebih berhati-hati dalam membelanjakan simpananya. PENGARUH DANA PIHAK KETIGA, LIKUIDITAS PENDANAAN, DAN RISIKO KREDIT TERHADAP PENYALURAN KREDIT BANK SELAMA PANDEMI COVID-19 DI INDONESIA Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. t tabel pada derajat bebas (df) = n-k = 62-4 = 58, dimana n = jumlah sampel, dan k = jumlah variabel, maka nilai t tabelnya sebesar 1,672. Sehimgga dapat diperoleh t-statistic > t tabel (12,54 > 1,672). Berdasarkan nilai t-statistic dan nilai prob yang diperoleh maka dapat disimpulkan bahwa Ha1 diterima dimana dana pihak ketiga berpengaruh positif terhadap penyaluran kredit bank. Hal ini menunjukkan semakin tinggi nilai dana pihak ketiga maka akan semakin tinggi juga tingkat penyaluran kredit bank. t tabel pada derajat bebas (df) = n-k = 62-4 = 58, dimana n = jumlah sampel, dan k = jumlah variabel, maka nilai t tabelnya sebesar 1,672. Sehimgga dapat diperoleh t-statistic > t tabel (12,54 > 1,672). Berdasarkan nilai t-statistic dan nilai prob yang diperoleh maka dapat disimpulkan bahwa Ha1 diterima dimana dana pihak ketiga berpengaruh positif terhadap penyaluran kredit bank. Hal ini menunjukkan semakin tinggi nilai dana pihak ketiga maka akan semakin tinggi juga tingkat penyaluran kredit bank. 2) 2) Likuiditas pendanaan dan penyaluran kredit bank Dari hasil penelitian diketahui bahwa likuiditas pndanaan memiliki nilai prob sebesar 0,051 < 0,1 dan nilai t-statistic sebesar -1,989. Jika dibandingkan dengan t tabel pada derajat bebas (df) = n-k = 62-4 = 58, dimana n = jumlah sampel, dan k = jumlah variabel, maka nilai t tabelnya sebesar 1,672. Sehingga dapat diperoleh t-statistic > t tabel (1,989 > 1,672). Berdasarkan nilai t hitung dan nilai prob maka dapat disimpulkan bahwa Ha2 diterima dimana likuiditas pendanaan berpengaruh negatif terhadap penyaluran kredit bank. Hal ini menunjukkan saat COVID-19 apabila terjadi peningkatan terhadap likuiditas pendanaan maka jumlah penyaluran kredit bank akan mengalami penurunan. 3) Risiko kredit dan penyaluran kredit bank Dari hasil penelitian diketahui bahwa risiko memiliki nilai t-statistic sebesar - 2,05 dan nilai prob sebesar 0.045 < 0,05. Jika dibandingkan dengan t tabel pada derajat bebas (df) = n-k = 62-4 = 58, dimana n = jumlah sampel, dan k = jumlah variabel, maka nilai t tabelnya sebesar 1,672. Sehingga dapat diperoleh t-statistic > t tabel (2,05 > 1,672). Berdasarkan nilai t hitung dan nilai prob yang diperoleh maka dapat disimpulkan bahwa Ha1 diterima dimana risiko kredit berpengaruh negatif terhadap penyaluran kredit bank. Hal ini menunjukkan semakin tinggi nilai risiko kredit maka penyaluran kredit bank akan mengalami penurunan dan juga sebaliknya jika risiko kredit mengalami penurunan maka penyaluran kredit bank akan mengalami peningkatan. y = 18,78 + 0,42x1 – 0,18x2 – 3,69x3 y = 18,78 + 0,42x1 – 0,18x2 – 3,69x3 b. Analisis Data Panel Model terbaik yang digunakan dari tiga model (Common Effect Model, Fixed Effect Model dan Random Effect Model) adalah Random Effect Model. 𝟐 b. Analisis Data Panel Model terbaik yang digunakan dari tiga model (Common Effect Model, Fixed Effect Model dan Random Effect Model) adalah Random Effect Model. 𝟐 c. Uji Koefisien Determinasi (𝐑𝟐) Dari tabel dalam hasil pengujian regresi Random Efect Model menunjukkan bahwa nilai Adjusted R-squared adalah sebesar 0,3859. Hal ini berarti bahwa dana pihak ketiga, likuiditas pendanaan dan risiko kredit mampu menjelaskan variasi penyaluran kredit bank sebesar 38,59%, sedangkan sisanya (61,41%) dijelaskan oleh variabel lain yang tidak dimasukkan ke dalam model. y g d. Uji F Dari tabel hasil regresi dengan metode Random Effect Model dalam penelitian ini diketahui bahwa nilai probabilitas F-statistic adalah sebesar 0,000001 lebih kecil dari 0,05. Hal ini berarti bahwa dana pihak ketiga likuiditas pendanaan dan risiko kredit secara bersama-sama mempunyai pengaruh signifikan terhadap penyaluran kredit. d. Uji F Dari tabel hasil regresi dengan metode Random Effect Model dalam penelitian ini diketahui bahwa nilai probabilitas F-statistic adalah sebesar 0,000001 lebih kecil dari 0,05. Hal ini berarti bahwa dana pihak ketiga likuiditas pendanaan dan risiko kredit secara bersama-sama mempunyai pengaruh signifikan terhadap penyaluran kredit. 4.1.4. Uji Hipotesis a. Uji T Dari tabel hasil regresi dengan metode Random Effect Model dalam penelitian ini dapat diinterpretasikan sebagai berikut: 1) Dana pihak ketiga dan penyaluran kredit bank Dari hasil penelitian diketahui bahwa dana pihak ketiga memiliki nilai prob b 0 00 0 05 d il i i i b 12 54 Jik dib di k d 4.1.4. Uji Hipotesis a. Uji T Dari tabel hasil regresi dengan metode Random Effect Model dalam penelitian ini dapat diinterpretasikan sebagai berikut: TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 681 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 681 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN 681 TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) peningkatan dalam bisnisnya sehingga penyaluran kredit bank mengalami penurunan. Situasi ini akan mendorong bank mencari cara lain dalam meningkatkan penyaluran kredit walaupun likuiditas pendanaan yang dimiliki bank cukup besar. peningkatan dalam bisnisnya sehingga penyaluran kredit bank mengalami penurunan. Situasi ini akan mendorong bank mencari cara lain dalam meningkatkan penyaluran kredit walaupun likuiditas pendanaan yang dimiliki bank cukup besar. 4.2.3. Risiko Kredit dan Penyaluran Kredit Bank Dalam penelitian ini menunjukan bahwa risiko kredit berpengaruh negatif teradap penyaluran kredit bank. Dalam situasi COVID-19 kredit yang disalurkan harus melakukan analisis terhadap kemampuan debitur untuk membayar kembali kewajibannya. Hal ini tidak terlepas dari situasi ekonomi yang menurun sehingga bank harus lebih selektif dalam menyalurkan kredit untuk bisnis yang dianggap akan mampu bertahan dan berkembang. Pandemi COVID-19 membuat risiko kredit macet meningkat karena bisnis yang dijalankan oleh masyarakat mengalami stagnan atau bahkan menurun karena kebijakan pembatasan sosial. Aktiitas masyarakat yang menurun membuat ekonomi juga mengalami penurunan akibatnya penyaluran kredit bank akan menurun. 4.2.2. Likuiditasn Pendanaan dan Penyaluran Kredit Bank Dalam peneitian ini menunjukan bahwa likuiditas pendanaan berpengaruh negatif terhadap penyaluran kredit bank Hal ini terjadi karena selama pandemi COVID-19, bisnis yang dijalankan masyarakat mengalami penurunan yang cukup drastis akibat kebijakan pemerintah dalam pembatasan sosial. Akibatnya walaupun bank memiiki likuiditas pendanaan yang meningkat, masyarakat tidak akan melakukan ekspansi maupun TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 TRANSEKONOMIKA: Akuntansi, Bisnis dan Keuangan VOLUME 3 NO. 4 (2023) PENGARUH DANA PIHAK KETIGA, LIKUIDITAS PENDANAAN, DAN RISIKO KREDIT TERHADAP PENYALURAN KREDIT BANK SELAMA PANDEMI COVID-19 DI INDONESIA Amril Muharyadi, Etty Gurendrawati, Dwi Handarini. mril Muharyadi, Etty Gurendrawati, Dwi Handarini. Fitri, L., Maulida, Y., & Indrawati, T. (2017). Pengaruh Suku Bunga Kredit, Dana Pihak Ketiga (DPK), dan Giro Wajib Minimum Terhadap Penyaluran Kredit Pada PT. Bank Central Asia, TBK. di Indonesia Tahun 2001-2015. Riau University. Horizons, E., & Dang, V. D. (2019). Funding liquidity and bank lending : Evidence from Vietnam. 15, 205–218. Jn, T. (2014). Credit Risk Management : Implications on Bank Performance and Lending Growth. 584–590. https://doi.org/10.21276/sjbms Mayasari, V. (2017). Pengaruh Likuiditas Terhadap Penyaluran Kredit Sektor Perbankan Dikota Palembang. Jurnal Adminika, 3(1), 80–92. Putra, I. G. O. P., & Rustariyuni, S. D. (2014). Penyaluran Kredit Modal Kerja Pada Bpr Di Provinsi Bali Tahun 2009-2014. E-Jurnal EP UNUD, 4(5), 451–464. Putu, N., & Puspita, I. (2018). E-Jurnal Akuntansi Universitas Udayana Pengaruh Tingkat Efisiensi , Risiko Kredit , dan Tingkat Penyaluran Kredit pada Profitabilitas Fakultas Ekonomi dan Bisnis Universitas Udayana ( Unud ), Bali , Indonesia Fakultas Ekonomi dan Bisnis Universitas Udayan. 24, 1164–1189. y Saifuddin, M., Scheule, H., & Wu, E. (2017). Funding liquidity and bank risk taking R. Journal of Banking and Finance, 82, 203–216. https://doi.org/10.1016/j.jbankfin.2016.09.005 5. KESIMPULAN Penelitian ini menyimpulkan bahwa dana pihak ketiga memiliki dampak positif yang signifikan terhadap penyaluran kredit oleh bank kategori BUKU III dan BUKU IV di Indonesia. Semakin besar dana yang berhasil dikumpulkan dari pihak ketiga, semakin besar pula jumlah kredit yang akan diberikan oleh bank. Namun, likuiditas pendanaan yang diukur dengan Net Stable Funding Ratio (NSFR) memiliki pengaruh negatif terhadap penyaluran kredit bank. Perubahan dalam likuiditas pendanaan tidak mempengaruhi penyaluran kredit. Selanjutnya, risiko kredit yang diukur dengan Non-Performing Loan (NPL) juga memiliki pengaruh negatif yang signifikan terhadap penyaluran kredit bank. Semakin tinggi risiko kredit, penyaluran kredit bank akan mengalami penurunan yang cukup besar, dan sebaliknya, jika risiko kredit menurun, penyaluran kredit bank akan meningkat secara signifikan. Meskipun penelitian ini memiliki keterbatasan dimana periode pengamatan yang relatif singkat, yaitu tahun 2020-2021, dan fokus pada bank kategori BUKU III dan BUKU IV mungkin belum mencakup semua faktor yang mempengaruhi penyaluran kredit bank. Oleh karena itu, penelitian selanjutnya dapat memperluas sampel penelitian untuk mencakup bank dari kategori lain dan mempertimbangkan faktor-faktor eksternal yang dapat mempengaruhi penyaluran kredit. Dengan demikian, penelitian mendatang dapat memberikan wawasan yang lebih komprehensif tentang dinamika penyaluran kredit bank di Indonesia, terutama dalam menghadapi tantangan masa pandemi COVID-19. DAFTAR PUSTAKA Adnan, A., Ridwan, R., & Fildzah, F. (2016). Pengaruh Ukuran Bank, Dana Pihak Ketiga, Capital Adequacy Ratio, dan Loan To Deposit Ratio Terhadap Penyaluran Kredit Pada Perusahaan Perbankan yang Terdaftar Di Bursa Efek Indonesia Tahun 2011-2015. Jurnal Dinamika Akuntansi Dan Bisnis, 3(2), 49–64. https://doi.org/10.24815/jdab.v3i2.5386 Ahmad, F. A., & Alham, F. (2016). Strategi Peningkatan Kinerja Pranata Laboratorium Pendidikan Di Institut Pertanian Bogor. MIX: Jurnal Ilmiah Manajemen, VI(1), 139– 151. TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 \| P-ISSN: 2809-6851 683 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN h // blik id/ j /i d h / k ik 683 Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/). 684 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 684 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851 TRANSEKONOMIKA \| AKUNTANSI, BISNIS DAN KEUANGAN https://transpublika.co.id/ojs/index.php/Transekonomika E-ISSN: 2809-6851 P-ISSN: 2809-6851
https://openalex.org/W2006626323	https://journals.plos.org/plospathogens/article/file?id=10.1371/journal.ppat.1001214&type=printable	English	null	Infectious Speciation Revisited: Impact of Symbiont-Depletion on Female Fitness and Mating Behavior of Drosophila paulistorum	PLOS pathogens	2,010	cc-by	17,790	Abstract This is an open-access article distributed under the terms of the Creative Commons Attributi unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: WJM and DS were partly supported by the research grant FWF P19206-B17 and P22634-B17 from the Austrian Science Fund and the EU-COST Action FA0701 ‘‘Arthropod Symbiosis: From Fundamental Studies to Pest and Disease Management’’. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: wolfgang.miller@meduniwien.ac.at Infectious Speciation Revisited: Impact of Symbiont- Depletion on Female Fitness and Mating Behavior of Drosophila paulistorum Wolfgang J. Miller1*, Lee Ehrman2, Daniela Schneider1 1 Laboratories of Genome Dynamics, Center of Anatomy and Cell Biology, Medical University of Vienna, Vienna, Austria, 2 Natural Sciences, State University of New York, Purchase College, Purchase, New York, United States of America mics, Center of Anatomy and Cell Biology, Medical University of Vienna, Vienna, Austria, 2 Natural Sciences, State University of New York w York, United States of America Abstract The neotropical Drosophila paulistorum superspecies, consisting of at least six geographically overlapping but reproductively isolated semispecies, has been the object of extensive research since at least 1955, when it was initially trapped mid-evolution in flagrant statu nascendi. In this classic system females express strong premating isolation patterns against mates belonging to any other semispecies, and yet uncharacterized microbial reproductive tract symbionts were described triggering hybrid inviability and male sterility. Based on theoretical models and limited experimental data, prime candidates fostering symbiont-driven speciation in arthropods are intracellular bacteria belonging to the genus Wolbachia. They are maternally inherited symbionts of many arthropods capable of manipulating host reproductive biology for their own benefits. However, it is an ongoing debate as to whether or not reproductive symbionts are capable of driving host speciation in nature and if so, to what extent. Here we have reevaluated this classic case of infectious speciation by means of present day molecular approaches and artificial symbiont depletion experiments. We have isolated the a-proteobacteria Wolbachia as the maternally transmitted core endosymbionts of all D. paulistorum semispecies that have coevolved towards obligate mutualism with their respective native hosts. In hybrids, however, these mutualists transform into pathogens by overreplication causing embryonic inviability and male sterility. We show that experimental reduction in native Wolbachia titer causes alterations in sex ratio, fecundity, and mate discrimination. Our results indicate that formerly designated Mycoplasma-like organisms are most likely Wolbachia that have evolved by becoming essential mutualistic symbionts in their respective natural hosts; they have the potential to trigger pre- and postmating isolation. Furthermore, in light of our new findings, we revisit the concept of infectious speciation and discuss potential mechanisms that can restrict or promote symbiont-induced speciation at post- and prezygotic levels in nature and under artificial laboratory conditions. Citation: Miller WJ, Ehrman L, Schneider D (2010) Infectious Speciation Revisited: Impact of Symbiont-Depletion on Female Fitness and Mating Behavior of Drosophila paulistorum. PLoS Pathog 6(12): e1001214. doi:10.1371/journal.ppat.1001214 Editor: Colin Parrish, Cornell University, United States of America Received November 9, 2009; Accepted October 27, 2010; Published December 2, 2010 er et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2010 Miller et al. Introduction hybrid incompatibilities and thereby to speciation [4–6]. In this speciation model, however, incompatibility genes are both nuclear genes that have evolved differently after an ancestral populations split. PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 Citation: Miller WJ, Ehrman L, Schneider D (2010) Infectious Speciation Revisited: Impact of Symbiont-Depletion on Female Fitness and Mating Behavior of Drosophila paulistorum. PLoS Pathog 6(12): e1001214. doi:10.1371/journal.ppat.1001214 PLoS Pathogens \| www.plospathogens.org The Neotropical Drosophila paulistorum Species Complex, a Classic Case of Speciation in statu nascendi Complex, a Classic Case of Speciation in statu nascendi The first case in literature suggesting that microbial reproduc- tive parasites can play a pivotal role in driving host-speciation dates back to the late 1960s when Ehrman and coworkers discovered endosymbionts that trigger incipient speciation via hybrid inviability and male sterility in a neotropical Drosophila group species [39–41] reviewed in [42]). The D. paulistorum species complex comprises at least six semispecies showing pronounced sexual isolation; matings between these semispecies succeed significantly less frequently than do those within a semispecies [35–38]. These are: Amazonian (AM), Andean-Brazilian (AB), Centroamerican (CA), Interior (IN), Orinocan (OR), and Transitional (TR) semispecies. The primary extrinsic mechanism differentiating the D. paulistorum semispecies is geographic isolation, albeit incompletely because of consistently overlapping distribu- tions [43]. Three intrinsic isolation mechanisms are operative here - sexual isolation via behavior and hybrid male sterility both in reciprocal crosses [43]. In addition to hybrid male sterility, abnormal pole cell development in early F1 hybrid embryos cause considerable high hybrid inviability [44,45]; depicted in Figure 1. Since mortality takes place during early stages of embryogenesis [44,46] and occurs at high frequencies in all hybrids derived from reciprocal matings between the six D. paulistorum semispecies [45,47], this phenotype can be best described as bidirectional cytoplasmic incompatibility, initially described by Laven in the Culex pipiens species complex [48]. parasitic to a compensatory mutualist relationship [14,15]. So it can be expected that repeated, localized episodes of antagonistic and compensatory coevolution may have the potential to diversify and to substructure host populations up to the point where hybrid incompatibilities emerge between them [16,17]. Increasing empirical evidence and numerous theoretical models predict that transovarially transmitted microbial symbionts that cause cyto- plasmic incompatibilities (CI), can have significant impacts as drivers of speciation processes in their natural hosts [12,16–22]. Over the last decades main CI work was focused on the a- proteobacteria Wolbachia, a frequent maternally-transmitted endo- symbiont of many arthropods and filarian nematodes (reviewed by [23–25]). As deduced from classical theoretical models, for example, Wolbachia cause unstable equilibrium in the prevalence of the symbiotic infection, rather than stable persistence of infected and uninfected populations [26,27]. More recent models, however, propose that under specific genetic and environmental conditions Wolbachia can promote host speciation in nature [19,21]. Nuclear vs. Symbiotic Speciation Mechanisms Nuclear vs. Symbiotic Speciation Mechanisms In contrast to prezygotic reproductive isolating mechanisms acting before fertilization via mating behavior, postzygotic isolation arises after mating when hybrids are less fit than their parents [1]. In the latter case, the Dobzhansky-Muller model proposes that hybrid incompatibilities, contributing to speciation, are caused by the interaction between nuclear genes that have functionally diverged over time in their respective hybridizing species [2,3]. Initial direct molecular and genetic evidence for the existence of such Dobzhansky-Muller incompatibility genes recently came from the two sister-species, D. melanogaster and D. simulans [4]. In this case one of the two incompatibility genes localizes to centromeric heterochromatin - an inherently dynamic part of all eukaryotic genomes known for its accelerated genomic turnover and molecular drive [5,6]. Genes located within or functionally interacting with hyperdynamic genomic regions, such as nuclear heterochromatin, evolve rapidly and so contribute to A second class of rapidly evolving genes possessive of high potential for driving accelerated functional divergence are those genes that are under antagonistic coevolution versus parasites, i.e. the reciprocal evolution of host defense and parasite counter- defense mechanisms [7–10]. Accordingly, one or more incompat- ibility genes is encoded by the host nucleus, and the second set by a parasite. In a consistent arms race between invading cytoplasma- tically inherited reproductive parasites and a local host population, interacting gene products encoded separately in both organisms evolve rapidly under strong diversifying selection driven by ongoing cyto-nuclear conflicts [11–13]. Dense interrelationships between a parasite and its host could functionally weld them together by favoring evolution towards a compensatory partner- ship where both biological systems finally depend on the presence of the other, and both progress evolutionarily from an antagonistic December 2010 \| Volume 6 \| Issue 12 \| e1001214 1 Infectious Speciation in Drosophila mycoplasma-like organisms (MLOs) of D. paulistorum [32] and this study, no empirical cases have been reported that endosymbionts are able to cause hybrid male sterility, a hallmark of an early stage in the evolution of postzygotic isolation [1]. PLoS Pathogens \| www.plospathogens.org Author Summary The Drosophila paulistorum species complex serves as a well-studied model system for evaluating the impact of symbiosis on host speciation since they evolve rapidly and comprise an ancestral, but highly dynamic, reservoir of microbial symbionts. Theory and some experimental evidence suggest that in evolutionary longterm host- symbiont interactions, reproductive parasites might evolve a more benign lifestyle towards mutualism, manipulate sexual mating behavior, and foster host speciation. However, it is an ongoing debate as to whether or not microbial symbionts are capable of driving host speciation in nature and if so, to what extent. Prime candidates are Wolbachia, inherited, endosymbiotic bacteria of many arthropods, presently attracting attention as potential biocontrol agents since they affect host reproductive biology. Here we document that all D. paulistorum semispecies harbor Wolbachia that provide significant fitness benefits to their natural hosts. In semispecies hybrids, however, mutualistic Wolbachia turn into patho- gens, triggering embryonic lethality and male sterility via overreplication. Besides their impacts on post-mating isolation, we show that in their native D. paulistorum hosts Wolbachia manipulate sexual behavior by triggering pre- mating isolation via selective mate avoidance, i.e. avoiding mates harboring another, incompatible symbiont variant. Our study reveals that endosymbionts can coevolve rapidly with their native hosts and play a significant role in driving natural host speciation. Considering the ubiquity of symbiotic interactions and their inherent coevolutionary potential, one could reasonably expect to find experimentally accessible model systems presently under incipient speciation in nature [17]. Based on theoretical consid- erations [1,12,16,17,19,20,29,33,34], requisite for experimentally testing this hypothesis, it is essential to choose a symbiotic model system where both biological components, the host and the symbiont, should fulfill the following main criteria: Firstly, an evolutionary young host species complex in statu nascendi ought to be chosen for studying the speciation process in nature. Secondly, this host system should be associated with a fixed endosymbiont with a perfect transmission rate capable of inducing strong bidirectional, postzygotic incompatibilities in hybrids. Finally, this endosymbiont should exert the capacity to trigger prezygotic isolation in order to avoid potential wasteful gene flow between different host infection types. Such candidate model systems are extensively documented in early literature on speciation mecha- nisms observed in neotropical Drosophila hosts [35–38]. The Neotropical Drosophila paulistorum Species Complex, a Classic Case of Speciation in statu nascendi December 2010 \| Volume 6 \| Issue 12 \| e1001214 Figure 1. Schematic presentation of incipient speciation among D. paulistorum semispecies. Crosses between semispecies give rise in both directions to high hybrid embryonic mortality and complete hybrid male sterility (after [35] as reviewed by [42]). doi:10.1371/journal.ppat.1001214.g001 hybrid male there is a rapid proliferation of MLO and a concomitant breakdown of spermatogenesis [49]. Indeed, hybrid male testes are ‘‘sperm-less’’ microorganismal factories. applying universal 16S rRNA PCR [50] on DNA of fertilized early embryos derived from sterilized eggs. This approach avoids contaminations with gut bacteria, focusing our survey on vertically-transmitted endosymbionts. In total, twenty amplified 16S rRNA consensus fragments of 437 bp derived from five independent PCR reactions were cloned and sequenced. All reads from early embryos of the OR semispecies were identical to the Wolbachia 16S rRNA gene of the sibling species D. willistoni (GenBank accession number DQ412086). Numerous efforts have failed to establish ‘‘MLO-free’’ D. paulistorum strains using diverse sets of antibiotic treatments [39,41]. Under mild antibiotic conditions however, Ehrman and colleagues were able to diminish MLO titer levels significantly and to partially rescue hybrid male sterility. Based on the observed hypersensitivity of D. paulistorum semispecies to different antibiotics, these authors suggested that MLOs are maternally transmitted, obligatory microbial symbionts of all D. paulistorum semispecies since no truly bacteria-free strain has or had ever been established [39,41]. Furthermore ‘‘MLOs’’ have been described earlier as consis- tently associated symbionts within adult reproductive tissues of all D. paulistorum semispecies, specifically over-replicating in testes of reciprocal hybrids [46,49]. Hence we have dissected under sterile conditions ovaries and testes from mature AM and OR semispecies, plus AxO hybrid flies for DNA extraction. Gonad- specific PCRs were performed with two sets of universal 16S rRNA primer [50,51] (for details see Materials and Methods) on parental and hybrid samples. As shown in Figure 2A (upper gel), gonads of both OR sexes harbor more microbial symbionts than AM semispecies, with weakest signal intensity in AM testes. In AxO hybrid testes, however, 16S rRNA signal intensities were significantly higher than in AM paternal ones (Figure 2A). Consensus 16S PCR fragments of 25 independent reactions derived from reproductive organs of native and hybrid samples (6, 14 and 5 reads for AM, OR and hybrids, respectively) were direct- sequenced. In the six reads derived from AM ovaries and testes double peaks were detected from which one sequence type was found with 100% homology to the 16S rRNA of wWil of D. willistoni (position 291–727 and 392–861 in DQ412086, for primer set of [50,51], respectively). The second type of 16S rRNA, however, uncovered the presence of a second bacteria, belonging to Acetobacter. On the other hand all reads of OR (14) and AxO hybrids (5) were identical to the 16S rRNA of D. willistoni (GenBank accession number HQ185362 and HQ185363, respec- tively) and no other bacteria could be recovered. The identifica- tion of Acetobacter in dissected gonads from AM samples demonstrates the sensitivity of our detection system since this semispecies habors Wolbachia at extreme low density compared to Here we report the isolation of close-related a-proteobacteria Wolbachia, belonging to the A supergroup, from the six D. paulistorum semispecies. Based on their intriguing morphological alikeness with MLOs, their tight germline associations, and their striking similarities in infection-titer dynamics in hybrids, Wolbachia are most likely identical to formerly designated ‘‘MLOs’’ by Ehrman and coworkers [41]. Furthermore repeated attempts via microbial universal 16S rRNA screening and direct sequencing have failed to isolate any other germline-associated microbes besides Wolbachia. Hence such ‘‘MLOs’’ of previous D. paulistorum literature should be reconsidered, from now onward, as members of the invertebrate-colonizing genus Wolbachia. We have furthermore reevaluated symbiotic interactions between Wolbachia and their natural D. paulistorum semispecies hosts by studying phenotypic traits appearing upon partial symbiont- depletion via antibiotics, traits such as female fecundity, sex ratios, and most significantly, female mating behavior. PLoS Pathogens \| www.plospathogens.org The Neotropical Drosophila paulistorum Species Complex, a Classic Case of Speciation in statu nascendi In addition, there are only a few empirical model systems from nature available in literature such as parasitoid wasps of the Nasonia sibling species group [28,29], and mushroom feeding Drosophila species [20,30]. These data provide clear evidence that Wolbachia-induced CI can promote speciation in nature acting in concert with other genetic and/or geographic isolation mechanisms (see Discussion). Transinfection studies have shown that extracts of testes from sterile hybrid males induce a similar sterility syndrome when injected into females of the paternal strain [32]. In 1970 pleomorphic cytoplasmatic symbionts were identified via ultra- structural analyses, and, based on morphological criteria, they were initially designated as mycoplasma-like microorganisms, MLOs [41]. These intracellular bacteria possess two and sometimes three or four enveloping membranes and were always found in testes, ovaries, and early embryos of both hybrid and nonhybrid D. paulistorum, though fewer were counted in nonhybrids. In the male reproductive organs of sterile F1 hybrids however, MLOs exist in greater numbers than they do in nonhybrid males [46,49]. It has been postulated that while each of these semispecies possessed its own MLO in a normally benign relationship, in the testes of the A recent experimental study, however, shows that artificial loss of Wolbachia by antibiotic treatment can have a significant impact on assortative mating behavior of some selected laboratory strains of D. melanogaster females, but the mechanism and biological significance in nature remains obscure [31]. Hence the actual role of endosymbionts in speciation in nature must be debated, since we lack appropriate natural and experimentally accessible model systems in early stages of speciation, where conditions that affect reinforcement can be directly tested in laboratories, as well as observed in nature. Finally, with the exception of so-called December 2010 \| Volume 6 \| Issue 12 \| e1001214 2 Infectious Speciation in Drosophila Infectious Speciation in Drosophila Figure 1. Schematic presentation of incipient speciation among D. paulistorum semispecies. Crosses between semispecies give rise in both directions to high hybrid embryonic mortality and complete hybrid male sterility (after [35] as reviewed by [42]). doi:10.1371/journal.ppat.1001214.g001 Infectious Speciation in Drosophila Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline Hybridizing PCR blots with a wsp-specific internal probe [53], however, definitely verified the presence of Wolbachia in adults of all six semispecies, i.e., AB, AM, CA, IN, OR and TR, although five of them harbor extremely low-titer infections, beyond the limit of standard PCR detection methods (Figure S1). OR and AxO hybrids (see below). In the latter two samples we never amplified Acetobacter from dissected gonads. Furthermore Acetobacter are free-living bacteria generally found as part of the gut flora in insects that have never been associated with any reproductive phenotype in literature. We therefore assume that the Acetobacter detected in AM samples can be regarded most likely as gut contaminant from gonad dissections. These data strongly implicate that Wolbachia are the unique germline associated symbiont in D. paulistorum semispecies. In addition we have performed PCR screens with a second Wolbachia-specific PCR marker set, the multi-copy transposon IS5 [54,55], on at least fifty single fly reactions per semispecies. DNA extractions and PCR reactions were set up under sterile conditions in independent replicates of ten males and ten females per semispecies, plus two simultaneously extracted single flies of the uninfected line D. simulans STC. Only such IS5 PCR data sets were counted where both negative samples were found uninfected. In total, all three hundred individuals of the six D. paulistorum semispecies were diagnosed as Wolbachia infected (data not shown). Hence the high infection status suggests perfect maternal transmis- sion of mainly low-titer Wolbachia in D. paulistorum superspecies. In order to correlate quantitative shifts of universal microbial rRNA titer levels in hybrid testes with our candidate Wolbachia, PCRs were performed on same DNA samples with a diagnostic wsp primer set. This approach is known as a robust and reliable Wolbachia-specific diagnostic marker targeting selectively the gene wsp-A encoding the Wolbachia surface protein-A [52]. Similar to observed 16S PCR patterns mentioned above, OR and AM semispecies differ dramatically in their gonad-associated symbiont titer levels (Figure 2A, lower gel). Whereas gonads of both OR sexes are highly infected with Wolbachia, via standard PCR methods no clear signals were obtained from ovaries and testes of AM semispecies. In contrast, ovaries of AxO hybrids exhibit weak but clear Wolbachia signals of expected size and much stronger ones in AxO testes. Both gonad-specific PCR assays were performed in independent triplicates exhibiting similar patterns of diagnostic Wolbachia intensities as mentioned above (data not shown). Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline Since ‘‘MLOs’’ have been diagnosed as maternally-transmitted endosymbionts that accumulate in testes, ovaries and early embryos [42,43,46,49] we have surveyed the germline-associated microbial diversity of the D. paulistorum OR semispecies by December 2010 \| Volume 6 \| Issue 12 \| e1001214 3 Infectious Speciation in Drosophila Figure 2. Presence of germline-associated microbes in D. paulistorum semispecies and their sterile hybrids. (A) PCR on total DNA extracted from each 15 pairs of ovaries (R) and testes (=) from ten day-old Amazonian (AM), Orinocan (OR) semispecies, and from hybrids (AxO) derived from matings between AM females and OR males. Controls are testes of D. simulans infected with wRi Wolbachia (+) and D. melanogaster strain w1118 uninfected (2). PCRs were performed with universal 16S rRNA (upper gel) and Wolbachia-specific wsp (lower gel) primer. (B) Wolbachia- specific wsp PCR on total DNA from ten flies of D. paulistorum semispecies and hybrids of Amazonian (AM), Centroamerican (CA), Orinocan (OR) semispecies, and F1 hybrids derived from crossings of AM females to OR males (AxO) and CA females to OR males (CxO). PCR controls are adults of D. simulans infected with wRi Wolbachia (+) and two uninfected strains (2), i.e., D. simulans STC and D. melanogaster strain w1118. DNA was extracted from 10-day-old females (f) and males (m). doi:10.1371/journal.ppat.1001214.g002 Figure 2. Presence of germline-associated microbes in D. paulistorum semispecies and their sterile hybrids. (A) PCR on total DNA extracted from each 15 pairs of ovaries (R) and testes (=) from ten day-old Amazonian (AM), Orinocan (OR) semispecies, and from hybrids (AxO) derived from matings between AM females and OR males. Controls are testes of D. simulans infected with wRi Wolbachia (+) and D. melanogaster strain w1118 uninfected (2). PCRs were performed with universal 16S rRNA (upper gel) and Wolbachia-specific wsp (lower gel) primer. (B) Wolbachia- specific wsp PCR on total DNA from ten flies of D. paulistorum semispecies and hybrids of Amazonian (AM), Centroamerican (CA), Orinocan (OR) semispecies, and F1 hybrids derived from crossings of AM females to OR males (AxO) and CA females to OR males (CxO). PCR controls are adults of D. simulans infected with wRi Wolbachia (+) and two uninfected strains (2), i.e., D. simulans STC and D. melanogaster strain w1118. DNA was extracted from 10-day-old females (f) and males (m). doi:10.1371/journal.ppat.1001214.g002 semispecies, whereas OR flies exert bright wsp PCR signals similar to the intensity of wRi of D. simulans (Figure 2B). PLoS Pathogens \| www.plospathogens.org Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline Whereas no single F1 fly emerged from control matings (both untreated) between 32 and 39 adult hybrids of both sexes emerged per assay from parents that were derived from artificially symbiont-depleted lines. For Wolbachia strain typing, wsp sequence data were obtained by PCRs from multiple independent samples per semispecies (embryonic DNA, adults and hybrids) of AM, CA, OR and TR semispecies. Within a semispecies group no sequence polymor- phism could be detected. The wsp locus of Wolbachia is encoding a rapidly evolving outer membrane protein that is under strong diversifying selection and hence permits distinguishing between even closely related Wolbachia strains [11,56,57,58]. Based on wsp data, CA, TR and OR Wolbachia (GenBank accession numbers GQ924888 to GQ924890) are almost identical to the wAu strain of D. simulans and to ancestral infections found in closely related neotropical Drosophila hosts, D. willistoni, D. prosaltans, and D. septentriosaltans [59]. As shown in Table S1 Wolbachia of CA and TR differ from OR at one unique diagnostic site in the hypervariable region HVR1 of the WSP protein [58]. Although wsp of OR semispecies is identical to the orthologous locus wSpt BCI1 of D. septentriosaltans and wCer2 of Rhagoletis cerasi, the OR infection harbors a unique variant of the hypervariable VNTR-105 [57] locus (GenBank accession number GQ924884; and unpublished data). Hence similar to earlier reports [39,41,49,65] longterm treatments at 0.01% Tetracycline and up to 0.2% Rifampicin in our laboratory are sufficient to accomplish partial rescue of hybrid male sterility and also to increase hatching rates among these D. paulistorum semispecies. Having determined sublethal dosages of these two antibiotics, Tetracycline and Rifampicin, we have observed that treated D. paulistorum females suffer low fecundity. In order to rule out unspecific side effects of antibiotics on fly oogenesis, we have applied the same treatment conditions to their closely related sister species D. willistoni, naturally infected with facultative Wolbachia [59]. Treated females of all D. paulistorum semispecies showed a significant reduction in mature egg numbers in their ovaries, whereas similar longterm treatments at 0.01% Tetracycline of closely related D. willistoni controls (Pan98) had no effect on oogenesis (Figure 4A). Scoring of mature eggs followed [66]. In D. willistoni average numbers of mature eggs per ovary before and after treatment were 18.861.30 and 19.662.69, respectively, but in the pool of the seven D. paulistorum strains tested numbers dropped down dramatically from averaged 16.6364.61 in untreated, to 4.6162.14 in treated ones. Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline Furthermore ultrastructural analyses on testes of OR (Figure 3C) AM (Figure 3F) and AxO hybrids (Figure 3I) confirmed the general presence of endosym- bionts during spermatogenesis with multiple membrane layers morphologically resembling Wolbachia and earlier descriptions [41,49]. Hence we conclude, that intracellular Wolbachia are identical to earlier designated ‘‘MLOs’’, the germline-associated core endosymbionts of D. paulistorum semispecies, and therefore pose most likely the unique microbial candidate as the causative agent driving incipient speciation in this Drosophila host system. single F1 progeny from curing assays of adults at 0.03% Tetracycline (unpublished data). Such results are in agreement with earlier observations that ‘‘MLOs’’ are obligatory microbial symbionts of all D. paulistorum semispecies and no true bacteria-free strain has ever been established [39,41,46,49,64,65]. On the other hand, mild antibiotic assay conditions, 0.01% Tetracycline or 0.01 to 0.2% Rifampicin, allowed us to establish lines of AM, CA, and OR semispecies for more than three generations on antibiotics. Furthermore we have managed to generate hybrid lines between different D. paulistorum semispecies that are stable on consistent antibiotic media (0.2% Rifampicin) for more than 12 consecutive generations under sibling mating conditions. These hybrid lines were originated from crosses between antibiotic-treated virgin AM and OR parents producing F1 hybrids of partially fertile males and fertile females. F1 progeny were allowed to mate randomly and have reached F12 at the time this manuscript was written. In contrast, control matings between untreated AM and OR semispecies have never resulted in any F2 hybrids. Detailed description of these ‘‘stabilized’’ hybrid lines will be presented elsewhere in a separate publication as soon as more generations are assayed. For evaluating the potential impact of artificial symbiont- depletion on rescuing hybrid mortality, we have crossed OR females and AM males. This combination is known to induce complete CI since we have never obtained a single F1 hybrid in repeated attempts in both of our laboratories. Antibiotic treatment of both parents, however, was sufficient to overcome complete CI and to partially rescue embryonic mortality. In these experiments interstrain matings were set up in three replicates of twenty virgin OR females and AM males each, allowed to lay hybrid eggs for 48 hrs. Parents were then discarded and hatching F1 adults were counted after two weeks. Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline Furthermore, Rifampicin treatment at 0.1% final concentration over three generations also resulted in a severe decrease of mature egg numbers in D. paulistorum where the vast majority of egg chambers were underdeveloped (Figure 4B, C). Closer inspections, using DAPI staining, showed abnormally shaped, and highly degenerate nurse cell nuclei within treated egg chambers (Figure 4D,4E), suggesting that the low fecundity observed in treated D. paulistorum females has been already provoked by defects in cell cycle and/or In AM semispecies however, a different Wolbachia wsp sequence identical to the one of wRi of D. simulans was found (GenBank accession number GQ924887). Canonical wRi Wolbachia were originally isolated from natural populations of D. simulans able to induce strong CI when infected males are mated to uninfected females [60]. The close phylogenetic relationship between wRi-like Wolbachia of AM D. paulistorum and wRi of D. simulans was corroborated by sequence data obtained from the non-coding VNTR-105 locus [57] with diagnostic base signatures characteristic of the wRi infection (GenBank accession number GQ924885). The phylogenetic relationships between Drosophila hosts and their respective Wolbachia symbionts are shown in Figure S2. Isolation and Characterization of Wolbachia as the Unique Endosymbiont Associated with the D. paulistorum Germline We therefore conclude that Wolbachia are most likely the exclusive germline associated symbiont that overreplicates in hybrid testes of D. paulistorum semispecies. As reported earlier, hybrid male sterility is caused by over- replication of maternally transmitted ‘‘MLOs’’ in the testes of interstrain hybrids that provoke breakdown of spermatogenesis and finally generate completely immotile gametes [41]. Therefore, loss of bacterial replication control in F1 testes should be correlated with a significant increase in Wolbachia infection titers in germlines of hybrids. Consequently, we have performed Wolbachia-specific wsp PCR on DNA from adult hybrid F1 females and males derived from crosses between low-titer AM and CA mothers and high-titer OR fathers. As shown in Figure 2B, Wolbachia levels shift quantitatively from extremely low-titer in AM and CA mothers, towards intermediate-titer infections in F1 hybrid of both sexes (AxO and CxO). Next we have monitored the presence of Wolbachia in individual adult females and males from AM, CA, and OR semispecies by wsp PCR. Furthermore, mature adult hybrids (AxO and CxO) obtained by mating AM and CA females to OR males respectively, were also assayed (Figure 2B). Adults of D. paulistorum semispecies, however, differ significantly in their respective Wolbachia titer levels. Under these assay conditions, no signals could be detected in adult males and females of AM and CA Furthermore ultrastructural analyses and whole mount immu- nostainings were performed applying the Wolbachia-specific WSP antibody on fixed embryos, testes and ovaries of native AM and OR hosts and AxO hybrids (Figure 3) Similar to data obtained December 2010 \| Volume 6 \| Issue 12 \| e1001214 4 Infectious Speciation in Drosophila from 16S rRNA and wsp PCR assays, OR and AM semispecies differ significantly in their respective symbiont titer levels during development. Compared to OR semispecies that exert strong Wolbachia signals in early blastodermal stages (Figure 3A), but further on concentrate specifically in primordial germline cells, PGC (Figure 3B), AM semispecies show only faint but universal staining signals during embryogenesis (Figure 3D,3E). Presence of Wolbachia was also undoubtedly detected in AM ovaries (Figure 3L). In contrast, in AxO hybrids, however, where the symbiont was inherited from low-titer AM mothers (Figure 3E), Wolbachia densities were significantly higher in blastodermal stages (Figure 3G,3H), and clear signals were detectable in hybrid testes within spermatids (Figure 3J,3K). PLoS Pathogens \| www.plospathogens.org PLoS Pathogens \| www.plospathogens.org Effect of Mild Antibiotic Treatments on D. paulistorum Female Fecundity Higher magnification of bastodermal AM embryos clearly shows low density of the symbiont (E). Signal intensities are clearly enhanced in F1s of AxO derived from matings between AM females and OR males, suggesting overreplication of the maternally-trasmitted symbiont in hybrids (G,H). Wolbachia immunostainings on cryosections of testes of AxO hybrids in transversal sections (J), and during spermatid development (K). Presence of Wolbachia during D. paulistorum oogenesis (L). doi:10.1371/journal.ppat.1001214.g003 of [59]. Transmission electron microscopy of rod-shaped pleomorphic Wolbachia cells (arrows) in the testes of fertile OR (C) and AM semispecies (F), plus sterile AxO hybrid males (I). Symbiont morphology and density in reproductive host tissues corroborate earlier pictures published in [41,49]. In OR semispecies high-titer Wolbachia accumulate in early blastodermal stages (A); in further differentiating embryos symbionts selectively target the primordial germ cells of OR (B). Presence of Wolbachia in OR testes associated with developing spermatids (C). In AM semispecies Wolbachia are present at very low-titer levels presumably in somatic and germline cells during blastodermal, gastrulating and late embryonic development (D). Higher magnification of bastodermal AM embryos clearly shows low density of the symbiont (E). Signal intensities are clearly enhanced in F1s of AxO derived from matings between AM females and OR males, suggesting overreplication of the maternally-trasmitted symbiont in hybrids (G,H). Wolbachia immunostainings on cryosections of testes of AxO hybrids in transversal sections (J), and during spermatid development (K). Presence of Wolbachia during D. paulistorum oogenesis (L). doi:10.1371/journal.ppat.1001214.g003 females = 1.01, x2 = 0.012; d.f. = 1; P = 0.91). However, higher dosages of Rifampicin (0.03% to 0.1%) were sufficient for inducing statistically significant male sex biases (P,0.0001) also in OR semispecies (Table 1). females = 1.01, x2 = 0.012; d.f. = 1; P = 0.91). However, higher dosages of Rifampicin (0.03% to 0.1%) were sufficient for inducing statistically significant male sex biases (P,0.0001) also in OR semispecies (Table 1). apoptosis during very early stages of oogenesis. Similar defects in oogenesis upon artificial symbiont depletion have been earlier reported in the parasitic wasp Asobara tabida [15]; and see Discussion). Hence, the two D. paulistorum semispecies tested on antibiotics are both highly permissive in shifting sex ratios towards males, but in drug dosage-dependent manners. Whereas sex ratio distortion becomes manifested in AM semispecies already at low concentra- tions of antibiotics, OR semispecies require more severe treatment conditions for inducing a similar phenotype. Emergence of Sex Ratio Distortions in Favor of Males upon Partial Symbiont Depletion In the course of our D. paulistorum semispecies depletion assays we have observed that mild antibiotic treatment conditions at 0.01% Tetracycline and 0.01% Rifampicin are capable of skewing sex ratios towards males in AM but not in OR semispecies. In accordance with earlier reports [42], untreated AM and OR semispecies have the usual 1:1 sex ratio (Table 1). Under mild Tetracycline treatment (0.01% over five generations) however, the AM semispecies produces a highly significant sex bias towards males (ratio males to females = 1.79, x2 = 43.27; d.f. = 1; P,0.0001), whereas the OR semispecies did not deviate from the usual 1:1 sex ratio (Table 1). Significant excess of males was also observed in AM flies kept for five generations on 0.01% Rifampicin (ratio males to females = 1.46, x2 = 26.60; d.f. = 1; P,0.0001). Like mild Tetracycline treatments, there was no effect of Rifampicin on sex ratios in OR semispecies (ratio males to Figure 4. Effect of mild antibiotic treatment on female Drosophila fecundity. (A) Mean numbers and standard deviations (SD) of mature eggs per pair of ovaries of eight-day-old females (blue bars) and after 15 generations of mild Tetracycline of 0.01% (red bars). Total n = 80. Seven D. paulistorum strains were assayed belonging to the six described semispecies Amazonian (AM), Centroamerican (CA), two Interior (IN and Ll), Andean- Brazilian (AB), Orinocan (OR), and Transitional (TR). D. willistoni (Pan98) is a control strain that was collected in 1998 in Panama naturally infected with wWil Wolbachia [59]. (B,C) Effect of mild antibiotic treatment on D. paulistorum oogenesis in 10-day-old females in (B) untreated OR control and (C) after three generations on 0.1% Rifampicin. (D,E) DAPI staining of egg chambers of (D) untreated and (E) treated OR females showing highly abnormally shaped nurse cells (white arrows). doi:10.1371/journal.ppat.1001214.g004 Figure 4. Effect of mild antibiotic treatment on female Drosophila fecundity. (A) Mean numbers and standard deviations (SD) of mature eggs per pair of ovaries of eight-day-old females (blue bars) and after 15 generations of mild Tetracycline of 0.01% (red bars). Total n = 80. Seven D. paulistorum strains were assayed belonging to the six described semispecies Amazonian (AM), Centroamerican (CA), two Interior (IN and Ll), Andean- Brazilian (AB), Orinocan (OR), and Transitional (TR). D. willistoni (Pan98) is a control strain that was collected in 1998 in Panama naturally infected with wWil Wolbachia [59]. (B,C) Effect of mild antibiotic treatment on D. Effect of Mild Antibiotic Treatments on D. paulistorum Female Fecundity Dosage-dependent induction of this sex ratio phenotype is directly correlated with our finding that OR semispecies harbor high-titer Wolbachia, while AM semispecies are only infected at low densities. We thus assume that in OR semispecies, higher dosages of antibiotics are necessary to overcome a specific threshold level, critical to affecting symbiont- density significantly, and capable for inducing sex ratio distortions in favor of males. Excess of adult males in treated D. paulistorum lines can potentially be explained by higher sensitivities of females than males to partial symbiont depletion, resulting in enhanced embryonic or larval mortality of females. Similar male-biased distortion phenotypes upon artificial microbial depletion have been reported earlier in host - symbiont systems where obligate, Effect of Mild Antibiotic Treatments on D. paulistorum Female Fecundity Currently, applying Tetracycline at 0.03% final concentration in regular food over up to three generations is the standard procedure for successfully curing a diverse range of Drosophila species of their natural symbionts without inducing significant negative fitness effects in their respective hosts [59,61,62]; but also see [63]). In contrast, D. paulistorum semispecies are hypersensitive to standard antibiotic treatments since we have never obtained a PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 5 Infectious Speciation in Drosophila Figure 3. Distribution of Wolbachia during development of D. paulistorum semispecies. OR (A–C), AM (D–F,L) and AxO hybrids (G–K AM females and OR males. Whole-mount immunostainings with the Wolbachia-specific anti-WSP antibody (yellow/green) were performed embryos (A,B,D,E, and G,H), testes (J,K), and ovaries (L); DNA is counter-stained with DAPI (blue), or propidium iodide (red) following the prot Infectious Speciation in Drosop PLoS Pathogens \| www.plospathogens.org 6 December 2010 \| Volume 6 \| Issue 12 \| e1001 Figure 3. Distribution of Wolbachia during development of D. paulistorum semispecies. OR (A–C), AM (D–F,L) and AxO hybrids (G–K) o AM females and OR males. Whole-mount immunostainings with the Wolbachia-specific anti-WSP antibody (yellow/green) were performed o embryos (A,B,D,E, and G,H), testes (J,K), and ovaries (L); DNA is counter-stained with DAPI (blue), or propidium iodide (red) following the protoco Figure 3. Distribution of Wolbachia during development of D. paulistorum semispecies. OR (A–C), AM (D–F,L) and AxO hybrids (G–K) of AM females and OR males. Whole-mount immunostainings with the Wolbachia-specific anti-WSP antibody (yellow/green) were performed on embryos (A,B,D,E, and G,H), testes (J,K), and ovaries (L); DNA is counter-stained with DAPI (blue), or propidium iodide (red) following the protocol PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 6 Infectious Speciation in Drosophila of [59]. Transmission electron microscopy of rod-shaped pleomorphic Wolbachia cells (arrows) in the testes of fertile OR (C) and AM semispecies (F), plus sterile AxO hybrid males (I). Symbiont morphology and density in reproductive host tissues corroborate earlier pictures published in [41,49]. In OR semispecies high-titer Wolbachia accumulate in early blastodermal stages (A); in further differentiating embryos symbionts selectively target the primordial germ cells of OR (B). Presence of Wolbachia in OR testes associated with developing spermatids (C). In AM semispecies Wolbachia are present at very low-titer levels presumably in somatic and germline cells during blastodermal, gastrulating and late embryonic development (D). Emergence of Sex Ratio Distortions in Favor of Males upon Partial Symbiont Depletion Untreated AM and OR semispecies have the usual 1:1 sex ratio (also see [42] and references therein). Both antibiotic treatments were performed either over five generations on 0.01% Tet or 0.01% Rif; or over three generations on 0.03% and 0.1% Rif. doi:10.1371/journal.ppat.1001214.t001 the D. paulistorum species complex, a cluster of semispecies in statu nascendi, serves as a perfect model system for elucidating the mechanisms and dynamics of incipient sexual isolation [42]. maternally-transmitted endosymbionts serve obligate mutualistic functions in filarian nematodes [67,68]; and see Discussion. Hence, this report is, to our knowledge, the first case in literature demonstrating the manifestation of a male-biased sex ratio distortion phenotype upon artificial symbiont depletion in arthropods. Based on our data, we propose that different semispecies of D. paulistorum harbor mutually incompatible fixed Wolbachia strains that cause postzygotic isolation by strong bidirectional CI and hybrid male sterility in the laboratory [42]. In reciprocal crosses between any of the six semispecies, F1 hybrid females are fertile and - under laboratory conditions - gene flow is possible between semispecies. In nature, however, incompatible matings between semispecies are avoided by female mating choice and courtship behavior [65,70]. In the D. paulistorum species-complex sexual isolation between sympatric semispecific strains is significantly greater than sexual isolation between allopatric semispecific strains [37,71]. Since all tested strains covering the six D. paulistorum semispecies are old laboratory lines, collected before the 1960s, it is possible that in the course of their long term rearing conditions these lines might have picked up Wolbachia artificially and/or that relaxed selection in the laboratory has led to artificial co-adaptations between host and symbiont. In order to overcome this criticism we have included two more recent D. paulistorum lines that were collected in Southern Brazil in 2003, (see Materials and Methods). These two lines, POA1 and POA10 were established from isofemales collected in Porto Alegre, Rio Grande. Both lines belong to the Andean-Brazilian (AB) semispecies group, since they give rise to fertile male hybrids when mated reciprocally to the Mesitas strain (AB); unpublished data. Based on wsp marker sequences both POA strains are infected with the same Wolbachia type (data not shown), but differ significantly in their respective titer levels (Figure S3). Whereas POA1 adults harbor high-titer infections similar to Orinocan semispecies, Wolbachia levels are very low in POA10 adults, similar to most other D. paulistorum semispecies. Emergence of Sex Ratio Distortions in Favor of Males upon Partial Symbiont Depletion We therefore conclude that both low- and high-titer Wolbachia are natural symbionts of D. paulistorum semispecies. More data from present-day samples from wild populations will be important to deepen our understanding in the temporal and spatial host-symbiont dynamics and their present impact on incipient speciation in nature. In order to evaluate the potential role of Wolbachia mutualists on sexual isolation we have performed mating choice assays between naturally infected and artificially depleted D. paulistorum strains of AB, AM, CA and OR semispecies. Sexual isolation indices (SIIs) were measured between combinations of heterogamic pairs of AM, CA, OR and AB (line POA1 and POA10) semispecies [71,72], before and after Rifampicin treatments for at least five consecutive generations. For AM x OR interstrain assays, lines derived from three different concentrations of the antibiotic, i.e., low-Rif at 0.01% and high-Rif at 0.1% and 0.2%, whereas AM x CA and AB x CA assays were performed at high-Rif only. As shown in Figure 5, control assays between untreated (U), naturally infected, pairs of AM and OR semispecies document high sexual isolation (assay 1, SIIAM U X OR U = 0.7060.004). Current SII of these two allopatric strains is slightly higher but similar to the index obtained in earlier assays (0.6160.07, in [37]). In contrast, sexual isolation of present-day mating choice assays between untreated AM and CA semispeciesis (assay 10, SIIAM U X CA U = 0.9060.002) is significantly stronger (P = 0.0164) than observed in earlier assays in the mid 1960s (0.7060.07, in [37]). PLoS Pathogens \| www.plospathogens.org Emergence of Sex Ratio Distortions in Favor of Males upon Partial Symbiont Depletion paulistorum oogenesis in 10-day-old females in (B) untreated OR control and (C) after three generations on 0.1% Rifampicin. (D,E) DAPI staining of egg chambers of (D) untreated and (E) treated OR females showing highly abnormally shaped nurse cells (white arrows). doi:10.1371/journal.ppat.1001214.g004 PLoS Pathogens \| www.plospathogens.org PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 7 Infectious Speciation in Drosophila Table 1. Effect of Mild Antibiotic Treatments on D. paulistorum Sex Ratios. Amazonian Males Females sex ratio1 x2 P untreated 498 470 1.06 0.81 0.36814 Tet 0.01% 347 194 1.79 43.27 ,0.0001 Rif 0.01% 456 313 1.46 26.60 ,0.0001 Rif 0.03% 597 430 1.39 27.16 ,0.0001 Rif 0.1% 544 372 1.46 32.30 ,0.0001 Orinocan Males Females sex ratio x2 P untreated 520 506 1.03 0.191 0.6621 Tet 0.01% 481 477 1.01 0.017 0.8972 Rif 0.01% 364 361 1.01 0.012 0.9113 Rif 0.03% 518 360 1.44 28.43 ,0.0001 Rif 0.1% 526 359 1.47 31.51 ,0.0001 1Sex ratio in the D. paulistorum semispecies Amazonian (AM) and Orinocan (OR) emerging from regular food (untreated) and two independent antibiotic-treatments with Tetracycline (Tet) and Rifampicin (Rif). Total number of flies counted for two weeks from initial day of hatching was 8,693 divided into 4,851 males and 3,842 females. Untreated AM and OR semispecies have the usual 1:1 sex ratio (also see [42] and references therein). Both antibiotic treatments were performed either over five generations on 0.01% Tet or 0.01% Rif; or over three generations on 0.03% and 0.1% Rif. doi:10.1371/journal.ppat.1001214.t001 1Sex ratio in the D. paulistorum semispecies Amazonian (AM) and Orinocan (OR) emerging from regular food (untreated) and two independent antibiotic-treatments with Tetracycline (Tet) and Rifampicin (Rif). Total number of flies counted for two weeks from initial day of hatching was 8,693 divided into 4,851 males and 3,842 females. Untreated AM and OR semispecies have the usual 1:1 sex ratio (also see [42] and references therein). Both antibiotic treatments were performed either over five generations on 0.01% Tet or 0.01% Rif; or over three generations on 0.03% and 0.1% Rif. doi:10.1371/journal.ppat.1001214.t001 1Sex ratio in the D. paulistorum semispecies Amazonian (AM) and Orinocan (OR) emerging from regular food (untreated) and two independent antibiotic-treatments with Tetracycline (Tet) and Rifampicin (Rif). Total number of flies counted for two weeks from initial day of hatching was 8,693 divided into 4,851 males and 3,842 females. Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance For each array five replicates and 120 matings were scored (12A RR +12B RR +12A == +12B == differentiated by rotated wing clips) for each row, totaling 2,160 matings. Two-tailed P values were calculated by comparing SIIs of untreated and treated pairs of mating choice assays by Fisher’s exact test. Statistically significant results are indicated by one, two or three asterisks, i.e., P,0.05; P,0.01 and P,0.001, respectively. Abbreviations: Amazonian (AM); Centroamerican (CA); Orinocan (OR) and Andean-Brazilian (AB) semispecies (lines POA1 and POA10). U = untreated; T = treated with antibiotics. doi:10.1371/journal.ppat.1001214.g005 Combinations between the two Wolbachia low-titer semispecies CA and AM (assays 10–13) showed that partial depletion of the symbiont via 0.1% Rifampicin decreases levels of mate discrim- ination by about 50% in all three directions, i.e. in double and both unilateral treatments. SII dropped from an original 0.9060.002 (both untreated) to 0.3860.007 (AMT x CAT), 0.4660.007 (AMT x CAU), and 0.4460.007 (AMU x CAT), respectively. For all three combinations reductions in mate discrimination were highly significant (z = 3.89, P,0.0001). isolation is not statistically significant under both treatment conditions (z = 1.88 for 0.01% Rif and 1.41 for 0.1% Rif, both P.0.05). Under 0.2% Rifampicin, however, double treatment of both partners induced almost random mating (assay 4; and Table S2, z = 3.89, P,0.0001). Unidirectional selective treatment of OR semispecies at low-Rif conditions slightly reduced sexual isolation when OR was mated with untreated AM semispecies, from SIIboth untreated = 0.706 0.004 to SIIAM U X OR T = 0.5460.004 (assay 5), although this reduction is not statistically significant (z = 0.18, P.0.05). More intense treatments of OR semispecies with 0.1% Rifampicin, however, revealed a 33% reduction of assortative matings (assay 6), compared to both untreated controls (SIIAM U X OR T = 0.4860.007 vs. SIIAM U X OR U = 0.7060.004). Although increased dosages of antibiotics in OR semispecies further reduce sexual isolation in AMU x ORT assays, Fisher’s Exact tests and t tests show only weak statistical significance (z = 1.93; P = 0.0539). Similar to double treatment assays (assays 2–4), in OR semispecies higher doses of Rifampicin up to 0.2% were necessary to significantly reduce sexual isolation (SII = 0.3360.007) in heterogamic mating choice assays (assay 7; Table S2, z = 3.19, P,0.0014). g ( ) Finally we have assayed the effect of partial Wolbachia depletion on assortative mating behavior in our most recently collected D. Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance Reinforcement is the process by which postmating isolation acts as a direct selective pressure fostering the evolution of premating isolation in areas of sympatry [2,69]. In this respect, reinforcement is judged the final most efficient stage in speciation processes that block gene flow to incipient species via mating discrimination. Understanding how sexual isolation evolves requires that we capture the process before it has reached completion. Therefore, Assays performed on AM and OR semispecies, where both partners were treated (AMT x ORT), resulted in a weak enhancement of SII from 0.7060.004 in both untreated to 0.8660.002 and 0.8460.003 under low- and high-Rif conditions, respectively (assays 1 to 3). However, this slight effect on sexual December 2010 \| Volume 6 \| Issue 12 \| e1001214 8 Infectious Speciation in Drosophila Figure 5. Mating preferences in combinations between untreated and treated heterogamic pairs of D. paulistorum semispecies. y- axis represents sexual isolation index (SII); number of mating assay (1–18) is shown on x-axis (corresponding to assay numbers in Table S2). Grey bars indicate untreated controls; black bars indicate assays with Rifampicin treated flies: Tested lines were kept on 0.01% Rifampicin for ten generations; or on 0.1% and 0.2% Rifampicin for five generations. SII and standard error (SE) were determined following [75]. For each array five replicates and 120 matings were scored (12A RR +12B RR +12A == +12B == differentiated by rotated wing clips) for each row, totaling 2,160 matings. Two-tailed P values were calculated by comparing SIIs of untreated and treated pairs of mating choice assays by Fisher’s exact test. Statistically significant results are indicated by one, two or three asterisks, i.e., P,0.05; P,0.01 and P,0.001, respectively. Abbreviations: Amazonian (AM); Centroamerican (CA); Orinocan (OR) and Andean-Brazilian (AB) semispecies (lines POA1 and POA10). U = untreated; T = treated with antibiotics. doi:10.1371/journal.ppat.1001214.g005 Figure 5. Mating preferences in combinations between untreated and treated heterogamic pairs of D. paulistorum semispecies. y- axis represents sexual isolation index (SII); number of mating assay (1–18) is shown on x-axis (corresponding to assay numbers in Table S2). Grey bars indicate untreated controls; black bars indicate assays with Rifampicin treated flies: Tested lines were kept on 0.01% Rifampicin for ten generations; or on 0.1% and 0.2% Rifampicin for five generations. SII and standard error (SE) were determined following [75]. PLoS Pathogens \| www.plospathogens.org Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance In addition, treated OR females mate with untreated AM males (T/U) significantly more frequently (P,0.05) than do untreated females (Figure 6F); higher concentra- tions of the antibiotic (0.2%) further enhanced effects on mate recognition patterns of OR females in a dose-dependent manner (Figure 6G). Similar to the ‘‘old’’ laboratory strains mentioned above, treated females of the two more recently established AB lines showed similar behavioral patterns (Figure 6H,6I). To sum up, our data imply that partial depletions of Wolbachia in females of AB, AM, CA and OR semispecies significantly impair their mate avoidance behavior; they lose proper mate recognition signals and finally accept improper heterogamic males. In addition there is a direct correlation between the natural Wolbachia titer level in a defined semispecies and their sensitivity to alterations in assortative mating behavior upon partial symbiont depletion. OR semispecies, harboring high-titer levels of their core-endosymbiont, seem more recalcitrant to antibiotic treat- ments than AM and CA. Similar to the sex ratio distortion phenotype (Table 1), OR semsipecies need higher dosages of antibiotics (up to 0.2% Rifampicin) in order to reach the critical threshold level sufficient for triggering statistically significant changes of mating behavior. A similar but statistically not significant correlation between natural symbiont titer level and drug competence was observed in POA lines belonging to AB semispecies. Whereas the high-titer line POA1 showed a moderate 46.7% reduction in mate avoidance after partial symbiont depletion, in the low-titer POA10 line a 63.2% reduction was observed (Figure 5, assays 14–17). Besides this effect observed in treated females, partial Wolbachia- depletion also affects male behavior. As shown in Figure 6C,6D treated AM and CA males succeed in copulating significantly more frequently with untreated heterogamic females (U/T) than do untreated males. This enhanced mating frequency on the part of treated males is statistically significant (P,0.05 for CA and P,0.001 for AM males, respectively). Enhanced mating success of treated AM males (P,0.05) was also observed in combinations with untreated OR females (Figure 6F), suggesting that partial symbiont-depletion is capable of also changing male sexual behavior in both low-titer semispecies. Our current Rifampicin conditions up to 0.2% applied to OR semispecies, however, were not sufficient to induce a similar change in male mating behavior, since mating success of treated OR males (U/T) equals the one observed for untreated males (U/U) in heterogamic AM assays (Figure 6E,6G). Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance paulistorum samples belonging to AB semispecies, i.e., high-titer POA1 and low-titer POA10 lines (Figure S3). Mating choice assays were performed in heterogametic combinations between untreated CA semispecies and both POA lines under high Rifampicin treatment conditions (Figure 5, assays 14–17). Similar to ‘‘old’’ laboratory strains of AM, CA and OR semispecies, both AB lines obtained from more recent collections showed statistically significant reduction in their assortative mating behavior after partial symbiont depletion in combinations with untreated CA semispecies (Table S2, z = 3.89, P,0.0001). SII dropped from original 0.9260.001 (POA1 in assay 14), and 0.9560.001 (POA10 in assay 16) to 0.4960.006 (assay 15), and to 0.3560.007 (assay 17), respectively. In contrast to OR harboring Wolbachia at high-titer, selective antibiotic treatments of low- titer AM semispecies trigger dramatic effects on assortative matings in AMT x ORU mating choice assays (assays 8 and 9). Combinations between AM treated and OR untreated lines resulted in the complete breakdown of sexual isolation towards random mating between these two D. paulistorum semispecies. Random mating was observed in combinations of AMT x ORU under conditions of low- (0.0660.008) and high Rifampicin (0.1860.008), where, in both cases, reductions were statistically highly significant (for both treatment conditions: z = 3.89, P,0.0001). In order to rule out potential toxic side effects of antibiotics on Drosophila mating behavior and sexual isolation, we have generated symbiont-depleted lines of the D. simulans Riverside strain naturally infected with wRi Wolbachia [60]. 0.2% Rifampicin treatments were performed for five consecutive generations in accordance to our depletion protocol for D. paulistorum lines (Materials and Methods). Standard mating choice assays were performed between untreated (D. simU) and treated (D. simT) PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 9 Infectious Speciation in Drosophila significant alterations were observed in treated high-titer OR females (Figure 6B). samples. As shown in Figure 5, high doses of Rifampicin have no effect on random mating behavior in native D. simulans-Wolbachia associations (assay 18; and Table S2). These data implicate that loss of assortative mating behavior in D. paulistorum semispecies was most likely caused by disturbing natural host-symbiont associa- tions, rather than by any unspecific toxic drug-effect. Under 0.1%-Rif conditions, however, treated AM and CA females (Figure 6C–6E) accepted reciprocal heterogamic males of untreated (T/U) and treated origins (T/T) significantly more often than do untreated females (P,0.01). Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance These data suggest that even our current 0.2% treatment conditions are not sufficient for altering mating behavior of OR males that naturally harbor high-titer Wolbachia. Summing up, the mating choice assays presented here imply strong effects of Wolbachia on premating isolation between D. paulistorum semispecies in a titer-dependent manner. Next we asked the question as to whether artificial symbiont depletion modifies mate avoidance patterns on the part of D. paulistorum females. For that purpose we have determined female mating frequencies of successful heterogamic matings obtained from mating choice assays itemized in Figure 5 and Table S2. As shown in Figure 6, untreated females of AB, AM, CA and OR semispecies strongly oppose heterogamic males, resulting in rare heterogamic mating events under our assay conditions. Under low-Rif conditions (0.01%) only AM-treated females lose mate discrimination by accepting heterogamic OR males (Figure 6A), whereas no Figure 6. Frequencies of successful heterogamic matings of D. paulistorum females. Box plots represent distribution of mating frequencies obtained from interstrain mating choice assays with Amazonian (AM), Centroamerican (CA), Orinocan (OR) and Andean-Brazilian (AB; lines POA1 and POA10, respectively) semispecies (females are first named). Tested lines were kept on normal food, or on 0.01% (A,B), 0.1% (C–F) and 0.2% Rifampicin (G–I). Mating frequency was determined by number of successful heterogamic matings out of twelve females in five replicas each (Table S2). Combinations of heterogamic mating choice pairs are indicated by U/U = both partners untreated; T/T = both treated; T/U = females treated, males untreated; and U/T = females untreated, males treated. Statistical significant values are indicated by one, two or three asterisks, i.e., P,0.05, P,0.01 and P,0.001, respectively; and statistical outliers by black stars. doi:10.1371/journal.ppat.1001214.g006 Figure 6. Frequencies of successful heterogamic matings of D. paulistorum females. Box plots represent distribution of mating frequencies obtained from interstrain mating choice assays with Amazonian (AM), Centroamerican (CA), Orinocan (OR) and Andean-Brazilian (AB; lines POA1 and POA10, respectively) semispecies (females are first named). Tested lines were kept on normal food, or on 0.01% (A,B), 0.1% (C–F) and 0.2% Rifampicin (G–I). Mating frequency was determined by number of successful heterogamic matings out of twelve females in five replicas each (Table S2). Combinations of heterogamic mating choice pairs are indicated by U/U = both partners untreated; T/T = both treated; T/U = females treated, males untreated; and U/T = females untreated, males treated. Discussion Uncovering Hidden Wolbachia Diversity in D. paulistorum Semispecies In agreement with earlier reports, D. paulistorum semispecies are highly sensitive to antibiotics [41]. In addition, ultrastructural analyses have shown that no D. paulistorum fly has ever been observed that is devoid of endosymbionts [39,41,49]. Since complete clearance of Wolbachia is obviously lethal for all D. paulistorum semispecies, it has been suggested that core-endosym- bionts of D. paulistorum are well-adapted obligate mutualists that serve essential vital functions benefiting to their hosts (reviewed in [42,65]). In this study we note that partial symbiont depletion under mild antibiotic condition is capable of triggering alterations in (i) female fecundity, (ii) male-biased sex ratio, and (iii) most significantly, female mating behavior in D. paulistorum. The presence of low-titer Wolbachia in five out of the six D. paulistorum semispecies corroborates earlier ultrastructural obser- vations, showing that ‘‘MLOs’’ reside at low densities in the reproductive organs of their natural hosts, but heavily over- replicate in hybrids [49]. This quantitative shift of Wolbachia densities from extremely low in native hosts to intermediate in interstrain hybrids, strongly suggests that the causative agent of incipient D. paulistorum speciation are Wolbachia. Moreover no other germ line-associated microbes were isolated by different means of experimental 16S consensus PCR approaches. In adult flies of AB, AM, CA, IN and TR semispecies, densities of native endosymbionts are too low for unambiguous detection by standard wsp PCR protocols. As such, earlier attempts have failed to isolate Wolbachia from members of the D. paulistorum species complex and hence were diagnosed as Wolbachia-free [59,73]. Importantly, none of these earlier studies reported surveys neither of the high-titer OR semispecies, nor of F1 interstrain hybrids, where Wolbachia is ostensible. In a recent publication we have applied our PCR-wsp-hybridization strategy [53] to uncovering multiple low-titer Wolbachia strains in the cherry fruit fly, Rhagoletis cerasi, not accessible by standard PCR methods in earlier studies. We hence encourage considering the broader existence of low-titer Wolbachia in nature, commonly escaping conventional detection approaches, before designating a sample or even a whole group under study as uninfected. The female fecundity phenotype, induced by artificial depletion of Wolbachia (Figure 4), engenders similar effects on oogenesis that have been detected in the parasitoid wasp Asobara tabida, where Wolbachia is an obligate symbiotic partner [80,81]. Female wasps that are cured of their Wolbachia, fail to produce any mature oocytes, rendering Wolbachia infection essential for female reproduction. Discussion In this case of mutualism, Wolbachia influence programmed cell death processes in A. tabida by modulating apoptosis during oogenesis [15]. The authors speculate that after a loss-of-function mutation in a key regulator gene of the A. tabida apoptosis network system, Wolbachia were capable of permanently taking over control of the host apoptotic pathway. Another example of the compensatory effect of Wolbachia on a host loss-of- function mutation, comes from laboratory strains of D. melanogaster, where the presence of Wolbachia is sufficient to restore the ovarian phenotype in Sxlf4 mutants [66]. The most recent example showing that Wolbachia can transform even in an obligate nutritional mutualist residing in a bacteriome was found in the bedbug Cimex lectularius, where artificial depletion of its symbiont results in sterility and retarded growth [82]. y Based on wsp and VNTR sequence analyses, Wolbachia of CA, OR and TR semispecies are host-specific and are closely related with, but not identical to, other wAu-like strains isolated from phylogenetically germane neotropical Drosophila hosts [59] that harbor unique diagnostic sites (Table S1). As shown in Figure S3, apparent conspecificity between native hosts and defined Wolbachia strains of D. paulistorum semispecies and their related sister species suggests an ancestral association assisted by stable vertical transmission. In order to elucidate the coevolutionary history of this host-symbiont interaction, more detailed molecular Wolbachia- strain-typing analysis will be necessary for D. paulistorum semi- species, from strains kept since the 1950s in the laboratory, as well as from recent samples of natural populations, by means of applying more informative, preferentially non-coding Wolbachia markers other than the wsp locus or other protein coding genes. As such, there are a growing number of cases in recent Wolbachia literature, where strains sharing identical wsp sequences induce clearly distinguishable phenotypes [74–77]. Importantly, these reported cases of phenotypic transitions of Wolbachia in fruit flies, aphids and butterflies, took place in extremely short evolutionary time scales, implying that these symbiotic associations are highly dynamic and rapidly evolving in natural and laboratory populations. The second phenotype, uncovered by mild antibiotic D. paulistorum treatments (Table 1), is, to our knowledge, the first report on a symbiont-induced male-biased sex ratio distortion in arthropods. In filarial nematodes, significant male-biased sex ratio distortion was observed after partial depletion of mutualistic Wolbachia by Tetracycline [67,68]. Effects of Wolbachia-Depletion on Premating Isolation and Female Mate Avoidance Statistical significant values are indicated by one, two or three asterisks, i.e., P,0.05, P,0.01 and P,0.001, respectively; and statistical outliers by black stars. doi:10.1371/journal.ppat.1001214.g006 PLoS Pathogens \| www.plospathogens.org PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 10 Infectious Speciation in Drosophila PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 Discussion Moreover, this criticism has been refuted in Bordenstein 2003 and is not Wolbachia-specific since it applies to any type of postzygotic incompatibilty [12]. In our Wolbachia-D. paulistorum model system however, theoretical caveats opposing most speciation models can be overruled by evolving ancestrally fixed, obligate mutualistic interrelations between the endosymbiont and its natural host, vital for both partners. So, even partial titer depletion of the obligate core symbiont below a critical Wolbachia threshold, results in severe pleiotropic and primarily female-biased phenotypes, which natural selection acts against. contributed by Wolbachia in order to develop properly and to mature gametes. In addition, immune responses of a host against a-proteobacteria frequently involve apoptosis of infected cells, and, in turn, symbionts can inhibit host cell suicide to circumvent the immune system [85]. As recently shown, the Wolbachia-encoded surface protein WSP is capable of inhibiting apoptosis in vitro [86]. Hence the permanent association between Wolbachia and D. paulistorum hosts is likely to be the result of a tight coevolution between an ancestral host responding to infection with apoptosis, the symbiont suppressing it, and the host compensating for this suppression because apoptosis is required for development. pp p p q p Alternatively but not exclusively, depletion of Wolbachia in D. paulistorum semispecies by mild antibiotic treatment might have unleashed a cryptic, but female-biased loss of function phenotype associated with the Insulin/IGF-like signaling (IIS) pathway. Mutations that reduce the activity of the IIS signaling cascade have pleiotropic effects on many traits, such as development, metabolic homeostasis, adult lifespan and fecundity in nematode worms, fruit flies, and mice (reviewed by [87]). As was recently shown, removal of symbiotic Wolbachia in naturally infected D. melanogaster reduces IIS and hence enhances mutant IIS pheno- types, suggesting that Wolbachia normally act to increase insulin signaling [88]. Since some of the key downstream effectors of the IIS cascade are also functionally implicated in regulating apoptosis [89,90], one may assume that these two models explaining the evolution of compensatory mutualistic Wolbachia in D. paulistorum semispecies, are not mutually exclusive. Under both evolutionary scenarios, however, closely related but host-specific Wolbachia strains might act as general suppressors of yet undefined, cryptic, female-biased mutations already present in the common ancestor of all the D. paulistorum superspecies. Discussion This assumption is corroborated by our findings that at least three out of the six infected semispecies (CA, OR and TR), from which we have already obtained sufficient information on Wolbachia-strain diag- nostics, harbor strain-specific and closely related wAu-like endosymbionts (Table S1). In addition, the TR semispecies is regarded as the relict ancestral semispecies of the D. paulistorum superspecies [78,79] further supporting their long term relation- ship and common ancestry with wAu-like Wolbachia symbionts. Second, the few empirical systems from present literature concerning roles of symbionts in speciation, such as parasitoid wasps of the Nasonia sibling species group, are highly informative in elucidating the evolutionary implications of endosymbionts [28,29]. Bordenstein and colleagues showed that in the youngest Nasonia species pair, N. giraulti and N. longicornis, Wolbachia-induced bidirectional CI was the primary form of reproductive isolation prior to the evolution of other isolating barriers. However, these species are not known to occur sympatrically in nature and premating isolation tested in the laboratory was weak and asymmetric [12,29]. Hence it has been suggested to focus on natural symbiotic systems of sympatric, hybridizing host species that are isolated because of CI [12]. Indeed, the mushroom feeding Drosophila species D. recens and D. succinea are found in a hybrid zone in central Canada. Moreover they express strong unidirectional CI triggered by Wolbachia in D. recens, plus asymmetrical premating isolation and hybrid male sterility. The latter two isolation mechanisms, however, are genetically based [20,30]. The assortment of symbiotic and genetically based isolation barriers provides clear evidence that Wolbachia can act in concert with genetic and/or geographic isolation mechanisms in nature. Our incipient speciation system presented here is in flagrant statu nascendi, occurs sympatrically in consistently overlap- ping geographic distributions in Middle and South America [43], exerts strong bidirectional CI [44–46], plus pronounced premating isolation in a symbiont-dependent manner (this study). p y y As such, regulation of host-programmed cell death during female development, as well as suppression of a cryptic IIS mutation in D. paulistorum semispecies, are both plausible mechanisms that can explain the evolutionary transition of Wolbachia from facultative parasitism towards obligate mutualism via compensatory evolution. Under both models, D. paulistorum Wolbachia can be regarded as obligate mutualists by compensating reduced female-specific fitness effects that were originally caused, either by a suppressor of IIS mutation, or by reproductive parasites such as Wolbachia. Discussion These authors suggest that if Wolbachia play a more active role in females than in males, this may imply a more biosynthetic activity of Wolbachia in female hosts and thus greater susceptibility to antibiotics. This assumption is based on evolutionary models [83,84], asserting that in a mutualistic relationship, selection on inherited microorganisms should (i) select directly for beneficial effects towards the host sex responsible for their transmission (in general, the female), and (ii) favor the spread of phenotypes that are detrimental to those hosts not involved in their transmission (males or uninfected females). In this respect, the profound sensitivity of D. paulistorum females to Wolbachia-depletion (loss-of-function) plus the hybrid male sterility phenotype caused by Wolbachia overreplication in testes (gain-of- function) corroborates this evolutionary model. In contrast to the wAu-like Wolbachia of CA, OR and TR, the core-endosymbiont of AM semispecies is unique in that it groups closely with the wRi infection of D. simulans, suggesting a more recent acquisition via horizontal transmission from a yet uncharacterized external source. Potentially, this new Wolbachia strain has replaced an ancestral wAu-like symbiont, similar to CA, OR and TR Wolbachia, already present in the AM progenitor lineage. The uniqueness of wRi-like Wolbachia in AM corroborates with earlier phylogenetic studies, designating AM as the most distal semispecies of the D. paulistorum complex [78,79]. Both phenotypes, low female fecundity and male-biased sex ratio distortions, induced by mild antibiotic treatments of D. paulistorum semispecies (Figure 4 and Table 1), can be explained by strong obligate mutualistic interactions between Wolbachia and their natural neotropical hosts, i.e., suppression of enhanced cellular mortality during female development or a direct effect on the Insulin/IGF-like signaling cascade. The first model assumes that the presence of Wolbachia above a specific threshold level plays an essential role in female, but not male, development by suppressing female lethality. For example, female larvae or pupae require a nutrient vitamin or hormone PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 11 Infectious Speciation in Drosophila incompatibilities [1,22]. First, as deduced from theoretical models, CI causes unstable equilibria in prevalence and transmission rate of symbiontic infections [16,26,27], but more recent models show that such stabilities can exist, depending on parameters such as host migration rate, CI induced gene flow reduction and the invasion success of mutants at a mate preference locus [21]. Discussion In each of the two scenarios, a former reproductive parasite became domesticated by its host system, thereby providing more recently evolved immunity against genomic loss of function mutations - and intracellular parasites, respectively. In summary, the coevolutionary interaction between D. paulistorum semispecies and Wolbachia infection has changed from antagonistic into reciprocally beneficial and thereby into an obligate mutualistic interaction, where the both systems are dependent on each other now. Finally, no empirical cases have been reported suggesting that endosymbionts are able to cause hybrid-male sterility, a hallmark of an early stage in the evolution of postzygotic isolation [1], but also see [32], and this study. However, the novel Wolbachia-induced phenotype of symbiont-induced hybrid male sterility reported here, can be regarded as currently residual and supportive, since all D. paulistorum semispecies have evolved strong prezygotic isolation mechanisms and intriguing female mate avoidance patterns by successfully preventing inter-semispecific matings [37]. Furthermore, natural hybrids between sympatric semispecies have never been reported in the field, nor has a trapped sperm- storing gravid female ever borne alien sperm (L.E. personal communication). After having demonstrated that Wolbachia are ancestrally fixed, obligate entities of all D. paulistorum semispecies, perfectly transmitted by the mother and vital for female oogenesis and female development, we have evaluated effects of mild symbiont depletion via antibiotics on female mating choice. Currently a growing number of reports emerge in our literature comparing specific behavioral traits possessed by symbiont-infected and uninfected individuals, such as mate avoidance, male promiscuity, longevity, pathogenic fungal resistance, and olfactory-cued PLoS Pathogens \| www.plospathogens.org The Impact of Wolbachia on Infectious Speciation Increasing empirical evidence and numerous theoretical models predict that transovarialy transmitted microbial symbionts can have significant impacts as drivers of speciation processes in their natural hosts [12,17–21]. However, there are substantial difficul- ties with speciation based solely on symbionts capable of inducing December 2010 \| Volume 6 \| Issue 12 \| e1001214 12 Infectious Speciation in Drosophila locomotion [31,91–95]. In spider mites, for just one example, Wolbachia-associated unidirectional CI can be avoided by females at the premating level [91]; they show spider-mite females exhibiting both precopulatory and ovipositional behaviors that increase odds of successful compatible matings. Furthermore, in mating choice experiments, uninfected females preferably mate with uninfected males, and in doing so, directly reduce opportunities for Wolbachia-induced unidirectional CI expression [91]. Colombia, TR. For full descriptions of origins and maintenance of all Drosophila paulistorum strains, see [42], and references therein. POA1 and POA10 strains, belonging to the Andean Brazilian (AB) semispecies are isofemale lines that were generated from collections in April 2003 in Porto Alegre City, Rio Grande do Sul State Brazil. Both lines were kindly provided by Yong-Kyu Kim, Emory University, Atlanta, GA, USA. The control strain Pan98, collected in Panama in 1998, belongs to the closely related sister species D. willistoni, naturally infected with wWil Wolbachia [59]; and two D. simulans lines, Riverside (RI) infected with wRi Wolbachia [60] plus the uninfected strain STC [100] serve as positive and negative controls, respectively. So according to the scenario observed in spider mites, we will propose the following model for understanding the biological role of Wolbachia driving sexual isolation between D. paulistorum semispecies: Obligate mutualistic Wolbachia, having evolved tight compensatory interrelations with their natural hosts, are capable of triggering female mate recognition patterns in favor of preferen- tially accepting males harboring the same type of Wolbachia - i.e., members of their natal semispecies - and to reject improper mates carrying an alien version of their obligate symbiont. In the mixed genetic background of artificially generated hybrids however, maternally transmitted, naturally benign mutualistic Wolbachia turn into reproductive parasites, capable of inducing strong embryonic bidirectional CI and complete male hybrid sterility via loss of their controlled replication. Although F1 hybrid females are fertile under laboratory conditions, they show an intriguingly altered sexual ‘‘old maid’’ behavior by hardly accepting any kind of mating partner [96]. Generation of D. paulistorum Interspecies Hybrids Generation of D. paulistorum Interspecies Hybrids Sexually mature virgin females of AB, AM, CA and TR semispecies were forced to mated en mass with a twofold excess of heterogamic OR males, respectively. These males had been isolated from females for several days. Flies were transferred into fresh vials every 48 hrs over a period of two weeks. Germline-Associated 16S rRNA PCR Screen Germline Associated 16S rRNA PCR Screen In order to determine the microbial diversity of germline associated symbionts of D. paulistorum OR semispecies, we have performed 16S rRNA consensus PCRs after [50] on five independent DNA samples, derived from sterilized 0–24 hrs embryo. From each sample, at least four amplified 16S rRNA consensus fragments were cloned and sequenced. All twenty sequence-reads were identical with the 16S rRNA gene of Wolbachia pipiens of D. willistoni (accession number DQ412086), and no other bacterial sequence was detected. To further confirm the presence of Wolbachia in reproductive organs of D. paulistorum we performed two independent 16S rRNA universal PCRs after [50,51] on sterile-dissected ovaries and testes of AM, OR semispecies and AxO hybrids derived from crossings of AM females with OR males. Two different primer sets amplifying overlapping universal microbial 16S rRNA fragments between 0.4 and 0.6 kb were used. Amplicons were purified using the Promega Wizard SV Gel PCR Clean-up system. Sequences were obtained via cycle sequencing with BigDye Terminator v3.1 performed at the Department of Marine Biology, University of Vienna, Austria. Sequences were analyzed using ApE plasmid editor v1.10.4 (M. Wayne Davis), ClustalX, (www.clustal.org), Mesquite v2.6 (www. mesquiteproject.org), and the BLAST algorithm (www.ncbi.nlm. nih.gov). Molecular Isolation and Strain Typing of D. paulistorum Semispecies Wolbachia High quality genomic DNA was extracted from a pool of ten adult flies, 15 ovaries/testes or from 30–50 mg embryos using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). Eggs were collected from apple juice agar plates in 24 hrs intervals and sterilized by extensive washes with 70% EtOH before DNA extraction. Testes and ovaries were dissected in sterile 1x PBS. All DNA samples were stored at 220uC until use. PLoS Pathogens \| www.plospathogens.org The Impact of Wolbachia on Infectious Speciation Hence we assume that symbiont-directed mate recognition could have evolved in order to prevent strong bidirectional CI and reduced sexual success of potential hybrids, thereby ensuring their continuing vertical transmission. Alterna- tively, native hosts might have evolved behavioral avoidance patterns that recognize potential mates carrying distinctive but incompatible Wolbachia variants [91]. As suggested by Koukou and colleagues Wolbachia might have evolved the capacity to modulate host pheromone expression and/or perception [31]. Recently, Albertson and colleagues [97] have shown that Wolbachia significantly concentrate in specific brain regions (i.e., the subesophageal ganglia, superior protocerebra and antennal lobe) of D. melanogaster and D. simulans; speculating that the presence of the endosymbiont could potentially affect insect behavior, such as olfactory-cued locomotion and mating behavior [95]. By analytical organic methods it was recently discovered that within the D. paulistorum complex the active male and female pheromonal compounds differ in proportions present in each of the six D. paulistorum semispecies - that is, quantitatively not qualitatively [79,98]. In accordance with theses data, we revealed that Wolbachia are also tightly associated with the antennal lobe of D. paulistorum semispecies (unpublished data). Hence, our D. paulis- torum-Wolbachia symbiosis system, affecting premating isolation between natural semispecies, provides a unique and excellent model system for studying the global impact of microbial endosymbionts on pheromone production, olfaction, and sexual behavior in insects. Measuring Sexual Isolation Mating choice assays, monitoring the effect of artificial symbiont depletion via Rifampicin at 0.01% and 0.1% for at least five generations, involved direct mating observations. All details concerning our experimental protocols have been itemized by [37], and reviewed and adjudicated by [104]. Briefly: All replicas were conducted mornings at room temperature in daylight facing north. Beforehand, virgin flies were aged two to three days after light ether anesthetization during which they were sexed (there are no sex differences in abdominal banding in this superspecies); and half were marked via minute distal wing clips (controlled in bioassays). These marks were rotated (wing to wing and treated to untreated). Such abrasions have tested neutral regarding behav- ioral influences [105] in this superspecies. Fly Strains Diagnostic Wolbachia PCR, Cloning and Sequencing wPau Wolbachia was detected using primer specifically targeting the wsp locus, the transposon IS5 as well as VNTR-105 and VNTR- 141 locus. wsp-primer sets were taken from [52] and primers for VNTRs and IS5 PCR were taken from [57]. In all reactions the uninfected D. simulans strain STC was included as negative controls. Diagnostic wsp-PCR reactions were performed under the following conditions: 10 ml reactions containing 1 ml of genomic The D. paulistorum species complex [99] is neotropical and comprises at least six semispecies in statu nascendi [35]. They are morphologically indistinguishable, but can be identified by allozymes, chromosomes and courtship behavior [42]. For this study, the following four strains, belonging to Amazonian (AM), Centroamerican (CA), Orinocan (OR), and Transitional (TR) D. paulistorum semispecies were used: Belem, Brazil, AM; Lancetilla, Honduras, CA; Georgetown, Guyana, OR; and Santa Marta, PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 13 Infectious Speciation in Drosophila DNA template in 1x polymerase reaction buffer (Promega, Madison, USA), 4.5 mM MgCl2, 0,3 mM of each primer, 150 mM of each dNTP and 0.4 U of Taq polymerase (Promega, Madison, USA). Diagnostic VNTR-PCRs were performed as follows: 10 ml reactions containing 1 ml of genomic DNA template in 1x polymerase reaction buffer (Promega, Madison, USA), 2,5 and 4 mM MgCl2 for VNTR-141 and 2105, respectively, 0,3 mM of each primer, 35 mM of each dNTP and 0.4 U of Taq polymerase (Promega, Madison, USA). PCR-profiles for wsp amplifications were taken from [52]. For VNTR-141 locus, amplifications consisted of 5 min at 94uC followed by 34 cycles of 1 min at 94uC, 1 min at 50uC and 1 min 30 sec at 72uC. Final elongation step was performed at 74uC for 8 min. For VNTR-105 a higher annealing temperature of 61uC was used and elongation was performed for 3 min instead of 1 min 30 sec. IS5 PCRs were applied as described in [55]. For all PCR reactions a Biometra T3000 Thermocycler (Biometra, Goettingen, Germany) was used. PCR products were purified using the peqGOLD Gel Extraction Kit (peqLab, Erlangen, Germany). For cloning, products were inserted into the pTZ57R/T vector (Fermentas, St. Leon-Rot, Germany), and used to transform competent DH5a Escherichia coli cells. Clones containing the insert were cycle sequenced with BigDye Terminator v3.1 at the Department of Marine Biology, University of Vienna, Austria. Sex Ratio Assays Sex ratio in the D. paulistorum semispecies Amazonian (AM) and Orinocan (OR) emerging from regular food (untreated) and two independent antibiotic-treatments with teracycline (Tet) and Rifampicin (Rif). Total number of flies counted for two weeks from initial day of hatching was 8,693 divided into 4,851 males and 3,842 females. Untreated AM and OR semispecies have the usual 1:1 sex ratio (also see [42]; and references therein). Both antibiotic treatments were performed either over five generations on 0.01% Tetracycline (Tet) or 0.01% Rifampicin (Rif); or over three generations on 0.03% and 0.1% Rifampicin (Rif). Ovary Phenotype Assay Analysis of ovary phenotype followed [66]. Females were raised on standard food with or without antibiotics (0.01% Tetracycline for fifteen generations and 0.1% Rifampicin for three generations) at 25uC and were dissected ten days after eclosion. Only eggs reaching stage 13 (after [101], indicated by filaments) were counted as mature eggs. For DAPI staining, ovaries were dissected ten days after eclosion in sterile Drosophila Ringer’s Solution. Ovaries were stained in a DAPI-TBST solution (1 mg/l) for 5 minutes under constant agitation and mounted in Citifluor Glycerol/PBS solution (Gro¨pl, Austria). Ovaries were analyzed using a Nicon Eclipse E800 fluorescent microscope. Sexual isolation index (SII) and standard error (SE) were measured as in previous studies using multiple mating choice assays, following two calculations, i.e., the isolation index after [72] and the joint isolation index of [107]. In our assays, both calculations result in very similar SII values (Table S2). In general, SIIs range from 21.00 (preference for unlikes, heterogamy) through 0 (random mating) to +1.00 (preference for likes, homogamy) between D. paulistorum semispecies. Fly Strains Sequences were analyzed as described above. sections of male tissues after [103]. Rabbit anti-wsp antibody was used at a 1:500 dilution overnight at 4uC and detected after incubation with a 1:500 dilution of Alexa Fluor 488 goat anti- rabbit IgG secondary antibody (Molecular Probes) at room temperature for 1 hour. Slides were stained for 3 min with 1 mg/ml DAPI (Molecular Probes), rinsed and stained with 5 mg/ ml propidium iodide (PI) (Molecular Probes) for 20 minutes, rinsed again and mounted with ProLong Antifade medium (Molecular Probes). Immunostainings of embryos and ovaries were examined by using a Zeiss Axiomot 2 Epifluorescence microscope. Images were processed using Photoshop 9.0.2 (Adobe). Detection of Low-Titer Wolbachia via Southern Hybridization Detection of Low-Titer Wolbachia via Southern Hybridization In order to detect low-titer Wolbachia infections in D. paulistorum semispecies, a digoxigenin (DIG, Roche Diagnostics, Germany) labeled probe, binding to the core sequence of the wsp gene, was applied on blots of standard wsp PCR [52] for DNA-DNA hybridization. For probe design, hybridization conditions and signal detection see our detailed protocol in [53]. Antibiotic Treatment Conditions Stock solutions of antibiotics were made by diluting Tetracy- cline (3 mg/ml) and Rifampicin (30 mg/ml) into ethanol 98%, stored at 220uC. Artificially Wolbachia-depleted strains of AB, AM, CA and OR D. paulistorum semispecies were established by adding aliquots of the respective stocks to regular fly food at final concentrations of 0.01 and 0.03% of Tetracycline and 0.01, 0.03, 0.1 and 0.2% of Rifampicin. Treated lines were kept for at least three generations on antibiotics for bioassays. For each of the eighteen interstrain mating choice assays monitoring the impact of Rifampicin treatment on mate avoidance between AB, AM, CA and OR semispecies and D. simulans controls, five replicas and 120 matings were scored (12A RR +12B RR +12A == +12B == differentiated by rotated wing clips) totaling 2,160 matings. Without repeated anesthetization at any time, 2-to-3-day old flies were then placed into glass mating chambers [37,106]. A four-power hand lens was employed for scoring partners: the time (from start of observations) each mating takes place; its sequence among other copulae which occur; where in the chamber the mating pair is located (for this purpose a grid constitutes the floor); the kind of female involved; and the kind of male involved, until all flies copulated, in approximately 30–40 minutes. If undisturbed, each copulation lasts approximately 15– 17 minutes; females mate only once in these arenas. Recording location of copula, even upside down, prevented scoring a copula more than once. PLoS Pathogens \| www.plospathogens.org Statistical Testing All statistical tests were performed using SPSS 16.0 for Mac and GraphPad Software (www.graphpad.com). Statistical significance of effects of antibiotic treatment on sex ratios in D. paulistorum semispecies was tested calculating single degree-of-freedom Chi- squares. We applied an unpaired t-test to mating frequencies in D. paulistorum combinations, and compared the SIIs obtained from mating choice assays using Fisher’s Exact test. Boxplot diagrams of mating frequency distributions were generated using SPSS. Acknowledgments We thank Bertha Green, Traude Kehrer and Monika Imb for excellent technical support and fly work; Ira Perelle for help with statistics and Silvia Bulgheresi, Harald Gruber and Johannes Rath for kindly providing sequencing facilities; and Michael Turelli and Markus Riegler for helpful discussions and encouragement. Accession Numbers Sequences were deposited in GenBank under the accession numbers GQ924884-GQ924891 and HQ185362-63. Supporting Information Figure S1 Wolbachia infection in adults of Drosophila paulistorum semispecies. Amazonian (AM); Centroamerican (CA); Interior (IN); Andean-Brazilian (AB); Transitional (TR); Orinocan (OR) semispecies and one F1 hybrid derived from crossings of CA females to OR males (CxO). (A) wsp-long PCR and (B) blot followed by hybridization with DIG-labeled wsp-probe (see Materials and Methods). Figure S1 Wolbachia infection in adults of Drosophila paulistorum semispecies. Amazonian (AM); Centroamerican (CA); Interior (IN); Andean-Brazilian (AB); Transitional (TR); Orinocan (OR) semispecies and one F1 hybrid derived from crossings of CA females to OR males (CxO). (A) wsp-long PCR and (B) blot followed by hybridization with DIG-labeled wsp-probe (see Materials and Methods). Table S2 Mating choice assays performed on D. paulistorum semispecies before and after Wolbachia depletion via rifampicin. Mating choice assays performed on Amazonian (AM), Centroa- merican (CA), Orinocan (OR) and the Andean Brazilian (AB; POA1 & POA10)) D. paulistorum semispecies before and after Wolbachia depletion via rifampicin. Found at: doi:10.1371/journal.ppat.1001214.s001 (0.53 MB TIF) Figure S2 Co-cladogenesis of Wolbachia and host. Host tree (left) was calculated from nine Adh sequences of Drosophila species (AY364506; X57362; AE014134.5; AF045118; DWU95252; EU532128; EU532123; EU532127; EU532121) with 405 sites including gaps; method = average linkage (UPGMA). Wolbachia tree (right) was generated from nine wsp sequences (AY897491; AF020070; GQ924887; AY620227; AF620218; AF020063; GQ924889-90) with 652 sites including gaps; same method. Alignments (L-INS-i strategy) and tree calculation were performed using MAFFT 6.0 (http://mafft.cbrc.jp/alignment/server/); trees were edited using FigTreev1.2 (http://tree.bio.ed.ac.uk/software/ figtree/). Asterisk indicates a region of rough resolution; for high resolution see SNP analysis in Table S1. Found at: doi:10.1371/journal.ppat.1001214.s005 (0.27 MB DOC) Detection of Wolbachia in D. paulistorum Hosts via Transmission Electron Microscopy To determine presence of Wolbachia within D. paulistorum hosts, living adult flies were sectioned using a vibrating blade microtome (Leica VT1000 S). Sections of 400–500 mm thickness were fixed in 2% glutaraldehyde, buffered with 0.1 M PO4 at pH 7.1. After postfixation with 1% OsO4, samples were dehydrated via an ethanol series [108]. Samples were then embedded in Epon and 70 nm ultrathin sections were produced. Following uranyl acetate and lead citrate staining of the sections, presence of Wolbachia was observed using a JEOL TEM. Found at: doi:10.1371/journal.ppat.1001214.s003 (0.42 MB TIF) Table S1 Variable nucleotide (A) and amino acid (B) sites in the wsp sequence of the closely related wAu-like Wolbachia strains of Drosophila and the Cherry fruit fly Rhagoletis cerasi (wCer2). a Position number 1 of the consensus sequence corresponds to position number 164 in the wsp sequence of wAu of D. simulans (AF020067). b Consensus wsp sequence obtained from the following Wolbachia strains: wAu of D. simulans Coffs Harbour [56]; wWil of D. willistoni; wPro SG1 & 2 of D. prosaltans; wSpt PLR1, 2 & BLI1 of D. septentriosaltans [59]; wPau of D. paulistorum CA, TR and OR semispecies (yellow, this study), wCer2 of Rhagoletis cerasi [109]; and wMel of D. melanogaster [56]. c Designation of hypervariable regions (HVRs) and conserved regions (CR) of the WSP protein after [55]. Found at: doi:10.1371/journal.ppat.1001214.s004 (0.14 MB DOC) Author Contributions Conceived and designed the experiments: WJM LE. Performed the experiments: WJM LE DS. Analyzed the data: WJM LE DS. Contributed reagents/materials/analysis tools: WJM LE. Wrote the paper: WJM LE. Conceived and designed the experiments: WJM LE. Performed the experiments: WJM LE DS. Analyzed the data: WJM LE DS. Contributed reagents/materials/analysis tools: WJM LE. Wrote the paper: WJM LE. Found at: doi:10.1371/journal.ppat.1001214.s002 (0.19 MB TIF) Infectious Speciation in Drosophila Figure S3 Natural presence of high- and low-titer Wolbachia in recent D. paulistorum samples from Southern Brazil. POA1 and POA10 are isofemale lines generated from collections in April 2003 in Porto Alegre City, Rio Grande do Sul State, kindly provided from Yong-Kyu Kim, Emory University, Atlanta, GA, USA. Wolbachia-specific wsp PCRs were performed on DNA of adults (a), 0–24 hrs embryos (e), and dissected ovaries (o). Similar to OR control flies (+), POA1 imagos harbor high-titer Wolbachia. Whereas standard wsp PCR detection systems are not sufficient to detect the symbiont in POA10 adults, they are clearly traceable in embryos (e) and ovaries (o). Intermediate wsp-signal intensity was obtained from hybrids of both sexes (AxO), derived from matings between low-titer AM females and high-titer OR males. The negative control was a Wolbachia-uninfected adult of the D. simulans strain STC. Figure S3 Natural presence of high- and low-titer Wolbachia in recent D. paulistorum samples from Southern Brazil. POA1 and POA10 are isofemale lines generated from collections in April 2003 in Porto Alegre City, Rio Grande do Sul State, kindly provided from Yong-Kyu Kim, Emory University, Atlanta, GA, USA. Wolbachia-specific wsp PCRs were performed on DNA of adults (a), 0–24 hrs embryos (e), and dissected ovaries (o). Similar to OR control flies (+), POA1 imagos harbor high-titer Wolbachia. Whereas standard wsp PCR detection systems are not sufficient to detect the symbiont in POA10 adults, they are clearly traceable in embryos (e) and ovaries (o). Intermediate wsp-signal intensity was obtained from hybrids of both sexes (AxO), derived from matings between low-titer AM females and high-titer OR males. The negative control was a Wolbachia-uninfected adult of the D. simulans strain STC. 10. Decaestecker E, Gaba S, Raeymaekers JA, Stoks R, Van Kerckhoven L, et al. (2007) Host-parasite ‘Red Queen’ dynamics archived in pond sediment. Nature 450: 870–873. 7. Hamilton WD, Axelrod R, Tanese R (2000) Sexual reproduction as an adaptation to resist parasites (a review). Proc Natl Acad Sci U S A 87: 3566–3573. 9. Harvell, D (2004) Ecology and evolution of host–pathogen interactions in nature. Am Nat 164: S1–S5. 8. Lively CM, Dybdahl, MF (2000) Parasite adaptation to locally common genotypes. Nature 405: 679–681. 6. Henikoff S, Ahmad K, Malik HS (2001) The centromere paradox: stable inheritance with rapidly evolving DNA. Science 294: 2478–2480. Immunohistochemistry Wolbachia density and tissue tropism in D. paulistorum semispecies were determined using the polyclonal WSP (Wolbachia surface protein) antibody [102]. WSP protein expression was analyzed in whole mount immunostainings on ovaries, staged embryos and on December 2010 \| Volume 6 \| Issue 12 \| e1001214 14 Infectious Speciation in Drosophila 6. Henikoff S, Ahmad K, Malik HS (2001) The centromere paradox: stable inheritance with rapidly evolving DNA. Science 294: 2478–2480. 7. Hamilton WD, Axelrod R, Tanese R (2000) Sexual reproduction as an adaptation to resist parasites (a review). Proc Natl Acad Sci U S A 87: 3566–3573. 8. Lively CM, Dybdahl, MF (2000) Parasite adaptation to locally common genotypes. Nature 405: 679–681. 9. Harvell, D (2004) Ecology and evolution of host–pathogen interactions in nature. Am Nat 164: S1–S5. 10. Decaestecker E, Gaba S, Raeymaekers JA, Stoks R, Van Kerckhoven L, et al. (2007) Host-parasite ‘Red Queen’ dynamics archived in pond sediment. Nature 450: 870–873. 3. Muller HJ (1942) Isolating mechanisms, evolution, and temperature. Biol Symp 6: 71–125. 4. Brideau NJ, Flores HA, Wang J, Maheshwari S, Wang X, et al. (2006) Two Dobzhansky-Muller genes interact to cause hybrid lethality in Drosophila. Science 314: 1238–1239. 6. Henikoff S, Ahmad K, Malik HS (2001) The centromere paradox: stable inheritance with rapidly evolving DNA. Science 294: 2478–2480. Infectious Speciation in Drosophila 28. Breeuwer JA, Werren JH (1990) Microorganisms associated with chromosome destruction and reproductive isolation between two insect species. Nature 346: 558–560. 61. Hoffmann, AA (1988) Partial cytoplasmic incompatibility between two Australian populations of Drosophila melanogaster. Entomol Exp Appl 48: 61–67. 62. Zabalou S, Charlat S, Nirgianaki A, Lachaise D, Merc¸ot H, et al. (2004) Natural Wolbachia infections in the Drosophila yakuba species complex do not induce cytoplasmic incompatibility but fully rescue the wRi modification. Genetics 167: 827–834. 29. Bordenstein SR, O’Hara FP, Werren JH (2001) Wolbachia-induced incompat- ibility precedes other hybrid incompatibilities in Nasonia. Nature 409: 707–710. 30. Shoemaker DD, Katju V, Jaenike J (1999) Wolbachia and the evolution of reproductive isolation between Drosophila recens and Drosophila subquinaria. Evolution 53: 1157–1164. 63. Ballard JW, Melvin RG (2007) Tetracycline treatment influences mitochondrial metabolism and mtDNA density two generations after treatment in Drosophila. Insect Mol Biol 16: 799–802. 31. Koukou K, Pavlikaki H, Kilias G, Werren JH, Bourtzis K, et al. (2006) Influence of antibiotic treatment and Wolbachia curing on sexual isolation among Drosophila melanogaster cage populations. Evolution 60: 87–96. 64. Ehrman L, Kernaghan RP (1971) The microoganismical basis of the infectious hybrid sterility in Drosophila paulistorum. J Hered 62: 66–71. 32. Williamson DL, Ehrman L, Kernaghan RP (1971) Induction of sterility in Drosophila paulistorum: effect of cytoplasmic factors. Proc Natl Acad Sci U S A 68: 2158–2160. 65. Ehrman L, Perelle I, Factor J (1995) Endosymbiotic infectivity in D. paulistorum semispecies. In: Levine L, ed. Genetics of Natural Populations: The continuing influence of Theodosius Dobzhansky. New York: Columbia University Press. pp 241–261. 33. Hurst GDD, Schilthuizen M (1998) Selfish genetic elements and speciation. Heredity 80: 2–8. 66. Starr DJ, Cline TW (2002) A host parasite interaction rescues Drosophila oogenesis defects. Nature 418: 76–79. 34. Ballard, JWO, Chernoff B James AC (2002) Divergence of mitochondrial DNA is not corroborated by nuclear DNA, morphology, or behavior in Drosophila simulans. Evolution 56: 527–545. 67. Casiraghi M, McCall JW, Simoncini L, Kramer LH, Sacchi L, et al. (2002) Tetracycline treatment and sex-ratio distortion: a role for Wolbachia in the moulting of filarial nematodes? Int J Parasitol 32: 1457–1468. 35. Dobzhansky T, Spassky B (1959) Drosophila paulistorum, a cluster of species in statu nascendi. Proc Natl Acad Sci U S A 45: 419–428. 68. Infectious Speciation in Drosophila 11. Jiggins FM, Hurst GD, Yang Z (2002) Host-symbiont conflicts: positive selection on an outer membrane protein of parasitic but not mutualistic Rickettsiaceae. Mol Biol Evol 19: 1341–1349. 45. Bates D, Ehrman L, Perelle I (1977) Intersemispecific hybridity and egg hatchability in D. paulistorum. Dros Inf Serv 52: 94. 46. Ehrman L, Daniels S (1975) Pole cells of Drosophila paulistorum: embryologic differentiation with symbionts. Aust J Biol Sci 28: 133–144. 12. Bordenstein, SR (2003) Symbiosis and the origin of species. In: Bourtzis K, Miller T, eds. Insect Symbiosis. New York: CRC Press. pp 283–304. 47. Daniels S, Ehrman L, Perelle IB (1977) Hybrid inviability: a third reproductive isolating mechanism in the Drosophila paulistorum species-complex. Biol Zbl 96: 385–406. 13. Brownlie JC, Adamski M, Slatko B, McGraw EA (2007) Diversifying selection and host adaptation in two endosymbiont genomes. BMC Evol Biol 7: 68. 14. Aanen DK, Hoekstra RF (2007) The evolution of obligate mutualism: if you can’t beat ’em, join ’em. Trends Ecol Evol 22: 506–509. 48. Laven H (1951) Crossing experiments with Culex strains. Evolution 5: 370–375. 15. Pannebakker BA, Loppin B, Elemans CP, Humblot L, Vavre F (2007) Parasitic inhibition of cell death facilitates symbiosis. Proc Natl Acad Sci U S A 104: 213–215. 49. Ehrman L, Kernaghan, RP (1972) Infectious heredity in Drosophila paulistorum. In: Ciba Foundation Symposium: Pathogenic Mycoplasmas. Amsterdam: Assoc. Scientific Publishers. pp 227–250. 16. Caspari E, Watson G (1959) On the evolutionary importance of cytoplasmic sterility in mosquitoes. Evolution 13: 568–570. 50. Winter C, Hein T, Kavka G, Mach RL, Farnleitner AH (2007) Longitudinal changes in the bacterial community composition of the Danube river: a whole- river approach. Appl Environ Microbiol 73: 421–431. 17. Werren, JH (1998) Wolbachia and speciation. In: Howard DJ, Berlocher SH, eds. Endless Forms: Species and Speciation. New York: Oxford University Press. pp 245–260. 51. Rousset F, Vautrin D, Solignac M (1992) Molecular identification of Wolbachia, the agent of cytoplasmic incompatibility in Drosophila simulans, and variability in relation with host mitochondrial types. Proc Biol Sci 247: 163–168. pp 18. Bordenstein SR, Werren JH (2007) Bidirectional incompatibility among divergent Wolbachia and incompatibility level differences among closely related Wolbachia in Nasonia. Heredity 99: 278–287. 52. Jeyaprakash A, Hoy MA (2000) Long PCR improves Wolbachia DNA amplification: wsp sequences found in 76% of sixty-three arthropod species. Insect Mol Biol 9: 393–405. 19. Infectious Speciation in Drosophila Telschow A, Hammerstein P, Werren JH (2005) The effect of Wolbachia versus genetic incompatibilities on reinforcement and speciation. Evolution 59: 1607–1619. 53. Arthofer W, Riegler M, Schneider D, Krammer M, Miller WJ, et al. (2009) Hidden Wolbachia diversity in field populations of the European cherry fruit fly Rhagoletis cerasi (Diptera, Tephritidae). Mol Ecol 18: 3816–3830. 20. Jaenike J, Dyer KA, Cornish C, Minhas MS (2006) Asymmetrical reinforcement and Wolbachia infection in Drosophila. PLoS Biol 4: e325. 54. Wu M, Sun LV, Vamathevan J, Riegler M, Deboy R, et al. (2004) Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: a streamlined genome overrun by mobile genetic elements. PLoS Biology 2: 327–341. 21. Telschow A, Flor M, Kobayashi Y, Hammerstein P, Werren JH (2007) Wolbachia-induced unidirectional cytoplasmic incompatibility and speciation: mainland-island model. PLoS ONE 2: e701. 22. Weeks AR, Reynolds KT, Hoffmann AA (2002) Wolbachia dynamics and host effects: What has (and has not) been demonstrated? Trends Ecol Evol 17: 257–262. 55. Iturbe-Ormaetxe I, Burke GR, Riegler M, O’Neill SL (2005) Distribution, expression, and motif variability of ankyrin domain genes in Wolbachia pipientis. J Bacteriol 187: 5136–5145. 23. Serbus LR, Casper-Lindley C, Landmann F, Sullivan W (2008) The genetics and cell biology of Wolbachia-host interactions. Annu Rev Genet 42: 683–707. 56. Zhou W, Rousset F, O’Neill SL (1998) Phylogeny and PCR based classification of Wolbachia strains using wsp gene sequences. Proc Royal Soc Lond B 265: 509–515. 24. Werren JH, Baldo L, Clark ME (2008) Wolbachia: master manipulators of invertebrate biology. Nat Rev Microbiol 6: 741–775. 57. Riegler M, Siduh M, Miller WJ, O’Neill SL (2005) Evidence for a global Wolbachia replacement in Drosophila melanogaster. Curr Biol 15: 1428–1433. 25. Merc¸ot H, Poinsot D (2009) Infection by Wolbachia: from passengers to residents. C R Biol 332: 284–297. p p g 58. Baldo L, Lo N, Werren JH (2005) Mosaic nature of the Wolbachia surface protein. J Bacteriol 187: 5406–5018. 26. Rousset F, Raymond M, Kjellberg F (1991) Cytoplasmic incompatibilities in the mosquito Culex-Pipiens – how to explain a cytotype polymorphism. J Evol Biol 4: 69–81. 59. Miller WJ, Riegler M (2006) Evolutionary dynamics of wAu-like Wolbachia variants in neotropical Drosophila spp. Appl Environ Microbiol 72: 826–835. 27. Turelli M (1994) Evolution of incompatibility-inducing microbes and their hosts. Evolution 48: 1500–1513. p pp pp 60. Turelli M, Hoffmann AA (1991) Rapid spread of an inherited incompatibility factor in California Drosophila. Nature 353: 440–442. References 1. Coyne JA, Orr HA (2004) Speciation. SunderlandMA: Sinauer Associates. 545 p. 2. Dobzhansky T (1937) Genetic and the Origin of Species. New York: Columbia University Press. 364 p. 3. Muller HJ (1942) Isolating mechanisms, evolution, and temperature. Biol Symp 6: 71–125. 4. Brideau NJ, Flores HA, Wang J, Maheshwari S, Wang X, et al. (2006) Two Dobzhansky-Muller genes interact to cause hybrid lethality in Drosophila. Science 314: 1238–1239. 5. Dover G (1982) Molecular drive: a cohesive mode of species evolution. Nature 299: 111–117. 5. Dover G (1982) Molecular drive: a cohesive mode of species evolution. Nature 299: 111–117. PLoS Pathogens \| www.plospathogens.org PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 15 Infectious Speciation in Drosophila Infectious Speciation in Drosophila Burla, H, Da Cunha, AB, Cordeiro AR, Dobzhansky T, Malogolowkin C, et al. (1949) The willistoni group of sibling species of Drosophila. Evolution 3: 300–314. 83. Werren JH, O’Neill SL (1997) The evolution of heritable symbionts. In: O’Neill SL, Hoffmann AA, Werren JH, eds. Influential passengers: inherited microorganisms and arthropod reproduction. Oxford, UK: Oxford University Press. pp 1–41. 100. Poinsot D, Montchamp-Moreau C, Merc¸ot H (2000) Wolbachia segregation rate in Drosophila simulans naturally bi-infected cytoplasmic lineages. Heredity 85: 191–198. 84. Bandi C, Trees AJ, Brattig NW (2001) Wolbachia in filarial nematodes: evolutionary aspects and implications for the pathogenesis and treatment of filarial diseases. Vet Parasitol 98: 215–238. 101. King RC (1970) Ovarian development of Drosophila melanogaster. New York: Academic Press. 102. Dobson SL, Bourtzis K, Braig HR, Jones BF, Zhou W, et al. (1999) Wolbachia infections are distributed throughout insect somatic and germ line tissues. Insect Biochem Mol Biol 29: 153–160. 85. Batut J, Andersson SG, O’Callaghan D (2004) The evolution of chronic infection strategies in the alpha-proteobacteria. Nat Rev Microbiol 2: 933–945. 86. Bazzocchi C, Comazzi S, Santoni R, Bandi C, Genchi C, et al. (2007) Wolbachia surface protein (WSP) inhibits apoptosis in human neutrophils. Parasite Immunol 29: 73–79. 103. Veneti Z, Clark ME, Karr TL, Savakis C, Bourtzis K (2004) Heads or tails: host parasite interactions in the Drosophila-Wolbachia system. Appl. Environ. Microbiol. 70: 5366–5372. 87. Broughton S, Partridge L (2009) Insulin/IGF-like signalling, the central nervous system and aging. Biochem J 418: 1–12. 104. Knoppien P (1985) Rare male mating advantage: A review. Biological Reviews 60: 881–117. 88. Ikeya T, Broughton S, Alic N, Grandison R, Partridge L (2009) The endosymbiont Wolbachia increases insulin/IGF-like signalling in Drosophila. Proc Biol Sci 276: 3799–3807. 105. Leonard JE, Ehrman L (1983) Does the rare male advantage result from faulty experimental design? Genetics 104: 713–716. 89. Stocker H, Andjelkovic M, Oldham S, Laffargue M, Wymann MP, et al. (2002) Living with lethal PIP3 levels: viability of flies lacking PTEN restored by a PH domain mutation in Akt/PKB. Science 295: 2088–2091. 106. Elens AA, Wattiaux JM (1964) Direct observation of sexual isolation. Drosophila Information Service 39: 118–119. 107. Malogolowikin-Cohen C, Simmons AS, Levene H (1965) A study of sexual isolation between certain strains of Drosophila paulistorum. Evolution 19: 95–103. 90. Junger MA, Rintelen F, Stocker H, Wasserman JD, Vegh M, et al. Infectious Speciation in Drosophila 76. Weeks AR, Turelli M, Harcombe WR, Reynolds KT, Hoffmann AA (2007) From parasite to mutualist: rapid evolution of Wolbachia in natural populations of Drosophila. PLoS Biol 5: e114. 93. Alexandrov ID, Alexandrova MV, Goryacheva II, Rochina NV, Shaikevich EV et al (2007) Removing endosymbiotic Wolbachia specifically decreases lifespan of females and competitiveness in a laboratory strain of Drosophila melanogaster. Genetika 43: 1372–1378. p 77. Hornett EA, Duplouy AM, Davies N, Roderick GK, Wedell N, et al. (2008) You can’t keep a good parasite down: evolution of a male-killer suppressor uncovers cytoplasmic incompatibility. Evolution 62: 1258–1263. 94. Panteleev Yu D, Goryacheva II, Andrianov BV, Reznik NL, Lazebny OE, et al. (2007) The endosymbiotic bacterium Wolbachia enhances the nonspecific resistance to insect pathogens and alters behavior of Drosophila melanogaster. Genetika 43: 1277–1280. uncovers cytoplasmic incompatibility. Evolution 62: 1258–1263 78. Gleason JM, Griffith EC, Powell JR (1998) A molecular phylogeny of the Drosophila willistoni group: conflicts between species concepts? Evolution 52: 1093–1103. 95. Peng Y, Nielsen JE, Cunningham JP, McGraw EA (2008) Wolbachia infection alters olfactory-cued locomotion in Drosophila spp. Appl Environ Microbiol 74: 3943–3948. 79. Kim YK, Phillips DR, Chao T, Ehrman L (2004) Developmental isolation and subsequent adult behavior of Drosophila paulistorum. VI. Quantitative variation in cuticular hydrocarbons. Behav Genet 34: 385–394. 96. Ehrman L (1960) A genetic constitution frustrating the sexual drive in Drosophila paulistorum. Science 131: 1381–1382. 80. Dedeine F, Vavre F, Fleury F, Loppin B, Hochberg ME, et al. (2001) Removing symbiotic Wolbachia bacteria specifically inhibits oogenesis in a parasitic wasp. Proc Natl Acad Sci U S A 98: 6247–6252. 97. Albertson R, Casper-Lindley C, Cao J, Tram U, Sullivan W (2009) Symmetric and asymmetric mitotic segregation patterns influence Wolbachia distribution in host somatic tissue. J Cell Sci 122: 4570–4583. 81. Dedeine F, Vavre F, Shoemaker DD, Boule´treau M (2004) Intra-individual coexistence of a Wolbachia strain required for host oogenesis with two strains inducing cytoplasmic incompatibility in the wasp Asobara tabida. Evolution 58: 2167–2174. 98. Chao TH, Ehrman L, Permaul R, Vincent R, Sattaur L, Brandt D (2010) Male-specific cuticular compounds of the six Drosophila paulistorum semispecies: structural identification and mating effect. Journal of Chemical Ecology. Aug 6 [Epub ahead of print]. 82. Hosokawa T, Koga R, Meng XY, Fukatsu T (2010) Wolbachia as a bacteriocyte- associated nutritional mutualist. Proc Natl Acad Sci U S A 107: 769–767. 99. Infectious Speciation in Drosophila Arumugam S, Pfarr KM, Hoerauf A (2008) Infection of the intermediate mite host with Wolbachia-depleted Litomosoides sigmodontis microfilariae: impaired L1 to L3 development and subsequent sex-ratio distortion in adult worms. Int J Parasitol 38: 981–987. 36. Ehrman L (1961) The genetics of sexual isolation in Drosophila paulistorum. Genetics 46: 1025–1038. 37. Ehrman L (1965) Direct observation of sexual isolation between allopatric and between sympatric strains of the different Drosophila paulistorum races. Evolution 19: 459–464. 69. Noor MA (1999) Reinforcement and other consequences of sympatry. Heredity 83: 503–508. 38. Dobzhansky T (1970) Genetics of Evolutionary Process. New York: Columbia University Press. 505 p. 70. Kim YK, Ehrman L (1998) Developmental isolation and subsequent adult behavior of Drosophila paulistorum. IV. Courtship. Behav Genet 28: 57–65. 39. Ehrman L (1968) Antibiotics and infections hybrid sterility in Drosophila paulistorum. Mol Gen Genet 103: 218–222. 71. Ehrman L (1966) Mating success and genotype frequency in Drosophila. Anim Behav 14: 332–339. 40. Ehrman L, Williamson DL (1969) On the etiology of the sterility of hybrids between certain strains of Drosophila paulistorum. Genetics 62: 193–199. 72. Ehrman L, Parsons RA (1981) Behavior Genetics and Evolution. New York: McGraw–Hill. 450 p. 41. Kernaghan RP, Ehrman L (1970) Antimycoplasmal antibiotics and hybrid sterility in Drosophila paulistorum. Science 169: 63–64. 73. Mateos M, Castrezana SJ, Nankivell BJ, Estes AM, Markow TA, et al. (2006) Heritable endosymbionts of Drosophila. Genetics 174: 363–376. 42. Ehrman L, Powell JR (1982) The Drosophila willistoni species group. In: Ashburner M, Carson HL, Thompson JN, eds. The genetics and biology of Drosophila. vol. 3b. New York: Academic Press. pp 193–225. 74. McGraw EA, Merritt DJ, Droller JN, O’Neill SL (2002) Wolbachia density and virulence attenuation after transfer into a novel host. Proc Natl Acad Sci U S A 99: 2918–2923. p J g Drosophila. vol. 3b. New York: Academic Press. pp 193–225. 43. Dobzhansky T, Pavlovsky O (1971) Experimentally created incipient species of Drosophila. Nature 230: 289–292. 75. Dunbar HE, Wilson AC, Ferguson NR, Moran NA (2007) Aphid thermal tolerance is governed by a point mutation in bacterial symbionts. PLoS Biol 5: e96. 44. Daniels S, Ehrman L (1974) Embryonic pole cells and mycoplasmalike symbionts in Drosophila paulistorum. J Invertebr Pathol 24: 14–19. PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 PLoS Pathogens \| www.plospathogens.org 16 Infectious Speciation in Drosophila Infectious Speciation in Drosophila (2003) The Drosophila forkhead transcription factor FOXO mediates the reduction in cell number associated with reduced insulin signaling. J Biol 2: 20. 108. Tazuke SI, Schulz C, Gilboa L, Fogarty M, Mahowald AP, et al. (2002) A germline-specific gap junction protein required for survival of differentiating early germ cells. Development 129: 2529–2539. 91. Vala F, Egas M, Breeuwer JA, Sabelis MW (2004) Wolbachia affects oviposition and mating behaviour of its spider mite host. J Evol Biol 17: 692–700. Ch i d C i i d ll d l i y g p 109. Riegler M, Stauffer C (2002) Wolbachia infections and superinfections in cytoplasmically incompatible populations of the European cherry fruit fly Rhagoletis cerasi (Diptera, Tephritidae). Mol Ecol 11: 2425–2434. 92. Champion de Crespigny FE, Pitt T, Wedell N (2006) Increased male mating rate in Drosophila is associated with Wolbachia infection. J Evol Biol 19: 1964–1972. PLoS Pathogens \| www.plospathogens.org December 2010 \| Volume 6 \| Issue 12 \| e1001214 17
https://openalex.org/W4224126831	https://journal.binus.ac.id/index.php/BBR/article/download/7898/4430	English	null	Role of Desire and Implementation of Intention in the Theory of Planned Behavior: A Bibliometric Analysis	Binus business review	2,022	cc-by-sa	6,576	Received: 18th November 2021/ Revised: 13th February 2022/ Accepted: 16th February 2022 How to Cite: Uturestantix, Purwanto, B. M., & Lukito-Budi, A. S. (2022). Role of Desire and Implementation of Intention in the Theory of Planned Behavior: A Bibliometric Analysis. Binus Business Review, 13(1), 97−107. https://doi.org/10.21512/bbr.v13i1.7898 How to Cite: Uturestantix, Purwanto, B. M., & Lukito-Budi, A. S. (2022). Role of Desire and Implementation of Intention in the Theory of Planned Behavior: A Bibliometric Analysis. Binus Business Review, 13(1), 97−107. https://doi.org/10.21512/bbr.v13i1.7898 ABSTRACT A weak finding of behavioral intention and behavior in the Theory of Planned Behavior (TPB) from a meta- analysis study by Armitage and Conner in 2001 has led to an increasing number of studies aiming to improve the TPB. Several researchers propose potential constructs to bridge the theoretical and empirical gap by suggesting that the construct of desire and the implementation of intentions can fill the theoretical and empirical gaps in the TPB model. The literature study aimed to retrospectively explore TPB studies with desire and implementation of intentions constructs in the behavioral science domain. The research retrieved 191 Scopus indexed papers (2012–2019) from the Google Scholars database. It summarized the descriptive data and produced visualization using VOSviewer. The results show that most studies developing the TPB model with other behavioral theories still focus on constructs to shape behavioral intentions because the behavioral intention construct is the best predictor of behavior. Most studies apply a quantitative approach with a cross-sectional survey design to collect primary data. In contrast, the experimental and longitudinal design approaches are relatively neglected in this TPB behavioral study. Furthermore, the analysis of 191 papers shows that the TPB model is often used in research in environmental, sustainable, and transport settings; health; psychological; hospitality and tourism; and innovation and technology. In addition, the use of the TPB model in research with entrepreneurial and legal settings is still limited. From these findings, the research proposes empirical research on TPB to implement further the relationship between desire and intention implementation to improve the TPB model by integrating the TPB model with several other theories, such as Mindset Theory of Action Phase and Model Goal-Directed Behaviour. Keywords: desire, implementation of intention, Theory of Planned Behavior, bibliometric analysis P-ISSN: 2087-1228 E-ISSN: 2476-9053 P-ISSN: 2087-1228 E-ISSN: 2476-9053 Binus Business Review, 13(1), March 2022, 97−107 DOI: 10.21512/bbr.v13i1.7898 Uturestantix1; Bernardinus Maria Purwanto2; Andy Susilo Lukito-Budi3 1,2Doctoral Program, Department of Management, Faculty of Economics and Business, Universitas Gadjah Mada Jln. Bulaksumur, Daerah Istimewa Yogyakarta 55281, Indonesia 1Department of Management, Faculty of Economics and Business, Universitas Cenderawasih Jln. Kamp Wolker, Jayapura 99224, Indonesia 3Department of Management, Faculty of Economics and Business, Atma Jaya Catholic University of Indonesia Jln. Jend. Sudirman No.51, Jakarta 12930, Indonesia 1uturestantix@yahoo.com; 2bm-purwanto@ugm.ac.id; 3andy.susilo@atmajaya.ac.id INTRODUCTION 2020; Nardi, Jardim, Ladeira, & Santini, 2019; Zaremohzzabieh et al., 2019). Furthermore, several studies that have integrated and compared TPB with several other behavioral theory models to improve the strength of the TPB model over the last three decades have shown inadequate results (Cheng, 2019; Choe, Kim, & Hwang, 2020; Feola, Vesci, Botti, & Parente, 2019; Li & Wu, 2019; Miller, Freimund, Metcalf, Nickerson, & Powell, 2019; Potard, Kubiszewski, Camus, Courtois, & Gaymard, 2018; Schuster, Tossan, & Drennan, 2017). The robustness of the TPB model 2020; Nardi, Jardim, Ladeira, & Santini, 2019; Zaremohzzabieh et al., 2019). Furthermore, several studies that have integrated and compared TPB with several other behavioral theory models to improve the strength of the TPB model over the last three decades have shown inadequate results (Cheng, 2019; Choe, Kim, & Hwang, 2020; Feola, Vesci, Botti, & Parente, 2019; Li & Wu, 2019; Miller, Freimund, Metcalf, Nickerson, & Powell, 2019; Potard, Kubiszewski, Camus, Courtois, & Gaymard, 2018; Schuster, Tossan, & Drennan, 2017). The robustness of the TPB model The Theory of Planned Behavior (TPB) explains and predicts individual behavior. This model has been widely applied in various fields of science (Carfora et al., 2019; Lee & Kim, 2018; Lin & Roberts, 2020; Moghimehfar, Halpenny, & Walker, 2018; Si et al., 2019a; Sun, 2020; Tiraieyari & Krauss, 2018). However, several studies with a meta- analytic approach show inadequate results (Armitage & Conner, 2001; Han & Stoel, 2017; Lin & Roberts, 97 Corresponding Author is still debated and discussed by behavioral researchers (Morren & Grinstein, 2021; Sun, 2020; Tornikoski & Maalaoui, 2019; Yuriev, Dahmen, Paillé, Boiral, & Guillaumie, 2020). on the weekends. If they conclude that swimming can be done on weekends and there are no other desires, they can form intentions under the desires. This situation is supported by Perugini and Bagozzi (2001) by showing that intention-forming antecedents are fully mediated by desire. In line with research conducted by Shin, Kim, and Severt (2018), desire is more proximal to predict behavioral intentions. , ) As explained in the Stimulus-Organism- Response (S-O-R) theory, individual behavior reflects individuals’ mental processes after receiving a stimulus and evaluating it, as explained in the Stimulus-Organism-Response (S-O-R) theory (Kim, Lee, & Jung, 2020). Then, the TPB model assumes that individuals’ behavior is always rational (Knauder & Koschmieder, 2019). Rational behavior reflects rational decisions. INTRODUCTION Individuals’ rational decisions are formed from beliefs-attitudes, subjective norms, and behavioral control that individuals perceive. However, many researchers have questioned this assumption. They argue that the TPB model ignores the decision- making process involves elements of internal motivation in the individual’s rational decision process (Bagozzi, Dholakia, & Basuroy, 2003; Perugini & Bagozzi, 2001). p p The subsequent weakness found in the TPB model is that a long enough period for the individual behavior process will allow the occurrence of events that are not expected to change behavioral intentions (Ajzen, 2015). Thus, it can be said that behavioral intention is not the most proximal predictor of individual behavior. The Rubicon model introduced in the Mindset of Action Phase Theory (Heckhausen, 2007) which was later corrected by Oliveira and Rua (2018), shows that intention implementation is more proximal than the intention to measure behavior. According to Orbell, Hodgkins, and Sheeran (1997), in the experimental study, the intentions are implemented in the form of actions that can encourage individuals to carry out the final behavior as the goal. g ) Although individuals have positive attitudes and subjective norms towards an object, it does not directly influence or generate individual behavioral intention (Malle and Knobe, 2001). Following the basic assumptions of the TPB model, the behavior that individuals perform is also influenced by their inner desires (Schuster et al., 2017). Desire is a part of an individual’s internal motivation (Ko, 2020). For example, confronted with an object of behavior (e.g., swimming behavior on the weekends), the individuals will not immediately decide or intend to swim on the weekends even though they have a positive attitude towards the behavior. They will first engage in some reasoning. Then, they must ensure that they can act in one way or another. Besides that, they must find out if they have other desires that exceed their desire to swim yi g Based on the initial literature, the researchers have conducted several studies by applying the TPB model and the integrated TPB model and comparing the TPB model. It can be concluded that the quantity of research that improves the TPB model by adding desire and intention implementation variables with satisfactory results is still limited. METHODS The research is divided into three parts. The first part covers the methods used to collect, copy, process, and analyze the reviewed papers presented. Then, the second part describes the results of data processing and analysis to answer the research questions. In the last part, a discussion of the results that lead to recommendations for further studies is summarized at the end of the research. The main objective of the research is to summarize and measure the history of behavioral research using TPB and reveal the overall status of TPB in the field of behavioral science from a cross- disciplinary perspective, journal sources, country and region, article citations, and keywords. Furthermore, the research proposes a new, integrated, and comprehensive knowledge framework for applying TPB in behavioral science research, including distribution of current topics, theory integration, extended factors, main methods, specific groups, and control variables. Then, the research also aims to provide an in-depth and critical analysis of state of the art and identify research, challenges, and directions for future behavior. The research is expected to provide more comprehensive information to assist future researchers in quickly understanding the current body of TPB knowledge and inspiring future researchers. The research applies bibliometric analysis. Bibliometric analysis is a general strategy for quantitatively elaborating documents published in specific areas (Ji, Liu, Huang, & Huang, 2018). It is adopted in the research to measure and capture the application of TPB in several cross-disciplinary behavioral science fields. Bibliometrics are also helpful for analyzing the characteristics of the literature on specific topics, such as investigating the performance of authors, organizations, countries or regions, and journals that are often used as research references and revealing research trends in the future (Xu, Yu, & Wang, 2018; Yanbing et al., 2020). Elements of this technique consist of co-authorship analysis, co-occurrence analysis, citation and co- citation analysis, and knowledge domain mapping. However, bibliometric analysis can extract and visualize critical information from many documents and provide an objective reference for retrospective analysis. Unfortunately, bibliometric analysis cannot replace a systematic manual review (Li, Wu, Shen, Wang, & Teng, 2017; Si et al., 2019b). The data in Figure 1 show the annual literature statistics for applying the TPB across disciplines from 2012 to 2019 in various fields of science. In its application, the previous research tries to understand the process of carrying out behavior in a planned and rational manner. INTRODUCTION The conclusion is obtained after looking at the number of studies that have been done to establish the strength of the TPB model (e.g., Nguyen, Nham, & Hoang, 2019; Scalco, Noventa, Sartori, & Ceschi, 2017). In addition, the progress of applications and the current status of TPB in the context of decision-making, which is valuable and essential for future behavioral research, remains Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 Figure 1 Number of TPB Publications in 2012‒2019 Figure 1 Number of TPB Publications in 2012‒2019 98 Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 unclear. Therefore, to strengthen the initial literature that the researchers have carried out, obtained conclusions, and approach the TPB model research in the future, in the literature study, the research applies the bibliometric analysis method to comprehensively understand the TPB model in the context of the decision process carried out by an individual. Bibliometric analysis methods have been widely used for knowledge bases in various fields (Indarti, Lukito- Budi, & Islam, 2020; Indarti, Hapsari, Lukito-Budi, & Virgosita, 2021; Si, Shi, Wu, Chen, & Zhao, 2019b). It can assist the retrospective analysis. said that until now, the application of the TPB model has still gotten attention from behavioral researchers in various disciplines. However, there has not been a systematic literature study approach like the research to the researchers’ knowledge. RESULTS AND DISCUSSIONS The results of the 191 studies are discussed to answer the research question by referring to theories, perspectives, research methodologies, research settings (across disciplines), and research theme groups. The result notes the emergence of integrating between TPB and other behavioral theories. Various behavioral theories (see Table 1) have been used to improve the explanatory and predictive power of TPB. Most studies still replicate and extend the TPB model in various contexts. Nevertheless, the six theories used show that these theories (number 2–7 in Table 1) are close to the TPB discussion, which seeks to explain the stages or processes of making individual decisions to behave. Table 1 shows the development of the application of the TPB Model in various cross- disciplines. y Figure 2 presents the initial step to find the relevant published research using the Google Scholars database to source the papers. Then, referring to the conclusions that the researchers obtain based on the initial literature review done previously, in the bibliometric analysis, the researchers use Publish or Perish (POP) software with four combinations of keywords and title words to capture various topics related to TPB. For example, in the first combination, the researchers use the initial keyword “Theory of Planned Behavior” or “Desire” or “Rubicon” or “Mindset Theory of Action Phase” or “Intention Implementation” and the title word “Theory of Planned Behavior”. The combination has resulted in 211 papers based on journals, proceedings, conferences, symposia, books, working papers/theses, and citations. Furthermore, Table 2 shows the method used in 191 papers. There are five studies of the TPB model in conceptual papers (i.e., traditional literature reviews), and most of the other studies are empirical papers. In terms of research methods, surveys (quantitative approach) dominate the findings (172 studies). The result is followed by an experimental method with 7 studies, a meta-analysis with 3 studies, and mixed methods (quantitative and qualitative approaches) with 2 studies. Meanwhile, case studies, and Focus Group Discussions (FGDs)/interviews are in the lowest position, with 1 study each. Then, many empirical studies combine more than one approach like survey and interview or surveys and FGDs. y p , , g p p , In the second step, the researchers establish specific criteria to select relevant papers for further analysis. Three criteria used to ensure the quality of the selected papers are language (English), index value (Scimagojr), and non-predatory journal (beallslist. weebly.com). RESULTS AND DISCUSSIONS As a result, the researchers obtain 191 papers for further analysis, consisting of 144 papers from Q1 (Scopus indexed), 22 papers from Q2, 14 papers from Q3, 2 papers from Q4, and 9 papers from non-Scopus indexed. p The third step is to enter data or systematically document 191 papers based on several central aspects of each paper, namely the theory used in the research, research methodology, research setting (i.e., country of origin), and research theme. This step is carried out with the help of Ms. Excel. The researchers detail some data from the collected papers (e.g., names of the variables used, the hypothesis testing results, and the proposed model proposition). In this step, cleaning the data and tidying the name of variables and other contents, such as theory, country, and others, are also done. The output of this third step is a clean database that is ready to be processed and analyzed. However, out of 191 studies over the past eight years, only 15 studies use the longitudinal survey method to capture the behavioral processes in the TPB model. At the same time, the essence of research with a longitudinal design is to increase the study’s validity, which cannot be achieved by using cross-sectional research (Stritch, 2017). The use of the longitudinal design is shown in Table 2 and Table 3. It indicates that there have been few prospective studies during the last eight years. It shows that the TPB model for prospective research in behavioral science is still very much needed compared to the application of the TPB model in several retrospective studies with a cross- sectional survey design. The fourth step is data processing or analysis and presentation. The researchers use descriptive analysis techniques to map the data based on some essential information in this analysis. The data are used theories and methods, year and type of study (longitudinal or not), country of research conducted, rate of citation of the selected papers, and thematic setting scientific fields that apply TPB. The researchers use the VOSviewer software to visualize the findings based on the network of relationships between keywords and titles for further analysis. The results of VOSviewer data processing present clusters that are formed based However, the TPB research model involving the FGD method is only found in one study during the last eight years (see Table 2). METHODS The data show that from 2012 to 2019, the average annual growth rate in the TPB publication has reached 9,51%. Moreover, in 2019, there was an increase in publications reaching 34 papers. It can be Role of Desire and Implementation ..... (Uturestantix et al.) Figure 2 Research Stage Figure 2 Research Stage Role of Desire and Implementation ..... (Uturestantix et al.) 99 In general, a comprehensive discussion based on the results of the bibliometric analysis should be undertaken to provide a critical review. The research uses VOSviewer to perform the bibliometric analysis. VOSviewer has the advantage of mapping the knowledge domain. In addition, its label layer structure can clearly display dense network interactions, making it suitable for analyzing complex networks and large- scale data generated by a large number of quotes and keywords used. i on the relationship between keywords (or nodes) and other keywords (or nodes). Last, the researchers interpret these results further to answer the research questions. Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 RESULTS AND DISCUSSIONS Therefore, it can be concluded that the lack of the FGD approach is one of the factors causing the TPB model that has been applied in various behavioral contexts to be debated until the research is conducted. The conclusions are based on Malle and Knobe (2001) and Perugini and Bagozzi (2001), distinguishing desire and intention. Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 100 Table 1 Development of Application (Replication/Expansion/Integration/Comparative Theory) of the TPB Model in Various Cross-Disciplines No Theory Number of Theory Usage 1 Theory of Planned Behavior 156 2 The Theory of Ethical Consumer Intention Formation 12 3 Mindset Theory of Action Phase 4 4 Norm Activation Theory or Model 4 5 Model of Goal Directed Behavior 2 6 Social Cognitive Theory 2 7 Self Determination Theory 2 8 Theory of Reasoned Action 1 9 Institutional Theory 1 10 Physical Activity Maintenance Theory 1 11 Neo-Socioanalytic Theory 1 12 Confirmation Interaction Model 1 13 Social Exchange Theory 1 14 Technology Acceptance Model 1 15 Health Belief Model 1 Table 2 Method Used in TPB Publications Research Methods Year 2012 2013 2014 2015 2016 2017 2018 2019 Total Conceptual 1 1 1 2 5 Meta-Analysis 1 1 1 3 Case Study 1 1 FGD/Interview 1 1 Survey 16 16 15 23 20 27 26 29 172 Experiment 2 2 1 2 7 Mixed Methods 1 1 2 Total 18 17 17 24 24 29 28 34 191 Table 2 Method Used in TPB Publications Meanwhile, if referring to the psychological view of the general public, it is known that intention and desire are used interchangeably for the same purpose. of three major nations (Malay, Chinese, and Indian). Furthermore, Indonesia has 1.340 ethnic groups and 1.100 regional languages (BPS, 2010). Furthermore, based on Table 4, the researchers group these countries and find that 64% of behavioral studies using the TPB model are often carried out in developed countries. On the other hand, around 31% are carried out in developing countries. The remaining 5% are unknown because the previous researchers do not mention the country where the research is conducted.i The implementation of the TPB model over the last eight years has come from 191 countries (see Table 4). Regarding the country of origin, the application of the TPB model has been dominated by the USA (50 studies) and China (23 studies) during the last eight years. RESULTS AND DISCUSSIONS Meanwhile, studies regarding individual behavior using the TPB are still limited in several Southeast Asian countries, such as Malaysia (4 studies), Singapore (3 studies), and Indonesia (2 studies). Therefore, this guide provides prospects for behavioral researchers who use the TPB model to examine Southeast Asian countries. These countries are rich in culture, customs, religion, ethnicity, and race. For example, Malaysia and Singapore consist Next, referring to the subject classification system in Scopus, the results analyze the research objectives of 191 studies that apply TPB. It results in 19 disciplines (see Figure 3). The proportion of the 19 disciplines area is 30% in business, management, and accounting (covering marketing, entrepreneurship, Role of Desire and Implementation ..... (Uturestantix et al.) 101 shown that intention often fails to predict behavior. Furthermore, the results from the visualization show that several studies with TPB from 2012 to 2019 have investigated some predictors of initial intentions, such as self-efficacy, social norms, Perceived Behavioral Control (PBC), and knowledge. Many studies are often carried out with survey designs with individual units of analysis level (such as students and consumers), and no unit of analysis at the group level is found. In several behavior studies, the decision-making process is caused by many considerations (groups), such as buying a house, marrying, taking a tour, and touring with the community. innovation, organization management, hospitality, travel, and tourism), 13% in the environmental area (such as the environment, recycling, landscape and urban planning, and clean production), and 10% in psychology; drugs; and agriculture and biology. Furthermore, about 5% of studies are in information, computers, and technology; public health; clinical; and communication. The rest is in sexology; pharmacy; laws, demographics; decision sciences; and nursing. These fields have the opportunity to be developed further in future research. Furthermore, in pharmacy, laws, and nursing, the research that adopts the TPB model tends to be replicative. Moreover, some studies adopt the TPB model but do not measure behavior. The studies only measure behavioral intention. y Finally, Figure 5 shows the trend of citations that apply the TPB from 2012 to 2019. There has been an increasing and decreasing trend from 2012 to 2016. Then, from 2017 to 2019, the citation trend decreased. A downward trend in the citation has occurred since 2017 because these studies have not maximized the number of research findings that prove the importance of the desire and intention implementation. RESULTS AND DISCUSSIONS Next, the researchers visualize potential networks between keywords and paper titles that reflect an article’s theme, method, and content using VOSviewer (see Figure 4). It shows that intention is the most frequently occurring word. It is undoubtedly the most commonly discussed variable of the TPB research since several meta-analytical studies have 102 Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 Table 3 Nature of Study by Timeline Longitudinal Year 2012 2013 2014 2015 2016 2017 2018 2019 Total Yes 1 3 3 2 1 5 15 No 17 14 17 21 22 29 27 29 176 Total 18 17 17 24 24 29 28 34 191 Table 4 Implementation of the TPB Model in Various Countries Countries Number of studies Countries Number of studies Australia 8 Iran 10 Combined Countries 6 Italy 4 Belgium 2 Korea 6 Brazil 1 Malaysia 4 Canada 5 Norway 3 China 23 Pakistan 4 The Netherlands 2 Portugal 1 East Africa 1 Romania 1 Finland 2 Saudi Arabia 1 France 7 Serbia 1 EU 1 Singapore 3 Germany 6 Spain 1 Greek 1 Taiwan 13 Hong Kong 1 UK 4 India 8 USA 50 Indonesia 2 Not Available 9 Table 3 Nature of Study by Timeline Table 3 Nature of Study by Timeline Table 3 Nature of Study by Timeline Longitudinal Year 2012 2013 2014 2015 2016 2017 2018 2019 Total Yes 1 3 3 2 1 5 15 No 17 14 17 21 22 29 27 29 176 Total 18 17 17 24 24 29 28 34 191 Table 4 Implementation of the TPB Model in Various Countries Countries Number of studies Countries Number of studies Australia 8 Iran 10 Combined Countries 6 Italy 4 Belgium 2 Korea 6 Brazil 1 Malaysia 4 Canada 5 Norway 3 China 23 Pakistan 4 The Netherlands 2 Portugal 1 East Africa 1 Romania 1 Finland 2 Saudi Arabia 1 France 7 Serbia 1 EU 1 Singapore 3 Germany 6 Spain 1 Greek 1 Taiwan 13 Hong Kong 1 UK 4 India 8 USA 50 Indonesia 2 Not Available 9 Table 4 Implementation of the TPB Model in Various Countries Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 Binus Business Review, Vol. 13 No. Role of Desire and Implementation ..... (Uturestantix et al.) RESULTS AND DISCUSSIONS 1, March 2022, 97−107 102 Figure 3 Research Setting in TPB Publications Figure 4 Network Visualization Figure 3 Research Setting in TPB Publications Figure 3 Research Setting in TPB Publications Figure 4 Network Visualization Figure 4 Network Visualization 103 Next, the researchers visualize potential networks between keywords and paper titles that reflect an article’s theme, method, and content using VOSviewer (see Figure 4). It shows that intention is the most frequently occurring word. It is undoubtedly the most commonly discussed variable of the TPB research since several meta-analytical studies have shown that intention often fails to predict behavior. Furthermore, the results from the visualization show that several studies with TPB from 2012 to 2019 have investigated some predictors of initial intentions, such as self-efficacy, social norms, Perceived Behavioral Control (PBC), and knowledge. Many studies are often carried out with survey designs with individual units of analysis level (such as students and consumers), and no unit of analysis at the group level is found. In several behavior studies, the decision-making process is caused by many considerations (groups), such as buying a house, marrying, taking a tour, and touring with the community. number of research findings that prove the importance of the desire and intention implementation. Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 REFERENCES Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi. org/10.1016/0749-5978(91)90020-T Ajzen, I. (2005). Attitudes, personality and behaviour. McGraw-Hill Education. Ajzen, I. (2015). The theory of planned behaviour is alive and well, and not ready to retire: A commentary on Sniehotta, Presseau, and Araújo-Soares. Health Psychology Review, 9(2), 131–137. https://doi.org/ 10.1080/17437199.2014.883474fi Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40(4), 471– 499. https://doi.org/10.1348/014466601164939 The results also show the utilization of the TPB model in agriculture, entrepreneurship, and the law is still limited. In addition, the setting of research locations in Southeast Asian countries combined with various relevant potential aspects can be considered for future research agendas. Then, the analysis results of 191 studies find that none of them uses the TPB model to predict behavior at the group level. Therefore, it is hoped that the next TPB model development will use the unit of analysis at the group level. Bagozzi, R. P., Dholakia, U. M., & Basuroy, S. (2003). How effortful decisions get enacted: The motivating role of decision processes, desires, and anticipated emotions. Journal of Behavioral Decision Making, 16(4), 273–295. https://doi.org/10.1002/bdm.446 BPS. (2010). Sensus penduduk 2010. Retrieved from https:// sp2010.bps.go.id/ Carfora, V., Cavallo, C., Caso, D., Del Giudice, T., De Devitiis, B., Viscecchia, R., ... & Cicia, G. (2019). Explaining consumer purchase behavior for organic milk: Including trust and green self-identity within the theory of planned behavior. Food Quality and Preference, 76, 1–9. https://doi.org/10.1016/j. foodqual.2019.03.006 Finally, this research summarizes three groups of research themes that can be developed in future research. Future research could explore in-depth the potential relationship between desire; intention; and implementation of intentions and develop further measurements of the three constructs to improve the TPB model. The results of the analysis of 191 studies show that the TPB model proposed by Ajzen (1991) is inadequate when measuring complex behaviors that require high involvement in the decision- making process because the TPB model proposed by Ajzen (1991) ignores internal motivations that can encourage individuals to intend perform the behavior. The internal motivation referred to in this study is desire. Furthermore, in complex behavioral settings, individuals will face other behaviors connected to the final behavior that is the goal. Thus, the implementation of intentions becomes an important part that bridges the relationship between behavioral intentions and behavior. CONCLUSIONS The results of the bibliometric analysis of 191 behavioral studies using the TPB model show that most of the studies conducted are still replicating the TPB model proposed by Ajzen (1991). Furthermore, a number of theories such as Theory of Ethical Consumer Intention Formation; Norm Activation Theory or Model; Mindset Theory of Action Phase; Model of Goal Directed Behavior; Social Cognitive Theory; Self Determination Theory used for the development of the TPB model focus on the constructs that shape behavioral intentions. As stated by Ajzen (2005) which states that behavioral intention is the best predictor of behavior. However, the results of a systematic review of 191 studies show that there are only four studies that use the Mindset Theory of Action Phase approach and two Goal-Directed Behavior models. These two theories demonstrate their ability to explain a complex behavior. Therefore, it is recommended for further research to select complex behaviors (such as traveling abroad, buying a house, and starting a medium-sized business) by choosing Finally, Figure 5 shows the trend of citations that apply the TPB from 2012 to 2019. There has been an increasing and decreasing trend from 2012 to 2016. Then, from 2017 to 2019, the citation trend decreased. A downward trend in the citation has occurred since 2017 because these studies have not maximized the Figure 5 Trends in Citation in Applying TPB Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 Figure 5 Trends in Citation in Applying TPB Figure 5 Trends in Citation in Applying TPB Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 104 Google Scholar database. Indeed, it is an extensive database, yet there are a lot of tidying-up works to prepare a clean dataset. For future research, it is suggested to use more exclusive databases, such as the Web of Science or Scopus, due to their better- structured data set. The following limitation is that the index criteria of the research paper only refer to the Scopus indexed journal. Although the research mitigates the index coverage by considering the citation level, it is still possible to miss recently good papers (non-Scopus indexed) due to their currently low citation index. Future bibliometric studies should consider recent publications in their dataset. one of the two theories in developing the TPB model. CONCLUSIONS In its implementation, it is necessary to review the role of the variables derived from the Mindset Theory of Action Phase and the Goal-Directed Behavior Model, i.e., each implementation of intentions and desires respectively. The desire variable plays an important role in determining behavioral intentions (Choi & Park, 2017; Ko, 2020; Schuster et al., 2017),while the intention implementation variable plays an important role in bridging the relationship between behavioral intentions and the final behavior that becomes an individual goal (Oliveira & Rua, 2018). From a methodological perspective, it is expected that future behavioral research will use an experimental and longitudinal approach rather than using a cross-sectional survey approach. However, in measuring individual behavior which is relatively complex and dynamic, researchers should use a longitudinal approach (Stritch, 2017) rather than experimentally, because if using an experimental approach to measure this behavior it will cause an erroneous causality effect, because there are so many variables and aspects that are difficult to control in the experimental process (Van Gelderen, Kautonen, Wincent, & Biniari, 2018). The longitudinal approach is essential for assessing change and causal relationships (Purwanto, Indarti, Lukito-Budi, & Uturestantix, 2020). In principle, theory and behavioral science are closely related to change. Thus, identifying change requires a longitudinal study (Purwanto et al., 2020). Role of Desire and Implementation ..... (Uturestantix et al.) REFERENCES However, this research has limitations. Cheng, E. W. L. (2019). Choosing between the Theory of Planned Behavior (TPB) and the Technology Acceptance Model (TAM). Educational Technology Research and Development, 67, 21–37. https://doi. org/10.1007/s11423-018-9598-6 Choe, J. Y., Kim, J. J., & Hwang, J. (2020). The environmentally friendly role of edible insect restaurants in the tourism industry: Applying an extended theory of planned behavior. International Journal of Contemporary Hospitality Management, 32(11), 3581–3600. https://doi.org/10.1108/ IJCHM-04-2020-0352 Choi, Y. J., & Park, J. W. (2017). The decision-making processes of duty-free shop users using a goal directed behavior model: The moderating effect As for the limitations, the researchers use the Role of Desire and Implementation ..... (Uturestantix et al.) 105 of gender. Journal of Hospitality and Tourism Management, 31, 152–162. https://doi.org/10.1016/j. jhtm.2016.11.001 Li, X., Wu, P., Shen, G. Q., Wang, X., & Teng, Y. (2017). Mapping the knowledge domains of Building Information Modeling (BIM): A bibliometric approach. Automation in Construction, 84(December), 195–206. https://doi.org/10.1016/j. autcon.2017.09.011 Feola, R., Vesci, M., Botti, A., & Parente, R. (2019). The determinants of entrepreneurial intention of young researchers: Combining the theory of planned behavior with the Triple Helix Model. Journal of Small Business Management, 57(4), 1424–1443. https://doi.org/10.1111/jsbm.12361 Lin, N., & Roberts, K. R. (2020). Using the theory of planned behavior to predict food safety behavioral intention: A systematic review and meta-analysis. International Journal of Hospitality Management, 90(September), 1–10. https://doi.org/10.1016/j. ijhm.2020.102612 Han, T. I., & Stoel, L. (2017). Explaining socially responsible consumer behavior: A meta-analytic review of theory of planned behavior. Journal of International Consumer Marketing. 29(2), 91–103. https://doi.org/10.1080/08961530.2016.1251870 Malle, B. F., & Knobe, J. (2001). The distinction between desire and intention: A folk- conceptual analysis. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions and intentionality: Foundations of social cognition (pp. 45–67). MIT Press. Heckhausen, J. (2007). The motivation-volition divide and its resolution in action-phase models of developmental regulation. Research in Human Development, 4(3–4), 163–180. https://doi. org/10.1080/15427600701662983 Miller, Z. D., Freimund, W., Metcalf, E. C., Nickerson, N., & Powell, R. B. (2019). Merging elaboration and the theory of planned behavior to understand bear spray behavior of day hikers in Yellowstone National Park. Environmental Management, 63, 366–378. https:// doi.org/10.1007/s00267-019-01139-w Indarti, N., Hapsari, N., Lukito-Budi, A. S., & Virgosita, R. (2021). Quo vadis, ethnic entrepreneurship? A bibliometric analysis of ethnic entrepreneurship in growing markets. Journal of Entrepreneurship in Emerging Economies, 13(3), 427–458. https://doi. org/10.1108/JEEE-04-2020-0080 Moghimehfar, F., Halpenny, E. REFERENCES A., & Walker, G. J. (2018). Front-country campers’ constraints, negotiation, and pro-environment behavioral intention: An extension to the theory of planned behavior. Leisure Sciences, 40(3), 174–193. https://doi.org/10.1080/01490400.2 017.1344163 Indarti, N., Lukito-Budi, A. S., & Islam, A. M. (2020). A systematic review of halal supply chain research: To where shall we go? Journal of Islamic Marketing, 12(9), 1930–1949. https://doi.org/10.1108/JIMA- 05-2020-0161 Morren, M., & Grinstein, A. (2021). The cross-cultural challenges of integrating personal norms into the theory of planned behavior: A Meta-Analytic Structural Equation Modeling (MASEM) approach. Journal of Environmental Psychology, 75(June), 1–16. https://doi.org/10.1016/j.jenvp.2021.101593 Ji, L., Liu, C., Huang, L., & Huang, G. (2018). The evolution of Resources Conservation and Recycling over the past 30 years: A bibliometric overview. Resources, Conservation and Recycling, 134(July), 34–43. https://doi.org/10.1016/j.resconrec.2018.03.005 Nardi, V. A. M., Jardim, W. C., Ladeira, W., & Santini, F. (2019). Predicting food choice: A meta-analysis based on the theory of planned behavior. British Food Journal, 121(10), 2250–2264. https://doi. org/10.1108/bfj-08-2018-0504 Knauder, H., & Koschmieder, C. (2019). Individualized student support in primary school teaching: A review of influencing factors using the Theory of Planned Behavior (TPB). Teaching and Teacher Education, 77(January), 66–76. https://doi.org/10.1016/j. tate.2018.09.012 Nguyen, T. M., Nham, P. T., & Hoang, V. N. (2019). The theory of planned behavior and knowledge sharing: A systematic review and meta-analytic structural equation modelling. VINE Journal of Information and Knowledge Management Systems, 49(1), 76–94. https://doi.org/10.1108/VJIKMS-10-2018-0086 Kim, M. J., Lee, C. K., & Jung, T. (2020). Exploring consumer behavior in virtual reality tourism using an extended stimulus-organism-response model. Journal of Travel Research, 59(1), 69–89. https:// doi.org/10.1177/0047287518818915 Oliveira, A., & Rua, O. L. (2018). From intention to entrepreneurial action: Assessing the impact of the barriers on the creation of new organizations. RAUSP Management Journal, 53(4), 507–534. https://doi. org/10.1108/RAUSP-07-2018-0039 Ko, H. C. (2020). Beyond browsing: Motivations for experiential browsing and goal-directed shopping intentions on social commerce websites. Journal of Internet Commerce, 19(2), 212–240. https://doi.org/ 10.1080/15332861.2020.1747219 Lee, S. J., & Kim, H. L. (2018). Roles of perceived behavioral control and self-efficacy to volunteer tourists’ intended participation via theory of planned behavior. International Journal of Tourism Research, 20(2), 182–190. https://doi.org/10.1002/jtr.2171 Orbell, S., Hodgkins, S., & Sheeran, P. (1997). Implementation intentions and the theory of planned behavior. Personality and Social Psychology Bulletin, 23(9), 945–954. https://doi. org/10.1177/0146167297239004 Li, C., & Wu, Y. (2019). Understanding voluntary intentions within the theories of self-determination and planned behavior. Journal of Nonprofit & Public Sector Marketing, 31(4), 378–389. org/10.1348/014466601164704 org/10.1348/014466601164704 Potard, C., Kubiszewski, V., Camus, G., Courtois, R., & Gaymard, S. (2018). Driving under the influence of alcohol and perceived invulnerability among young adults: An extension of the theory of planned behavior. Transportation Research Part F: Traffic Psychology and Behaviour, 55(May), 38–46. https:// doi.org/10.1016/j.trf.2018.02.033 Tiraieyari, N., & Krauss, S. E. (2018). Predicting youth participation in urban agriculture in Malaysia: Insights from the theory of planned behavior and the functional approach to volunteer motivation. Agriculture and Human Values, 35, 637–650. https:// doi.org/10.1007/s10460-018-9854-8 Purwanto, B. M., Indarti, N., Lukito-Budi, A. S., & Uturestantix. (2020). Berbagai bias dalam studi Longitudinal. In H. M. Jogiyanto (Ed.), Bias di penelitian dan cara mitigasinya. Yogyakarta: Penerbit Andi. Tornikoski, E., & Maalaoui, A. (2019). Critical reflections – The theory of planned behaviour: An interview with Icek Ajzen with implications for entrepreneurship research. International Small Business Journal, 37(5), 536–550. https://doi. org/10.1177/0266242619829681 Scalco, A., Noventa, S., Sartori, R., & Ceschi, A. (2017). Predicting organic food consumption: A meta- analytic structural equation model based on the theory of planned behavior. Appetite, 112, 235–248. https://doi.org/10.1016/j.appet.2017.02.007 Van Gelderen, M., Kautonen, T., Wincent, J., & Biniari, M. (2018). Implementation intentions in the entrepreneurial process: Concept, empirical findings, and research agenda. Small Business Economics. 51, 923–941. https://doi.org/10.1007/s11187-017-9971- 6 Schuster, L., Tossan, V., & Drennan, J. (2017). Consumer acceptance of mhealth services: A comparison of behavioral intention models. Services Marketing Quarterly, 38(2), 115–128. https://doi.org/10.1080/ 15332969.2017.1289791 Xu, Z., Yu, D., & Wang, X. (2018). A bibliometric overview of International Journal of Machine Learning and Cybernetics between 2010 and 2017. International Journal of Machine Learning and Cybernetics. 10, 2375–2387. https://doi.org/10.1007/s13042-018- 0875-9 Shin, Y. H., Kim, H., & Severt, K. (2018). Antecedents of consumers’ intention to visit food trucks. Journal of Foodservice Business Research, 21(3), 239–256. https://doi.org/10.1080/15378020.2017.1368810 Yanbing, S., Zhu, R., Wang, C., Han, S., Liu, H., & Duan, Z. (2020). Bibliometric analysis of Journal of Nursing Management from 1993 to 2018. Journal of Nursing Management. 28(2), 317–331. https://doi. org/10.1111/jonm.12925 Si, H., Shi, J. G., Tang, D., Wen, S., Miao, W., & Duan, K. (2019a). Application of the theory of planned behavior in environmental science: A comprehensive bibliometric analysis. International Journal of Environmental Research and Public Health, 16(15), 1–26. https://doi.org/10.3390/ijerph16152788 Yuriev, A., Dahmen, M., Paillé, P., Boiral, O., & Guillaumie, L. (2020). Pro-environmental behaviors through the lens of the theory of planned behavior: A scoping review. Resources, Conservation and Recycling, 155, 1‒12. https://doi.org/10.1016/j. REFERENCES https://doi.org/10.1080 /10495142.2018.1526745 Perugini, M., & Bagozzi, R. P. (2001). The role of desires and anticipated emotions in goal-directed behaviours: Broadening and deepening the theory of planned behaviour. British Journal of Social Psychology, 40(1), 79–98. https://doi. 106 Binus Business Review, Vol. 13 No. 1, March 2022, 97−107 intention formation: Re-extending the theory of planned behavior. AMS Review, 10, 260–278. https:// doi.org/10.1007/s13162-019-00156-6 org/10.1348/014466601164704 resconrec.2019.104660 Si, H., Shi, J. G., Wu, G., Chen, J., & Zhao, X. (2019b). Mapping the bike sharing research published from 2010 to 2018: A scientometric review. Journal of Cleaner Production, 213, 415–427. https://doi. org/10.1016/j.jclepro.2018.12.157 Stritch, J. M. (2017). Minding the time: A critical look at longitudinal design and data analysis in quantitative public management research. Review of Public Personnel Administration, 37(2), 219–244. https:// doi.org/10.1177/0734371X17697117 Zaremohzzabieh, Z., Ahrari, S., Krauss, S. E., Samah, A. A., Meng, L. K., & Ariffin, Z. (2019). Predicting social entrepreneurial intention: A meta-analytic path analysis based on the theory of planned behavior. Journal of Business Research, 96(March), 264–276. https://doi.org/10.1016/j.jbusres.2018.11.030 Sun, W. (2020). Toward a theory of ethical consumer 107 Role of Desire and Implementation ..... (Uturestantix et al.)
https://openalex.org/W4389043994	https://www.life-science-alliance.org/content/lsa/7/2/e202201870.full.pdf	English	null	Endothelial cells adopt a pro-reparative immune responsive signature during cardiac injury	Life science alliance	2,023	cc-by	13,747	Endothelial cells adopt a pro-reparative immune responsive signature during cardiac injury Hali Long1,2, Jeffrey D Steimle2, Francisco Jose Grisanti Canozo2, Jong Hwan Kim2,3, Xiao Li3 , Yuka Morikawa3, Minjun Park2, Diwakar Turaga4, Iki Adachi5, Joshua D Wythe2,6, Md Abul Hassan Samee2, James F Martin1,2,3,6,7 Modulation of the heart’s immune microenvironment is crucial for recovery after ischemic events such as myocardial infarction (MI). Endothelial cells (ECs) can have immune regulatory functions; however, interactions between ECs and the immune environment in the heart after MI remain poorly understood. We identiﬁed an EC-speciﬁc IFN responsive and immune regulatory gene signature in adult and pediatric heart failure (HF) tissues. Single-cell tran- scriptomic analysis of murine hearts subjected to MI uncovered an EC population (IFN-ECs) with immunologic gene signatures similar to those in human HF. IFN-ECs were enriched in regenerative-stage mouse hearts and expressed genes encoding immune responsive transcription factors (Irf7, Batf2, and Stat1). Single-cell chromatin accessibility studies revealed an enrichment of these TF motifs at IFN-EC signature genes. Expression of immune regulatory ligand genes by IFN-ECs suggests bidirectional signaling between IFN-ECs and macrophages in regenerative-stage hearts. Our data suggest that ECs may adopt immune regulatory signatures after cardiac injury to accompany the reparative response. The presence of these signatures in human HF and murine MI models suggests a potential role for EC-mediated immune regulation in responding to stress induced by acute injury in MI and chronic adverse remod- eling in HF. acute or chronic ischemia. Improved understanding of the basic cellular and molecular mechanisms underlying the cardiac injury response may reveal novel therapeutic targets for treating heart disease. In the 1st wk of life, neonatal mice can achieve full cardiac functional recovery after acute injury, such as coronary artery occlusion or apex resection; however, this transient regenerative capacity is lost after postnatal day seven (Porrello et al, 2011). The loss of regenerative potential is because of a decrease in car- diomyocyte proliferation and the subsequent formation of a ﬁbrotic scar as an imperfect and poorly functional replacement (Song et al, 2012; Porrello et al, 2013; Senyo et al, 2013; Chong et al, 2014; Kong et al, 2014; van Berlo & Molkentin, 2014; Foglia & Poss, 2016; Travers et al, 2016; Deshmukh et al, 2019). 1Interdepartmental Program in Integrative Molecular and Biomedical Sciences, Baylor College of Medicine, Houston, TX, USA 2Department of Integrative Physiology, Baylor College of Medicine, Houston, TX, USA 3Cardiomyocyte Renewal Laboratory, The Texas Heart Institute, Houston, TX, USA 4Section of Critical Care Medicine, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA 5Section of Cardiothoracic Surgery, Department of Surgery, Baylor College of Medicine, Houston, TX, USA 6Cardiovascular Research Institute, Baylor College of Medicine, Houston, TX, USA 7Center for Organ Repair and Renewal, Baylor College of Medicine, Houston, TX, USA on 23 October, 2024 life-science-alliance.org Downloaded from http://doi.org/10.26508/lsa.202201870 Published Online: 27 November, 2023 \| Supp Info: on 23 October, 2024 life-science-alliance.org Downloaded from http://doi.org/10.26508/lsa.202201870 Published Online: 27 November, 2023 \| Supp Info: on 23 October, 2024 life-science-alliance.org Downloaded from http://doi.org/10.26508/lsa.202201870 Published Online: 27 November, 2023 \| Supp Info: Endothelial cells adopt a pro-reparative immune responsive signature during cardiac injury Although most research efforts have focused on cardiomyocyte replenishment as the therapeutic remedy for combating heart disease, recent studies have uncov- ered a critical role for noncardiomyocytes in modulating the local environment to support regeneration (Aurora et al, 2014; Lavine et al, 2014). The cardiac injury response is a highly dynamic and complex multicellular process in which noncardiomyocyte cell types such as immune cells and ECs play crucial roles to aid in recovery of the ischemic heart (Aurora et al, 2014; Lavine et al, 2014; Bajpai et al, 2018, 2019; Liao et al, 2018; Nahrendorf, 2018; Das et al, 2019; Dick et al, 2019; Lai et al, 2019). During acute myocardial infarction (MI), different cell types including cardiomyocytes, cardiac ﬁbroblasts, and ECs promote changes in the cardiac immune environment to facilitate repair (Rainer et al, 2014; Frieler & Mortensen, 2015). The immune environment during the cardiac injury response is es- sential for debris clearance, inﬂammation resolution, and angio- genesis (Lai et al, 2019). Although studies on ECs have largely focused on their roles in angiogenesis and vascularization (Mar´ın- Juez et al, 2016; Ingason et al, 2018; Das et al, 2019; Lu et al, 2021), ECs can also modulate their local immune environment (Seternes et al, 2002; Ding et al, 2012; Wohlleber & Knolle, 2016; Qiu et al, 2018; Amersfoort et al, 2022). Improved understanding of the molecular DOI 10.26508/lsa.202201870 \| Received 9 December 2022 \| Revised 11 November 2023 \| Accepted 14 November 2023 \| Published online 27 November 2023 © 2023 Long et al. signals that may dampen the local inﬂammatory microenviron- ment in pediatric HF. regulation of ECs and immune modulation during the cardiac injury response is essential for effective therapeutic development against heart disease. To investigate whether EC-speciﬁc immune regulatory signatures exist in adult HF tissues, we reanalyzed published single-cell datasets from heart tissues of patients with dilated cardiomyop- athy (DCM), congestive HF or no HF (Calcagno et al, 2020; Wang et al, 2020). Our analysis revealed that CD274 (PD-L1) was enriched in adult heart failure tissues compared with non-HF tissues (Figs 1D and S1F and G). Because the expression of PD-L1 and CD73 on ECs is associated with type 1 and type 2 IFN signaling in other contexts (Lucas et al, 2018; Eichin et al, 2021), we developed an IFN- stimulated gene (ISG) score to identify cells transcriptionally re- sponsive to IFN stimulus (Fig S1D). The ISG score is determined using several well-characterized and canonical targets of IFN signaling, including, BST2, IRF7, and STAT2. We found that ECs from HF tissues had higher ISG scores than ECs from non-HF tissues (Figs 1D and S1E–G; Table S3). These data show that ECs express IFN and immune regulatory signatures to interact with the cardiac immune environment in both pediatric and adult human HF. Together, these ﬁndings suggest that EC-mediated immune regulation may play a role during the stress response to chronic inﬂammation and ad- verse remodeling in human HF. Here, we provide evidence of an EC-speciﬁc immune regulatory signature in chronic human heart disease obtained by performing imaging mass cytometry (IMC) of pediatric heart failure (HF) tissues and computational analysis of adult HF single-cell transcriptomic datasets. To further dissect the role of EC-mediated immune regulation in the cardiac injury response, we performed single-cell proﬁling of regenerative and nonregenerative murine MI models. In regenerative-stage hearts, we observed EC-speciﬁc transcriptomic changes associated with cell proliferation, the IFN response, and immune regulation. Furthermore, we identiﬁed distinct subpopu- lations of ECs enriched for these gene signatures. Binding motifs of IFN-responsive transcription factors (TFs), IRF7, BATF2, and STAT1, were enriched near accessible chromatin regions of immune regulatory genes, highlighting potential transcriptional regulators of EC immune regulatory function during the cardiac injury re- sponse. These ﬁndings support that EC proliferation and immune cell crosstalk occur in response to IFN signaling during the cardiac injury response. The observation of these features in both human HF and murine MI models suggests that the immune regulatory function of ECs may be involved in resolving cardiac stress induced by both acute injury in MI and adverse remodeling in chronic HF. Results To investigate the role of EC-mediated immune regulation in fa- cilitating cardiac repair, we used a murine model of regenerative and nonregenerative ischemic injury induced by left anterior descending artery occlusion (LAD-O) (Porrello et al, 2011). LAD-O was performed at postnatal day one (P1MI, regenerative stage) or eight (P8MI, nonregenerative stage), and cells were isolated from the left ventricle 4 d later at P5 or P12, respectively (Fig 2A). We then performed single-cell RNA sequencing (scRNA-seq) on the isolated cells to compare the transcriptomes between regenerative- and nonregenerative-stage neonatal mouse hearts at single-cell res- olution. After discarding low-quality cells and doublets (see the Materials and Methods section and Table S4), we identiﬁed 11,333 cells that were separated into nine distinct clusters based on gene expression (Figs 2B and S2A and B). We manually annotated these clusters using well-deﬁned gene expression signatures (Table S5) and identiﬁed cells corresponding to ECs, cardiomyocytes, cardiac ﬁbroblasts, macrophages, T cells, pericytes, epicardial cells, en- docardial cells, and smooth muscle cells (SMCs). Endothelial immune responsive signature in heart injury Long et al. Introduction Heart disease, the leading cause of mortality in the world, results from irreversible cardiomyocyte loss after tissue damage such as ischemic injury (Roth et al, 2020; Virani et al, 2021). Extensive basic, translational, and clinical research efforts have been dedicated to developing therapies to combat heart disease. To date, however, no available therapy can permanently restore cardiac function after https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 1 of 16 Human heart failure tissues exhibit EC-speciﬁc immune regulatory signatures Previous studies have uncovered associations between immune- response genes (such as PD-L1, CD73, MHC-I, and IRF8) and human heart diseases including cardiomyopathy, atrial ﬁbrillation, and atherosclerosis (Fig 1A) (Hennecke & Wiley, 2001; St Hilaire et al, 2011; Leonard et al, 2013; Meder et al, 2014; Johnson et al, 2016; Esfahani et al, 2019; Bracamonte-Baran et al, 2021). Moreover, recent ﬁndings of decreased macrophage proliferation in human adult and pediatric heart disease tissues suggest immune inhibitory signaling from cardiac cells in response to injury and stress (Hill et al, 2022; Koenig et al, 2022). To identify potential cell types that engage with the immune microenvironment in human heart dis- ease, we performed IMC on pediatric HF samples. Tissue samples were collected from four pediatric patients with end-stage HF who were undergoing either ventricular assist device placement or heart transplantation (Hill et al, 2022). Patient diagnoses included hy- pertrophic obstructive cardiomyopathy and chemotherapy-induced cardiomyopathy (Fig 1B and Table S1). A panel of 23 markers, in- cluding proteins associated with immune cells, vessels, extracellular matrix, and immune inhibitory signaling molecules, was used to spatially resolve the cellular organization of these tissue samples (Table S2). IMC revealed the expression of PD-L1 and CD73, which functions in immunosuppressive and anti-inﬂammatory signal- ing, in vessels with perivascular macrophage inﬁltration (Figs 1C and S1A) (Chen et al, 2018; Eichin et al, 2021). This observation supports the hypothesis that ECs express anti-inﬂammatory Next, we explored EC-speciﬁc differences between the tran- scriptomes of regenerative- (P1MI) and nonregenerative-stage (P8MI) mouse hearts (Fig S2C and Table S6). We identiﬁed 1,728 differentially expressed genes (DEGs) (adjusted P-value < 0.01, log2 fold change ≥0.2) that were more highly expressed in regenerative- stage ECs than in nonregenerative-stage ECs (Table S6). To in- vestigate potential pathways activated during cardiac repair, we performed gene ontology (GO) enrichment analysis using the 50 most signiﬁcant DEGs (Ge et al, 2020). GO analysis revealed gene sets associated with cell proliferation (e.g., Top2a, Ccnd1), antigen presentation (e.g., β2m, H2-Q4), the IFN response (e.g., Iigp1, Iﬁtm3), and immune processes (e.g., Serpina3g, Lgals9) (Figs 2C and S2C and https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 2 of 16 Endothelial immune responsive signature in heart injury Long et al. 2 of 16 thelial immune responsive signature in heart injury Long et al. IFN-responsive TF binding motif networks are enriched at IFN-EC genes To map the chromatin accessibility landscape in regenerative- and nonregenerative-stage mouse hearts under ischemic stress, we performed single-nuclei assay for transposase-accessible chro- matin followed by single-nuclei ATAC sequencing (snATAC-seq) on nuclei isolated from regenerative- and nonregenerative-stage mouse hearts after LAD-O (P1MI and P8MI) (Fig 2A). After remov- ing low-quality nuclei (see the Materials and Methods section and Table S3), 37,881 nuclei remained, with 120,401 accessible chromatin regions identiﬁed. Using scRNA-seq data to inform cluster identi- ﬁcation (Fig 2B) in the snATAC-seq data, we identiﬁed nine clusters representing ECs, CMs, cardiac ﬁbroblasts, macrophages, T cells, pericytes, epicardial cells, endocardial cells, and SMCs (Figs 4A and S4A). g To further explore the characteristics of the three specialized EC clusters, we identiﬁed the DEGs that distinguish the three clusters (adjusted P < 0.01, log2 fold change ≥1.5). We also examined the enriched GO terms for each cluster (Fig 2E, Table S7). All three clusters were enriched with distinct cell cycle IFN-response sig- natures (e.g., Iigp1, Iﬁtm3) that corresponded with the genes more highly expressed in regenerative-stage than in nonregenerative- stage ECs from our previous whole heart dataset (Figs 2C and 3A, Table S6). Two of the clusters were enriched in cell cycle processes and were therefore annotated as proliferating EC clusters ProEC-1 and ProEC-2 (Fig S3A and Table S7). Cell cycle analysis of ProEC-1 and ProEC-2 cells revealed an enrichment of cells in S or G2M phases, respectively (Fig S3B). In addition, ProEC-1 and ProEC-2 were most closely associated with angiogenic ECs, which is consistent with EC proliferation observed during angiogenesis and vessel growth in regenerating hearts (Schaper, 1996; Dor & Keshet, 1997; Ware & Simons, 1997; Das et al, 2019). The remaining specialized EC cluster, which appeared similar to capillary ECs, showed speciﬁc enrich- ment of IFN-γ signaling (e.g., Iigp1, Iﬁtm3) and antigen-presentation pathways (e.g., H2-D1, H2-K1, Psmb8) and was termed the IFN-EC cluster (Figs 2E and 3B, and Table S7). RNA-scope imaging showing RNA expression of Iigp1 in ECs of regenerative-stage heart tissue To explore potential EC-speciﬁc changes in chromatin accessi- bility, we also performed snATAC-seq on CD31-enriched cells from regenerative- and nonregenerative-stage mouse hearts harvested 4 d after injury at P1 and P8 (Fig 2A). A total of 11,289 cells with 100,155 open chromatin regions were obtained. Table S6). These data reveal a novel EC-speciﬁc immune responsive gene signature that was not identiﬁed previously in the single-cell proﬁling data of regenerative- and nonregenerative-stage hearts (Wang et al, 2020). Taken together, our ﬁndings reveal that tran- scriptomic differences between regenerative- and nonregenerative- stage ECs are associated with cell proliferation, antigen presentation, the IFN response, and immune processes. supports observations from both scRNA-seq datasets (Fig 3C). Furthermore, the IFN-EC cluster featured higher expression of IFN- associated TFs such as Batf2, Irf7, Irf8, Irf9, Stat1, and Stat2 than did most other EC populations (Fig 3D). RNA-scope imaging of EC-speciﬁc Stat1 RNA expression in regenerative-stage hearts supports this observation (Fig 3E). We next used the CD31+-enriched datasets and compared cell composition between regenerative-stage (P1MI) and nonregenerative- stage (P8MI) mouse hearts after ischemic injury. The Pro-EC and IFN-EC populations were signiﬁcantly enriched in regenerative-stage hearts (P < 0.001, log fold difference > 1; Fig S3C and D, Table S7). Therefore, consistent with our scRNA-seq dataset revealing an up-regulation of cell cycle, IFN response, and immune regulatory genes in the ECs of regenerative-stage hearts versus nonregenerative-stage hearts after ischemic injury (Figs 2C and 3A), we highlight distinct EC subpopulations with these gene signatures (Fig 3B). Our data also suggest that the unique IFN-EC cluster, in addition to the two proliferating EC clusters, may be important for the regenerative response after MI. Cardiac EC-speciﬁc transcriptomic proﬁling reveals distinct EC states with cell cycle and IFN response signatures To further study the transcriptomic heterogeneity of ECs from regenerative-stage and nonregenerative-stage hearts in greater detail, we performed scRNA-seq of ECs (CD31+) enriched from murine hearts injured during regenerative (P1MI) and nonregenerative stages (P8MI) using magnetic-activated sorting (MACs) (Fig 2A and see the Materials and Methods section). After ﬁltering out low- quality cells and doublets (see the Materials and Methods sec- tion), we detected 21,414 cells that were clustered into 10 unique CD31+ subpopulations (Fig 2D). The 10 distinct clusters were tran- scriptionally heterogeneous (Fig S2D, Table S7). Using established marker gene expression (Schiller et al, 2019; Goveia et al, 2020; Kalucka et al, 2020; Travaglini et al, 2020; Schupp et al, 2021), we identiﬁed arterial EC, capillary EC, capillary venous EC, venous EC, angiogenic EC, lymphatic EC, and endocardial cell clusters (Fig S3A and Table S7). In addition, we identiﬁed three specialized EC clusters (Figs 2D and S2D, Table S7). Figure 1. IFN-responsive and immune regulatory signaling by endothelial cells in human heart failure tissues. Human heart failure tissues exhibit EC-speciﬁc immune regulatory signatures. (A) Table of genes implicated in human heart disease. (B) Hematoxylin and eosin (left) and trichrome (right) staining of cardiac tissue from a pediatric patient diagnosed with dilated cardiomyopathy (DCM). (B, C) Imaging mass cytometry showing the expression of the selected markers in pediatric DCM cardiac tissue shown in panel (B). The lower panel is a zoomed-in view of the region inside the dashed box in the upper panel. Arrows indicate the expression of PD-L1 (blue) and CD73 (green) in vessels with nearby macrophage inﬁltration. (D) Dot plot showing the IFN-stimulated gene score and expression of IFN response and immune regulatory gene signatures in cardiac endothelial cells (ECs) from adult patients with DCM or congestive heart failure compared with those from normal donors. Human heart failure tissues exhibit EC-speciﬁc immune regulatory signatures https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 3 of https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 3 of 16 Endothelial immune responsive signature in heart injury Long et al. Endothelial immune responsive signature in heart injury Long et al. Endothelial immune responsive signature in heart injury Long et al. IFN-responsive TF binding motif networks are enriched at IFN-EC genes Label transferring from the scRNA-seq data (Fig 2C) identiﬁed and annotated three distinct clusters as EC, EndoC, and lymphatic endothelial cell (LEC) (Figs 4B and S4B). We investigated whether chromatin accessibility differences underlie the difference in gene expression between regenerative- and nonregenerative-stage ECs. Although we did not observe signiﬁcant chromatin accessibility differences (adjusted P < 0.01, log2 fold change ≥0.5) associated with IFN-EC DEGs, such as Iigp1 Figure 1. IFN-responsive and immune regulatory signaling by endothelial cells in human heart failure tissues. Human heart failure tissues exhibit EC-speciﬁc immune regulatory signatures. (A) Table of genes implicated in human heart disease. (B) Hematoxylin and eosin (left) and trichrome (right) staining of cardiac tissue from a pediatric patient diagnosed with dilated cardiomyopathy (DCM). (B, C) Imaging mass cytometry showing the expression of the selected markers in pediatric DCM cardiac tissue shown in panel (B). The lower panel is a zoomed-in view of the region inside the dashed box in the upper panel. Arrows indicate the expression of PD-L1 (blue) and CD73 (green) in vessels with nearby macrophage inﬁltration. (D) Dot plot showing the IFN-stimulated gene score and expression of IFN response and immune regulatory gene signatures in cardiac endothelial cells (ECs) from adult patients with DCM or congestive heart failure compared with those from normal donors. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 4 of 16 https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 4 of 16 Endothelial immune responsive signature in heart injury Long et al. 4 of 16 diac cells and CD31+-enriched cells from regenerative- and nonregenerative-stage murine 2. Single-cell RNA sequencing (scRNA-seq) of all cardiac cells and CD31+-enriched cells from regenerative- and nonregenerative-stage murine ardial infarction (MI). -seq of ECs reveal a subpopulation enriched with immune responsive signatures. (A) Schematic showing the experimental design for the scRNA-seq ATAC sequencing analysis of regenerative mouse hearts (MI at P1 and harvested at P5) and nonregenerative mouse hearts (MI at P8 and harvested at P t ti f di ll l t id tiﬁd f RNA l i Cl t l d d di g t ll t (l ft) CM di t EC Figure 2. IFN-responsive TF binding motif networks are enriched at IFN-EC genes Single-cell transcriptomic analysis reveals an immune responsive gene signature in endothelial cells from regenerative-s mouse hearts after myocardial infarction (MI). Single-cell RNA sequencing of ECs rev subpopulation enriched with immun responsive signatures. (A) Violin plot showing expression levels of the indic genes in endothelial cells (ECs) from single-cell RNA sequencing data of a cardiac cells. (B) Violin plots showing expression levels of indicated genes the CD31+ cell cluster. (C) Representa RNA-scope images showing RNA expression of the indicated genes in E of regenerative-stage P1MI (left) and nonregenerative-stage P8MI (right) he (D) Dot plot showing gene expressio levels of the indicated transcription factors in each CD31+ cell cluster. (E) Representative RNA-scope image showing RNA expression of the indica transcription factors in ECs of regenerative-stage P1MI (left) and nonregenerative-stage P8MI (right) hearts. Figure 3. Single-cell transcriptomic analysis reveals an immune responsive gene signature in endothelial cells from regenerative-stage mouse hearts after myocardial infarction (MI). Single-cell RNA sequencing of ECs reveal a subpopulation enriched with immune responsive signatures. (A) Violin plots showing expression levels of the indicated genes in endothelial cells (ECs) from single-cell RNA sequencing data of all cardiac cells. (B) Violin plots showing expression levels of indicated genes in the CD31+ cell cluster. (C) Representative RNA-scope images showing RNA expression of the indicated genes in ECs of regenerative-stage P1MI (left) and nonregenerative-stage P8MI (right) hearts. (D) Dot plot showing gene expression levels of the indicated transcription factors in each CD31+ cell cluster. (E) Representative RNA-scope images showing RNA expression of the indicated transcription factors in ECs of regenerative-stage P1MI (left) and nonregenerative-stage P8MI (right) hearts. Together, these data suggest that IRFs, BATFs, and STATs may be in- volved with the transcription of IFN-EC genes. accessible chromatin (Felsenfeld et al, 1996; Thurman et al, 2012; Tsompana & Buck, 2014; Klemm et al, 2019), we focused on the TF motifs in accessible chromatin that may be regulating transcription. We ﬁrst developed a normalized motif score (NMS) for every gene using the three classes of TFs that were enriched in the IFN-EC cluster (Fig 3D), namely the IRFs, BATFs, and STATs (see the Materials and Methods section). When applying the NMS for the IRFs, BATFs, and STATs to the IFN-EC marker genes, we found that the NMS was signiﬁcantly higher for all three TF families at IFN-EC marker genes than at other expressed genes (Fig 4C and Table S8). IFN-responsive TF binding motif networks are enriched at IFN-EC genes Single-cell RNA sequencing (scRNA-seq) of all cardiac cells and CD31+-enriched cells from regenerative- and nonregenerative-stage murine hearts after myocardial infarction (MI) RNA-seq) of all cardiac cells and CD31+-enriched cells from regenerative- and nonregenerative-stage murine hearts after e-cell RNA sequencing (scRNA-seq) of all cardiac cells and CD31+-enriched cells from regenerative- and nonregenerative-stage ction (MI). Figure 2. Single-cell RNA sequencing (scRNA-seq) of all cardiac cells and CD31+-enriched cells from regenerative- and nonre myocardial infarction (MI). y scRNA-seq of ECs reveal a subpopulation enriched with immune responsive signatures. (A) Schematic showing the experimental design for the scRNA-seq and single- nuclei ATAC sequencing analysis of regenerative mouse hearts (MI at P1 and harvested at P5) and nonregenerative mouse hearts (MI at P8 and harvested at P12). (B) UMAP representation of cardiac cell clusters identiﬁed from scRNA-seq analysis. Clusters were color coded according to cell types (left). CM, cardiomyocytes; EC, endothelial cells; CF, ﬁbroblasts; PeriC, pericytes; SMC, smooth muscle cells; EndoC, endocardial cells; EpiC, epicardial cells; MF, macrophages; TC, T-cells. (C) Gene ontology (GO) analysis of differentially expressed genes up-regulated in ECs of regenerative-stage hearts versus nonregenerative-stage hearts after MI. (D) UMAP representation of CD31+-enriched cell clusters from single-cell RNA sequencing (scRNA-seq), color coded according to cell type. CEC, capillary endothelial cell; Angio-EC, angiogenic endothelial cell; Pro-EC1, proliferative endothelial cell 1; Pro-EC2, proliferative endothelial cell 2; IFN-EC, interferon capillary endothelial cell; CVEC, capillary venous endothelial cell; VEC, venous endothelial cell; AEC, arterial endothelial cell; LEC, lymphatic endothelial cell; EndoC, endocardial cell. (E) Gene ontology (GO) analysis of differentially expressed genes up-regulated in the IFN-EC cluster of regenerative-stage hearts versus nonregenerative-stage hearts after MI. and H2-D1 (Fig S4C–F and Table S6), changes in chromatin accessibility do not always accompany changes in gene expression. This has been seen in some contexts of IFN signaling in which other mechanisms of regulation, such as histone modiﬁcations, may play a role (Hogan et al, 2017; Kamada et al, 2018; Hota et al, 2022; Platanitis et al, 2022). Because gene expression is regulated by the combinatorial binding of TFs to regulation, such as histone modiﬁcations, may play a role (Hogan et al, 2017; Kamada et al, 2018; Hota et al, 2022; Platanitis et al, 2022). Because gene expression is regulated by the combinatorial binding of TFs to Endothelial immune responsive signature in heart injury Long et al. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 5 of 16 Figure 3. Endothelial immune responsive signature in heart injury Long et al. IFN-responsive TF binding motif networks are enriched at IFN-EC genes RNA-scope imaging of cardiac tissue from mice with the heterozygous knockout of BATF2 (BATF+/−) suggest a decrease in Iigp1 and Stat1 (IFN-EC genes) RNA expression (Fig 4D and E). We then used network analysis to identify DNA-binding motifs of all TF superclasses that co-occur with IRF motifs within accessible chro- matin regions of IFN-EC marker genes (Fig S4G). Spectral network analysis classiﬁed the immunoglobulin fold superclass, basic domain class, and helix-turn-helix domain class TFs, which includes the STATs, BATFs, and IRFs, respectively, as members of the same community for IFN-EC marker genes, that is, these factors co-occur in more similar patterns (Fig S4E). However, STATs and BATFs no longer formed com- munities with IRFs within the accessible chromatin regions at other expressed genes, supporting that the co-occurrence of TFs enriched in https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 6 of 16 https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 6 of 16 Endothelial immune responsive signature in heart injury Long et al. Endothelial immune responsive signature in heart injury Long et al. 6 of 16 IFN-ECs (i.e., BATF2, IRF7/8/9, and STAT1/2) at accessible chromatin regions of IFN-EC marker genes may be important for their transcription. Figure 4. Single-nuclei chromatin accessibility proﬁling of cardiac cells from regenerative- and nonregenerative-stage mouse hearts after myocardial infarction (MI). Single-nuclei ATAC sequencing (snATAC-seq) of ECs reveal potential transcription factor regulators of IFN-EC gene signature. (A) UMAP representation of cardiac cell clusters from snATAC-seq data. Clusters are color coded according to cell type. EC, endothelial cells; LEC, lymphatic endothelial cell; EndoC, endocardial cells. (B) UMAP representation of CD31+ cell clusters from snATAC-seq data. Clusters are color coded according to the cell type. (C) Normalized motif score plots showing the occurrence of binding motifs for the indicated transcription factors correlated with chromatin regions of IFN-EC differentially expressed (DE) differentially expressed genes versus non-differentially expressed genes. (D) Representative RNA- scope images showing Stat1 RNA expression in ECs of BATF+/−knockout mouse hearts. (E) Representative RNA-scope images showing Iigp1 RNA expression in ECs of BATF+/−knockout mouse hearts. (E) Representative RNA-scope images showing Iigp1 RNA expression in ECs of BATF+/−knockout mouse hearts. IFN-ECs (i.e., BATF2, IRF7/8/9, and STAT1/2) at accessible chromatin regions of IFN-EC marker genes may be important for their transcription. with nonregenerative-stage hearts (Fig S5E and F) (Jin et al, 2021). IFN-responsive TF binding motif networks are enriched at IFN-EC genes Our analysis revealed that BST2 signaling, which is related to type I IFN signaling and inhibits the production of IFNs and pro-inﬂammatory cytokines from immune cells, is increased between IFN-ECs and macrophages in regenerative-stage hearts (P1MI) compared with nonregenerative-stage hearts (P8MI) (Fig 5C) (Cao et al, 2009). We also found that the CD39 signaling pathway, which is implicated in im- munosuppression in cancer and myocardial protection after is- chemic injury, is preferentially enriched between IFN-ECs and immune cells in regenerative-stage mouse hearts (P1MI) and signals from macrophages to IFN-ECs (Figs 5C and S5F) (K¨ohler et al, 2007; Saldanha-Araujo et al, 2011; Wheeler et al, 2012; Allard et al, 2017). These ﬁndings highlight transcriptomic changes associated with immunosuppressive signaling from ECs in regenerative-stage hearts. IFN-ECs express genes encoding immune regulatory ligands during cardiac regeneration Because IFN-ECs appear to be involved in immune signaling like the ECs in human HF tissues (Fig 1), we investigated the ligands expressed by IFN-ECs. By used the immune cells captured during CD31 en- richment, we characterized the signaling networks between ECs and immune cells of regenerative- and nonregenerative-stage mouse hearts after ischemic injury (Fig S5A and B). Examination of the IFN-EC marker genes revealed several signaling ligands with corresponding receptors on immune cells (Fig S5C and D). In IFN-ECs, enriched li- gands included Bst2, Cd274, Cxcl9, Lgals9, H2-T22, H2-T24, and C4a (Fig 5A). Bst2, Cd274, and Lgals9, which are ligands for BST2, PD-L1, and GALECTIN signaling pathways, respectively, are related to type I IFN signaling and have immunosuppressive functions (Cao et al, 2009; Wu et al, 2014; Chen et al, 2018). Some receptors for the IFN-EC–enriched ligands, speciﬁcally Pira2, C3ar1, Cd44, and Havcr2 (receptors for Bst2, C4a, and Lgals9, respectively), are expressed on macrophages (Fig 5B). In addition, some receptors for the IFN-EC–enriched ligands, speciﬁcally Cxcr3, Pdcd1, and Cd8a/Cd8b1 (receptors for Cxcl9, Cd274, H2-T22, and T2-T24, respectively), are expressed on T cells (Fig 5B). As a secondary approach, we used CellChat to identify ligand– receptor interactions between IFN-ECs and other cardiac cell types that were enriched in regenerative-stage mouse hearts compared Endothelial immune responsive signature in heart injury Long et al. Discussion The balance between activation and attenuation of inﬂammation during the injury response is critical for cardiac regeneration (Steffens et al, 2022). In human HF tissues, we uncovered an EC- speciﬁc IFN response and immune regulatory signature. IFN re- sponse and immune regulatory signatures in ECs have been described across different organs, but not in the context of cardiac regeneration after ischemic injury (Seternes et al, 2002; Qiu et al, 2018; Goveia et al, 2020; Kalucka et al, 2020; Amersfoort As a secondary approach, we used CellChat to identify ligand– receptor interactions between IFN-ECs and other cardiac cell types that were enriched in regenerative-stage mouse hearts compared Endothelial immune responsive signature in heart injury Long et al. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 7 of 16 https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 7 of 16 Figure 5. Ligand-receptor analysis of CD31-enriched cells from regenerative-stage and nonregenerative-stage mouse hearts after myocardial infarction (MI). EC subpopulations express genes encoding immune regulatory ligands. (A) Dot plot showing the indicated ligands enriched in the IFN–endothelial cell (EC) cluster. (B) Dot plot showing immune cell expression of receptors for ligands expressed in the IFN-EC cluster. (C) Chord plots showing the indicated signaling pathway network between the IFN-EC cluster and macrophages from regenerative-stage and nonregenerative- stage mouse hearts after MI. CEC, capillary endothelial cell; CAEC, capillary arterial endothelial cell; AEC, arterial endothelial cell; CVEC, capillary venous endothelial cell; VEC, venous endothelial cell; IFN-EC, interferon capillary endothelial cell; angio-EC, angiogenic endothelial cell; Pro-EC1, proliferative endothelial cell 1; Pro- EC2, proliferative endothelial cell 2; Pro-EC3, proliferative endothelial cell 13; Pro-EC4, proliferative endothelial cell 4; LEC, lymphatic endothelial cell; BC, B-cells; TC, T-cells; MF1, macrophage 1; MF2, macrophage 2; EndoC, endocardial cells. (D) Hypothetical model of IFN-induced immune regulation by ECs in cardiac injury. et al, 2022). Using single-cell transcriptomic proﬁling of cardiac ECs from regenerative-stage mouse hearts subjected to cardiac injury, we identiﬁed a subpopulation of capillary ECs, namely IFN- ECs, that showed a gene expression signature similar to that of ECs in human HF tissues (Fig 5D). These data revealed an enrichment of transcripts encoding IFN response TFs and immune regulatory ligands in IFN-ECs. Furthermore, an enrichment of binding motifs for IFN response TFs were enriched in IFN-EC markers genes, suggesting that these TFs may be upstream transcriptional acti- vators of the immune regulatory signatures in ECs (Fig 5D). Endothelial immune responsive signature in heart injury Long et al. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 the expression of IFN-responsive TFs IRF7, STAT2, and BST2. Previous studies have shown that the inhibition of PD-L1 (CD274) signaling to treat cancer leads to subsequent myocarditis in humans, which can be alleviated with immunosuppression (Johnson et al, 2016; Mahmood et al, 2018; Moslehi et al, 2018; Salem et al, 2018; Esfahani et al, 2019). Furthermore, the expression of PD-L1 is important for engraftment after transplantation (Bracamonte-Baran et al, 2021). Mutations in NT5E, which encodes CD73, have been identiﬁed in patients with arterial calciﬁcation, an indicator of cardiovascular risk and inﬂammation (Lehto et al, 1996; Shaw et al, 2003; St Hilaire et al, 2011). These data, in addition to our ﬁndings, suggest that immunosuppressive signaling via PD-L1 and CD39/CD73 pathways by ECs may limit the deleterious effects of adverse inﬂammation in HF. H2-T22. H2-T22 and H2-T24 are part of the MHC Class I family and are important for signaling to T cells (Hennecke & Wiley, 2001). Genes encoding MHC Class I and II proteins have been associated with a risk locus for DCM (Meder et al, 2014). Increased Bst2, Cd274, and CD39/CD73 signaling in regenerative stage hearts is consistent with our hypothesis that the EC-speciﬁc IFN response and immune regulation may mediate the immune in human HF and cardiac regeneration. In conclusion, our ﬁndings reveal a novel role for ECs in immune regulation in human HF and in mice after cardiac ischemic injury (Fig 5D). Our single-cell sequencing datasets can be used for further study into EC function, signaling pathways, and transcriptional regulation that may be related to heart repair. Future studies are needed to understand how ECs interact with immune components during the reparative process after cardiac injury and during hu- man HF. Further elucidating the role of ECs in regulating the im- mune response to cardiac injury in murine and human HF may provide novel insights into cellular mechanisms for alleviating both acute and chronic cardiac inﬂammation. This will not only provide a better understanding of the underlying biology of human heart disease, but it will pave the way for developing novel therapeutic targets. The datasets we obtained from murine models of MI revealed a gene expression signature characterized by the up-regulation of cell cycle, IFN response, and immune regulatory signatures in regenerative-stage hearts, suggesting that EC-mediated immune regulation, in addition to EC proliferation, is related to cardiac regeneration. Importantly, the up-regulation of immune responsive genes in regenerative-stage hearts may be attributed to a speciﬁc EC subpopulation, IFN-ECs. Our multi-omic analysis revealed the enrichment of transcripts and binding motifs of IFN response TFs, such as certain IRFs, BATFs, and STATs, in IFN-ECs and in regenerative-stage hearts, suggesting that these IFN response TF families may regulate the gene expression of IFN-EC marker genes. Batf2 expression is related to type 1 and type 2 IFN sig- naling, modulates the immune response, and has been linked to Irf activity in immune cells (Schraml et al, 2009; Betz et al, 2010; Ise et al, 2011; Li et al, 2012; Jabeen et al, 2013; Roy et al, 2015; Matatall et al, 2016). Irf7, 8, and 9 are important for immune regulation and are protective against various cardiac pathologies (Biron, 2001; Lohoff & Mak, 2005; Tamura et al, 2008; D¨oring et al, 2012; Jiang et al, 2014a, 2014b, 2014c; Zhang et al, 2014; McNab et al, 2015). STAT1 and STAT2 have been shown to bind IRF9 to form the trimer complex ISGF3, which binds the IFN-stimulated response element and induces transcription of genes such as Irf7 (Honda & Taniguchi, 2006; Au- Yeung et al, 2013; Cheon et al, 2013; McComb et al, 2014). Importantly, we found IRF7 to be enriched in ECs from adult HF tissues. Notably, gene variants in IRF8 have been identiﬁed as risk factors for coronary heart disease (Leonard et al, 2013). These ﬁndings suggest that IFN- response TFs, which are typically active in immune cells, may regulate the transcription of a gene expression program promoting the IFN response and immune regulation in ECs. Ethics approval for the use of donated tissues Cardiac tissues and blood samples used in this study were collected during cardiothoracic surgeries performed at Texas Children’s Hospital (Houston, Texas). The protocols for the procurement and use of these patient samples were approved by the Institutional Review Board for Baylor College of Medicine and Afﬁliated Hospitals (Protocol Number H-26502). With the help of the Heart Center Biorepository at Texas Children’s Hospital, consent was obtained from patients with various forms of pediatric heart disease, including cardiomyopathies. The anatomic location of the tissue collected was based on the speciﬁc surgical repair being performed. This information, along with more speciﬁc patient information, can be found in Table S1. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 9 of 16 Endothelial immune responsive signature in heart injury Long et al. Discussion To- gether, our data hypothesizes a model in which EC-speciﬁc immune regulation be part of a central biologic response to acute and chronic cardiac inﬂammation induced by MI and HF, re- spectively (Fig 5D). et al, 2022). Using single-cell transcriptomic proﬁling of cardiac ECs from regenerative-stage mouse hearts subjected to cardiac injury, we identiﬁed a subpopulation of capillary ECs, namely IFN- ECs, that showed a gene expression signature similar to that of ECs in human HF tissues (Fig 5D). These data revealed an enrichment of transcripts encoding IFN response TFs and immune regulatory ligands in IFN-ECs. Furthermore, an enrichment of binding motifs for IFN response TFs were enriched in IFN-EC markers genes, suggesting that these TFs may be upstream transcriptional acti- vators of the immune regulatory signatures in ECs (Fig 5D). To- gether, our data hypothesizes a model in which EC-speciﬁc immune regulation be part of a central biologic response to acute and chronic cardiac inﬂammation induced by MI and HF, re- spectively (Fig 5D). The response of the immune system to ischemic cardiac injury has been described as both beneﬁcial in the initial phase and deleterious in the long-term injury response (Rurik et al, 2021). Inﬂammatory immune cells are important for debris clearance in the early stages of cardiac injury but can contribute to chronic inﬂammation and progressive pathologic remodeling in human HF. Therefore, regulation of the cardiac immune environment to maintain an immunosuppressed state may be an adaptive re- sponse in human HF, although ultimately an inadequate one. Our IMC data of pediatric HF tissue revealed expression of CD274 (PD-L1) and CD73 (part of the CD39/CD73 signaling pathway) in ECs with immune inﬁltration. scRNA-seq analysis of adult HF tissues revealed a similar enrichment in ECs of CD274 compared with those of donor tissues. Furthermore, this signature was accompanied by https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 Endothelial immune responsive signature in heart injury Long et al. 8 of 16 Sample collection and preservation Cardiac tissue and blood samples were collected in the operating room during various pediatric cardiovascular surgeries. Cardiac tissue samples were kept in cold saline on ice during transfer to the laboratory for preservation. Before cardiac bypass was initiated, blood samples were collected into EDTA-coated vacutainers and were then transferred to the laboratory on ice. Cardiac tissue samples were carefully dissected into multiple aliquots, some of which were ﬂash-frozen and stored at −80°C and others of which were ﬁxed in 10% neutral buffered formalin for 16–24 h at 4°C. Formalin-ﬁxed samples were serially dehydrated and embedded in parafﬁn blocks for histology. Formalin-ﬁxed parafﬁn-embedded (FFPE) samples were then used to make a tissue microarray (2- mm cores) for high-throughput image analysis. Our analysis revealed that multiple cardiac EC populations ex- press ligands involved in immune regulatory pathways. Studies have shown that LECs and venous ECs can participate in immune regulatory signaling with immune cells (Lee et al, 2014; Takeda et al, 2019; Brulois et al, 2020; Xiang et al, 2020). Furthermore, LECs can signal to macrophages, T cells, and dendritic cells via Ccl21, a lymphoid homing chemokine (Nagira et al, 1997; Willimann et al, 1998; Saeki et al, 1999; Luther et al, 2000; Shields et al, 2010; Card et al, 2014). Venous ECs have gene expression signatures associated with leukocyte recruitment and chemokine signaling (Thiriot et al, 2017; Goveia et al, 2020; Kalucka et al, 2020; Schupp et al, 2021). We found that the capillary EC subpopulation of IFN-ECs also expressed immune modulatory ligand genes such as Bst2, Cd274, CD39, and https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 9 of 16 Endothelial immune responsive signature in heart injury Long et al. 9 of 16 The ISG analysis was based on previously reported ISGs (Schneider et al, 2014; Calcagno et al, 2020). A total of 96 ISGs were included for scoring. A detailed gene list is provided in Table S3. The Seurat “AddModuleScore” function was used to calculate ISG scores. This method scores the expression of ISGs by normalizing ex- pression levels of ISGs against those of randomly selected background genes with similar expression levels. The size of the background gene pool is 10 times that of the input gene set. ISG scores that ﬁt Gaussian distribution across cells were trans- formed into Z-scores for plotting and statistical signiﬁcance testing. LAD-O of neonatal mice Neonatal mice were subjected to MI via LAD-O surgery as previously described (Porrello et al, 2011). Pups were subjected to hypothermic anesthetization before procedures. Feet pinching was used to evaluate proper anesthetization. Nylon sutures (8-0 nonabsorb- able) were used for LAD occlusion. Proper occlusion was indicated by blanching of the affected myocardium. The thoracic cavity was closed with vicryl sutures (6-0 absorbable). Duration of the surgery from hypothermic induction to recovery was around 10 min. Sham procedures are identical in all aspects except for occlusion of the LAD. Mice were euthanized 4 d after surgery. Hearts were harvested and digested into single-cell suspensions to be used for single-cell RNA and single-nuclei ATAC sequencing. IMC data analysis For scRNA-seq and snATAC-seq, the atria and aorta were removed from the heart, and the remaining tissue was used for single-cell digestion. Cells were harvested at either P5 or P12, from pools of ﬁve hearts subjected to MI at P1 or P8, respectively. Tissues were minced into small pieces in digestion buffer (1 mg/ml collagenase A; HBSS) and incubated at 37°C with agitation for 30 min. The suspension was triturated with a 5-ml pipette every 10 min. Proper digestion was veriﬁed by the appearance of single cells under a microscope. The digestion was then diluted 1:1 with HBSS, and the resulting cell suspension was centrifuged at 300g for 6 min to remove car- diomyocytes. Supernatant was passed through a 40-micron ﬁlter to remove tissue fragments and cardiomyocytes, and then centrifuged at 500g for 6 min. The cell pellet was resuspended in 1 ml of MACs buffer (PBS, 0.5% BSA, 2 mM EDTA). The resuspended cells were MCD Viewer was used to process and convert data to TIFF format. Images were further enhanced using the median noise and sharpen functions in Adobe Photoshop. Sample collection and preservation Differences of ISG scores and PD-L1 expression between control ECs and ECs from patients with DCM or congestive heart failure were tested for signiﬁcance using the nonparametric Wilcoxon signed-rank test. Experimental animals The tissue microarray of FFPE cardiac samples was used for IMC analysis. Tissue sections were warmed at 60°C for 1 h and dewaxed in three separate xylene washes for 10 min each. The tissue sections were then rehydrated in a graded series of alcohol (ethanol: deionized water 100:0, 100:0, 96:4, 90:10, 80:20, 70:30) for 5 min each and then in PBS for 10 min. Epitope retrieval was performed in Tris EDTA retrieval buffer (pH 9; GeneMed) at 95°C for 20 min, after which, the slides were immediately cooled in TBS for 20 min. Samples were blocked in 3% BSA and 10% donkey serum in PBS with Tween (PBST) for 2 h at RT. Incubation with the antibody panel was performed in blocking buffer overnight at 4°C. Tissue samples were washed twice with TBST and twice with TBS. Slides were incubated with Intercalator-Ir solution for 5 min at RT and washed twice with TBS. The samples were dipped in water and dried before obtaining IMC measurements. The antibody panel targeted func- tional markers for DNA damage, immune regulation, cell cycle, and phenotypic markers to identify epithelial, endothelial, mesenchy- mal, and immune cell types. All conjugated antibodies used in this study can be found in Table S7. The Hyperion Imaging System (Fluidigm) was used for IMC image acquisition. The largest square area was selected from the center of each tissue microarray core for laser ablation. Commercial acquisition software (Fluidigm) was used to preprocess raw data and monitor acquisition quality. All murine experiments were performed under the Baylor College of Medicine Institutional Animal Care and Use Committee protocol number 5713. ICR mice were acquired from the Center for Com- parative Medicine at Baylor College of Medicine and used for all surgical experiments. Female mice were set up for timed pregnancy to deliver pups for surgical procedures at postnatal day 1 or 8 (P denotes postnatal day, i.e., days after birth). Histology Tissue sections were deparafﬁnized at 60°C for 1 h and then dewaxed and rehydrated in a graded series of alcohol. For he- matoxylin and eosin (H&E) staining, tissue sections were incubated with hematoxylin for 15 min, followed by incubation with acidiﬁed Eosin Y for 3 min at RT. The samples were re-parafﬁnized, dried, and mounted before imaging. Masson’s Trichrome staining was per- formed according to manufacturer’s instructions (HT15; Sigma- Aldrich). Images were acquired using the Cytation 5 Cell Imaging Multi-Mode Reader (Biotek). Contrast of both H&E and Masson’s Trichrome images were enhanced using the auto-contrast function in Adobe Photoshop, with the same contrast settings applied to each image for consistency. Isolation of single cells and nuclei from murine cardiac tissue Neonatal hearts were harvested and washed with ice-cold Tyrode’s solution. The left ventricle was removed and diced with curved scissors. The tissue was then digested with Collagenase A at 37°C until single-cell dissociation was achieved (~20 min). Dissociated cells were diluted to a concentration of 200 cells per μl in PBS with 0.01% BSA. Drop-seq was then performed as previously described (Macosko et al, 2015). Cells were co-encapsulated into nano- liter–sized droplets containing barcoded microparticles (catalog number Macosko201110; ChemGenes) and lysis buffer using a custom microﬂuidics device (FlowJEM). After droplet breakage, reverse transcription (Thermo Fisher Scientiﬁc), and exonuclease treatment (NEB), all cDNA was PCR-ampliﬁed (KAPA), pooled, pu- riﬁed with Ampure XP beads (Beckman Coulter), and run on a fragment analyzer (Advanced Analytical Technologies, Inc.) for quality control, quantiﬁcation, and size determination. Library preparation was performed with the Illumina Nextera XT kit, and libraries were triple-puriﬁed with Ampure XP beads (Beckman Coulter). All libraries were sequenced on an Illumina NextSeq500 instrument. The FindVariableFeatures function was used to calculate the top 5,000 most variable features for each sequenced library. The Seurat function FindIntegrationAnchors was used to ﬁnd mutual nearest neighbors across subsets. Seurat’s IntegrateData function was used to construct an integrated matrix for each cell with a correction vector (based on anchor and similarity scores). The integrated data were scaled and the principal components were calculated (n = 50). Dimensional reduction (UMAP) and clustering were used to identify clusters (dimensions = 20, cluster resolution = 1.5). Nuclear isolation was performed as previously described, with the following speciﬁcations (Mo et al, 2015). Brieﬂy, fresh cardiac tissue was harvested on ice and was immediately homogenized with a Biogen Series PRO200 (PRO Scientiﬁc) before performing Dounce homogenization in HB buffer (0.25 M sucrose, 25 mM KCl, 5 mM MgCl2, 20 mM Tricine-KOH, pH 7.8; with protease inhibitors; 1 mM DTT; 0.15 mM spermine; 0.5 mM spermidine, and RNAse in- hibitors) with 5% IGEPAL CA-630. Nuclei were isolated via density gradient centrifugation with optiprep density gradient medium after mixing homogenate 1:1 with a 50 iodoxinal (5 vol Optiprep [Cat#D1556; Sigma-Aldrich] with 1 vol diluent [150 mM KCl; 30 mM MgCl2; 120 mM Tricine-KOH; pH 7.8]). After centrifugation at 10,000G for 18 min, all nuclei isolated from the 30–40% interface were recovered and washed before performing snATAC-seq. Isolation of single cells and nuclei from murine cardiac tissue For the whole heart dataset, 13,504 cells were generated across 4 experiments, with an average sequencing depth of 1,736 reads per cell, 1,031 genes per cell (Table S2). We identiﬁed cardiomyocytes based on expression of myoﬁbrillar markers Tnnt2 and Myl6 and cardiac ﬁbroblasts based on expression of ﬁbroblast and extra- cellular matrix markers Col1a1, Tcf21, and Dcn. SMCs were distin- guished by the expression of markers Acta2, Eln, and Rgs5, whereas pericytes were identiﬁed by Pdgfrb expression. Immune cells were identiﬁed by expression of Lyz2 and Cx3cr1 in macrophages, and Cd52, Trbc2, and Cd3d in T-cells. ECs were enriched for markers such as Fabp4 and Pecam-1, whereas endocardial cells expressed Nrp3 and Igf2. Epicardial cells were deﬁned by the expression of Saa3 and other unique markers. For the CD31+-enriched dataset, we excluded contaminating nonendothelial cells, which were characterized by the expression of ﬁbroblast (Col1a1, Tcf21), cardiomyocyte (Tnnt2, Myh6), red blood cell (Hbb-bt, Hba-a1, Hbb-bs), macrophage (Lyz2, Cx3cr1), T cell (Cd52, Trbc2, Cd3d), B cell (Ighm, Igkc), pericyte (Pdgfrb) or SMC (Acta2) markers. To further remove doublets, FindAllMarkers was used to identify clusters of cells that express more than 1 major cell type marker. We identiﬁed endocardial cells based on the distinct expression of Npr3, Dcn, and Postn, and lymphatic ECs based on the expression of canonical markers Ccl21a, Prox1, Pdpn, and Fgl2 (Fig S3A and Table S5). ECs were identiﬁed by the expression of pan- endothelial markers, such as Fabp4 and Pecam-1, which encodes CD31. Arterial ECs were positive for Gja4, Gja5, Sox17, and Hey1 (Fig College of Medicine or on a HiSeq Xten Illumina platform (Novaseq Illumina platform) by Novogene. centrifuged at 500g for 6 min to remove residual buffer and resuspended in a ﬁnal volume of 80 μl for incubation with 20 μl of CD31 MACs beads (Miltenyi Biotech). MACs enrichment for CD31+ cells was performed according to the Miltenyi enrichment protocol. Live cells were quantiﬁed with Trypan Blue and in Cyto C-Chip DHC- F01. For single-nuclei ATAC sequencing, nuclei were isolated from cells according to the 10x protocol. Live nuclei were quantiﬁed with DAPI and in Cyto C-Chip DHC-F01. scRNA-sequencing analysis The 10x Genomics Cell Ranger software (cellranger 3.0.1, www.10xgenomics.com) was used to handle raw sequencing data. Fastq ﬁles were aligned to the mouse mm10 genome. A pre-mRNA version of the reference genome was provided for gene count quantiﬁcation with the Cellranger count function. CellBender v0.1.0 was used to remove background from the Cellranger output with the remove-background module in default settings. The h5 expression matrices from each data source were then imported into Seurat 3 for downstream processing and annotation to account for source- speciﬁc quality differences (Stuart et al, 2019). Single cells were ﬁl- tered for UMI (unique molecular identiﬁers) counts with a minimum of 500 and a maximum of 25,000 per cell. We also imposed a cutoff of 0.5% for mitochondrial reads and a cutoff of 0.25% for doublet score. Seurat toolkit version v.4.0 in R v.3.8 was used to perform downstream analyses. The processed UMI count matrices of ﬁltered cells were normalized with the Seurat function SCTransform to regress out additional variation. ISG analysis We directly downloaded publicly available human data that were previously processed and annotated (Wang et al, 2020) and performed standard data normalization and visualization using the Seurat software suite. Speciﬁcally, we computed 50 principal components (PCs) from the top 2,000 highly variable genes. We then performed batch effect correction on the PC dimension by using RunHarmony function and projected cells on the UMAP dimension for visualization by using RunUMAP. Endothelial immune responsive signature in heart injury Long et al. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 10 of 16 Endothelial immune responsive signature in heart injury Long et al. Supplementary Information is available at https://doi.org/10.26508/lsa. 202201870. Supplementary Information is available at https://doi.org/10.26508/lsa. 202201870. TF–TF co-localization and network analysis S3A, Table S5), whereas venous ECs expressed markers such as Vwf, Nr2f2, and CouptfII (Fig S3A, Table S5). Capillary venous ECs positive for Egr1, Fos, and JunB were also detected, although capillary ECs showed the least overall speciﬁcity in terms of unique gene ex- pression and were broadly enriched for markers such as Gpihbp1, Rgcc, and Car4 (Fig S3A, Table S5). For the TFs having motif matches in the ATAC peaks, we obtained their families from the CisBP (v2.0.0) database. We then computed the Pearson’s correlation coefﬁcient between every pair of families’ motif counts across the ATAC peaks and created a network where each node indicates a motif family, and the edges represent the Pearson’s correlation coefﬁcient values between families. To identify the groups of colocalizing TF families, we then applied the igraph package’s spectral community detection algorithm on these networks (https://igraph.org/). DEGs for each cell cluster were calculated with the Findmarkers function, Seurat implementation of the Wilcoxon rank-sum test (min.pct = 0.01, thresh.use = 0.25). Gene ontology analysis was performed with ShinyGO v0.75 (Ge et al, 2020). The CellChat v.1.1.2 R package was used to perform ligand–receptor analysis (Jin et al, 2021). Data were visualized by the DoHeatmap and Dotplot func- tions from Seurat. Study approval All murine experiments were performed under the Baylor Col- lege of Medicine Institutional Animal Care and Use Committee protocol number 5713. Human cardiac tissue collection was approved by the Institutional Review Board for Baylor College of Medicine and Afﬁliated Hospitals (Protocol Number H-26502) and collection occurred during cardiothoracic surgeries per- formed at Texas Children’s Hospital. To identify clusters in the snATAC-seq data, we transferred cluster labels from scRNA-seq data. To accomplish this, we ﬁrst identiﬁed common correlation patterns in the gene activity matrix, which is the summation of fragments intersecting with the gene body and promoter region, from both datasets. Mapping of frag- ments of each cell to the 2-kb upstream region of gene coordinates is performed with the GeneActivity function in Seurat. The Find- TransferAnchors function was used to identify patterns between the two experimental conditions for integration. Dimensionality reduction was performed with Harmony to correct principal com- ponent analysis embeddings. Each snATAC-seq nuclei was assigned scores on the basis of cluster labels from the scRNA-seq data. Data Availability All sequencing data generated in this study will be deposited to the National Center for Biotechnology Information (NCBI) Gene Ex- pression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/). Computing NMS The motif scanning tool FIMO was used with motifs from the CisBP (v2.0.0) database to identify peaks with motif matches (Grant et al, 2011; Weirauch et al, 2014). We used the standard P-value cutoff of 1 × 10−4 for motif matching in FIMO (Maurano et al, 2012; Whyte et al, 2013; Kheradpour & Kellis, 2014). The Seurat::AverageExpression (slot = “counts”) function was used to calculate the average ac- cessibility count within each population. For a given TF, a gene’s NMS (i.e., NMS) was computed as a weighted summation of the TF’s motif counts in the peaks assigned to the gene. In particular, the NMS for gene and TF was computed as follows: for the -th peak assigned to the gene, denotes the peak’s accessibility score, de- notes the motif count of TF in the peak, and denotes the peak’s length in kilobases. snATAC-sequencing analysis FFPE murine cardiac tissue samples were sectioned into 5-μm thick slides. The slides were stained using the RNAscope 2.5 HD Assay- RED protocol (ACD) according to the manufacturer’s instructions. Nuclei were counterstained for 10 min at RT with DAP. Endothelial cells were stained with Isolectin B4 (Vector Laboratories). The following probes were used for analysis: Probe-Mm-Iigp1 (Cat No. 520771-C2) and Probe-Mm-Stat1 (Cat No. 479611-C2). Slides were imaged with a Zeiss LSM880 confocal microscope. Visualization and image processing was performed with FIJI/ImageJ software. The 10x Genomics Cell Ranger software (cellranger-atac-1.2.0) was used to handle raw sequencing data. Reads were mapped to the mouse mm10 genome. The peak outputs from each sample were aligned with cellranger-atac count and aggregated with cellranger- atac aggr (–nosecondary –normalize = none). MACS2 callpeak (v2.1.1.20160309) was used for initial clustering and per cluster and per sample peak calling (-f BAMPE -g mm -B -q 0.1). We then reaggregated samples with cellranger-atac aggr using a union set of peaks and imported them in Seurat for subsequent analysis. Low- quality nuclei with <15% of fragments (reads) in peaks were re- moved. Nuclei with a minimum of 30 K (few reads) and a maximum of 1,500 K (doublets/nuclei clumps) peak region fragments were also removed. scRNA and snATAC sequencing scRNA-seq libraries were generated using the 10x Chromium Single Cell 39 Library Kit v3 (10x Genomics) according to the manufacturer’s instructions. We diluted cell suspensions for a target of 10,000 cells for each library. snATAC-seq libraries were generated using the 10x Chromium Single Cell ATAC Library Kit v1 according to the manufacturer’s instructions. Final quantiﬁcation and quality con- trol were determined by using Quibit and Fragment Analyzer. Nuclei suspensions were diluted for a target of 10,000 nuclei per library. Libraries were sequenced on an Illumina NextSeq 500 at the Baylor https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 Endothelial immune responsive signature in heart injury Long et al. 11 of 16 Endothelial immune responsive signature in heart injury Long et al. References Das S, Goldstone AB, Wang H, Farry J, D’Amato G, Paulsen MJ, Eskandari A, Hironaka CE, Phansalkar R, Sharma B, et al (2019) A unique collateral artery development program promotes neonatal heart regeneration. Cell 176: 1128–1142 e18. doi:10.1016/j.cell.2018.12.023 Allard B, Longhi MS, Robson SC, Stagg J (2017) The ectonucleotidases CD39 and CD73: Novel checkpoint inhibitor targets. Immunol Rev 276: 121–144. doi:10.1111/imr.12528 Deshmukh V, Wang J, Martin JF (2019) Leading progress in heart regeneration and repair. Curr Opin Cell Biol 61: 79–85. doi:10.1016/j.ceb.2019.07.005 Amersfoort J, Eelen G, Carmeliet P (2022) Immunomodulation by endothelial cells - partnering up with the immune system? Nat Rev Immunol 22: 576–588. doi:10.1038/s41577-022-00694-4 Dick SA, Macklin JA, Nejat S, Momen A, Clemente-Casares X, AlthagaﬁMG, Chen J, Kantores C, Hosseinzadeh S, Aronoff L, et al (2019) Self- renewing resident cardiac macrophages limit adverse remodeling following myocardial infarction. Nat Immunol 20: 29–39. doi:10.1038/ s41590-018-0272-2 Au-Yeung N, Mandhana R, Horvath CM (2013) Transcriptional regulation by STAT1 and STAT2 in the interferon JAK-STAT pathway. JAKSTAT 2: e23931. doi:10.4161/jkst.23931 Ding L, Saunders TL, Enikolopov G, Morrison SJ (2012) Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 481: 457–462. doi:10.1038/nature10783 Aurora AB, Porrello ER, Tan W, Mahmoud AI, Hill JA, Bassel-Duby R, Sadek HA, Olson EN (2014) Macrophages are required for neonatal heart regeneration. J Clin Invest 124: 1382–1392. doi:10.1172/JCI72181 Dor Y, Keshet E (1997) Ischemia-driven angiogenesis. Trends Cardiovasc Med 7: 289–294. doi:10.1016/S1050-1738(97)00091-1 Bajpai G, Schneider C, Wong N, Bredemeyer A, Hulsmans M, Nahrendorf M, Epelman S, Kreisel D, Liu Y, Itoh A, et al (2018) The human heart contains distinct macrophage subsets with divergent origins and functions. Nat Med 24: 1234–1245. doi:10.1038/s41591-018-0059-x D¨oring Y, Soehnlein O, Drechsler M, Shagdarsuren E, Chaudhari SM, Meiler S, Hartwig H, Hristov M, Koenen RR, Hieronymus T, et al (2012) Hematopoietic interferon regulatory factor 8-deﬁciency accelerates atherosclerosis in mice. Arterioscler Thromb Vasc Biol 32: 1613–1623. doi:10.1161/ATVBAHA.111.236539 Bajpai G, Bredemeyer A, Li W, Zaitsev K, Koenig AL, Lokshina I, Mohan J, Ivey B, Hsiao HM, Weinheimer C, et al (2019) Tissue resident CCR2- and CCR2+ cardiac macrophages differentially orchestrate monocyte recruitment and fate speciﬁcation following myocardial injury. Circ Res 124: 263–278. doi:10.1161/CIRCRESAHA.118.314028 Eichin D, Pessia A, Takeda A, Laakkonen J, Bellmann L, Kankainen M, Imhof BA, Stoitzner P, Tang J, Salmi M, et al (2021) CD73 contributes to anti- inﬂammatory properties of afferent lymphatic endothelial cells in humans and mice. Eur J Immunol 51: 231–246. Author Contributions Brulois K, Rajaraman A, Szade A, Nordling S, Bogoslowski A, Dermadi D, Rahman M, Kiefel H, O’Hara E, Koning JJ, et al (2020) A molecular map of murine lymph node blood vascular endothelium at single cell resolution. Nat Commun 11: 3798. doi:10.1038/s41467-020-17291-5 H Long: conceptualization, data curation, formal analysis, valida- tion, investigation, visualization, methodology, and writing—original draft, review, and editing. JD Steimle: writing—original draft and writing—review and editing. FJ Grisanti Canozo: formal analysis, visualization, and methodology. JD Steimle: writing—original draft and writing—review and editing. FJ Grisanti Canozo: formal analysis, visualization, and methodology. Calcagno DM, Ng RP, Jr, Toomu A, Zhang C, Huang K, Aguirre AD, Weissleder R, Daniels LB, Fu Z, King KR (2020) The myeloid type I interferon response to myocardial infarction begins in bone marrow and is regulated by Nrf2-activated macrophages. Sci Immunol 5: eaaz1974. doi:10.1126/ sciimmunol.aaz1974 JH Kim: formal analysis, visualization, and methodology. X Li: data curation, formal analysis, methodology, and wri- ting—original draft, review, and editing. ting—original draft, review, and editing. Cao W, Bover L, Cho M, Wen X, Hanabuchi S, Bao M, Rosen DB, Wang YH, Shaw JL, Du Q, et al (2009) Regulation of TLR7/9 responses in plasmacytoid dendritic cells by BST2 and ILT7 receptor interaction. J Exp Med 206: 1603–1614. doi:10.1084/jem.20090547 Y Morikawa: data curation and writing—original draft. M Park: formal analysis and methodology. D Turaga: data curation. I Adachi: data curation. Card CM, Yu SS, Swartz MA (2014) Emerging roles of lymphatic endothelium in regulating adaptive immunity. J Clin Invest 124: 943–952. doi:10.1172/ JCI73316 JD Wythe: supervision and methodology. MAH Samee: formal analysis, supervision, visualization, and methodology. Chen G, Huang AC, Zhang W, Zhang G, Wu M, Xu W, Yu Z, Yang J, Wang B, Sun H, et al (2018) Exosomal PD-L1 contributes to immunosuppression and is associated with anti-PD-1 response. Nature 560: 382–386. doi:10.1038/ s41586-018-0392-8 JF Martin: conceptualization, supervision, funding acquisition, and writing—original draft, review, and editing. JF Martin: conceptualization, supervision, funding acquisition, and writing—original draft, review, and editing. writing—original draft, review, and editing. writing—original draft, review, and editing. Cheon H, Holvey-Bates EG, Schoggins JW, Forster S, Hertzog P, Imanaka N, Rice CM, Jackson MW, Junk DJ, Stark GR (2013) IFNβ-dependent increases in STAT1, STAT2, and IRF9 mediate resistance to viruses and DNA damage. EMBO J 32: 2751–2763. doi:10.1038/emboj.2013.203 F31HL156681. JD Steimle was supported by NIH K99HL169742, NIH T32HL139430, and NIH F32HL156465. Special thanks to Matthew C. Hill (currently at Broad Institute) for assisting with experimentation. Bracamonte-Baran W, Gilotra NA, Won T, Rodriguez KM, Talor MV, Oh BC, Grifﬁn J, Wittstein I, Sharma K, Skinner J, et al (2021) Endothelial stromal PD-L1 (programmed death ligand 1) modulates CD8(+) T-cell inﬁltration after heart transplantation. Circ Heart Fail 14: e007982. doi:10.1161/CIRCHEARTFAILURE.120.007982 Conﬂict of Interest Statement JF Martin is a cofounder of and owns shares in Yap Therapeutics. The remaining authors declare no competing interests related to the topic of this article. Chong JJ, Yang X, Don CW, Minami E, Liu YW, Weyers JJ, Mahoney WM, Van Biber B, Cook SM, Palpant NJ, et al (2014) Human embryonic-stem-cell- derived cardiomyocytes regenerate non-human primate hearts. Nature 510: 273–277. doi:10.1038/nature13233 Acknowledgements This project was supported by grants from the National Institutes of Health (NIH) through R01HL127717, R01HL118761, and R01HL169511 (JF Martin). Ad- ditional project funding was provided by the Leducq Foundation Transat- lantic Network of Excellence Award 14CVD01 (JF Martin). JF Martin was supported by the Vivian L Smith Foundation. H Long was supported by NIH https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 12 of 16 Endothelial immune responsive signature in heart injury Long et al. 12 of 16 Felsenfeld G, Boyes J, Chung J, Clark D, Studitsky V (1996) Chromatin structure and gene expression. Proc Natl Acad Sci U S A 93: 9384–9388. doi:10.1073/pnas.93.18.9384 Johnson DB, Balko JM, Compton ML, Chalkias S, Gorham J, Xu Y, Hicks M, Puzanov I, Alexander MR, Bloomer TL, et al (2016) Fulminant myocarditis with combination immune checkpoint blockade. N Engl J Med 375: 1749–1755. doi:10.1056/NEJMoa1609214 Foglia MJ, Poss KD (2016) Building and re-building the heart by cardiomyocyte proliferation. Development 143: 729–740. doi:10.1242/dev.132910 Kalucka J, de Rooij L, Goveia J, Rohlenova K, Dumas SJ, Meta E, Conchinha NV, Taverna F, Teuwen LA, Veys K, et al (2020) Single-cell transcriptome atlas of murine endothelial cells. Cell 180: 764–779 e20. doi:10.1016/ j.cell.2020.01.015 Frieler RA, Mortensen RM (2015) Immune cell and other noncardiomyocyte regulation of cardiac hypertrophy and remodeling. Circulation 131: 1019–1030. doi:10.1161/CIRCULATIONAHA.114.008788 Kamada R, Yang W, Zhang Y, Patel MC, Yang Y, Ouda R, Dey A, Wakabayashi Y, Sakaguchi K, Fujita T, et al (2018) Interferon stimulation creates chromatin marks and establishes transcriptional memory. Proc Natl Acad Sci U S A 115: E9162–E9171. doi:10.1073/pnas.1720930115 Ge SX, Jung D, Yao R (2020) ShinyGO: A graphical gene-set enrichment tool for animals and plants. Bioinformatics 36: 2628–2629. doi:10.1093/ bioinformatics/btz931 Goveia J, Rohlenova K, Taverna F, Treps L, Conradi LC, Pircher A, Geldhof V, de Rooij LPMH, Kalucka J, Sokol L, et al (2020) An integrated gene expression landscape proﬁling approach to identify lung tumor endothelial cell heterogeneity and angiogenic candidates. Cancer Cell 37: 421–436 e13. doi:10.1016/j.ccell.2020.03.002 Kheradpour P, Kellis M (2014) Systematic discovery and characterization of regulatory motifs in ENCODE TF binding experiments. Nucleic Acids Res 42: 2976–2987. doi:10.1093/nar/gkt1249 Klemm SL, Shipony Z, Greenleaf WJ (2019) Chromatin accessibility and the regulatory epigenome. Nat Rev Genet 20: 207–220. doi:10.1038/s41576- 018-0089-8 Grant CE, Bailey TL, Noble WS (2011) FIMO: Scanning for occurrences of a given motif. Bioinformatics 27: 1017–1018. doi:10.1093/bioinformatics/btr064 Koenig AL, Shchukina I, Amrute J, Andhey PS, Zaitsev K, Lai L, Bajpai G, Bredemeyer A, Smith G, Jones C, et al (2022) Single-cell transcriptomics reveals cell-type-speciﬁc diversiﬁcation in human heart failure. Nat Cardiovasc Res 1: 263–280. doi:10.1038/s44161-022- 00028-6 Hennecke J, Wiley DC (2001) T cell receptor-MHC interactions up close. Cell 104: 1–4. doi:10.1016/s0092-8674(01)00185-4 Hill MC, Kadow ZA, Long H, Morikawa Y, Martin TJ, Birks EJ, Campbell KS, Nerbonne J, Lavine K, Wadhwa L, et al (2022) Integrated multi-omic characterization of congenital heart disease. Nature 608: 181–191. doi:10.1038/s41586-022-04989-3 K¨ohler D, Eckle T, Faigle M, Grenz A, Mittelbronn M, Laucher S, Hart ML, Robson SC, Müller CE, Eltzschig HK (2007) CD39/ectonucleoside triphosphate diphosphohydrolase 1 provides myocardial protection during cardiac ischemia/reperfusion injury. Circulation 116: 1784–1794. doi:10.1161/ CIRCULATIONAHA.107.690180 Hogan NT, Whalen MB, Stolze LK, Hadeli NK, Lam MT, Springstead JR, Glass CK, Romanoski CE (2017) Transcriptional networks specifying homeostatic and inﬂammatory programs of gene expression in human aortic endothelial cells. Elife 6: e22536. doi:10.7554/eLife.22536 Kong P, Christia P, Frangogiannis NG (2014) The pathogenesis of cardiac ﬁbrosis. Cell Mol Life Sci 71: 549–574. doi:10.1007/s00018-013-1349-6 Honda K, Taniguchi T (2006) IRFs: Master regulators of signalling by toll-like receptors and cytosolic pattern-recognition receptors. Nat Rev Immunol 6: 644–658. doi:10.1038/nri1900 Lai SL, Mar´ın-Juez R, Stainier DYR (2019) Immune responses in cardiac repair and regeneration: A comparative point of view. Cell Mol Life Sci 76: 1365–1380. doi:10.1007/s00018-018-2995-5 Hota SK, Rao KS, Blair AP, Khalilimeybodi A, Hu KM, Thomas R, So K, Kameswaran V, Xu J, Polacco BJ, et al (2022) Brahma safeguards canalization of cardiac mesoderm differentiation. Nature 602: 129–134. doi:10.1038/s41586-021-04336-y Lavine KJ, Epelman S, Uchida K, Weber KJ, Nichols CG, Schilling JD, Ornitz DM, Randolph GJ, Mann DL (2014) Distinct macrophage lineages contribute to disparate patterns of cardiac recovery and remodeling in the neonatal and adult heart. Proc Natl Acad Sci U S A 111: 16029–16034. doi:10.1073/pnas.1406508111 Ingason AB, Goldstone AB, Paulsen MJ, Thakore AD, Truong VN, Edwards BB, Eskandari A, Bollig T, Steele AN, Woo YJ (2018) Angiogenesis precedes cardiomyocyte migration in regenerating mammalian hearts. J Thorac Cardiovasc Surg 155: 1118–1127 e1. doi:10.1016/j.jtcvs.2017.08.127 Lee M, Kiefel H, LaJevic MD, Macauley MS, Kawashima H, O’Hara E, Pan J, Paulson JC, Butcher EC (2014) Transcriptional programs of lymphoid tissue capillary and high endothelium reveal control mechanisms for lymphocyte homing. Nat Immunol 15: 982–995. doi:10.1038/ni.2983 Ise W, Kohyama M, Schraml BU, Zhang T, Schwer B, Basu U, Alt FW, Tang J, Oltz EM, Murphy TL, et al (2011) The transcription factor BATF controls the global regulators of class-switch recombination in both B cells and T cells. Nat Immunol 12: 536–543. doi:10.1038/ni.2037 Lehto S, Niskanen L, Suhonen M, R¨onnemaa T, Laakso M (1996) Medial artery calciﬁcation. A neglected harbinger of cardiovascular complications in non-insulin-dependent diabetes mellitus. Arterioscler Thromb Vasc Biol 16: 978–983. References doi:10.1002/ eji.201948432 Betz BC, Jordan-Williams KL, Wang C, Kang SG, Liao J, Logan MR, Kim CH, Taparowsky EJ (2010) Batf coordinates multiple aspects of B and T cell function required for normal antibody responses. J Exp Med 207: 933–942. doi:10.1084/jem.20091548 Esfahani K, Buhlaiga N, Th´ebault P, Lapointe R, Johnson NA, Miller WH, Jr (2019) Alemtuzumab for immune-related myocarditis due to PD-1 therapy. N Engl J Med 380: 2375–2376. doi:10.1056/NEJMc1903064 Biron CA (2001) Interferons alpha and beta as immune regulators–a new look. Immunity 14: 661–664. doi:10.1016/s1074-7613(01)00154-6 https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 13 of 16 Endothelial immune responsive signature in heart injury Long et al. 13 of 16 Porrello ER, Mahmoud AI, Simpson E, Johnson BA, Grinsfelder D, Canseco D, Mammen PP, Rothermel BA, Olson EN, Sadek HA (2013) Regulation of neonatal and adult mammalian heart regeneration by the miR-15 family. Proc Natl Acad Sci U S A 110: 187–192. doi:10.1073/ pnas.1208863110 Lucas ED, Finlon JM, Burchill MA, McCarthy MK, Morrison TE, Colpitts TM, Tamburini BAJ (2018) Type 1 IFN and PD-L1 coordinate lymphatic endothelial cell expansion and contraction during an inﬂammatory immune response. J Immunol 201: 1735–1747. doi:10.4049/ jimmunol.1800271 Qiu J, Salama ME, Hu CS, Li Y, Wang X, Hoffman R (2018) The characteristics of vessel lining cells in normal spleens and their role in the pathobiology of myeloﬁbrosis. Blood Adv 2: 1130–1145. doi:10.1182/ bloodadvances.2017015073 Luther SA, Tang HL, Hyman PL, Farr AG, Cyster JG (2000) Coexpression of the chemokines ELC and SLC by T zone stromal cells and deletion of the ELC gene in the plt/plt mouse. Proc Natl Acad Sci U S A 97: 12694–12699. doi:10.1073/pnas.97.23.12694 Macosko EZ, Basu A, Satija R, Nemesh J, Shekhar K, Goldman M, Tirosh I, Bialas AR, Kamitaki N, Martersteck EM, et al (2015) Highly parallel genome- wide expression proﬁling of individual cells using nanoliter droplets. Cell 161: 1202–1214. doi:10.1016/j.cell.2015.05.002 Rainer PP, Hao S, Vanhoutte D, Lee DI, Koitabashi N, Molkentin JD, Kass DA (2014) Cardiomyocyte-speciﬁc transforming growth factor beta suppression blocks neutrophil inﬁltration, augments multiple cytoprotective cascades, and reduces early mortality after myocardial infarction. Circ Res 114: 1246–1257. doi:10.1161/CIRCRESAHA.114.302653 Mahmood SS, Fradley MG, Cohen JV, Nohria A, Reynolds KL, Heinzerling LM, Sullivan RJ, Damrongwatanasuk R, Chen CL, Gupta D, et al (2018) Myocarditis in patients treated with immune checkpoint inhibitors. J Am Coll Cardiol 71: 1755–1764. doi:10.1016/j.jacc.2018.02.037 Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, Barengo NC, Beaton AZ, Benjamin EJ, Benziger CP, et al (2020) Global burden of cardiovascular diseases and risk factors, 1990-2019: Update from the GBD 2019 study. J Am Coll Cardiol 76: 2982–3021. doi:10.1016/ j.jacc.2020.11.010 Mar´ın-Juez R, Marass M, Gauvrit S, Rossi A, Lai SL, Materna SC, Black BL, Stainier DYR (2016) Fast revascularization of the injured area is essential to support zebraﬁsh heart regeneration. Proc Natl Acad Sci U S A 113: 11237–11242. doi:10.1073/pnas.1605431113 Roy S, Guler R, Parihar SP, Schmeier S, Kaczkowski B, Nishimura H, Shin JW, Negishi Y, Ozturk M, Hurdayal R, et al (2015) Batf2/Irf1 induces inﬂammatory responses in classically activated macrophages, lipopolysaccharides, and mycobacterial infection. J Immunol 194: 6035–6044. doi:10.4049/jimmunol.1402521 Matatall KA, Jeong M, Chen S, Sun D, Chen F, Mo Q, Kimmel M, King KY (2016) Chronic infection depletes hematopoietic stem cells through stress- induced terminal differentiation. Cell Rep 17: 2584–2595. doi:10.1016/ j.celrep.2016.11.031 Rurik JG, Aghajanian H, Epstein JA (2021) Immune cells and immunotherapy for cardiac injury and repair. Circ Res 128: 1766–1779. doi:10.1161/ CIRCRESAHA.121.318005 Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, Reynolds AP, Sandstrom R, Qu H, Brody J, et al (2012) Systematic localization of common disease-associated variation in regulatory DNA. Science 337: 1190–1195. doi:10.1126/science.1222794 Saeki H, Moore AM, Brown MJ, Hwang ST (1999) Cutting edge: Secondary lymphoid-tissue chemokine (SLC) and CC chemokine receptor 7 (CCR7) participate in the emigration pathway of mature dendritic cells from the skin to regional lymph nodes. J Immunol 162: 2472–2475. doi:10.4049/jimmunol.162.5.2472 McComb S, Cessford E, Alturki NA, Joseph J, Shutinoski B, Startek JB, Gamero AM, Mossman KL, Sad S (2014) Type-I interferon signaling through ISGF3 complex is required for sustained Rip3 activation and necroptosis in macrophages. Proc Natl Acad Sci U S A 111: E3206–E3213. doi:10.1073/pnas.1407068111 Saldanha-Araujo F, Ferreira FI, Palma PV, Araujo AG, Queiroz RHC, Covas DT, Zago MA, Panepucci RA (2011) Mesenchymal stromal cells up- regulate CD39 and increase adenosine production to suppress activated T-lymphocytes. Stem Cell Res 7: 66–74. doi:10.1016/ j.scr.2011.04.001 McNab F, Mayer-Barber K, Sher A, Wack A, O’Garra A (2015) Type I interferons in infectious disease. Nat Rev Immunol 15: 87–103. doi:10.1038/nri3787 Meder B, Rühle F, Weis T, Homuth G, Keller A, Franke J, Peil B, Lorenzo Bermejo J, Frese K, Huge A, et al (2014) A genome-wide association study identiﬁes 6p21 as novel risk locus for dilated cardiomyopathy. Eur Heart J 35: 1069–1077. doi:10.1093/eurheartj/eht251 Salem JE, Manouchehri A, Moey M, Lebrun-Vignes B, Bastarache L, Pariente A, Gobert A, Spano JP, Balko JM, Bonaca MP, et al (2018) Cardiovascular toxicities associated with immune checkpoint inhibitors: An observational, retrospective, pharmacovigilance study. Lancet Oncol 19: 1579–1589. doi:10.1016/S1470-2045(18)30608-9 Mo A, Mukamel EA, Davis FP, Luo C, Henry GL, Picard S, Urich MA, Nery JR, Sejnowski TJ, Lister R, et al (2015) Epigenomic signatures of neuronal diversity in the mammalian brain. Neuron 86: 1369–1384. doi:10.1016/ j.neuron.2015.05.018 Schaper W (1996) Collateral vessel growth in the human heart. Role of ﬁbroblast growth factor-2. Circulation 94: 600–601. doi:10.1161/01.atv.16.8.978 Jabeen R, Goswami R, Awe O, Kulkarni A, Nguyen ET, Attenasio A, Walsh D, Olson MR, Kim MH, Tepper RS, et al (2013) Th9 cell development requires a BATF-regulated transcriptional network. J Clin Invest 123: 4641–4653. doi:10.1172/JCI69489 Leonard D, Svenungsson E, Sandling JK, Berggren O, J¨onsen A, Bengtsson C, Wang C, Jensen-Urstad K, Granstam SO, Bengtsson AA, et al (2013) Coronary heart disease in systemic lupus erythematosus is associated with interferon regulatory factor-8 gene variants. Circ Cardiovasc Genet 6: 255–263. doi:10.1161/CIRCGENETICS.113.000044 Jiang DS, Liu Y, Zhou H, Zhang Y, Zhang XD, Zhang XF, Chen K, Gao L, Peng J, Gong H, et al (2014a) Interferon regulatory factor 7 functions as a novel negative regulator of pathological cardiac hypertrophy. Hypertension 63: 713–722. doi:10.1161/HYPERTENSIONAHA.113.02653 Li P, Spolski R, Liao W, Wang L, Murphy TL, Murphy KM, Leonard WJ (2012) BATF- JUN is critical for IRF4-mediated transcription in T cells. Nature 490: 543–546. doi:10.1038/nature11530 Jiang DS, Luo YX, Zhang R, Zhang XD, Chen HZ, Zhang Y, Chen K, Zhang SM, Fan GC, Liu PP, et al (2014b) Interferon regulatory factor 9 protects against cardiac hypertrophy by targeting myocardin. Hypertension 63: 119–127. doi:10.1161/HYPERTENSIONAHA.113.02083 Liao X, Shen Y, Zhang R, Sugi K, Vasudevan NT, Alaiti MA, Sweet DR, Zhou L, Qing Y, Gerson SL, et al (2018) Distinct roles of resident and nonresident macrophages in nonischemic cardiomyopathy. Proc Natl Acad Sci U S A 115: E4661–E4669. doi:10.1073/pnas.1720065115 Jiang DS, Wei X, Zhang XF, Liu Y, Zhang Y, Chen K, Gao L, Zhou H, Zhu XH, Liu PP, et al (2014c) IRF8 suppresses pathological cardiac remodelling by inhibiting calcineurin signalling. Nat Commun 5: 3303. doi:10.1038/ ncomms4303 Lohoff M, Mak TW (2005) Roles of interferon-regulatory factors in T-helper- cell differentiation. Nat Rev Immunol 5: 125–135. doi:10.1038/nri1552 Jin S, Guerrero-Juarez CF, Zhang L, Chang I, Ramos R, Kuan CH, Myung P, Plikus MV, Nie Q (2021) Inference and analysis of cell-cell communication using CellChat. Nat Commun 12: 1088. doi:10.1038/s41467-021-21246-9 Lu P, Wang Y, Liu Y, Wang Y, Wu B, Zheng D, Harvey RP, Zhou B (2021) Perinatal angiogenesis from pre-existing coronary vessels via DLL4-NOTCH1 signalling. Nat Cell Biol 23: 967–977. doi:10.1038/s41556-021-00747-1 https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 14 of 1 Endothelial immune responsive signature in heart injury Long et al. 14 of 16 renewal by pre-existing cardiomyocytes. Nature 493: 433–436. doi:10.1038/nature11682 renewal by pre-existing cardiomyocytes. Nature 493: 433–436. doi:10.1038/nature11682 van Berlo JH, Molkentin JD (2014) An emerging consensus on cardiac regeneration. Nat Med 20: 1386–1393. doi:10.1038/nm.3764 Seternes T, Sørensen K, Smedsrød B (2002) Scavenger endothelial cells of vertebrates: A nonperipheral leukocyte system for high-capacity elimination of waste macromolecules. Proc Natl Acad Sci U S A 99: 7594–7597. doi:10.1073/pnas.102173299 Virani SS, Alonso A, Aparicio HJ, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, Chamberlain AM, Cheng S, Delling FN, et al (2021) Heart disease and stroke statistics-2021 update: A report from the American heart association. Circulation 143: e254–e743. doi:10.1161/ CIR.0000000000000950 Shaw LJ, Raggi P, Schisterman E, Berman DS, Callister TQ (2003) Prognostic value of cardiac risk factors and coronary artery calcium screening for all-cause mortality. Radiology 228: 826–833. doi:10.1148/ radiol.2283021006 Wang L, Yu P, Zhou B, Song J, Li Z, Zhang M, Guo G, Wang Y, Chen X, Han L, et al (2020) Single-cell reconstruction of the adult human heart during heart failure and recovery reveals the cellular landscape underlying cardiac function. Nat Cell Biol 22: 108–119. doi:10.1038/s41556-019- 0446-7 Shields JD, Kourtis IC, Tomei AA, Roberts JM, Swartz MA (2010) Induction of lymphoidlike stroma and immune escape by tumors that express the chemokine CCL21. Science 328: 749–752. doi:10.1126/science.1185837 Ware JA, Simons M (1997) Angiogenesis in ischemic heart disease. Nat Med 3: 158–164. doi:10.1038/nm0297-158 Song K, Nam YJ, Luo X, Qi X, Tan W, Huang GN, Acharya A, Smith CL, Tallquist MD, Neilson EG, et al (2012) Heart repair by reprogramming non- myocytes with cardiac transcription factors. Nature 485: 599–604. doi:10.1038/nature11139 Weirauch MT, Yang A, Albu M, Cote AG, Montenegro-Montero A, Drewe P, Najafabadi HS, Lambert SA, Mann I, Cook K, et al (2014) Determination and inference of eukaryotic transcription factor sequence speciﬁcity. Cell 158: 1431–1443. doi:10.1016/j.cell.2014.08.009 St Hilaire C, Ziegler SG, Markello TC, Brusco A, Groden C, Gill F, Carlson- Donohoe H, Lederman RJ, Chen MY, Yang D, et al (2011) NT5E mutations and arterial calciﬁcations. N Engl J Med 364: 432–442. doi:10.1056/ NEJMoa0912923 Wheeler DG, Joseph ME, Mahamud SD, Aurand WL, Mohler PJ, Pompili VJ, Dwyer KM, Nottle MB, Harrison SJ, d’Apice AJF, et al (2012) Transgenic swine: Expression of human CD39 protects against myocardial injury. J Mol Cell Cardiol 52: 958–961. doi:10.1016/j.yjmcc.2012.01.002 Steffens S, Nahrendorf M, Madonna R (2022) Immune cells in cardiac homeostasis and disease: Emerging insights from novel technologies. Eur Heart J 43: 1533–1541. renewal by pre-existing cardiomyocytes. Nature 493: 433–436. doi:10.1038/nature11682 doi:10.1093/eurheartj/ehab842 Whyte WA, Orlando DA, Hnisz D, Abraham BJ, Lin CY, Kagey MH, Rahl PB, Lee TI, Young RA (2013) Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell 153: 307–319. doi:10.1016/j.cell.2013.03.035 Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM 3rd, Hao Y, Stoeckius M, Smibert P, Satija R (2019) Comprehensive integration of single-cell data. Cell 177: 1888–1902 e21. doi:10.1016/j.cell.2019.05.031 Willimann K, Legler DF, Loetscher M, Roos RS, Delgado MB, Clark-Lewis I, Baggiolini M, Moser B (1998) The chemokine SLC is expressed in T cell areas of lymph nodes and mucosal lymphoid tissues and attracts activated T cells via CCR7. Eur J Immunol 28: 2025–2034. doi:10.1002/ (SICI)1521-4141(199806)28:06<2025::AID-IMMU2025>3.0.CO;2-C Takeda A, Hollm´en M, Dermadi D, Pan J, Brulois KF, Kaukonen R, L¨onnberg T, Bostr¨om P, Koskivuo I, Irjala H, et al (2019) Single-cell survey of human lymphatics unveils marked endothelial cell heterogeneity and mechanisms of homing for neutrophils. Immunity 51: 561–572 e5. doi:10.1016/j.immuni.2019.06.027 Wohlleber D, Knolle PA (2016) The role of liver sinusoidal cells in local hepatic immune surveillance. Clin Transl Immunol 5: e117. doi:10.1038/ cti.2016.74 Tamura T, Yanai H, Savitsky D, Taniguchi T (2008) The IRF family transcription factors in immunity and oncogenesis. Annu Rev Immunol 26: 535–584. doi:10.1146/annurev.immunol.26.021607.090400 Wu C, Thalhamer T, Franca RF, Xiao S, Wang C, Hotta C, Zhu C, Hirashima M, Anderson AC, Kuchroo VK (2014) Galectin-9-CD44 interaction enhances stability and function of adaptive regulatory T cells. Immunity 41: 270–282. doi:10.1016/j.immuni.2014.06.011 Thiriot A, Perdomo C, Cheng G, Novitzky-Basso I, McArdle S, Kishimoto JK, Barreiro O, Mazo I, Triboulet R, Ley K, et al (2017) Differential DARC/ ACKR1 expression distinguishes venular from non-venular endothelial cells in murine tissues. BMC Biol 15: 45. doi:10.1186/s12915- 017-0381-7 Xiang M, Grosso RA, Takeda A, Pan J, Bekkhus T, Brulois K, Dermadi D, Nordling S, Vanlandewijck M, Jalkanen S, et al (2020) A single-cell transcriptional roadmap of the mouse and human lymph node lymphatic vasculature. Front Cardiovasc Med 7: 52. doi:10.3389/ fcvm.2020.00052 Thurman RE, Rynes E, Humbert R, Vierstra J, Maurano MT, Haugen E, Shefﬁeld NC, Stergachis AB, Wang H, Vernot B, et al (2012) The accessible chromatin landscape of the human genome. Nature 489: 75–82. doi:10.1161/ 01.cir.94.4.600 Moslehi JJ, Salem JE, Sosman JA, Lebrun-Vignes B, Johnson DB (2018) Increased reporting of fatal immune checkpoint inhibitor-associated myocarditis. Lancet 391: 933. doi:10.1016/S0140-6736(18)30533-6 Schiller HB, Montoro DT, Simon LM, Rawlins EL, Meyer KB, Strunz M, Vieira Braga FA, Timens W, Koppelman GH, Budinger GRS, et al (2019) The human lung cell atlas: A high-resolution reference map of the human lung in health and disease. Am J Respir Cell Mol Biol 61: 31–41. doi:10.1165/rcmb.2018-0416TR Nagira M, Imai T, Hieshima K, Kusuda J, Ridanp¨a¨a M, Takagi S, Nishimura M, Kakizaki M, Nomiyama H, Yoshie O (1997) Molecular cloning of a novel human CC chemokine secondary lymphoid-tissue chemokine that is a potent chemoattractant for lymphocytes and mapped to chromosome 9p13. J Biol Chem 272: 19518–19524. doi:10.1074/ jbc.272.31.19518 Schneider WM, Chevillotte MD, Rice CM (2014) Interferon-stimulated genes: A complex web of host defenses. Annu Rev Immunol 32: 513–545. doi:10.1146/annurev-immunol-032713-120231 Nahrendorf M (2018) Myeloid cell contributions to cardiovascular health and disease. Nat Med 24: 711–720. doi:10.1038/s41591-018-0064-0 Schraml BU, Hildner K, Ise W, Lee WL, Smith WAE, Solomon B, Sahota G, Sim J, Mukasa R, Cemerski S, et al (2009) The AP-1 transcription factor Batf controls T(H)17 differentiation. Nature 460: 405–409. doi:10.1038/ nature08114 Platanitis E, Gruener S, Ravi Sundar Jose Geetha A, Boccuni L, Vogt A, Novatchkova M, Sommer A, Barozzi I, Müller M, Decker T (2022) Interferons reshape the 3D conformation and accessibility of macrophage chromatin. iScience 25: 103840. doi:10.1016/ j.isci.2022.103840 Schupp JC, Adams TS, Cosme C, Jr, Raredon MSB, Yuan Y, Omote N, Poli S, Chioccioli M, Rose KA, Manning EP, et al (2021) Integrated single-cell atlas of endothelial cells of the human lung. Circulation 144: 286–302. doi:10.1161/CIRCULATIONAHA.120.052318 Porrello ER, Mahmoud AI, Simpson E, Hill JA, Richardson JA, Olson EN, Sadek HA (2011) Transient regenerative potential of the neonatal mouse heart. Science 331: 1078–1080. doi:10.1126/science.1200708 Senyo SE, Steinhauser ML, Pizzimenti CL, Yang VK, Cai L, Wang M, Wu TD, Guerquin-Kern JL, Lechene CP, Lee RT (2013) Mammalian heart https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 Endothelial immune responsive signature in heart injury Long et al. 15 of 16 License: This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/ licenses/by/4.0/). renewal by pre-existing cardiomyocytes. Nature 493: 433–436. doi:10.1038/nature11682 doi:10.1038/nature11232 Zhang SM, Gao L, Zhang XF, Zhang R, Zhu LH, Wang PX, Tian S, Yang D, Chen K, Huang L, et al (2014) Interferon regulatory factor 8 modulates phenotypic switching of smooth muscle cells by regulating the activity of myocardin. Mol Cell Biol 34: 400–414. doi:10.1128/MCB.01070-13 Travaglini KJ, Nabhan AN, Penland L, Sinha R, Gillich A, Sit RV, Chang S, Conley SD, Mori Y, Seita J, et al (2020) A molecular cell atlas of the human lung from single-cell RNA sequencing. Nature 587: 619–625. doi:10.1038/ s41586-020-2922-4 Travers JG, Kamal FA, Robbins J, Yutzey KE, Blaxall BC (2016) Cardiac ﬁbrosis: The ﬁbroblast awakens. Circ Res 118: 1021–1040. doi:10.1161/ CIRCRESAHA.115.306565 License: This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/ licenses/by/4.0/). License: This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/ licenses/by/4.0/). Tsompana M, Buck MJ (2014) Chromatin accessibility: A window into the genome. Epigenetics Chromatin 7: 33. doi:10.1186/1756-8935-7-33 Endothelial immune responsive signature in heart injury Long et al. https://doi.org/10.26508/lsa.202201870 vol 7 \| no 2 \| e202201870 16 of 16
https://openalex.org/W3009022002	http://old.scielo.br/pdf/ccrh/v32n87/0103-4979-ccrh-32-87-0505.pdf	Portuguese	null	INDÚSTRIA CULTURAL E IDEOLOGIA	Caderno CRH	2,019	cc-by	7,402	* Universidade Federal de Rondônia (UNIR). Departamen to de Ciências Sociais. BR 364, Km 9,5. CEP: 76801-059. Porto Velho – Rondônia – Brasil. humbertoalvesj9@gmail.com https://orcid.org/0000-0002-5503-5484 INDÚSTRIA CULTURAL E IDEOLOGIA DOSSIÊ DOSSIÊ Humberto Alves Silva Junior* Humberto Alves Silva Junior* O trabalho parte da análise de conteúdo para abordar as discussões sobre o conceito de indústria cultural cunhado e analisado por Theodor Adorno e Max Horkheimer e seu desdobramento no livro Teoria Estética de Adorno, em especial sobre o cinema. O conceito compreende o caráter comercial e o modo de produção industrial das produções culturais no capitalismo, tratadas, inclusive, como mercadorias e suas consequ ências sobre o público, atuando principalmente como instrumento de manipulação ideológica na visão adorniana. Entretanto, em Teoria Estética, o autor avança a discussão e admite que, apesar da presença da ideologia, a indústria cultural poderia também desenvolver um espaço alternativo para produções massifi cadas. Posteriormente, Frederic Jameson, inspirado no trabalho de Adorno, traçou linha semelhante, ao per ceber que os produtos da indústria cultural não seriam apenas ideológicos. Para além de Adorno, afirmava que eles também poderiam ser utópicos, pois a cultura de massa atrai o público com promessas coletivas e individuais de um futuro melhor. Palavras-chave: Ideologia. Emancipação. Cinema. Cultura de massa. Escola de Frankfurt. INDÚSTRIA CULTURAL E IDEOLOGIA industrial, a nova complexidade do tráfego das ruas, as intervenções e as demolições urbanas e o aumento vertiginoso da população. dernidade, torna-se o contrário: dissemina ainda mais seu caráter tenso. O gosto do grande públi co pelas novas atividades culturais do século XIX estava assentado na “impressão de realidade” e nos estímulos e choques provenientes desses novos meios de diversão. Muitos autores, como Adorno, consideravam que esse mecanismo do entretenimento mantinha o trabalhador preso ao ritmo veloz das fábricas e dos centros urbanos modernos, e, mesmo no ócio, o ritmo assemelha -se ao da produção. De outro lado, os produtos ar tísticos supostamente buscam semelhanças cada vez mais estreitas com essa realidade, apenas me diante um naturalismo. A profusão dessas imagens espelha, num grau acentuado, o ritmo, a velocidade e a mudança que marcam a experiência moder na, o que, por sua vez, condiciona e transfor ma inexoravelmente os aspectos psicológicos e fisiológicos dos indivíduos expostos a essa gama variada de estímulos, tornando-os seres angustiados, ansiosos, ciclotímicos e nervosos. O cinema também contribui para a formação desse estado neurológico. Especialmente em Hollywood, a produção se assemelha à organi zação de uma indústria, que inclui, dentre ou tros elementos, uma divisão de trabalho defi nida, cujo objetivo principal é o lucro. Ou seja, a realização de um filme segue às exigências do ritmo do capital, com a rapidez das produ ções, que atinge toda a equipe de trabalhado res: atores, diretores, montadores etc. Por outro lado, o público experimenta sensações típicas da vida moderna, como atentam Leo Charney e Vanessa Schwartz: A “impressão de realidade” e os cho ques intencionalmente veiculados pela indús tria cultural são recursos frequentes, que se tornaram norma nesse tipo de produção, em especial no cinema. O que, de fato, está por trás desse mecanismo é o padrão da cultura de massa, que, assim como os outros produ tos comerciais, necessitam dessa padronização como meio imprescindível de garantir o lucro. Esse resultado advém, sobretudo, do ca ráter da cultura de massa, fundamentado na economia capitalista mais ampla. Desse modo, assim como a indústria de eletrodomésticos padroniza todos os âmbitos da produção, a indústria cultural também necessita padroni zar todo o seu sistema, seja na produção, na distribuição, ou mesmo no próprio conteúdo do produto. INTRODUÇÃO velocidade, a transformação, a efemeridade, a instabilidade, a tecnologia, a ciência, o espetá culo, o consumo e a perda da identidade, entre outros. Esses signos, que comumente atuam na vida cotidiana moderna de forma aguda, ex cessiva, e de vários modos, estão inseridos no cinema. Por esse motivo, Charney e Schawrtz (2001) consideram o cinema como a arte que melhor define e sintetiza a modernidade. Afir mam que “a cultura moderna foi cinematogra fia antes do cinema”, pois ele surgiu no final do século XIX, seguindo as linhas da contur bada realidade social do capitalismo moderno. O acelerado incremento científico e tec nológico, principalmente a partir do século XIX, com a Revolução Industrial e a crescente difusão da cultura de massa no século XX, pro piciou o crescimento vertiginoso do consumo do lazer, não apenas facultando o nascimento do cinema, mas sustentando-o até hoje como meio de expressão artística, documental ou de entretenimento comercial. Sob o aspecto cien tífico, pode-se observar ainda que a “invenção” do cinema tanto foi um produto desse tipo de conhecimento, através do cruzamento de di versas ciências – matemática, física, química, mecânica óptica e eletricidade –, como tam bém foi um instrumento destinado ao estudo, dentre outros objetos, da fisiologia animal, dos processos da visão e da fotografia. C CRH S l d 32 87 505 515 S O cinema acompanhou a velocidade dos novos tempos, apresentando várias regiões do mundo em apenas alguns segundos, atra vés da mudança rápida das imagens, compri mindo, assim, as distâncias e o tempo, como fazem os novos meios de transporte. A rápida sucessão de imagens expostas ao espectador provoca a sensação de não pertencimento a lugar algum, apresentando, de forma cabal, a identidade descentrada do sujeito moderno, para o qual não há referências seguras dian te da intensificação das mudanças. Esse ritmo veloz acompanha também a pressa do trabalho Por outro lado, o cinema enfeixa, em seu interior, vários signos fundamentais que normalmente caracterizam a modernidade: a 505 http://dx.doi.org/10.9771/ccrh.v32i87.32099 INDÚSTRIA CULTURAL E IDEOLOGIA Portanto, a inserção constante da tensão, do choque e da chamada “impressão de realidade” nos produtos artísticos da cultu ra de massa segue o interesse calculado do em presário da área cultural em padronizar seus produtos como meio de apreender a atenção do espectador e, assim, não arriscar os seus negócios. O padrão estipulado pela indústria cultural ocorre exatamente com o fito de agra dar e acostumar o grande público ao modelo, garantindo que os produtos culturais de hoje e os de amanhã serão facilmente vendidos, des de que sigam a fórmula. Os primeiros filmes de atualidades apresentavam com frequência um simulacro de viagem não apenas ao apresentar paisagens estrangeiras, mas também “pas seios fantasmas, que eram filmados da parte dianteira de trens ou da proa de barcos e que davam aos espec tadores, sentados e parados, uma sensação palpável de movimento (Charney; Schwartz, 2001, p.17). Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 RH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. Esse estado neurológico provocado pela modernidade é representado e incitado não apenas pelo cinema, mas por vários modos de entretenimento e meios de comunicação que, na virada do século XIX, tinham como princi pais motes o choque, o escândalo, o surpreen dente, o sórdido, São exemplos dessas caracte rísticas, observa Singer (2001), os cartuns de revistas e de jornais sensacionalistas norte-a mericanos, que comumente, em suas páginas, apresentavam imagens de acidentes domésti cos e automobilísticos. Caderno CR Ao contrário do que se poderia imaginar, o entretenimento, ao invés de servir como um bálsamo para os problemas nevrálgicos da mo Para Adorno e Horkheimer (1985), o pa drão da indústria cultural é consolidado pela 506 Humberto Alves Silva Junior técnica, responsável pelo poder de sedução que imprime sobre os espectadores. Através do seu aperfeiçoamento constante, a sensação do real reproduzido pelo modelo é sempre renovada. repetem insistentemente os mesmo clichês – conforma o espectador aos mecanismos de ma nipulação, inculcando noções maniqueístas de certo e de errado, de bom e de mau, as quais, frequentemente, estão de acordo com as pers pectivas da moral dominante. O padrão exerce também uma função ideológica e estabelece diretrizes não só para a forma do produto artís tico comercial, para a qual o espectador deve ser previamente preparado, mas igualmente difunde ideias, valores, normas e regras de conduta dominantes. Nesse sentido, a atração exercida pela cultura de massa, principalmente o cinema co mercial, está intimamente ligada ao poder de levar os espectadores a um estado de superex citarão, no qual eles passam a se sentir como se, de fato, estivessem vivenciando aquilo que é representado, entregando-se ao deleite das sensações e da emoção. Esses estados “sensa cionalistas” produzidos pela indústria cultural tornam cada vez mais perfeita a ilusão de que o produto artístico é similar ao mundo real, ou que uma possível realidade futura, como no caso dos filmes de ficção, encontra-se a nosso alcance nos dias atuais. O conteúdo das obras da cultura de mas sa está, portanto, impreterivelmente de acordo com o establishment, pois seus produtores ten tam convencer que a ordem social, defendida tão bem nos filmes, é a única possível, elimi nando qualquer fator que coloque em risco a moral e os interesses econômicos do capitalis mo. INDÚSTRIA CULTURAL E IDEOLOGIA entre alta cultura e cultura de massa, no âmbi to mais amplo da indústria cultural. nome de um novo conceito de arte que, além de estar em consonância com o dinamismo da mo dernidade, deveria instaurar “originalidade e espírito de pesquisa” como “elementos centrais da produção artística ” (Xavier, 1978, p. 62). Esses movimentos pretendiam produ zir filmes que escapassem da padronização da narrativa clássica de Hollywood e que, ao mesmo tempo, iniciassem uma nova expres são artística. Estavam interessados em empres tar ao cinema um caráter artístico, tal como se verifica no teatro, na pintura, na música, na escultura e na poesia. Seus membros passa ram então a formular teorias estéticas para dar fundamentação à nova arte, cujo suporte é a estrutura cinematográfica. Os filmes, segundo essa concepção, de veriam conduzir os espectadores, através da imersão nas imagens, a um estado de liberta ção dos mecanismos da lógica, percorrendo os caminhos do fluxo imagético. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 Se a cultura de massa não aniquila os es paços destinados à reflexão e ao livre pensar do espectador, tais espaços ficam bastante reduzi dos por conta das características do capital. Ao padrão inerente das produções co merciais corresponde – presumem seus ide alizadores – uma recepção também padroni zada. Devido à difusão em massa do produto comercial, busca-se construir uma espécie de espectador adaptado a um estilo estereotipa do, ansioso por se reconhecer na mesma forma estética de sempre. No cinema, o padrão de termina também o modo de olhar do público, pois a lente normalmente é colocada em um ponto no qual possa causar o efeito de se estar vendo uma “realidade objetiva” (Xavier, 1978, p. 22). Por isso, a narrativa se torna o modelo mais comum nesse tipo de expressão. A racionalidade capitalista incorporada ao fazer artístico interfere no conteúdo da obra de arte, e é por esse motivo que surgem vários movimentos com a pretensão de resgatar, ou até mesmo fundar, novos princípios defini dores da arte. Esses grupos normalmente vão além desse objetivo ao criticarem a sociedade capitalista e seus mecanismos de controle, em princípio incompatíveis com a tendência liber tadora da arte. O grande público da indústria cultural e do cinema, acostumado com a linguagem do cinema de Hollywood, rejeita o discurso discrepante que possa destoar da linguagem simplista dos vários gêneros comerciais, su bordinando-se ao sensacionalismo repetitivo que embala as produções da cultura de massa. O propósito é apenas o divertimento, através da distensão emocional induzida por mecanis mos técnicos padronizados que, aparentemen te, aproximam a representação do real. C C S l d S No caso do cinema, os primeiros movi mentos artísticos surgem nas primeiras déca das do século XX, como o Cinema Soviético. Teóricos, críticos e artistas, de um modo geral, passam a elaborar filmes que tentam fugir das padronizações impostas pela estética indus trial, pretendendo trazer, também para o cine ma, a aura artística já configurada em outras expressões estéticas. É a partir de então que se fomenta a oposição entre filme comercial e filme de arte, ou filme de autor (termo elabo rado pelos críticos da Cahiers du Cinéma na década de 1950), que corresponde à oposição Há, por último, para Adorno e Horkhei mer (1985), o padrão do próprio conteúdo – com um número limitado de gêneros que 507 INDÚSTRIA CULTURAL E IDEOLOGIA ADORNO, HORKHEIMER E A CRÍ TICA À INDÚSTRIA CULTURAL Os primeiros passos nesse sentido foram dados ainda no cinema mudo. Grupos de intelec tuais vinculados à renovação artística moderna viram no cinema, até então considerado como mero entretenimento, não só o meio ideal para a expressão de uma nova arte, mas também a con sideraram a arte moderna por excelência, pela ampla capacidade expressiva do novo suporte artístico, que possibilita ao artista experimentar uma gama maior de recursos estéticos. No campo da teoria social o texto A In dústria cultural – o esclarecimento como mis tificação das massas, parte integrante do li vro Dialética do esclarecimento, de Adorno e Horkheimer, lançado em 1947, aposta numa vi são pessimista em relação aos meios de expres são representados pela indústria cultural, ape sar da existência de uma crítica em relação às produções da cultura de massa, principalmente no interior do cinema. Como foi mostrado an teriormente, esses autores aparentemente a ig noraram, e elaboraram uma contundente crítica aos “produtos artísticos industrializados”. Por outro lado, o cinema, que tem sua origem na ciência e na técnica, é a arte que me lhor traduz, através da imagem em movimen to, a mudança, a efemeridade e a fragmentação que formam a base material da vida moderna. O artista moderno tem, diante de si, um instru mento privilegiado, que lhe permite conformar sua obra às condições de vida produzida pelo espírito na modernidade. Por conseguinte, para Adorno e Horkhei mer, à medida que se ampliavam os eventos artís ticos de um modo geral, difundia-se também um tipo de produção artística voltada para o grande público, tendo por meta principal o lucro. O cres cimento dessa perspectiva promove a absorção dos bens culturais pela lógica empresarial, que organiza todas as etapas de construção da obra. É a partir, portanto, desses elementos que se funda menta o produto artístico comercial, tão caracte rístico da cultura de massa. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 or, v. 32, n. 87, p. 505-515, Set./D Dessa forma, o autor de cinema alcança aquele ideal segundo Charles Baudelaire, que definia o artista moderno como “alguém capaz de concentrar a visão em elementos comuns da vida da cidade”, compreendendo “suas qualida des fugidias e, ainda assim, extrair, do momen to fugaz, todas as sugestões da eternidade nele contida” (Baudelaire, apud Harvey, 1999, p. 29). Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./D Dessa forma, o autor de cinema alcança aquele ideal segundo Charles Baudelaire, que definia o artista moderno como “alguém capaz de concentrar a visão em elementos comuns da vida da cidade”, compreendendo “suas qualida des fugidias e, ainda assim, extrair, do momen to fugaz, todas as sugestões da eternidade nele contida” (Baudelaire, apud Harvey, 1999, p. 29). A proposta dos primeiros teóricos do ci nema segue, de perto, as orientações das van guardas de outras expressões artísticas moder nas, vinculadas, principalmente, às artes plás ticas e à literatura, que combatiam os antigos ideais de beleza e perfeição da arte clássica, em Em virtude disso, o status da arte muda de perfil, devido ao impacto causado pela co mercialização dos bens culturais. Um primei ro aspecto a observar é que, no capitalismo, as relações sociais que envolvem o fazer artísti co passam, como outras relações sociais, a ser mediadas pelo dinheiro; portanto, mesmo as 508 Humberto Alves Silva Junior invés de, inversamente, formar o público mais am plo numa cultura intacta em substância (Habermas, 2003, p. 195). obras de arte que, de fato, pretendem testemu nhar uma crítica à economia capitalista, de vem passar pela chancela do capital. Com isso, as obras de arte, que, outrora, eram considera das como elementos de veneração e respeito, perdem sua aura e tornam-se mercadorias. Um segundo aspecto significativo é o fato de o ar tista se tornar um trabalhador comum, atento às vicissitudes do mercado para poder garan tir sua própria sobrevivência. Como observam Marx e Engels, no Manifesto Comunista: O caráter comercial dos produtos da cul tura de massa acaba por determinar a forma e o conteúdo desses mesmos produtos, com a intenção de que eles sejam facilmente absorvi dos por um número cada vez maior de espec tadores (consumidores). É por esse motivo que as produções comerciais seguem determinadas fórmulas predefinidas, exatamente para atingir o grande público, fazendo com que ele se fa miliarize com o modelo. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 É nesse sentido que Adorno e Horkheimer se referiam à importân cia dos clichês nas produções cinematográfi cas da cultura de massa: A burguesia despojou de sua auréola todas as ativida des até então reputadas veneráveis e encaradas com piedoso respeito. Transformou em seus trabalhado res assalariados o médico, o jurista, o padre, o poeta, o homem de ciência (Marx; Engels, 1986, p.24). As obras de arte, assim como a ciência, já não são mais objetos de intensos debates como ocorria antes nos cafés e salões; em seu lugar, assume a cultura de massa, que não fa vorece, como afirmam Adorno e Horkheimer (1985), o livre curso do raciocínio político e ar tístico, pois a informação é dada a partir de um centro, sem que o espectador possa replicar. Uma prova de que, no capitalismo, a arte pas sa a ser uma mercadoria como outra qualquer e se insere na lógica mais ampla do mercado está no fato de que sua difusão está atrelada ao aparato do comércio geral, visando, funda mentalmente, ao lucro. Não somente os tipos das canções de sucesso, os astros, as novelas ressurgem ciclicamente como in variantes fixos, mas o conteúdo do espetáculo é ele próprio derivado deles e só varia na aparência. Os detalhes se tornam fungíveis. A breve sequência de intervalos, fácil de memorizar, como mostrou a can ção de sucesso; o fracasso temporário do herói, que ele sabe suportar como good sport que é; a boa pal mada que a namorada recebe da mão forte do astro; sua rude reserva em face da herdeira mimada são, como todos os detalhes, clichês prontos para serem empregados arbitrariamente aqui e ali e completa mente definidos pela finalidade que lhes cabe no esquema (Adorno; Horkheimer, 1985, p. 117-118). H Salvador v 32 n 87 p 505-515 Set /Dez 2019 A configuração da arte como um bem de consumo a partir da indústria cultural, colo ca em foco a tensão existente entre a esfera da produção e da circulação eminentemente mer cantil relacionada ao entretenimento, e uma outra comumente chamada de modernismo, “erudita”, de “arte”, ou ainda da alta cultura, mais preocupada com os aspectos estéticos das obras. A primeira se insere de forma mais completa nos parâmetros da indústria cultural, enquanto a segunda tenta se estabelecer como uma resistência ao processo de mercantiliza ção total da arte. INDÚSTRIA CULTURAL E IDEOLOGIA Por conseguinte, ocorre uma dupla contradição em relação às produções da alta cultura. Por um lado, a arte corre o risco de ter seu valor intrínseco subjugado pelo valor mo netário; por outro, isso torna vazio o protesto político comumente relacionado a essa tendên cia, vazio, pois as obras dependem do sistema político e econômico que seus autores criticam. Há também uma outra similaridade en tre os dois tipos de produção, que é a busca in cessante pelo novo. Na cultura de massa, isso aparece, por exemplo, na frequente criação de gêneros e subgêneros: suspense, terror, fic ção científica, pornografia, western. Contudo a mudança é aparente, pois os estilos sempre retornam com poucas modificações, para ape nas chamar a atenção do público, que, assim, se sente como estivesse, de fato, consumindo um produto novo. Na alta cultura, a busca per manente pelo novo também ocorre, mas com Portanto até mesmo as artes mais tradi cionais, como a pintura, necessitam ser inter mediadas pelo comércio até que alcancem o público. Por conseguinte, ocorre uma dupla contradição em relação às produções da alta cultura. Por um lado, a arte corre o risco de ter seu valor intrínseco subjugado pelo valor mo netário; por outro, isso torna vazio o protesto político comumente relacionado a essa tendên cia, vazio, pois as obras dependem do sistema político e econômico que seus autores criticam. Ainda que, de fato, essas afirmações se refiram à cultura de massa, é preciso fazer al gumas ressalvas, pois elas acabam conceden do um poder absoluto à ideologia dominante, compreendendo a indústria cultural apenas como manipulação coletiva. Entretanto, em contraponto ao sentido pessimista contido no texto Indústria Cultu ral em relação à cultura de massa, Adorno, na Teoria Estética, já teria apontado, em algumas passagens, elementos utópicos na arte contem porânea por ele considerada até então como reificada. Adorno, nesse último texto, ressalva que o segmento crítico da arte moderna repre senta um lócus de contestação no interior das relações reificadas do capital. Caderno CRH, Salvador, v. 32, n. 87, p. 505-51 Há também uma outra similaridade en tre os dois tipos de produção, que é a busca in cessante pelo novo. Na cultura de massa, isso aparece, por exemplo, na frequente criação de gêneros e subgêneros: suspense, terror, fic ção científica, pornografia, western. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 Dessa forma, o fazer artístico, no capi talismo, se destina prioritariamente às neces sidades do capital, sendo controlado por ele e não mais correspondendo às necessidades do espírito, como afirma Habermas: […] as leis do mercado já penetram na substância das obras, tornando-se imanentes a elas como leis estruturais. Não mais apenas a difusão e escolha, a apresentação e embalagem das obras – mas a pró pria criação delas enquanto tais se orienta nos seto res amplos da cultura dos consumidores, conforme pontos de vista da estratégia de vendas no merca do. Sim, a cultura de massa recebe o seu duvidoso nome exatamente por conformar-se às necessidades de distração e diversão de grupos de consumidores com um nível de formação relativamente baixo, ao Entretanto a linha que delimita essas duas tendências é extremamente precária, pois existem pontos de convergência entre elas. Isso 509 INDÚSTRIA CULTURAL E IDEOLOGIA se deve ao fato de tanto as produções comer ciais como as produções da alta cultura se en contrarem inseridas no complexo mais amplo da economia capitalista. Uma está mais direta mente ligada ao caráter mercantil, fundamen tada em uma estrutura burocrática capitalista, a outra se estabelece como contraponto à pri meira, na tentativa de garantir a preponderân cia do valor artístico sobre o valor econômico. Contudo as duas vertentes estão relacionadas à lógica capitalista; seus produtos precisam ser elaborados, distribuídos e comercializados dentro do processo econômico existente. uma outra intenção, a de criticar a sociedade de consumo e romper com o formalismo ideo lógico da cultura de massa. Entretanto as ino vações alcançam um patamar-limite, devido à própria banalização do recurso, estiolando o processo criativo; a pretensão de causar es cândalo se esgota, e o recurso é, em si, neutra lizado. Assim, o projeto estético inovador da cultura moderna é derrotado. Um último ponto a ser observado em relação a uma possível interpenetração entre alta cultura e cultura de massa se refere à ide ologia. No texto Industria Cultural, Adorno e Horkheimer definiam as produções da indús tria cultural como ideológicas, apontando uma suposta manipulação total do público, como fica evidente na seguinte passagem: Uma segunda observação comumente feita entre os estudiosos da cultura refere-se ao aspecto mercantil, pois quase nenhuma pro dução cultural, de “arte” ou “comercial”, pode escapar da esfera da circulação na sociedade capitalista. É esse segundo dilema que demar ca a oposição entre alta cultura e cultura de massa. Uma manifestação estética somente po derá completar seu significado estético se for colocada para a apreciação pública. Os consumidores são os trabalhadores e os empre gados, os lavradores e os pequenos burgueses. A produção capitalista os mantém presos em corpo e alma e eles sucumbem sem resistência ao que lhes é oferecido. Assim como os dominados sempre leva ram mais a sério do que os dominadores a moral que deles recebiam, hoje em dia, as massas logradas su cumbem mais facilmente ao mito do sucesso do que os bem-sucedidos. Elas têm os desejos deles. Obsti nadamente, insistem na ideologia que as escraviza (Adorno; Horkheimer,1985 p. 125). Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 colocada para a apreciação pública. Portanto até mesmo as artes mais tradi cionais, como a pintura, necessitam ser inter mediadas pelo comércio até que alcancem o público. 1 Benjamin foi um dos primeiros autores da Escola de Frankfurt que concebia de forma positiva as novas ex pressões artísticas fundamentadas na reprodução técnica (principalmente no texto A obra de arte na época de sua reprodutibilidade técnica) e, nesse sentido, avaliava que os conceitos que definiam a arte em um período anterior à modernidade teriam sido superados; ao mesmo tempo, argumenta a favor de um suposto potencial político dessa nova arte. INDÚSTRIA CULTURAL E IDEOLOGIA Contudo a mudança é aparente, pois os estilos sempre retornam com poucas modificações, para ape nas chamar a atenção do público, que, assim, se sente como estivesse, de fato, consumindo um produto novo. Na alta cultura, a busca per manente pelo novo também ocorre, mas com Depois de vinte e três anos entre a publi cação de Indústria Cultural e o texto póstumo 510 Humberto Alves Silva Junior Teoria Estética, Adorno modificava em parte suas impressões sobre a indústria cultural, reconhecendo o pioneirismo de Walter Ben jamin1 ao observar possíveis utilizações dos meios de comunicação de massa modernos a favor da emancipação humana. uma arte contemporânea compromete-se no plano teleológico, político e prático com a rea lidade social, mesmo que inserida na lógica da indústria cultural, admitindo que isso é pos sível através da conjunção dos meios técnicos avançados com as experiências estéticas tam bém avançadas, Adorno discute a resistência da arte em relação aos aspectos reificantes da indústria cultural: Toda obra, enquanto destinada a uma pluralidade, é já, segundo sua ideia, a sua reprodução. Que Benja min, na dicotomia da obra de arte aurática e da obra de arte tecnológica, reprimisse este momento de unidade em favor da diferença, que seria de fato a crítica dialética de sua teoria (Adorno, p. 59, 2008). No entanto, quando a arte autônoma absorveu seria mente os procedimentos técnicos industriais, estes permaneceram-lhe exteriores. A reprodutibilidade em massa de nenhum modo se tornou lei formal imanente, como a identificação com o agressor se compraz em afirmar. No próprio cinema, os momen tos industriais e estético-artesanais divergem sob pressão socioeconômica da industrialização radical da arte, a sua adaptação integral aos padrões técni cos alcançados colide com o que na arte se recusa à integração (Adorno, 2008, p. 327). Assim, Adorno observava que o moder nismo e o aspecto técnico de perfil industrial, na arte contemporânea, eram fenômenos reais, que não destruiriam a arte pelo fato de ela ter sido impelida pelos condicionamentos técni cos e econômicos da indústria cultural. É antes o postulado rimbaudiano mais progressista, no qual os procedimentos técnicos mais avançados e mais diferenciados se interpenetram com as ex periências mais avançadas e mais diferenciadas. Mas estas, enquanto sociais, são críticas. Esta arte moderna deve mostrar-se adulta à grande indústria, não a manipulando apenas. INDÚSTRIA CULTURAL E IDEOLOGIA Além disso, ele também reconhece que a mar ginalização da arte moderna radical seria um sintoma da reação do status quo da sociedade em relação às inovações propostas. da logicidade como elemento essencial na sua definição. Para Adorno, as obras modernas ten dem a ser abertas e, assim, dissociarem a forma do conteúdo. Mas mesmo os arroubos revolu cionários e inovadores dos mais variados tipos de modernismo, na tentativa de estabelecer uma arte inquieta, instável, desestabilizadora da percepção, não são capazes de excluir a ló gica do interior dessas novas formas. Ele cha ma a atenção para definição da forma a partir das ideias de simetria e repetição, mas que, mesmo quando se tenta destruir essa “harmo nia”, a lógica em si não desaparece, como ele defende na Teoria Estética: O caráter técnico da arte moderna está de acordo com o estilo de vida contemporâneo e os meios de produção. No caso da arte mo derna, segundo Adorno, os vestígios técnicos se destacam e aderem a toda a obra nova, como cicatrizes. O caráter técnico da arte moderna está de acordo com o estilo de vida contemporâneo e os meios de produção. No caso da arte mo derna, segundo Adorno, os vestígios técnicos se destacam e aderem a toda a obra nova, como cicatrizes. No entanto, essa mesma obra se opõe ao mundo por ela representado, exatamente por configurá-lo de modo distinto, corresponden do à interioridade dos homens como represen tação, e, por conseguinte, representação de um período. Portanto o conceito marxista de ideo logia é capaz de dar pistas para decifrar o cará ter concreto da relação entre a arte e a socieda de, que constitui o estatuto das representações sociais e das representações cinematográficas. As análises musicais, por exemplo, mostram que mesmo nas obras mais desorganizadas e mais opos tas à repetição existem analogias, que numerosas partes correspondem a outras em quaisquer carac terísticas e que apenas pela referência a elementos idênticos é que se realiza a não identidade procura da; sem nenhuma semelhança o caos permaneceria por seu turno uma invariante (Adorno, 2008, p. 216). RH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 Adorno também reconhecia que a arte está imbuída do desejo de se construir um mundo melhor. Entretanto ele era pessimis ta em relação às possibilidades utópicas das obras em relação à sociedade. INDÚSTRIA CULTURAL E IDEOLOGIA O seu próprio compor tamento e a sua linguagem formal devem reagir es pontaneamente à situação objetiva; a reação espon tânea, enquanto norma, circunscreve em um para doxo eterno da arte (Adorno, p. 59, 2008). Mesmo mantendo reservas em relação ao cinema, Adorno considerava como exemplo que não seria possível, mesmo na “industriali zada”, uma integração completa entre o aspec to técnico e o estético, apagando-se os rastros desse último em uma arte relacionada com a reprodutibilidade técnica. Pelo contrário, para o autor, há uma resistência do aspecto estético diante da técnica, como é caso do cinema, no qual há uma tensão entre os fatores “estéticos” e os “industriais”, a partir da pressão exercida pelos interesses econômicos e políticos dos grupos envolvidos na produção cinematográfi ca, apesar de reconhecer, na Teoria Estética, as possibilidades da arte em um novo contexto das condições sociais e econômicas de produção. Adorno flexibilizava seu posicionamen to no texto posterior, reconhecendo a interpe netração da técnica com as experiências esté ticas na arte contemporânea e a possibilidade de aproveitar ao máximo dessa conjunção para elaborar uma reação à própria indústria cul tural, com sua tendência homogeneizadora e manipuladora. Uma reação que se inscreve na tradição da arte: “reagir à situação objetiva”. Adorno compreendia as relações intrínsecas entre arte e a situação de seu tempo. Para ele, Hoje em dia, é já possível, na eletrônica, produzir artisticamente a partir da natureza específica de meios de origem extra-artística. O salto qualitativo é evidente entre a mão que desenha um animal na parede da caverna e a câmera, que permite o apare cimento simultâneo das imagens em inúmeros luga res (Adorno, 2008, p.59). Ca o CR Assim, o autor reconhece o caráter qua litativo diferenciado dessa nova arte com a reprodução das imagens em locais distintos. 511 INDÚSTRIA CULTURAL E IDEOLOGIA INDÚSTRIA CULTURAL E IDEOLOGIA Segundo o autor, Freud defendia que as obras não são satisfa ções imediatas do desejo, mas servem de meio para transformar a pulsão libidinal em pro dução social, apresentando, assim, o caráter acrítico da obra na sociedade, e, desse modo, concebendo a arte como aceitação conformis ta. Assim, a psicanálise, que concebe a obra de arte como um bem cultural agradável, acaba por excluir a negatividade da arte, desperdi çando os conflitos pulsionais que estão em sua gênese e que poderiam indicar o seu potencial de desvelação. Nesse sentido, Adorno conce bia que uma arte com essas características faz sumir a admiração para mergulhar o indivíduo na obra, não mantendo a devida distância, o desinteresse que permite estabelecer a diferen ça entre a arte e a sociedade, e, assim, manter o seu caráter de negatividade. Estendendo essa análise aos filmes pro duzidos por algumas vanguardas, como o Ci nema Soviético de 1920, é possível perceber que, por mais que se utilizem recursos para destruir a diegese do filme, sua forma mesma não é destruída, e o esforço em apagar os li mites tradicionais da representação fílmica se torna uma constante; ele, por si só, estabele ce uma constância, uma coerência, uma lógi ca. Para Adorno, a associação entre a obra e o real, que o artista e o crítico realizam, é uma obrigação que faz parte da legitimação social de qualquer arte. A intenção de desconstruir do setor crítico da arte moderna e massificada serve mais para cobrir do que para se destacar como uma crítica verdadeira. JAMESON E AS POSSIBILIDADE POLÍTICAS DA INDÚSTRIA CUL TURAL A partir do texto clássico de Adorno so bre a indústria cultural, o autor contemporâneo Frederic Jameson elaborou reflexões fundamen tais para a compreensão do fenômeno da cultura de massa. Crítico do primeiro texto de Adorno, Indústria cultural – o esclarecimento como mis tificação das massas, esse autor defende uma concepção mais pretensiosa para as produções artísticas de caráter industrial. Passa a nutrir es peranças em relação à arte contemporânea, de fendendo que mesmo as obras mais massificadas, inseridas por completo no sistema de produção vigente, podem obter um grau de criatividade e até mesmo um conteúdo político. A abordagem de Jameson relaciona es sas duas instâncias opostas, ideologia e utopia, como fundamentais para se compreender a complexidade da relação entre a produção ar tística na modernidade, em especial o cinema, e sua recepção pelo público. O autor ressalta também a relação entre, de um lado, cinema, interesses dominantes do capital, produção de ideologia no contexto cinematográfico, e, de outro, suas implicações sociais. Assim, Jame son percebia que a dimensão utópica da cons ciência é indissociável da dimensão ideológi ca, observando que há uma troca compensató ria, uma gratificação, através de um vislumbre positivo de um sentimento de comunidade em troca da passividade. Jameson destaca que, na indústria cultu ral, a arte política sobrevive concomitantemen te com a ideologia dominante, inserida na pró pria obra crítica. Ele afirma que o mesmo lugar de atuação da ideologia na produção artística é o lugar de sua crítica, defendendo que manipu lação e utopia estão imbricadas no cinema. De modo semelhante, Jameson ao se contrapor à ideia da arte apenas como manipulação, colo ca outro conceito freudiano, o de recalque. Ao analisar as obras da cultura de massa, aponta que elas exerceriam um poder de contenção em relação aos sentimentos negativos, como: “trau ma, memória culpada, desejo culpado ou inti midador, angústia” (1995, p.25), em que o dese jo recalcado é aplacado por um preenchimento simbólico, como afirma na seguinte passagem do livro O Inconsciente Político: Compreendendo desse modo essa jun ção, vê-se que a ideologia não é apenas coer ção, mas também sedução, e, por isso ela não é algo mecânico, que surge apenas como um epifenômeno da infraestrutura. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 Adorno, em Teoria Estética, defendia que a arte crítica (parte do modernismo) tem de se adequar às exigências do status quo, pois, se gundo esse autor, as obras verdadeiramente au tênticas são incompatíveis com a figuração que a própria sociedade tem de si mesma, a pon to de colocar em risco a sua autoconservação. Por isso cabe ao artista, sob essa perspectiva, ir Caderno CR Outra observação adorniana sobre a arte moderna, que é extremamente fecunda nas análises cinematográficas, é sobre a presença 512 Humberto Alves Silva Junior jetos espúrios, então um passo preliminar também deve ser teorizado em que esses mesmos impulsos – na matéria prima sobre qual age o processo – são inicialmente despertados dentro do próprio texto que busca silenciá-los (Jameson, 1988, p. 297). além. Ele defende que o artista precisa ir ao ex tremo em sua criação contando com o material que possui, no interior nas condições de exis tência em que se encontra, como a tecnologia. Jameson, referindo-se ao cinema, afirma que o recalque e a satisfação dos desejos corres pondem à unidade de um mesmo mecanismo no interior do aspecto imaginário da obra de arte. Mas não apenas isso: ele observa a dimen são utópica no interior do próprio recalque, da própria ideologia, pois esses mecanismos tam bém estão associados às esperanças e fantasias positivas da coletividade, que são expressas de forma independente das distorções ideológicas do cinema. INDÚSTRIA CULTURAL E IDEOLOGIA e do esmero estético que os caracteriza, eram veiculados nos mesmos canais de divulgação dos filmes comerciais. Adorno, antes de Jameson, concebia que a arte não é desprovida de ideologia e de verdade, como se fossem cão e gato. Por isso mesmo, reforça a ideologia do material estéti co da arte, pois ideologia e verdade aparecem indivisíveis, na totalidade administrada pela sociedade. Como elemento superestrutural, a in dústria cultural é dependente das condições de seu tempo, da economia e da história, e aponta para elementos da realidade social contempo rânea, tanto na técnica como no conteúdo. O conceito adorniano demonstra sua capacidade de suscitar novas discussões, tanto que o pró prio Adorno o retoma vinte anos depois, em uma abordagem mais ampla do significado da indústria cultural. Por conseguinte, é possível perceber que a representação fílmica se encontra no in terior da representação ideológica. No sentido social mais amplo, o cinema, como as outras instituições sociais, são veículos da ideologia dominante, mas, ao mesmo tempo, podem ser espaços da crítica à ideologia. A respeito da cultura de massa e da in dústria cultural, Jameson afirmava que esse binômio continua como referência principal de crítica por parte do cinema alternativo, no esforço de cineastas que não concordam com o modo industrial e a ideologia dessas produções comerciais, apesar do constante crescimento do entrecruzamento de aspectos dos dois tipos de cinema, tornando mais aguda a dificuldade em classificar um filme em uma das duas cate gorias. O que não impede que elas sejam abor dadas, discutidas e redefinidas ao se investigar a realidade da arte contemporânea, em espe cial o cinema e suas relações com a sociedade. É nesse sentido que assim afirma Jameson: Recebido para publicação em 10 de junho de 2019 Aceito em 18 de outubro de 2019 JAMESON E AS POSSIBILIDADE POLÍTICAS DA INDÚSTRIA CUL TURAL Para Jameson, analogamente, a consci ência de classe é ideológica e utópica, pois car rega em si a esperança, a alegria e o desejo de se viver em comunidade, ao mesmo tempo em que contém forte tendência à manipulação e à acomodação. Do mesmo modo poderia ser per cebida a recepção no cinema, pois ela é uma expressão alegórica de uma solidariedade que se encontra no imaginário coletivo. […] a função ideológica da cultura de massa é enten dida como um processo pelo qual impulsos de outra forma perigosos e protopolíticos são ‘administrados’ e desativados, racionalizados e se lhes oferecem ob 513 XAVIER, I. Sétima arte: um culto moderno. São Paulo: Perspectiva, 1978. Recebido para publicação em 10 de junho de 2019 Aceito em 18 de outubro de 2019 REFERÊNCIAS ADORNO, T.; HORKHEIMER, M. Dialética do esclarecimento. Rio de Janeiro: Jorge Zahar, 1985. ADORNO, T. W. Teoria estética. Lisboa: Edições 70, 2008. CHARNEY, L.; SCHWARTZ, V. O Cinema e a invenção da vida moderna. São Paulo: Cosac & Naif, 2001. BENJAMIM, W. Magia e técnica, arte e política. São Paulo: Brasiliense, 1994. HABERMAS, J. Mudança estrutural na esfera pública. Rio de Janeiro: Tempo Brasileiro, 2003. HARVEY, D. Condição pós-moderna. São Paulo: Loyola, 1999. JAMESON, F. O inconsciente político. São Paulo: Ática, 1992. Humberto Alves Silva Junior Humberto Alves Silva Junior De l’analyse de contenu, analyse les discussions sur le concept d’industrie culturelle inventé et analysé par Theodor Adorno et Max Horkheimer et son déploiement dans le livre La théorie esthétique d’Adorno, en particulier le cinéma. L’industrie culturelle ou culture de masse comprend le caractère commercial et le mode de production industriel des productions culturelles du capitalisme, même traitées comme des marchandises et leurs conséquences pour le public, agissant principalement comme un instrument de manipulation idéologique selon la vision adornienne. Cependant, dans la théorie de l’esthétique, Adorno fait avancer la discussion et admet que, malgré la présence d’une idéologie, l’industrie culturelle pourrait également développer un espace alternatif aux productions de masse. Plus tard, Frederic Jameson, inspiré par le travail d’Adorno, tire un trait similaire dans lequel il réalise que les produits de l’industrie culturelle ne seraient pas seulement idéologiques. En plus d’Adorno, il a affirmé qu’ils pourraient aussi être utopiques promesses collectives et individuelles d’un avenir meilleur. This article, based on content analysis, analyzes the discussions about the concept of cultural industry coined and analyzed by Theodor Adorno and Max Horkheimer and its unfolding in the book Adorno’s Aesthetic Theory, especially cinema. The cultural industry or mass culture comprises the commercial character and industrial mode of production of cultural productions in capitalism, treated even as commodities and their consequences to the public, acting mainly as an instrument of ideological manipulation in the adornian view. However, in Aesthetic Theory, Adorno advances the discussion and admits that despite the presence of ideology, the cultural industry could also develop an alternative space to mass productions. Later Frederic Jameson, inspired by Adorno’s work, draws a similar line in which he realizes that the products of the cultural industry would not only be ideological, and in addition to Adorno, he asserted that they could also be utopian, as mass culture attracts the public with collective and individual promises of a better future. Mots-clés: Idéologie. Émancipation. Cinéma. Culture de masse. École de Frankfurt. Mots-clés: Idéologie. Émancipation. Cinéma. Culture de masse. École de Frankfurt. Mots-clés: Idéologie. Émancipation. Cinéma. Culture de masse. École de Frankfurt. Keywords: Ideology. Emancipation. Movie theater. Mass culture. Frankfurt School. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 Humberto Alves Silva Junior – Doutor em Ciências Sociais pela UFBA. Professor do Departamento de Ciências Sociais da Universidade Federal de Rondônia. Keywords: Ideology. Emancipation. Movie theater. Mass culture. Frankfurt School. Caderno CRH, Salvador, v. 32, n. 87, p. 505-515, Set./Dez. 2019 t t c q p g r C m 5-515, Set./Dez. 2019 Tal aproximação exige que se leia a alta cultura e a cultura de massa como fenômenos objetivamen te relacionados e dialeticamente interdependentes, como formas gêmeas e inseparáveis da fissão da produção estética sob o capitalismo (Jameson, 1995, p. 25). JAMESON, F. As marcas do visível. Rio de Janeiro: Graal, 1995. JAMESON, F. Pós-modernismo – a lógica cultural do capitalismo tardio. Ática, 1996. MARX, K.; ENGELS, F. O manifesto do partido comunista. Petrópolis: Vozes, 1986. PIZZINI, J. Cinemais. Rio de Janeiro: Aeroplano, jan./fev. 2003. p.10-13. n. 33, Caderno CRH, Salvador, v. 32, n. 87, p. 5 A aproximação se estabelece na própria tentativa de se classificar uma obra como per tencente a um dos lados da oposição entre alta cultura e cultura de massa, pois existem obras que possuem tanto as características de um produto comercial como aos traços que confi guram uma produção “artística” da alta cultu ra. Um exemplo disso são os filmes de Charles Chaplin, que, apesar do tom denunciatório das mazelas, seja do capitalismo, seja do fascismo, SINGER, B. Modernidade, hiperestímulo e o início do sensacionalismo. In: CHARNEY, L.; SCHWARTZ, V. O cinema e a invenção da vida moderna. São Paulo: Cosac & Naif, 2001. XAVIER, I. Sétima arte: um culto moderno. São Paulo: Perspectiva, 1978. 514 Humberto Alves Silva Junior Mots-clés: Idéologie. Émancipation. Cinéma. Culture de masse. École de Frankfurt. Humberto Alves Silva Junior Integra o Núcleo de Pesquisa em Sociologia da Arte (NUCLEART/UFBA) Grupo de Pesquisa em Sociologia da Arte (SOAR/UNIR), desenvolvendo pesquisas na área de Sociologia do Cinema. Humberto Alves Silva Junior – Doutor em Ciências Sociais pela UFBA. Professor do Departamento de Ciências Sociais da Universidade Federal de Rondônia. Integra o Núcleo de Pesquisa em Sociologia da Arte (NUCLEART/UFBA) Grupo de Pesquisa em Sociologia da Arte (SOAR/UNIR), desenvolvendo pesquisas na área de Sociologia do Cinema. 515
https://openalex.org/W2574160333	https://revista.univap.br/index.php/revistaunivap/article/download/853/603	Portuguese	null	MODELAGEM DA COPA DE ÁRVORES DE EUCALIPTO EMPREGANDO REDES NEURAIS ARTIFICIAIS	Revista UniVap	2,017	cc-by	471	243 243 Revista Univap – revista.univap.br São José dos Campos SP Brasil v 22 n 40 Edição Especial 2016 ISSN 2237 1753 1 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. Email: rosalia.trindade22@gmail.com. 2 Engenharia Florestal/Instituto Federal de Educação, Ciência e Tecnologia de Minas Gerais, Brasil. Email: vinicius.aguiar.agr@gmail.com. 3 Engenharia Florestal/Instituto Federal de Educação, Ciência e Tecnologia de Minas Gerais, Brasil. Email: eduardohma1996@gmail.com. 4 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. Email: penidotma@gmail.com. 5 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: victorduarte.vdv@gmail.com. 6 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: bruno.lafeta@ifmg.edu.br. 7 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: marcio.takeshi@ifmg.edu.br. Revista Univap – revista.univap.br São José dos Campos-SP-Brasil, v. 22, n. 40, Edição Especial 2016. ISSN 2237-1753 1 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. Email: rosalia.trindade22@gmail.com. 2 Rosália Nazareth Rosa Trindade1 Vinícius Faúla Aguiar2 Eduardo Henrique Mendes Alves3 Tamires Mousslech Andrade Penido4 Victor Duarte Vieira5 Bruno Oliveira Lafetá6 Márcio Takeshi Sugawara7 Resumo: O objetivo foi avaliar a eficiência da estimação simultânea de índices morfométricos da copa de árvores de eucalipto em um plantio adensado empregando RNA. Foram estabelecidas três parcelas experimentais em uma plantação de eucalipto localizada no IFMG – campus de São João Evangelista. Realizou-se o inventário florestal aos 37 meses de idade. Calculou-se para cada árvore três índices morfométricos de copa. Foram treinadas 400 RNA para estimar simultaneamente os índices, sendo 200 Multilayer Perceptron (MLP) e 200 Radial Basis Function (RBF). Uma desvantagem observada foi a perda na precisão das estimativas de formal da copa, fato que pode ser atribuído a maior variabilidade dos valores observados (CV de 48 %) e à sua tendência linear constante. Conclui-se que a modelagem por RNA com arquiteturas MLP e RBF demonstraram adequabilidade e precisão para estimar os índices morfométricos proporção da copa e índice de abrangência, respectivamente. Palavras-chave: Formal da copa; Índice de abrangência; Inteligência artificial; Proporção da copa. Revista Univap – revista.univap.br 1 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. Email: rosalia.trindade22@gmail.com. 2 Engenharia Florestal/Instituto Federal de Educação, Ciência e Tecnologia de Minas Gerais, Brasil. Email: vinicius.aguiar.agr@gmail.com. 3 Engenharia Florestal/Instituto Federal de Educação, Ciência e Tecnologia de Minas Gerais, Brasil. Email: eduardohma1996@gmail.com. 4 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. Email: penidotma@gmail.com. 5 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: victorduarte.vdv@gmail.com. 6 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: bruno.lafeta@ifmg.edu.br. 7 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. E- mail: marcio.takeshi@ifmg.edu.br. 7 Departamento de Engenharia Florestal/Universidade Federal dos Vales do Jequitinhonha e Mucuri, Brasil. mail: marcio.takeshi@ifmg.edu.br.
https://openalex.org/W2963085086	https://pulmonarychronicles.com/index.php/pulmonarychronicles/article/download/562/1240	English	null	Acquired cystic fibrosis transmembrane conductance regulator protein (CFTR) dysfunction and cigarette smoking	The Southwest respiratory and critical care chronicles	2,019	cc-by-sa	1,947	Cystic fibrosis background which alters the ionic composition of extracellular fluid and reduces water secretion leading to thickened mucus on airway surfaces.1 According to the “low volume” hypothesis, the reduced airways surface vol ume decreases the mucociliary clearance function to eliminate inhaled bacteria on lung epithelial surfaces.1 The increased bacterial load and increased inflamma tory response of the lung damages lung tissue and thereby reducing respiratory function.1 Furthermore, CF causes meconium ileus, fat soluble vitamin defi ciencies, and male infertility.1 Despite diagnostic and therapeutic advances, CF patients have a median sur vival of 50 years.1 Cystic fibrosis (CF) remains a prevalent genetically inherited disease in Caucasian populations, affecting 1 in 2500 newborns.1 It is an autosomal recessive dis ease caused by mutations in the CF transmembrane conductance regulator gene (CFTR).1 Specifically, deletion of phenylalanine at codon 508, known as ΔF508, occurs in 70% of CF patients; hundreds of rarer CFTR mutations also occur and produce less severe clinical presentations.1 The CFTR protein is expressed throughout the body and regulates fluid volume on epi thelial surfaces by modulating chloride secretion and sodium absorption.1 In CF, CFTR mutations prevent the secretion of chloride ions on epithelial surfaces, In the clinic In the clinic Abstract Cystic fibrosis (CF) remains a prevalent genetically inherited disease in Caucasian populations. Investigation of the respiratory symptoms which occur in patients with CF helps us understand the pathophysiology of chronic lung disease. Environmental insults, such as cigarette smoke, can reduce the cystic fibrosis transmembrane receptor (CFTR) function or expression leading to an acquired CF phenotype and could contribute to the development and progression of smoking-related lung disease. However, it is uncertain if the acquired CF phenotype can be diagnosed with the same methods, such as the sweat chloride test and the measurement of nasal potential difference, used for genetically-acquired CF. More studies are needed to investigate the prevalence of acquired CFTR dysfunction and the differences between acquired and genetically-inherited CFTR dysfunction. Overall, acquired CFTR dysfunction challenges the distinction between genetic and acquired disorders, suggesting that environmental agents may modulate the functions of genes and the increase risk for pulmonary disease. Keywords: Cystic fibrosis, acquired cystic fibrosis, smoking, mucociliary clearance Acquired cystic fibrosis transmembrane conductance regulator protein (CFTR) dysfunction and cigarette smoking Jonathan Kopel BS Cigarette smoking and acquired cystic fibrosis Corresponding author: Jonathan Kopel Contact Information: Jonathan.kopel@ttuhsc.edu DOI: 10.12746/swrccc.v7i30.562 Chronic exposure to cigarette smoke damages the upper and lower airways and reduces mucus clear ance, which is strongly associated with COPD.2–4 COPD includes several clinical manifestations resulting Corresponding author: Jonathan Kopel Contact Information: Jonathan.kopel@ttuhsc.edu DOI: 10.12746/swrccc.v7i30.562 43 The Southwest Respiratory and Critical Care Chronicles 2019;7(30):43–46 Acquired Cystic Fibrosis Transmembrane Conductance Regulator Protein (CFTR) Dysfunction and Cigarette Smokin Kopel CFTR function and expression were observed with e-cigarette vapors, which contain acrolein and other toxins.15 Pathological studies of upper and lower air way epithelial cells demonstrate that cigarette smoke exposure decreases ciliary beating frequency through reduced CFTR and calcium activated chloride chan nel, which directly inhibits mucociliary clearance.2–4,14 Furthermore, studies measuring nasal potential differ ence and lower airway potential difference and stud ies with endobronchial biopsies from subjects without CFTR genetic mutations showed significant decreases in CFTR protein function and expression with acute and chronic exposure of cigarette smoke.2–4,16–18 Chemical components in cigarette smoke also enter systemic circulation and can produce pancreatitis, infertility, and cachexia.2–4 Overall, these studies sug gest cigarette smoke oxidants reduce CFTR mRNA transcript expression, accelerate CFTR protein deg radation, and alter CFTR channel gating leading to acquired CFTR dysfunction.4 from large-airway inflammation and remodeling with increased airway resistance and from alveolar wall destruction with decreased alveolar lung surface area available for gas exchange.5 Studies in patients with tobacco smoke related COPD (confirmed homozygote wild-type CFTR genotype) have demonstrated the pres ence of acquired CFTR dysfunction in some COPD patients and that this persists despite smoking cessa tion and is associated with chronic bronchitis sever ity.6 Environmental insults, such as cigarette smoke, can reduce CFTR function or expression leading to an acquired CFTR dysfunction, possibly contributing to the development and progression of COPD.2–4 The population distribution of sweat chloride lev els is unknown and may vary by country and ethnic ity. Studies have focused on patients with respiratory symptoms or disease. Cigarette smoking and acquired cystic fibrosis An epidemiology study at a Brazilian referral CF center examining 5,721 patients (0–85.58 years old) suspected of cystic fibrosis, over a 30 year period, reported that 61.6% had a sweat chloride level less than 30 mmol/L chloride, 25.1% had a sweat chloride level between 30 and 60 mmol/L chloride, and 13.3% had a sweat chloride level greater than 60 mmol/L chloride.7 Another study examining 182 patients (0.2–23 years old) at a Northern India hospital with suspected cystic fibrosis with one or more clinical features of cystic fibrosis found 5% had borderline sweat chloride levels (sweat chloride levels 40-59 mmol/L chloride) and 22.5% had ele vated sweat chloride levels (>60 mmol/L chloride).8 Among current and former smokers, 20–31% have elevated sweat chloride levels (≥40 mmol/L chloride), which indicates a reduced function and expression of CFTR.9,10 Similarly, a clinical report examining former or current smokers with COPD (excluding those with CFTR gene mutation) showed 40% of had an ele vated sweat chloride test.10,11 However, it is uncertain whether acquired CFTR dysfunction can be diagnosed with the same meth ods, such as quantitative sweat chloride concentra tions and measurement of nasal potential differences, used for genetically inherited cystic fibrosis.19 It remains unknown whether acquired CFTR dysfunc tion in COPD patients increases the severity of cough, sputum production, and deterioration in lung function beyond the usual effects of cigarette smoking alone. Cigarette smoking and acquired cystic fibrosis Since a significant proportion of COPD patients have been shown to have CFTR dysfunction, this suggests the CFTR protein dysfunction might be an important contributor to the pathogenesis of COPD and other pulmonary diseases.6 More studies are needed to investigate clinical or phenotypic expressive differ ences between a patient population with acquired CFTR dysfunction and one with genetically inherited CF.19 For example, smokers with acquired CFTR dysfunction do not typically present with pancreatic enzyme deficiencies or obstruction of the vas def erens seen in those with genetically inherited CF.19 Currently, N-acetylcysteine, ivacaftor, and roflumilast can reverse decreases in CFTR function and expres sion.12,14,20–22 However, ivacaftor is effective in CF patients with a less common CFTR gene mutation, G551D, and requires more clinical trials to determine Among the various components in cigarette smoke, acrolein and cadmium produce the greatest decrease in CFTR function and expression along the respiratory epithelium.2–4 Specifically, acrolein and cadmium produce CFTR dysfunction through reduc ing CFTR expression and function, destabilizing CFTR proteins, and chemically modifying cysteine and lysine residues.2–4,12–14 Similar alterations in The Southwest Respiratory and Critical Care Chronicles 2019;7(30):43–46 44 cquired Cystic Fibrosis Transmembrane Conductance Regulator Protein (CFTR) Dysfunction and Cigarette Smokin Kopel Kopel 8. Raina MA. Assessment of correlation between sweat chlo ride levels and clinical features of cystic fibrosis patients. J Clin Diag Res 2016. whether it will be effective against tobacco smoke induced acquired CFTR dysfunction.19 Despite limited therapeutic options, a clinical report examining smok ing cessation showed improved CFTR-dependent sweat secretion rate over a 21-day period.23 Overall, acquired CFTR dysfunction challenges the distinction between genetic and acquired disorders and sug gests that environmental agents may modulate the function of genes and the risk for disease.19 9. Griesenbach U. Faculty of 1000 evaluation for cigarette smoke induces systemic defects in cystic fibrosis transmem brane conductance regulator function. In. Post-Publication Peer Rev of the Biomed Lit: F1000 ( Faculty of 1000 Ltd); 2015. 10. Raju SV, Jackson PL, Courville CA, et al. Cigarette smoke induces systemic defects in cystic fibrosis transmembrane conductance regulator function. Ameri J Resp Crit Care Med 2013;188(11):1321–1330. Article citation: Kopel J. Acquired cystic fibrosis transmembrane conductance regulator protein (CFTR) dysfunction and cigarette smoking. The Southwest Respiratory and Critical Care Chronicles 2019;7(30): 43–46 11. Courville CA, Tidwell S, Liu B, et al. Acquired defects in CFTR-dependent ββ-adrenergic sweat secretion in chronic obstructive pulmonary disease. Resp Res 2014;15(1):25. 12. Cigarette smoking and acquired cystic fibrosis Raju SV, Lin VY, Liu L, et al. The cystic fibrosis transmem brane conductance regulator potentiator ivacaftor augments mucociliary clearance abrogating cystic fibrosis transmem brane conductance tegulator inhibition by cigarette smoke. Am J Respir Cell Mol Biol 2017;56(1):99–108. From: The School of Medicine, Texas Tech University Health Sciences Center, Lubbock, Texas Submitted: 3/9/2019 Accepted: 6/19/2019 Reviewer: Adaobi Kanu MD Conflicts of interest: none This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. From: The School of Medicine, Texas Tech University Health Sciences Center, Lubbock, Texas Submitted: 3/9/2019 Accepted: 6/19/2019 Reviewer: Adaobi Kanu MD Conflicts of interest: none This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. 13. Cantin AM, Hanrahan JW, Bilodeau G, et al. Cystic fibro sis transmembrane conductance regulator function is sup pressed in cigarette smokers. Am J Respir Crit Care Med 2006;173(10):1139–1144. 14. Sloane PA, Shastry S, Wilhelm A, et al. A pharmacologic approach to acquired cystic fibrosis transmembrane con ductance regulator dysfunction in smoking related lung dis ease. PloS One 2012;7(6):e39809–e39809. 22. Lambert JA, Raju SV, Tang LP, et al. Cystic fibrosis trans membrane conductance regulator activation by roflumilast contributes to therapeutic benefit in chronic bronchitis. Amer J Resp Cell Molec Biol 2014;50(3):549–558. 23. Courville CA, Raju SV, Liu B, et al. Recovery of acquired cystic fibrosis transmembrane conductance regulator dys function after smoking cessation. Amer J Resp Crit Care Med 2015;192(12):1521–1524. The Southwest Respiratory and Critical Care Chronicles 2019;7(30):43–46 The Southwest Respiratory and Critical Care Chronicles 2019;7(30):43–46 References 15. Lin VY, Fain MD, Jackson PL, et al. Vaporized e-cigarette liquids induce ion transport dysfunction in airway epithelia. Am J Respir Cell Mol Biol 2018. 1. Bush A, Davies JC. Cystic fibrosis. Resp Med 2013;355: 315–326. 16. Savitski AN, Mesaros C, Blair IA, et al. Secondhand smoke inhibits both Cl− and K+ conductances in normal human bronchial epithelial cells. Resp Res 2009;10(1):120–120. 2. Banks C, Freeman L, Cho DY, et al. Acquired cystic fibrosis transmembrane conductance regulator dysfunction. World J Otorhinolaryngol Head Neck Surg 2018;4(3):193–199. 17. Kreindler JL, Jackson AD, Kemp PA, et al. Inhibition of chloride secretion in human bronchial epithelial cells by cig arette smoke extract. Amer J Phys 2005;288(5):L894–L902. 3. Cho DY, Woodworth BA. Acquired cystic fibrosis transmem brane conductance regulator deficiency. Adv Otorhinolaryn gol 2016;79:78–85. 18. Clunes LA, Davies CM, Coakley RD, et al. Cigarette smoke exposure induces CFTR internalization and insolubility, leading to airway surface liquid dehydration. FASEB J 2012; 26(2):533–545. 4. Shi J, Li H, Yuan C, et al. Cigarette smoke-induced acquired dysfunction of cystic fibrosis transmembrane conductance regulator in the pathogenesis of chronic obstructive pulmo nary disease. Oxid Med Cell Longev 2018;2018:1–13. 19. Crystal RG. Are the smoking-induced diseases an acquired form of cystic fibrosis? Am J Respir Crit Care Med 2013;188(11): 1277–1278. 5. Tuder RM, Petrache I. Pathogenesis of chronic obstructive pulmonary disease. J Clin Inves 2012;122(8):2749–2755.i 6. Raju SV, Solomon GM, Dransfield MT, et al. Acquired cystic fibrosis transmembrane conductance regulator dysfunction in chronic bronchitis and other diseases of mucus clearance. Clin Chest Med 2016;37(1):147–158. 20. Raju SV, Rasmussen L, Sloane PA, et al. Roflumilast reverses CFTR-mediated ion transport dysfunction in ciga rette smoke-exposed mice. Resp Res 2017;18(1):173–173. 21. Solomon GM, Fu L, Rowe SM, et al. The therapeutic poten tial of CFTR modulators for COPD and other airway dis eases. Curr Opini in Pharm 2017;34:132–139. 7. Faria AG, Marson FAL, Gomez CCS, et al. Thirty years of sweat chloride testing at one referral center. Frontiers in Ped 2017;5. 45 The Southwest Respiratory and Critical Care Chronicles 2019;7(30):43–46 Acquired Cystic Fibrosis Transmembrane Conductance Regulator Protein (CFTR) Dysfunction and Cigarette Smokin Kopel 46
https://openalex.org/W2885183257	https://europepmc.org/articles/pmc6089339?pdf=render	English	null	Visualizing BDNF Transcript Usage During Sound-Induced Memory Linked Plasticity	Frontiers in molecular neuroscience	2,018	cc-by	19,515	ORIGINAL RESEARCH published: 31 July 2018 doi: 10.3389/fnmol.2018.00260 Visualizing BDNF Transcript Usage During Sound-Induced Memory Linked Plasticity Lucas Matt 1†, Philipp Eckert 2†, Rama Panford-Walsh 2, Hyun-Soon Geisler 2, Anne E. Bausch 1, Marie Manthey 2, Nicolas I. C. Müller 3, Csaba Harasztosi 4, Karin Rohbock 2, Peter Ruth 1, Eckhard Friauf 3, Thomas Ott 5, Ulrike Zimmermann 2, Lukas Rüttiger 2, Thomas Schimmang 6, Marlies Knipper 2* and Wibke Singer 2 1 Department of Pharmacology, Institute of Pharmacy, Toxicology, and Clinical Pharmacy, University of Tübingen, Tübingen, Germany, 2 Department of Otolaryngology, Tübingen Hearing Research Centre, Molecular Physiology of Hearing, University of Tübingen, Tübingen, Germany, 3 Animal Physiology Group, Department of Biology, University of Kaiserslautern, Kaiserslautern, Germany, 4 Section of Physiological Acoustics and Communication, Department of Otolaryngology, Tübingen Hearing Research Center, University of Tübingen, Tübingen, Germany, 5 Transgenic Facility Tübingen, University of Tübingen, Tübingen, Germany, 6 Instituto de Biologíay Genética Molecular, Universidad de Valladolid, Consejo Superior de Investigaciones Cientíﬁcas (CSIC), Valladolid, Spain Activity-dependent BDNF (brain-derived neurotrophic factor) expression is hypothesized to be a cue for the context-speciﬁcity of memory formation. So far, activity-dependent BDNF cannot be explicitly monitored independently of basal BDNF levels. We used the BLEV (BDNF-live-exon-visualization) reporter mouse to speciﬁcally detect activity-dependent usage of Bdnf exon-IV and -VI promoters through bi-cistronic co-expression of CFP and YFP, respectively. Enriching acoustic stimuli led to improved peripheral and central auditory brainstem responses, increased Schaffer collateral LTP, and enhanced performance in the Morris water maze. Within the brainstem, neuronal activity was increased and accompanied by a trend for higher expression levels of Bdnf exon-IV-CFP and exon-VI-YFP transcripts. In the hippocampus BDNF transcripts were clearly increased parallel to changes in parvalbumin expression and were localized to speciﬁc neurons and capillaries. Severe acoustic trauma, in contrast, elevated neither Bdnf transcript levels, nor auditory responses, parvalbumin or LTP. Together, this suggests that critical sensory input is essential for recruitment of activity-dependent auditory-speciﬁc BDNF expression that may shape network adaptation. Keywords: Bdnf exon-IV, Bdnf exon-VI, LTP, memory acquisition, feed-forward inhibition, parvalbumin, vasculature, sound-accentuation Edited by: Isabel Varela-Nieto, Consejo Superior de Investigaciones Cientíﬁcas (CSIC), Spain Reviewed by: José María Frade, Instituto Cajal (IC), Spain Anthony John Hannan, Florey Institute of Neuroscience and Mental Health, Australia Reviewed by: José María Frade, Instituto Cajal (IC), Spain Anthony John Hannan, Florey Institute of Neuroscience and Mental Health, Australia *Correspondence: Marlies Knipper marlies.knipper@uni-tuebingen.de †These authors have contributed equally to this work. †These authors have contributed equally to this work. Received: 18 April 2018 Accepted: 12 July 2018 Published: 31 July 2018 Received: 18 April 2018 Accepted: 12 July 2018 INTRODUCTION Matt L, Eckert P, Panford-Walsh R, Geisler H-S, Bausch AE, Manthey M, Müller NIC, Harasztosi C, Rohbock K, Ruth P, Friauf E, Ott T, Zimmermann U, Rüttiger L, Schimmang T, Knipper M and Singer W (2018) Visualizing BDNF Transcript Usage During Sound-Induced Memory Linked Plasticity. Brain-derived neurotrophic factor (BDNF), identiﬁed in 1982 (Barde et al., 1982), is recognized as key modulator of synaptic plasticity during homeostatic readjustment processes and a master regulator of energy homeostasis (for review see: Bramham and Messaoudi, 2005; Rauskolb et al., 2010; Park and Poo, 2013; Marosi and Mattson, 2014; Nahmani and Turrigiano, 2014; Jeanneteau and Arango-Lievano, 2016; Mitre et al., 2017). BDNF is well-known for its involvement in Schaﬀer collateral long-term potentiation (LTP) (Minichiello, 2009) as well as in inhibition (Huang and Reichardt, 2001; Wardle and Poo, 2003; Lu et al., 2009; Waterhouse and Xu, 2009; Duguid et al., 2012; Park and Poo, 2013; Parkhurst et al., 2013; Nahmani and Turrigiano, 2014). Nevertheless, July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org Visualizing BDNF During Sound-Induced Plasticity Matt et al. we do not yet understand BDNF’s inﬂuence on circuit stabilization in the adult system or its function in platelets (Chacón-Fernández et al., 2016), capillary endothelial cells (Donovan et al., 2000), microglia, and astrocytes (Ferrini and De Koninck, 2013; Parkhurst et al., 2013) during homeostatic readjustment processes (Nahmani and Turrigiano, 2014). Among other reasons, this is due to the diﬃculty of detecting the very low (Dieni et al., 2012) endogenous expression of BDNF in the mature central nervous system (CNS) (Dieni et al., 2012). Furthermore, many BDNF mutant mouse lines present severe phenotypes that complicate analysis of BDNF under normal physiological conditions (Ernfors et al., 1994; Rios et al., 2001; Postigo et al., 2002; Chourbaji et al., 2008; Hong et al., 2008; Sakata et al., 2009; Rauskolb et al., 2010; Lyons and West, 2011; Zuccotti et al., 2012; Vanevski and Xu, 2013; Mallei et al., 2015; Hill et al., 2016; Maynard et al., 2016). co-expression of cyan- and yellow-ﬂuorescent-protein (CFP and YFP), respectively (Singer et al., submitted). BLEV reporter mice are viable without any BDNF-related mutant phenotype (Singer et al., submitted). Importantly, they allow the detection of Bdnf exon-IV-CFP and Bdnf exon-VI-YFP at sites of BDNF protein expression in neurons, microglia, astrocytes, and capillaries (Singer et al., submitted), cell types previously shown to express BDNF (Edelmann et al., 2014; Serra-Millàs, 2016). INTRODUCTION Furthermore, in BLEV mice it is possible to observe activity-dependent hippocampal BDNF expression in response to glutamate receptor activation through either CFP and YFP ﬂuorescence or quantiﬁcation of protein tags by Western blot (Singer et al., submitted). To test if BLEV mice also allow monitoring of activity-dependent BDNF during behaviorally relevant long-term adaptive processes in response to external cues, we monitored Bdnf -transcript changes in BLEV reporter mice following diﬀerent sound exposure conditions. Previous experiments demonstrated that exposures to deﬁned enriching, mild traumatic, or severe traumatic sound pressure levels (SPL) caused long-lasting alterations to sound-sensitivity and diﬀerentially induced hippocampal plasticity (Singer et al., 2013). Likewise, acoustically induced diﬀerences in central sound sensitivity correlated with sensitivity to stress as shown through a social stress paradigm or pharmacological inhibition of stress receptors (Singer et al., 2018). y ) Part of the multifaceted functions of BDNF might be obfuscated by the complex structure of the Bdnf gene, which is comprised of eight independently transcribed non-coding exons (I-VIII), each of which is spliced to a common protein encoding exon (IX) (Timmusk et al., 1993; Aid et al., 2007), resulting in multiple transcripts that display diﬀerent stability, targeting, and translatability (Vaghi et al., 2014). BDNF expression from each of these eight diﬀerent promoters is independently regulated (Vaghi et al., 2014). Of particular interest are exon-IV and exon- VI, both containing promoters directly or indirectly regulated by neuronal activity (Hong et al., 2008; West et al., 2014; Tuvikene et al., 2016). Dysfunction of these two Bdnf transcripts is associated with deﬁcits in sleep, fear, and memory (Hill et al., 2016), as well as depression (Sakata et al., 2010), cortical inhibition deﬁciency (Hong et al., 2008), and cognitive decline (Vaghi et al., 2014; Mallei et al., 2015). Moreover, downregulation of activity-dependent BDNF was observed in stress-related neuropsychiatric disorders (Pariante, 2009; Castrén and Rantamäki, 2010) or in states of increased glucocorticoid resistance, for example during chronic stress (Bath et al., 2013; Gray et al., 2013; Jeanneteau and Arango-Lievano, 2016). Any stress reaction that may result in behavioral changes related to external cues requires an activity-dependent signal to provide context speciﬁcity to the otherwise ubiquitous glucocorticoid receptor (GR)-mediated stress response (de Kloet, 2014). For example, metabolic support for behaviorally relevant motor learning can only be provided through activity- dependent reallocation of GR eﬀects (Liston et al., 2013; Arango- Lievano et al., 2015; Jeanneteau and Arango-Lievano, 2016). INTRODUCTION In this context, BDNF was previously hypothesized to provide the corresponding signal (Jeanneteau and Arango-Lievano, 2016). In the present study we observed persistent sound sensitivity changes after these sound exposure paradigms in BLEV mice. In the auditory brainstem, these changes were accompanied by increased expression of VGLUT1, an established marker of excitation in auditory ﬁbers, and a tendency to upregulate Bdnf exon-IV and exon-VI. These observations were speciﬁc for the auditory system and could not be detected in the olfactory bulb. Furthermore, diﬀerent exposure levels correlated with altered Bdnf exon-IV and exon-VI expression in speciﬁc hippocampal neurons and vascular cells. In the hippocampus, this was paralleled by altered levels of GluA2 and parvalbumin expression and associated with a changed balance between excitatory vs. inhibitory inputs to CA1 pyramidal cells, underlined by altered Schaﬀer collateral LTP and spatial memory performance. These ﬁndings can only be explained by a critical sensory input that drives activity-dependent BDNF expression toward long-term adaptive responses. Animals Animal care and use and experimental protocols correspond to national and institutional guidelines and were reviewed and approved by the animal welfare commissioner and the regional board for animal experimentation. All experiments were performed according to the European Union Directive 2010/63/EU for the protection of animals used for experimental and other scientiﬁc purposes. Mice were kept according to national guidelines for animal care in an SPF animal facility at 25◦C on a 12/12 h light/dark cycle with average noise levels of around 50–60 dB. So far, however, it was technically impossible to detect increased activity-dependent BDNF expression above the background of basal BDNF levels. We have generated BDNF- live-exon-visualization (BLEV) knock-in reporter mice to speciﬁcally detect BDNF in response to Bdnf exon-IV and -VI promoter usage (Singer et al., submitted). The generation and validation of this new reporter mouse line is described in detail in Singer et al. (submitted). In BLEV mice, Bdnf exon-IV and -VI mRNA translation sites are tagged by bi-cistronic July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 2 Visualizing BDNF During Sound-Induced Plasticity Matt et al. Immunohistochemistry and Ribbon Counting Cochleae were isolated, ﬁxed, cryosectioned, and stained as described (Tan et al., 2007; Singer et al., 2016). Image acquisition and CtBP2/RIBEYE-immunopositive spot counting were carried out as previously described (Zuccotti et al., 2012; Singer et al., 2016). Vector Construct for a Transgenic BDNF Mouse overnight at 4◦C with the primary antibody, see Supplementary Table 1. Membranes for one approach were incubated with several primary antibodies at the same time, see originals in Supplementary Figures 2, 3. On the second day, the membrane was washed three times with Tris buﬀer/0.1% Tween 20; the secondary antibody (see Supplementary Table 1) was incubated for 1 h at room temperature in a sealed envelope. The membrane was washed again three times with Tris buﬀer/0.1% Tween 20. Proteins were visualized with ECL Prime WB Detection Reagent (GE Healthcare) using the Proxima 2700 (Isogen Life Science). For protein sizes, see Supplementary Table 1. For a detailed description of the generation of the new mouse model please see Singer et al., (submitted). In brief, the Bdnf exon-IV and -VI sequence, both including the corresponding promoter sequences, were extended by CFP or YFP, respectively, both containing a stop codon. A HA-tag was added to Bdnf exon- IV-CFP and a cMyc-tag to Bdnf exon-VI-YFP. The translation of the protein-coding Bdnf exon-IX is enabled by an IRES sequence, which keeps the mRNA at the ribosome, despite the presence of a stop codon. Additionally, the growth-associated protein 43 (GAP43), is added to anchor the ﬂuorescent proteins at the site of translation. This allows diﬀerential monitoring of the non-coding Bdnf exon-IV and Bdnf exon-VI by the ﬂuorescent proteins CFP and YFP without interfering with Bdnf exon-IX. Field Excitatory Postsynaptic Potential (fEPSP) Recordings in Hippocampal Slices Extracellular fEPSP recordings were performed according to standard methods as previously described (Matt et al., 2011; Chenaux et al., 2016). Field Excitatory Postsynaptic Potential (fEPSP) Recordings in Hippocampal Slices Extracellular fEPSP recordings were performed according to standard methods as previously described (Matt et al., 2011; Chenaux et al., 2016). Field Excitatory Postsynaptic Potential (fEPSP) Recordings in Hippocampal Slices Extracellular fEPSP recordings were performed according to standard methods as previously described (Matt et al., 2011; Chenaux et al., 2016). In brief: 400 µm thick slices were cut on a vibratome (Leica VT 1000S) while submerged in ice-cold dissection buﬀer (composition in mM) 127 NaCl, 1.9 KCl, 1.2 KH2PO4, 26 NaHCO3, 10 D-glucose, 23 MgSO4, and 1.1 CaCl2, saturated with 5% CO2 and 95% O2 (pH 7.4). Slices were incubated in oxygenated artiﬁcial cerebrospinal ﬂuid (ACSF, in mM: 127 NaCl, 26 NaHCO3, 1.2 KH2PO4, 1.9 KCl, 2.2 CaCl2, 1 MgSO4 and 10 D-glucose; pH 7.4) for 1 h at 30◦C and then stored at room temperature. Recordings were performed in a submerged- type recording chamber (Warner Instruments). Stimulation (TM53CCINS, WPI) and recording electrodes (ACSF-ﬁlled glass pipettes, 2–3 M) were positioned in the stratum radiatum to record Schaﬀer collateral fEPSPs. Signals were ampliﬁed with an Axopatch 200B (Molecular Devices), digitized at 5 kHz with an ITC-16 (HEKA) and recorded using WinWCP from the Strathclyde Electrophysiology Suite. Stimuli (100 µs) were delivered through a stimulus isolator (WPI). The same stimulus intensity was used during baseline recording (0.067 Hz) and induction of long-term potentiation (LTP) using 100 stimuli given at 100 Hz (1 s). The baseline was determined by the average of fEPSP initial slopes from the period before the tetanus. The level of LTP was determined by the average of fEPSP initial slopes from the period between 50 and 60 min after the tetanus. For wash-in experiments with 50 µM picrotoxin (Sigma), the level of potentiation was determined by the average of fEPSP initial slopes from the period between 15 and 20 min after the beginning of wash-in. Before tetanic stimulation or wash-in, each Tissue Preparation For protein isolation, brains were dissected with small forceps and immediately frozen in liquid nitrogen and stored at −80◦C before use. Brain and cochlear tissue for immunohistochemistry was prepared as previously described (Singer et al., 2016). Frontiers in Molecular Neuroscience \| www.frontiersin.org Hearing Measurements and Noise Exposure p The hearing function of 2–3 months old BLEV reporter mice of both sexes was studied by measuring auditory brainstem responses (ABRs), as previously described (Zuccotti et al., 2012; Rüttiger et al., 2013). For noise exposure, animals were exposed to 10 kHz for 40 min at 80, 100, or 120 dB SPL while under anesthesia. For ABR recordings and noise exposure, we anesthetized the animals with an intraperitoneal injection of a mixture of ketamine-hydrochloride (75 mg/kg body weight, Ketavet, Pharmacia, Erlangen, Germany) and xylazine hydrochloride (5 mg/kg body weight, Rompun, Bayer, Leverkusen, Germany). Additional doses of anesthetics were administered if needed. Sham-exposed animals were anesthetized and placed in the reverberating chamber but not exposed to acoustic stimulus (i.e., the speaker remained turned oﬀ). Mice were randomly allocated to the diﬀerent treatment groups. Immunohistochemistry on brain sections was carried out as previously described (Singer et al., 2016). Antibodies are described in Supplementary Table 1. Ribbon Counting Ribbons are shown as average ribbon number per IHC (±SD). Statistical analysis was performed using 2-way ANOVA followed by a 1-tailed Student’s t-test with α = 0.05. Morris Water Maze Data were analyzed using Smart tracking software (Panlab) and Microsoft Excel by an experimenter unaware to the treatment of each mouse. Statistics were performed with IBM SPSS with α = 0.05. One-way repeated measures ANOVA was used to test for eﬀect of training over time within the two treatment groups, separately. Pairwise comparisons were further performed to test for signiﬁcant diﬀerences between the ﬁrst training day and the following days within the two treatment groups. Two mice that did not enter the platform at least once during acquisition phase were excluded from analysis. Morris Water Maze (MWM) The Morris water maze test was performed as previously described (Bausch et al., 2015) using 3.4–4.5 and 1.7–2.7-month- old homozygous and heterozygous female BLEV reporter mice 10 days after exposure to sham or 80 dB SPL. We included as control group 11 homozygous and seven heterozygous mice (four males, 14 females) and as 80 dB SPL exposure group 10 homozygous and seven heterozygous mice (two males, 15 females). Age and sex was equally distributed among the groups. In mice, no signiﬁcant diﬀerence is observed in MWM performance between males and females (Jonasson, 2005). Furthermore, the chosen age ranges are well below the age in which declining memory performance is to be expected (see e.g., de Fiebre et al., 2006). Therefore, we expected either sex or age diﬀerences to signiﬁcantly inﬂuence variability. Hearing Measurements Click- and noise-ABR measurements were analyzed by 1-way ANOVA with α = 0.05, post-test: Bonferroni-Holms. F-ABR measurements were group analyzed by 2-way ANOVA with α = 0.05, post-test: Bonferroni-Holms (GraphPad Prism). Data are shown as mean ± SD. Electrophysiology Data were analyzed and processed using Clampﬁt 10 (Molecular Devices) and Microsoft Excel. Statistics and visualization were performed with GraphPad Prism. Results between conditions were statistically compared using 1-way ANOVA and Bonferroni’s Multiple Comparison Test to compare baseline vs. LTP or wash-in for both genotypes as well as LTP or wash-in between genotypes. y p The circular pool, 112 cm in diameter, (Stoelting) was located in a room surrounded by extra-maze (distal) cues. Water was made opaque by addition of powdered milk and water temperature was maintained at 21 ± 1◦C. A cylindrical escape platform, 12 cm in diameter, made of clear plastic, was submerged 0.5 cm beneath the water surface. The maze was virtually divided by two axes (N to S and W to E) into four quadrants (NE, SE, SW, NW). The hidden platform was positioned in the middle of the SW quadrant. During six acquisition trainings mice, starting from diﬀerent, pseudo-random locations around the perimeter of the tank, received two swim trials a day (max. 60 s) with an inter- trial-interval of 15 s. A probe trial (60 s) without platform was performed 24 h after the last acquisition training day. Latencies of the two daily trials are averaged. ABR Analysis For each individual ear, the peak input-output function (peak I/O) of the noise-ABR measurements was analyzed as previously described (Chumak et al., 2016). Two peak classes were selected: (1) early peaks (at 1.2–1.8 ms, wave I) interpreted as the sum of the ﬁrst stimulus-related action potential within the auditory nerve, and (2) delayed peaks (at 4.1–4.9 ms, wave IV), the response from the auditory midbrain. Data were analyzed by 2-way ANOVA with α = 0.05 (GraphPad Prism). The learning paradigm was speciﬁcally designed to demonstrate improved learning performance. We therefore opted for a very challenging task consisting of only two learning trials per day, as the acquisition of the MWM task with less than four trials per day is very challenging for mice (Vorhees and Williams, 2006). A previous study demonstrated improved learning performance in the MWM after environmental enrichment, which was only detected by a 2-trial a day, but not a 4-trial a day paradigm (van Praag et al., 1999). The diﬃculty of the task was additionally exacerbated by the presence of only very sparse visual cues. Additionally a waveform correlation analysis was performed as described in (Rüttiger et al., 2013; Singer et al., 2013). Data were analyzed by 1-way ANOVA with α = 0.05 (GraphPad Prism). Protein Isolation and Western Blot For isolation of cMyc-tagged proteins and HA-tagged proteins, the Mild Puriﬁcation kit and the HA-tagged Protein Puriﬁcation kit were used, respectively (BiozolDiagnostica). In brief, tissues were dissolved in a lysis buﬀer (CelLytic M, Sigma-Aldrich) and incubated for 1 h with anti-cMyc or anti-HA tag beads suspension. The suspension was then centrifuged and washed; cMyc- and HA-tagged proteins were eluted with Elution Peptide Solution from the kit. All samples underwent the same procedures. The ﬂow-through of the IP after extraction of the HA- and cMyc-tagged proteins was loaded on the gel to detect GAPDH (see Supplementary Table 1) and all other antibodies (expect of RCFP). Equal amounts of proteins are loaded for each lane. Proteins were separated by electrophoresis and placed on a transfer membrane; non-speciﬁc epitopes of the membrane were blocked with 5% milk powder solution and incubated July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 3 Visualizing BDNF During Sound-Induced Plasticity Matt et al. performed on original blots. For original blots see Supplementary Figures 2, 3. slice was used to record input-output relation (IOR) and paired- pulse facilitation (PPF) at the same stimulation strength as LTP recordings. Four traces were averaged for each data point. Values for single animals and mean per treatment group are shown. Additionally the 95% conﬁdence interval of the control group is marked as dotted lines. Data Analyses Statistics and Numbers All statistical information and n numbers can be found in the results part and in Supplementary Tables 2, 3. In ﬁgures, signiﬁcance is indicated by asterisks (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p< 0.001). n.s. denotes non-signiﬁcant results (p > 0.05). Western Blot The intensity of the bands was analyzed using the TotalLab Quant software. Band intensities of the genes of interest were normalized to housekeeping gene GAPDH. The analyses were Pictures acquired from brain sections stained for parvalbumin (PV), were analyzed using the free software Image J. (NIH, July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 4 Visualizing BDNF During Sound-Induced Plasticity Matt et al. p < 0.0001] directly after exposure, but only animals subjected to 120 dB SPL developed a permanent threshold shift [PTS; Figures 1A–C right panel; PTS: 1-way ANOVA: F(3, 144) = 54.72, p < 0.0001, post-hoc test Bonferroni’s test: con vs. 120 dB SPL p <0.0001; noise-ABR: 1-way ANOVA: F(3, 142) = 75.45, p < 0.0001, post-hoc test Bonferroni’s test: con vs. 120 dB SPL p < 0.0001; f-ABR: 2-way ANOVA: F(3, 597) = 79.9, p < 0.0001, post-hoc test Bonferroni’s test: con vs. 120 dB SPL p < 0.05]. As observed before in rats (Singer et al., 2013), the diﬀerent sound exposure paradigms manifested as long-lasting adaptations along the ascending auditory pathway (Figure 1D). These included (i) elevated (80 dB SPL, middle turn), moderately reduced (100 dB SPL, midbasal turn), or considerably reduced (120 dB SPL) numbers of CtBP2/RIBEYE-positive active release sites in the ﬁrst inner hair cell (IHC) synapse [Figure 1E; 2-way ANOVA: F(3, 60) = 11.08, p < 0.0001, post-hoc test 1-tailed unpaired Student’s t-tests: middle turn: control/80 dB SPL p < 0.05; control/120 dB SPL p < 0.01; midbasal turn: control/100 dB SPL p < 0.05; control/120 dB SPL p < 0.01; n = 6 ears from four animals per group, 1–3 repetitions each, 8–24 IHCs per turn and group]. (ii) that the overall ABR waves’ ﬁne structure showed a loss of ABR waveform with increasing exposure levels before and 2 weeks after exposure [Figure 1F; 1-way ANOVA: F(3, 131) = 17.51, p < 0.0001, post-hoc test Tukey’s Multiple Comparison test: control/120 dB SPL p < 0.001, 80/120 dB SPL p < 0.001, 100/120 dB SPL p < 0.01; n = 8 animals, 15 ears (control), n = 9 animals, 18 ears (80 dB SPL), n = 5 animals, 10 ears (100 dB SPL), n = 9 animals, 17 ears (120 dB SPL)], which could be conﬁrmed by detailed analyses of supra-threshold ABR waves. Western Blot The amplitudes of the early supra-threshold ABR wave I [Figure 1G; ABR wave I: 2-way ANOVA; con: F(1, 1031) = 0.003, p = 0.955; 80 dB SPL: F(1, 890) = 6.02, p= 0.0143; 100 dB SPL: F(1, 836) = 28.59, p < 0.0001; 120 dB SPL: F(1, 396) = 185.8, p < 0.0001] and late ABR wave IV [Figure 1H; ABR wave IV: 2-way ANOVA, con: F(1, 1034) = 1.296, p = 0.2551; 80 dB SPL F(1, 951) = 0.89, p= 0.3446; 100 dB SPL: F(1, 743) = 0.09, p = 0.7706; 120 dB SPL: F(1, 452) = 82.88, p < 0.0001] were elevated after enriching 80 dB SPL exposure, reduced but centrally compensated after mildly traumatic 100 dB SPL exposure, and reduced for both early and late ABR waves after severely traumatic 120 dB SPL exposure (Figures 1G,H). Bethesda, MD; USA). For each section, three pictures for each single channel (YFP, CFP, PV) were saved and analyzed independently. For the 10× magniﬁed pictures, after conversion to an 8-bit image, background was reduced using the rolling ball algorithm (included in Image J) with standard parameters and a region of interest (ROI) of 300 × 100 µm was created and placed on the CA3 region in each single channel picture. Afterwards the average ﬂuorescence intensity within the ROI was calculated as single pixel intensity (0–255)/no. of pixel. Data are shown as mean pixel intensity (±SEM). Data were analyzed by 1-way ANOVA with α = 0.05, post-test: Bonferroni-Holms (GraphPad Prism). To generate representative ﬂuorescence proﬁle plots, a ROI of 390 × 100 µm was created and the speciﬁc Image J built-in function was used. For the 60× magniﬁed pictures, the same procedure was applied as for the 10× magniﬁed pictures within a ROI of 300 × 650 µm. To generate representative ﬂuorescence proﬁle plots, a ROI of 350 × 100 µm was created and the speciﬁc Image J built-in function was used. In pictures acquired from brain sections stained for δGABAA- R, or α1GABAA-R ﬂuorescence puncta numbers were counted using ImageJ. A 300 × 300 µm ROI was cropped, individual channels were separated and binary masks created using an appropriate thresholding algorithm for each channel. Binary particles were counted using the inbuilt Analyze Particles function. Western Blot In monochannel pictures of brain sections stained for VGLUT1 in the VCN with 100× magniﬁcation, the mean ﬂuorescence light intensity was measured in a frame of 85 × 60 µm (total picture) using build-in function in ImageJ. Data Availability The datasets generated during and/or analyzed during the current study are available from the corresponding author upon request. Different Sound Exposure Conditions Indicate Changes in Bdnf Exon-IV and -VI Transcription That Reﬂect Sound Sensitivity in the Brainstem This indicates that sound sensitivity is persistently increased after acoustic enrichment (80 dB SPL), preserved despite reduced auditory input after mild trauma (100 dB SPL), or decreased after severe acoustic trauma (120 dB SPL). In particular the elevated IHC ribbon number and ABR wave I and IV in response to 80 dB SPL or the compensated ABR wave IV despite a reduced ABR wave I in response to 100 dB SPL exposure, can only be explained through an adaptive response that permanently alters neuronal activity in auditory pathways. Following severe auditory trauma, this adaptive response consistently failed to occur (120 dB SPL) (Figures 1G,H) as seen before in the rat model (Singer et al., 2013; Knipper et al., 2015). BLEV reporter mice (described in detail in Singer et al., submitted), in which translation of Bdnf exon-IV and -VI can be monitored by co-expression of CFP and YFP, were exposed to diﬀerent sound pressure levels (SPL) inducing acoustic enrichment (80 dB SPL) and mild (100 dB SPL) or severe (120 dB SPL) acoustic trauma (Knipper et al., 2013; Singer et al., 2013). 2 weeks after exposure we observed no (80, 100 dB SPL) or moderate (120 dB SPL) loss of hearing thresholds in click-, noise-burst and frequency-speciﬁc auditory brainstem responses (ABR, Figures 1A–C). Animals exposed to 80, 100, and 120 dB SPL show a temporary threshold shift in click-ABR [TTS; Figure 1A middle panel; TTS: 1-way ANOVA: F(3, 129) = 92.67, p < 0.0001, post-hoc test Bonferroni’s est: con vs. 80 dB SPL p < 0.0001, con vs. 100 dB SPL p < 0.0001; con vs. 120 dB SPL In case activity-dependent Bdnf transcription may alter sound responsiveness by strengthening synapses via BDNF-TrkB receptor signaling, as shown in diﬀerent brain regions (Kellner et al., 2014), we might expect changes in sound sensitivity to July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 5 Matt et al. Visualizing BDNF During Sound-Induced Plasticity FIGURE 1 \| Acoustic enrichment increases hearing acuity through central adaptation while acoustic trauma leads to uncompensated hearing loss. Homozygous BLEV reporter mice were sham exposed (control) or exposed to a 10 kHz tone of 80, 100, or 120 dB SPL for 40 min. Different Sound Exposure Conditions Indicate Changes in Bdnf Exon-IV and -VI Transcription That Reﬂect Sound Sensitivity in the Brainstem (E) IHC ribbon numbers of BLEV reporter mice for the middle and midbasal cochlear turn, representing higher frequency areas, show an increase (80 dB SPL) or decline (100 and 120 dB SPL) 14 days after sound exposure. Data represented as mean ± SD. (F) Analyses of ABR waveform of control animals or mice exposed to 80, 100, or 120 dB SPL 14 days after exposure. The changes in waveforms and signals were calculated as correlation factor (CorF) (Rüttiger et al., 2013). Dashed lines indicate the 95% conﬁdence interval for the controls. ABR waveform after 120 dB SPL exposure is signiﬁcantly reduced. Data represented as mean ± SEM. (G,H) Mean peak growth input/output function of noise burst stimulus for early peaks (G, ABR wave I) and late peaks (H, ABR wave IV) before (open circles) and 14 days after exposure (red circles) were signiﬁcantly increased after 80 dB SPL exposure, signiﬁcantly decreased for early peaks, but non-signiﬁcantly different for late peaks after 100 dB SPL, and massively decreased after 120 dB SPL. Red crosses: early and late peaks in mice exposed to 120 dB SPL (ABR wave IV) that could not be recorded anymore. Data represented as mean ± S.E.M. correlate with Bdnf transcript levels and neuronal activity. We chose vesicular glutamate transporter-1 (VGLUT1) as a marker for activity levels, as it is the predominant glutamate transporter in auditory brainstem synapses (Zhou et al., 2007). auditory responses, but not in animals with critically reduced supra-threshold auditory nerve responses that failed to be centrally restored. This is consistent with the idea that a relevant auditory input drives Bdnf transcription and subsequently elevates levels of BDNF in the brainstem, which might alter VGLUT1 levels, as previously suggested for hippocampal neurons. This could strengthen synapses which depend on the activity of BDNF (Kellner et al., 2014). Comparable changes in Bdnf transcription were not detected in other sensory regions like the olfactory bulb, indicating speciﬁcity of activity-dependent BDNF transcription following auditory inputs. Western blot (WB) analysis of Bdnf exon-IV-CFP and exon- VI-YFP expression levels, following immunoprecipitation with HA and cMyc tags, respectively, indicated a qualitative increase of CFP and YFP levels after acoustic enrichment (80 dB SPL), but not after acoustic trauma in the auditory brainstem (120 dB SPL; Figure 2A, left panel, see Supplementary Figure 1A for a representative selection of WBs). Different Sound Exposure Conditions Indicate Changes in Bdnf Exon-IV and -VI Transcription That Reﬂect Sound Sensitivity in the Brainstem Consistent with elevated neuronal activity, VGLUT1 levels were qualitatively increased 2 weeks after acoustic enrichment (80 dB SPL) but not after severe acoustic trauma (120 dB SPL, Figure 2A, 2nd panel). This suggests that pathological reduction of relevant auditory input persistently prevents activity-dependent elevation of BDNF and VGLUT1. We observed a trend for mobilization of Bdnf exon-IV-CFP and exon-VI-YFP and VGLUT1 not only in the auditory brainstem, but also in the inferior colliculus (Figure 2A, 3rd panel, see Supplementary Figure 1B for a representative selection of WBs). However, due to the highly variable immunoprecipitation yield, quantiﬁcation of WB analyses did not reach statistical signiﬁcance (Figure 2B). Importantly, following enriched sound no indications for mobilization of transcript-speciﬁc BDNF in the olfactory bulb was observed, as shown for Bdnf exon-IV-CFP and Bdnf exon-VI-YFP (Figure 2A, right panel). To quantitatively verify long-lasting changes of neuronal activity in the target cells of the auditory nerve, we additionally analyzed VGLUT1 immunoreactivity (IR) in bushy cells of the cochlear nucleus (CN) within the auditory brainstem 2 weeks after acoustic enrichment [80 dB SPL; Figures 2C,D; two-tailed Student’s t- test: t = 3.63 df = 10 p = 0.0046; n = 6 mice/group 2–3 repetitions]. A signiﬁcantly elevated punctate VGLUT1 IR at the level of CN bushy cells, exemplarily shown in Figure 2C and quantiﬁed in Figure 2D, was observed in comparison to sham (control) exposure. This suggested that 2 weeks following 80 dB SPL exposure, auditory nerve synapses may exhibit larger numbers of active release sites or greater spike ﬁdelity, considered as a functional correlate of elevated VGLUT1 following sound exposure (Ngodup et al., 2015). Different Sound Exposure Conditions Indicate Changes in Bdnf Exon-IV and -VI Transcription That Reﬂect Sound Sensitivity in the Brainstem Mean ABR thresholds ± SD for click stimuli (A, click-ABR) before (pre sound exposure), directly after (TTS) and 14 d after sound exposure (PTS), for noise burst stimuli (B, noise-ABR) and for frequency-speciﬁc stimuli (C, f-ABR) 14 d after exposure demonstrate a signiﬁcant temporary threshold shift (TTS) for all paradigms directly after exposure becoming permanent (PTS) (Continued) Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 1 \| Acoustic enrichment increases hearing acuity through central adaptation while acoustic trauma leads to uncompensated hearing loss. Homozygous BLEV reporter mice were sham exposed (control) or exposed to a 10 kHz tone of 80, 100, or 120 dB SPL for 40 min. Mean ABR thresholds ± SD for click stimuli (A, click-ABR) before (pre sound exposure), directly after (TTS) and 14 d after sound exposure (PTS), for noise burst stimuli (B, noise-ABR) and for frequency-speciﬁc stimuli (C, f-ABR) 14 d after exposure demonstrate a signiﬁcant temporary threshold shift (TTS) for all paradigms directly after exposure becoming permanent (PTS) (Continued) FIGURE 1 \| Acoustic enrichment increases hearing acuity through central adaptation while acoustic trauma leads to uncompensated hearing loss. Homozygous BLEV reporter mice were sham exposed (control) or exposed to a 10 kHz tone of 80, 100, or 120 dB SPL for 40 min. Mean ABR thresholds ± SD for click stimuli (A, click-ABR) before (pre sound exposure), directly after (TTS) and 14 d after sound exposure (PTS), for noise burst stimuli (B, noise-ABR) and for frequency-speciﬁc stimuli (C, f-ABR) 14 d after exposure demonstrate a signiﬁcant temporary threshold shift (TTS) for all paradigms directly after exposure becoming permanent (PTS) (Continued) July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 6 Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 1 \| only after 120 dB SPL sound exposure. A–C; con n = 19 animals; 80 dB SPL; n = 19 animals; 100 dB SPL, n = 16 animals; 120 dB SPL, n = 19 animals. (D) Schematic drawing of an ABR waveform in relation to the corresponding auditory nuclei in the ascending auditory pathway (green arrows) starting at the auditory nerve (AN), cochlear nucleus (CN), superior olivary complex (SOC), inferior colliculus (IC), medial geniculate body (MGB), auditory cortex (AC), and entorhinal cortex (EC) to the hippocampus as well as hippocampal projections to cortical areas (red). Frontiers in Molecular Neuroscience \| www.frontiersin.org Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition q We previously observed altered levels of Arc (activity-regulated cytoskeletal protein) in the hippocampal CA1 region after acoustic enrichment, as well as after mild and severe acoustic trauma in rats (Singer et al., 2013). Arc plays a key role in determining synaptic strength through facilitation of AMPA receptor (AMPAR) endocytosis in response to BDNF signaling (Bramham et al., 2008; Wall and Corrêa, 2017). Similarly, as previously reported for Arc (Singer et al., 2013), we observed a trend for an increase in the hippocampal expression of another excitability marker, the GluA2 subunit of the AMPAR (Tanaka et al., 2000; Singer et al., 2013) 2 weeks after acoustic enrichment, but not after severe acoustic trauma in BLEV reporter mice (Singer et al., 2013). This elevation of hippocampal excitability was paralleled by a trend for higher levels of Bdnf exon-IV-CFP and exon-VI-YFP (Figure 3A, ﬁrst panel, Supplementary Figure 1C for a representative selection of WBs) which was likewise only observed after acoustic enrichment, but not after severe acoustic trauma. These ﬁndings are in line with diﬀerent sound exposure conditions driving adaptations of Bdnf transcript levels and glutamatergic neuronal activity not only in the brainstem but also in the hippocampus. Modulation of BDNF-dependent Arc expression was previously associated with increased hippocampal synaptic plasticity (Kuipers et al., 2016), particularly with formation of long-term potentiation (LTP) (Messaoudi et al., 2007). Therefore, we went on to test if the altered hippocampal BDNF expression levels and neuronal excitability changes In summary, we observed increased expression of VGLUT1 together with a trend for upregulation of Bdnf exon-IV-CFP and exon-VI-YFP in the auditory brainstem in animals with elevated early (auditory nerve) or late (brainstem) supra-threshold July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 7 Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 2 \| Acoustic enrichment leads to a trend for higher expression of VGLUT1 in the brainstem and inferior colliculus together with Bdnf exon-IV-CFP and exon-VI-YFP. (A) Western blot analysis of CFP and YFP in the brainstem and olfactory bulb as well as VGLUT1 in the brainstem and inferior colliculus in homozygous BLEV reporter mice 2 weeks after exposure to either sham, 80, or 120 dB SPL. The housekeeping gene GAPDH is used for normalization between conditions. For originals see Supplementary Figures 2A–C. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition (B) Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5–6 mice/group 2–10 Western blots each). (C) VGLUT1 immunostaining (green) in cochlear nucleus mouse slices 2 weeks after exposure to sham or 80 dB SPL. After exposure to 80 dB SPL immunoreactivity (IR) for VGLUT1 is increased in the dorsal and ventral cochlear nucleus (DCN, VCN). Nuclei stained with DAPI (blue). Arrows indicate VGLUT1 localization. Scale bars: 100 µm; 5 µm. (D) Quantiﬁcation of VGLUT1 IR in an 85 × 60 µm frame in the VCN following sham and 80 dB SPL exposure. Data represented as mean ± SD. FIGURE 2 \| Acoustic enrichment leads to a trend for higher expression of VGLUT1 in the brainstem and inferior colliculus together with Bdnf exon-IV-CFP and exon-VI-YFP. (A) Western blot analysis of CFP and YFP in the brainstem and olfactory bulb as well as VGLUT1 in the brainstem and inferior colliculus in homozygous BLEV reporter mice 2 weeks after exposure to either sham, 80, or 120 dB SPL. The housekeeping gene GAPDH is used for normalization between conditions. For originals see Supplementary Figures 2A–C. (B) Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5–6 mice/group 2–10 Western blots each). (C) VGLUT1 immunostaining (green) in cochlear nucleus mouse slices 2 weeks after exposure to sham or 80 dB SPL. After exposure to 80 dB SPL immunoreactivity (IR) for VGLUT1 is increased in the dorsal and ventral cochlear nucleus (DCN, VCN). Nuclei stained with DAPI (blue). Arrows indicate VGLUT1 localization. Scale bars: 100 µm; 5 µm. (D) Quantiﬁcation of VGLUT1 IR in an 85 × 60 µm frame in the VCN following sham and 80 dB SPL exposure. Data represented as mean ± SD. FIGURE 2 \| Acoustic enrichment leads to a trend for higher expression of VGLUT1 in the brainstem and inferior colliculus together with Bdnf exon-IV-CFP and exon-VI-YFP. (A) Western blot analysis of CFP and YFP in the brainstem and olfactory bulb as well as VGLUT1 in the brainstem and inferior colliculus in homozygous BLEV reporter mice 2 weeks after exposure to either sham, 80, or 120 dB SPL. The housekeeping gene GAPDH is used for normalization between conditions. For originals see Supplementary Figures 2A–C. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition (B) Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5–6 mice/group 2–10 Western blots each). (C) VGLUT1 immunostaining (green) in cochlear nucleus mouse slices 2 weeks after exposure to sham or 80 dB SPL. After exposure to 80 dB SPL immunoreactivity (IR) for VGLUT1 is increased in the dorsal and ventral cochlear nucleus (DCN, VCN). Nuclei stained with DAPI (blue). Arrows indicate VGLUT1 localization. Scale bars: 100 µm; 5 µm. (D) Quantiﬁcation of VGLUT1 IR in an 85 × 60 µm frame in the VCN following sham and 80 dB SPL exposure. Data represented as mean ± SD. FIGURE 2 \| Acoustic enrichment leads to a trend for higher expression of VGLUT1 in the brainstem and inferior colliculus together with Bdnf exon-IV-CFP and exon-VI-YFP. (A) Western blot analysis of CFP and YFP in the brainstem and olfactory bulb as well as VGLUT1 in the brainstem and inferior colliculus in homozygous BLEV reporter mice 2 weeks after exposure to either sham, 80, or 120 dB SPL. The housekeeping gene GAPDH is used for normalization between conditions. For originals see Supplementary Figures 2A–C. (B) Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5–6 mice/group 2–10 Western blots each). (C) VGLUT1 immunostaining (green) in cochlear nucleus mouse slices 2 weeks after exposure to sham or 80 dB SPL. After exposure to 80 dB SPL immunoreactivity (IR) for VGLUT1 is increased in the dorsal and ventral cochlear nucleus (DCN, VCN). Nuclei stained with DAPI (blue). Arrows indicate VGLUT1 localization. Scale bars: 100 µm; 5 µm. (D) Quantiﬁcation of VGLUT1 IR in an 85 × 60 µm frame in the VCN following sham and 80 dB SPL exposure. Data represented as mean ± SD. July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 8 Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 3 \| Acoustic enrichment enhances hippocampal BDNF expression and LTP as well as memory acquisition in the Morris water maze (MWM). (A) Left panel: Western blot analysis of Bdnf exon-IV-CFP and exon-VI-YFP (left blot) as well as GluA2 (right blot) in the hippocampus of BLEV reporter mice exposed to sham, 80, or 120 dB SPL. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition Right panel: Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5 mice/group 2–10 Western blots each). For originals see Supplementary Figure 2D. (B) Left Panel: Averaged time courses of fEPSP slopes in acute brain slices from BLEV reporter mice 2 weeks after exposure to sham, 80, 100, or 120 dB SPL. Representative traces before (black) and after (gray) induction of LTP are shown on top. Animals from all conditions readily show signiﬁcant SC LTP. Right panel: Signiﬁcantly increased LTP was observed for 80 (171 ± 16% of baseline) and 100 (172 ± 8% of baseline) but not 120 dB SPL (136 ± 4% of baseline) compared to controls (135 ± 7% of baseline). Data are represented as mean ± SEM. (C) Latency to target during acquisition learning. After 3 days mice exposed to 80 dB SPL need less time to ﬁnd the platform in comparison to the ﬁrst experimental day than mice in the control group. Data are expressed as mean ± SEM. FIGURE 3 \| Acoustic enrichment enhances hippocampal BDNF expression and LTP as well as memory acquisition in the Morris water maze (MWM). (A) Left panel: Western blot analysis of Bdnf exon-IV-CFP and exon-VI-YFP (left blot) as well as GluA2 (right blot) in the hippocampus of BLEV reporter mice exposed to sham, 80, or 120 dB SPL. Right panel: Densitometric analyses of Western blots (dots: averaged results per animal, bars: means, dashed lines: 95% conﬁdence interval for the controls; n = 5 mice/group 2–10 Western blots each). For originals see Supplementary Figure 2D. (B) Left Panel: Averaged time courses of fEPSP slopes in acute brain slices from BLEV reporter mice 2 weeks after exposure to sham, 80, 100, or 120 dB SPL. Representative traces before (black) and after (gray) induction of LTP are shown on top. Animals from all conditions readily show signiﬁcant SC LTP. Right panel: Signiﬁcantly increased LTP was observed for 80 (171 ± 16% of baseline) and 100 (172 ± 8% of baseline) but not 120 dB SPL (136 ± 4% of baseline) compared to controls (135 ± 7% of baseline). Data are represented as mean ± SEM. (C) Latency to target during acquisition learning. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition 120 dB SPL p < 0.01; con: n = 4 animals; 80 dB SPL n = 3 animals; 100 dB SPL n = 3 animals; 120 dB SPL n = 5 animals, 3 repetitions each]. Fluorescence intensity proﬁles taken through the SL and SP of CA3 of a control mouse revealed that YFP ﬂuorescence was predominantly found in the SL, while CFP and PV ﬂuorescence was restricted to the SP (representative proﬁle plot in Figure 4F). We wanted to know next if elevated hippocampal BDNF expression not only correlates with increased GluA2 levels and SC LTP, but also with performance in hippocampus-dependent learning. To test this, starting 10 days after sound exposure, we subjected mice exposed to sham treatment (control) or acoustic enrichment (80 dB SPL) to a Morris water maze (MWM) task, which is a hippocampus-dependent learning paradigm. In order to identify increased learning performance (van Praag et al., 1999), we opted for a very challenging paradigm that included only two trials per day (for details, see Methods). While control mice improved slightly but not signiﬁcantly over time performing the task, acoustically enriched mice signiﬁcantly improved their performance upon training. This was evidenced by reduced latencies in ﬁnding the platform as compared to the ﬁrst day [Figure 3C; repeated measure ANOVA: con: F = 2.56 DF = 5 p = 0.033; post-hoc test Bonferroni’s test: day 1 vs. 2 n.s., day 1 vs. 3 n.s., day 1 vs. 4 n.s., day 1 vs. 5 n.s., day 1 vs. 6 n.s.; 80 dB SPL: F = 5.85 DF = 5 p < 0.001; Bonferroni’s test: day 1 vs. 2 n.s., day 1 vs. 3 p = 0.013, day 1 vs. 4 p = 0.054, day 1 vs. 5 p = 0.006, day 1 vs. 6 p < 0.001; con n = 18 animals 80 dB SPL n = 17]. We could not detect any correlation between MWM performance and gender or age. So far, the presented data indicate that acoustic enrichment leads to elevated expression of Bdnf exon-IV-CFP and exon-VI- YFP together with excitatory markers in the auditory brainstem and the hippocampus, paralleled by increased hippocampal LTP and improved performance in a hippocampus-dependent learning paradigm. Neither sound-induced improval of auditory performance nor increased hippocampal LTP are observed in animals exposed to severe acoustic trauma, in which activity- dependent Bdnf transcript levels are not elevated in response to auditory sound exposure. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition After 3 days mice exposed to 80 dB SPL need less time to ﬁnd the platform in comparison to the ﬁrst experimental day than mice in the control group. Data are expressed as mean ± SEM. of input strengths (Supplementary Figure 4A). Additionally, similar levels of paired-pulse facilitation in all four conditions indicated no changes in presynaptic function (Supplementary Figure 4B). We were able to observe signiﬁcant LTP in acute brain slices from mice under all four conditions in response to tetanic stimulation (1 s, 100 Hz). This potentiation, however, was signiﬁcantly stronger in animals exposed to acoustic enrichment or mild acoustic trauma as compared to animals exposed to sham or severe acoustic trauma (Figure 3B). This ﬁnding suggests that persistently improved (after acoustic enrichment) or restored (after mild acoustic trauma) sound responses leads to altered Bdnf exon-IV-CFP and exon-VI-YFP expression as well as synaptic excitability and plasticity in the brainstem and hippocampus. Reduced sound responses after severe acoustic trauma, on the other hand, forestall activity- dependent BDNF expression, improved excitability, and synaptic plasticity. observed after acoustic enrichment or acoustic trauma also inﬂuenced hippocampal synaptic plasticity in the BLEV reporter mouse. To this end, we recorded ﬁeld excitatory postsynaptic potentials (fEPSP) in the CA3 to CA1 Schaﬀer collateral (SC) synapses in the stratum radiatum (SR) from acute forebrain slices of BLEV reporter mice 2 weeks after sham (control) treatment, acoustic enrichment, as well as mild and severe acoustic trauma [Figure 3B, Supplementary Figure 4; 1-way ANOVA: F(3, 19) = 4.99, p = 0.01, post-hoc test Bonferroni’s test baseline/tetanized (b/t) n = 4 animals/group con: 7 slices, 80 dB SPL: 7 slices, 100 dB SPL: 6 slices, 120 dB SPL: 5 slices, control p < 0.01, 80 dB SPL p < 0.001, 100 dB SPL p < 0.001, 120 dB SPL p < 0.01; tetanized/tetanized (t/t) con vs. 80 dB SPL p < 0.01 con vs. 100 dB SPL p < 0.01 con vs. 120 dB SPL n.s.]. None of the sound exposure paradigms led to changes in basal synaptic transmission, as all four conditions displayed similar fEPSP amplitudes in response to a range July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 9 Visualizing BDNF During Sound-Induced Plasticity Matt et al. gyrus (Figure 4C) of the hippocampus. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition Compared to sham exposed animals (control), we observed a general upregulation of CFP, YFP, and PV (red) in all regions of the hippocampus after acoustic enrichment and mild acoustic trauma, the two conditions associated with increased hippocampal synaptic plasticity (Figure 3B, Supplementary Figure 4). In control animals, CFP ﬂuorescence is mostly seen in capillary vessels of the highly vascularized ﬁssura hippocampalis (FH) (Figure 4A). These CFP levels in the FH were signiﬁcantly higher after acoustic enrichment and mild acoustic trauma but not after severe acoustic trauma [Figure 4D; 1-way ANOVA: F(3, 19) = 10.5, p < 0.0003, post-hoc test Bonferroni’s test: con vs. 80 dB SPL, p = 0.0013, con vs. 100 dB SPL n.s. con vs. 120 dB SPL n.s., 80 dB SPL vs. 120 dB SPL p < 0.0001, 100 dB SPL vs. 120 dB SPL p = 0.0072; con: n = 6 animals, 1–4 repetitions; 80 dB SPL n = 6 animals, 1–6 repetitions; 100 dB SPL n = 5 animals, 2–5 repetitions; 120 dB SPL n = 6 animals, 1–3 repetitions]. In the CA1 region of animals exposed to acoustic enrichment or mild acoustic trauma, we observed increased PV labeling in the CA1 pyramidal layer (Figure 4A, SP). In the CA3, we found strongly ampliﬁed signals of both PV and Bdnf exon-VI-YFP in the area where mossy ﬁbers are predicted to target CA3 dendrites (Figures 4B,E). PV labeling was also increased in the supra- pyramidal blade of the DG (Figure 4C). This generalized pattern of up-regulation was not observed in animals exposed to 120 dB SPL (Figures 4A–E). Particularly within the stratum lucidum (SL) of the CA3 region (Figure 4B), the increase in Bdnf exon- VI-YFP expression, together with PV, was either robust (80 dB SPL), moderate (100 dB SPL), or absent [120 dB SPL, Figure 4E; YFP: 1-way ANOVA: F(3, 11) = 6.96, p = 0.0068, post-hoc test Bonferroni’s test con vs. 80 dB SPL p < 0.01, 80 dB SPL vs. 120 dB SPL p < 0.01; CFP: 1-way ANOVA F(3, 11) = 2.37, p = 0.13; PV: 1-way ANOVA F(3, 11) = 10.07, p = 0.0017, post-hoc test Bonferroni’s test con vs. 80 dB SPL p < 0.01, 80 dB SPL vs. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition Acoustic Enrichment and Mild, but Not Severe Trauma, Synchronously Alter Hippocampal Bdnf Exon-IV and Exon-VI Transcription in Neuronal and Vascular Cells in Parallel to Parvalbumin Expression We subsequently took advantage of the unique utility of the BLEV reporter mouse to identify the cell types where activity- dependent Bdnf transcription occurs in response to sound exposure. As inhibitory transmission is a major regulator of hippocampal synaptic plasticity (Matt et al., 2011) and BDNF is known to inﬂuence the function of parvalbumin (PV)-positive inhibitory interneurons (Marty, 2000; Yamada and Nabeshima, 2003; Messaoudi et al., 2007; Hong et al., 2008; Minichiello, 2009; Waterhouse et al., 2012; Park and Poo, 2013), we asked if changes in activity-dependent Bdnf transcript usage might be observed in neurons or cells known to target PV-positive interneurons. For example, mossy ﬁbers express BDNF and are suggested to target PV-positive interneurons (Danzer et al., 2008; Dieni et al., 2012). We ﬁrst examined Bdnf exon-IV- CFP, exon-VI-YFP, and PV expression in deconvoluted high- resolution ﬂuorescence stacks of low-magniﬁcation (Figure 4) in the CA1 region (Figure 4A; Supplementary Video 1), the CA3 region (Figure 4B; Supplementary Video 1), and the dentate Consistent with our previous observations in terms of ABR wave amplitudes, the expression of excitatory markers in the auditory brainstem, the inferior colliculus, and the hippocampus, as well as hippocampal LTP and learning, we thus also conﬁrmed a correlation between PV expression patterns and changes of Bdnf exon-IV-CFP and exon-VI-YFP levels in the tri-synaptic pathway. This ﬁnding led us to perform a more detailed analysis of CFP, YFP, and PV expression in the hippocampal CA3 region at a high-magniﬁcation. CFP ﬂuorescence was mainly restricted to blood vessels in the SL (Figure 5A, lower panels, open arrows) and to perisomatic regions in the SP (Figure 5B, open arrows). YFP ﬂuorescence was predominantly found in the SL of the CA3 region (Figure 5A, lower panels; Figure 5B, closed arrows). In all these areas, ﬂuorescence intensity was increased after acoustic enrichment and mild acoustic trauma, but not after severe acoustic trauma. Many of the YFP-positive puncta in the SL July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 10 Matt et al. Visualizing BDNF During Sound-Induced Plasticity FIGURE 4 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin upon sound exposure in hippocampal circuits (Supplementary Video 1). Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition (A–C) CFP/YFP ﬂuorescence and parvalbumin immunostaining (PV, red) in dorsal hippocampal mouse slices 2 weeks after exposure to sham, 80, 100, or 120 dB SPL. Note a prominent dynamic change of PV (red) expression in the stratum pyramidale (SP) of the CA1 region (A), in the stratum lucidum (SL) of the CA3 region (n = 6 animals/group) (B), and stratum granulosum (SG) of the dentate gyrus (DG) (C) concomitant to distinct changes in YFP and CFP. SM, stratum moleculare; SR, stratum radiatum; FH, ﬁssura hippocampalis. Scale bars, 100 µm. (D) Quantiﬁcation of Bdnf exon-IV-CFP in capillary outlines at the level of the FH revealed a peak of ﬂuorescence in 80 dB SPL exposed animals that declined below control levels in 120 dB SPL exposed animals. Data represented as mean ± SD. (E) Quantiﬁcation of YFP, CFP, and PV intensities in the CA3 region in an area of 300 × 100 µm following different sound exposure paradigms. Data represented as mean ± SEM. (F) Intensity proﬁle of CFP, YFP, and PV along a line through the CA3 of a control mouse, perpendicular to the SP (length: approximately 390 µm). Representative for Matt et al. Visualizing BDNF During Sound-Induced Plasticity FIGURE 4 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin upon sound exposure in hippocampal circuits (Supplementary Video 1). (A–C) CFP/YFP ﬂuorescence and parvalbumin immunostaining (PV, red) in dorsal hippocampal mouse slices 2 weeks after exposure to sham, 80, 100, or 120 dB SPL. Note a prominent dynamic change of PV (red) expression in the stratum pyramidale (SP) of the CA1 region (A), in the stratum lucidum (SL) of the CA3 region (n = 6 animals/group) (B), and stratum granulosum (SG) of the dentate gyrus (DG) (C) concomitant to distinct changes in YFP and CFP. SM, stratum moleculare; SR, stratum radiatum; FH, ﬁssura hippocampalis. Scale bars, 100 µm. (D) Quantiﬁcation of Bdnf exon-IV-CFP in capillary outlines at the level of the FH revealed a peak of ﬂuorescence in 80 dB SPL exposed animals that declined below control levels in 120 dB SPL exposed animals. Data represented as mean ± SD. (E) Quantiﬁcation of YFP, CFP, and PV intensities in the CA3 region in an area of 300 × 100 µm following different sound exposure paradigms. Data represented as mean ± SEM. Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition For YFP intensity, a signiﬁcant reduction between enriching and severe acoustic trauma was observed [Figure 5D; CFP: 1-way ANOVA: F(3, 25) = 22.44, p < 0.0001, post-hoc test Bonferroni’s test con vs. 80 dB SPL p < 0.001 80 dB SPL vs. 120 dB SPL p < 0.0001; YFP: 1-way ANOVA: F(3, 25) = 3.32, p = 0.036 post- hoc test Bonferroni’s test 80 dB SPL vs. 120 dB SPL p < 0.05 n = 6 animals/group; 4–6 repetition each]. These ﬁndings suggest that BLEV mice can be used to identify mossy ﬁber terminals and CA3 projection neurons that respond to deﬁned behaviorally relevant sensory stimuli. interneurons and a simultaneous decrease of such connections in the SR. PV-positive interneurons target CA1 PC through perisomatic δ subunit containing GABAA-receptors and its dendrites through α1 subunit containing GABAA-receptors (Klausberger et al., 2003; Glykys et al., 2008) (Figure 6F). Indeed, we found that puncta of PV IR in the CA1 SP often co-localized with IR for the δ subunit containing GABAA- receptor (Supplementary Figures 5A,B) and that α1 subunit containing GABAA-receptor IR was localized in the CA1 SR (Supplementary Figure 5D). After acoustic enrichment and mild acoustic trauma, however, no signiﬁcant changes of δGABAA receptor-immunopositive dots (Supplementary Figures 5C,E) or of α1GABAA-receptor IR in the CA1 SR (Supplementary Figures 5D,F) were detected. As α1 subunit containing GABAA-receptors in the CA1 SR are suggested to regulate excitability and action potential thresholds of PC dendrites (Willadt et al., 2013) a functional test of dendritic inhibition in the CA1 SR was performed [Figure 6G; 1-way ANOVA: F(3, 19) = 5.24, p = 0.005, post-hoc test Bonferroni’s test: con p < 0.001, 80 dB SPL n.s., 100 dB SPL n.s., 120 dB SPL p < 0.001; con vs. 120 dB SPL p < 0.01, con vs. 80 dB SPL p < 0.01, con vs. 100 dB SPL n.s.; baseline/wash-in n = 4 animals/group; 9 slices/group]. We observed SC fEPSP during wash-in of the GABAA receptor antagonist picrotoxin. The ensuing disinhibition was signiﬁcantly decreased in mice exposed to acoustic enrichment and mild acoustic trauma (Figure 6G), which is consistent with the downregulation of PV- positive terminals (Figures 6C,E) and a slight decrease in α1 subunit containing GABAA-receptor expression in the CA1 SR (Supplementary Figure 5D, green), compared to animals exposed to sham treatment or severe acoustic trauma. Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells y We next examined high-magniﬁcation images of the CA1 area. Following acoustic enrichment and mild trauma, CFP ﬂuorescence increased in the FH (Figure 6A, open arrows), in perisomatic areas of the SP (Figure 6B, open arrows), and in blood vessels of the SR (Figure 6C, open arrows). YFP ﬂuorescence on the other hand was increased in the FH and the SR after acoustic enrichment and mild acoustic trauma close to capillaries (Figures 6A,C, arrowheads) (Marosi and Mattson, 2014; Miyamoto et al., 2014). Similar to CA3, perisomatic PV IR (red) in the SP of CA1 was signiﬁcantly increased after acoustic enrichment and mild acoustic trauma [Figure 6A; Figure 6B, closed arrows; Figure 6D, quantiﬁcation; 1-way ANOVA: F(3, 19) = 5.96, p = 0.0049, post-hoc test Bonferroni’s test: con vs. 80 dB SPL n.s.; con vs. 100 dB SPL, p = 0.0301; con vs. 120 dB SPL n.s; 100 vs. 120 dB SPL p = 0.01; con n = 6 animals, 80 dB SPL n = 6 animals, 100 dB SPL n = 5 animals, 120 dB SPL n = 6 animals; 4–6 repetitions]. In the SR, however, we found a signiﬁcant decrease of PV IR after acoustic enrichment and mild acoustic trauma [Figure 6C, red IR; Figure 6E, quantiﬁcation; 1-way ANOVA: F(3, 19) = 4.61, p = 0.0138, post-hoc test Bonferroni’s test: con vs. 80 dB SPL p < 0.05, con vs. 100 dB SPL p < 0.05, con n = 6 animals 80 dB SPL n = 5 animals, 100 dB SPL n = 6 animals, 120 dB SPL, n = 6 animals; 2 repetitions each]. After severe acoustic trauma, we once more could not detect any diﬀerences in CFP and YFP ﬂuorescence as well as PV IR in comparison to controls (Figures 6A–E). g Taken together, the present ﬁndings suggest that exposure to acoustic enrichment (80 dB SPL) and mild acoustic trauma (100 dB SPL), conditions that increase sound sensitivity, hippocampal LTP, and learning appears to correlate with an elevation of Bdnf exon-IV-CFP and Bdnf exon-VI-YFP in the brainstem and hippocampus. In the hippocampus, sound-driven increase of Bdnf exon-VI-YFP levels reaches mossy ﬁber terminals and hippocampal capillaries where an elevation of Bdnf exon-IV-CFP is observed. In addition, Bdnf exon-IV-CFP in the pyramidal layer is activated in response to sound and within CA1 PC that receive more perisomatic, but a reduced number of dendritic PV-positive contacts (Figures 7A–C). Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition Additionally, this decrease in dendritic inhibition may serve as an explanation for increased LTP and learning capability observed after acoustic enrichment and mild acoustic trauma but not after severe acoustic trauma (Figures 3B,C, 6F). Enriching Sound Exposure Indicates Changes in Hippocampal Bdnf Transcription Correlating With Increased Synaptic Plasticity and Improved Memory Acquisition (F) Intensity proﬁle of CFP, YFP, and PV along a line through the CA3 of a control mouse, perpendicular to the SP (length: approximately 390 µm). Representative for four experiments. FIGURE 4 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin upon sound exposure in hippocampal circuits (Supplementary Video 1). (A–C) CFP/YFP ﬂuorescence and parvalbumin immunostaining (PV, red) in dorsal hippocampal mouse slices 2 weeks after exposure to sham, 80, 100, or 120 dB SPL. Note a prominent dynamic change of PV (red) expression in the stratum pyramidale (SP) of the CA1 region (A), in the stratum lucidum (SL) of the CA3 region (n = 6 animals/group) (B), and stratum granulosum (SG) of the dentate gyrus (DG) (C) concomitant to distinct changes in YFP and CFP. SM, stratum moleculare; SR, stratum radiatum; FH, ﬁssura hippocampalis. Scale bars, 100 µm. (D) Quantiﬁcation of Bdnf exon-IV-CFP in capillary outlines at the level of the FH revealed a peak of ﬂuorescence in 80 dB SPL exposed animals that declined below control levels in 120 dB SPL exposed animals. Data represented as mean ± SD. (E) Quantiﬁcation of YFP, CFP, and PV intensities in the CA3 region in an area of 300 × 100 µm following different sound exposure paradigms. Data represented as mean ± SEM. (F) Intensity proﬁle of CFP, YFP, and PV along a line through the CA3 of a control mouse, perpendicular to the SP (length: approximately 390 µm). Representative for four experiments. Frontiers in Molecular Neuroscience \| www.frontiersin.org July 2018 \| Volume 11 \| Article 260 11 Visualizing BDNF During Sound-Induced Plasticity Matt et al. overlapped with PV IR (Figure 5C), indicating that YFP positive terminals not only contact dendrites of pyramidal cells (PCs) but also those of PV-positive interneurons (Figure 5E). This suggests that Bdnf exon-IV translation in mossy ﬁber terminals (Danzer et al., 2008; Zheng et al., 2011; Dieni et al., 2012) in response to enriched or mild traumatic sound is linked to elevated levels of PV in the perisomatic area of CA3 pyramidal neurons (Figure 5B, red). Similarly as observed above, PV IR was strongly increased after acoustic enrichment and mild acoustic trauma, but not after severe acoustic trauma (Figure 5B). Quantiﬁcation of ﬂuorescence intensity in the SP revealed an increase of CFP ﬂuorescence after acoustic enrichment in comparison to controls and a reduction between enriching and severe acoustic trauma. Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells In contrast, a reduction of auditory input after severe acoustic trauma (Figures 7D–F, discontinuous red lined arrow) showed no apparent mobilization of Bdnf exon-VI-YFP and Bdnf exon-IV-CFP transcripts, neither in the brainstem, nor in neuronal or capillary hippocampal cells. Under these conditions, we could not observe compensatory adaptation of sound sensitivity, nor increased hippocampal LTP (Figures 7D–F). These ﬁndings suggest that a crucial level of auditory input drives activity-dependent transcription of BDNF restricted to the ascending auditory pathway and the hippocampus to establish persistent adaptation of the auditory These observations point to an increased perisomatic inhibitory connection in the SP derived from PV-positive July 2018 \| Volume 11 \| Article 260 July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 12 Matt et al. Visualizing BDNF During Sound-Induced Plasticity FIGURE 5 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin expression in the CA3 region following sound exposure. (A,B) Hippocampal CA3 region with magniﬁcation of the stratum lucidum (SL) and stratum pyramidale (SP). (A) Immunostaining of parvalbumin (PV, red). Note the difference in PV expression in the SP in dependence of the applied sound exposure. Scale bars: 10 µm. Lower panel: Triple staining of the same CA3 region. Open arrows indicate Bdnf exon-IV-CFP in blood vessels. Scale bars: 10 µm. (B) Magniﬁed view from solid frame in (A). CFP (open arrows) and PV (red) expression at the level of the SP. Closed arrows indicate Bdnf exon-VI-YFP. Nuclei stained with DAPI (blues). Scale bars: 5 µm. (C) Magniﬁed view from dashed frame in (A). Bdnf exon-VI-positive-YFP expression in mossy ﬁber terminals contacts overlaps with PV-positive (red) interneurons in the SL region of the CA3 and increases (80 dB SPL) or declines (120 dB SPL) with sound. Scale bars: 5 µm. (A–C) n = 6 animals/group. (D) Quantiﬁcation of CFP, YFP, and PV ﬂuorescence averaged over the SL and SP region of the CA3 region. Bdnf exon-IV-CFP ﬂuorescence intensity peaks in animals subjected to 80 dB SPL. Bdnf exon-VI-YFP and Bdnf exon-IV-CFP after 80 dB SPL sound exposure were also signiﬁcantly elevated when compared to 120 dB SPL-exposed animals. Data represented as mean ± SD. (E) Model depicting assumed locations of altered CFP, YFP, and PV expression after 80 dB SPL sound exposure. Mossy ﬁber terminals (yellow) terminate on dendrites of excitatory CA3 pyramidal cells (PC) and PV-positive inhibitory interneuron (IN, red). Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells CFP is found in the perisomatic areas of CA3 PCs (blue). SO, stratum oriens. FIGURE 5 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin expression in the CA3 region following sound exposure. (A,B) Hippocampal CA3 region with magniﬁcation of the stratum lucidum (SL) and stratum pyramidale (SP). (A) Immunostaining of parvalbumin (PV, red). Note the difference in PV expression in the SP in dependence of the applied sound exposure. Scale bars: 10 µm. Lower panel: Triple staining of the same CA3 region. Open arrows indicate Bdnf exon-IV-CFP in blood vessels. Scale bars: 10 µm. (B) Magniﬁed view from solid frame in (A). CFP (open arrows) and PV (red) expression at the level of the SP. Closed arrows indicate Bdnf exon-VI-YFP. Nuclei stained with DAPI (blues). Scale bars: 5 µm. (C) Magniﬁed view from dashed frame in (A). Bdnf exon-VI-positive-YFP expression in mossy ﬁber terminals contacts overlaps with PV-positive (red) interneurons in the SL region of the CA3 and increases (80 dB SPL) or declines (120 dB SPL) with sound. Scale bars: 5 µm. (A–C) n = 6 animals/group. (D) Quantiﬁcation of CFP, YFP, and PV ﬂuorescence averaged over the SL and SP region of the CA3 region. Bdnf exon-IV-CFP ﬂuorescence intensity peaks in animals subjected to 80 dB SPL. Bdnf exon-VI-YFP and Bdnf exon-IV-CFP after 80 dB SPL sound exposure were also signiﬁcantly elevated when compared to 120 dB SPL-exposed animals. Data represented as mean ± SD. (E) Model depicting assumed locations of altered CFP, YFP, and PV expression after 80 dB SPL sound exposure. Mossy ﬁber terminals (yellow) terminate on dendrites of excitatory CA3 pyramidal cells (PC) and PV-positive inhibitory interneuron (IN, red). CFP is found in the perisomatic areas of CA3 PCs (blue). SO, stratum oriens. FIGURE 5 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin expression in the CA3 region following sound exposure. (A,B) Hippocampal CA3 region with magniﬁcation of the stratum lucidum (SL) and stratum pyramidale (SP). (A) Immunostaining of parvalbumin (PV, red). Note the difference in PV expression in the SP in dependence of the applied sound exposure. Scale bars: 10 µm. Lower panel: Triple staining of the same CA3 region. Open arrows indicate Bdnf exon-IV-CFP in blood vessels. Scale bars: 10 µm. (B) Magniﬁed view from solid frame in (A). CFP (open arrows) and PV (red) expression at the level of the SP. Closed arrows indicate Bdnf exon-VI-YFP. Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells (B) High-power magniﬁcation of PV IR in regions framed exemplarily in (A) unravels an increase in PV-immunopositive perisomatic contacts (closed arrows) concomitant to Bdnf exon-IV-CFP expression (open arrows) in a somatic localization at the SP level. Scale bars: 20 µm. (C) PV IR in the SR, see frames in exemplarily region in (A), indicating reduced PV-immunopositive puncta in BLEV reporter mice 2 weeks after exposure to 80 and 100 dB SPL in comparison to control or 120 dB SPL. Scale bars: 100 µm. (A–C) n = 6 animals/group. (D) Quantiﬁcation of PV-immunopositive puncta at the SP level are elevated for 80 and 100 dB SPL exposed animals. Data are represented as mean ± SD. (E) The quantiﬁcation of PV-immunopositive puncta within the SR revealed a signiﬁcant decline of PV-immunopositive dots for 80 and 100 dB SPL- and unchanged levels in 120 dB SPL-exposed animals. Data are represented as mean ± SEM. (F) Abstract ﬁgure of the CA1 region indicating the expression pattern of Arc, PV, CFP, and YFP. L Indicates increased activity. FH, Fissura hippocampalis; SO, Stratum oriens; SP, Stratum pyramidale; SR, Stratum radiatum; PC, pyramidal cell; BC, basket cell; BS, bistratiﬁed cell, SC, Schaffer collaterals. (G) Averaged time courses of picrotoxin wash-in experiments. Representative traces before (black) and after (gray) wash-in of 50 µM picrotoxin are shown on top. Wash-in of picrotoxin leads to an increase in fEPSP amplitude. This disinhibition is signiﬁcantly stronger in controls (272 ± 43% of baseline) and 120 dB SPL (292 ± 59% of baseline) compared to 80 dB SPL (139 ± 7% of baseline) and 100 dB SPL (134 ± 8% of baseline). Data are represented as mean ± SEM. acute behavioral stress paradigm and glucocorticoid receptor antagonists inﬂuence sound-induced suprathreshold auditory nerve responses, indicating that circulating cortisol reaches the cochlea via the bloodstream (Singer et al., 2013, 2018). Moreover, the same behavioral stress paradigm triggered sound-enrichment driven elevated levels of the LTP-associated activity-regulated cytoskeletal protein (Arc) in the hippocampus (Singer et al., 2013). On the other hand, stress paradigms that lead to high corticosterone blood levels, elevate the risk to reduce central auditory responses after acoustic trauma and reduced Arc levels (Singer et al., 2013). Overall, these previous studies support the hypothesis that various kinds of sound exposures inﬂuence long- lasting sensitivity-changes of responses through stress eﬀects on the auditory nerve level. DISCUSSION We here provide evidence that BLEV mice allow speciﬁc identiﬁcation of neurons and capillaries that respond to a critical behaviorally relevant sensory input by elevation of exon-IV and -VI derived Bdnf expression. Our data indicate that a sound- induced and memory-dependent improvement or restoration of sound sensitivity may take place in the auditory pathway and hippocampus. In contrast, an impairment of the critical auditory input after severe auditory trauma apparently leads to a failure to elevate Bdnf transcripts and to adapt to sound sensitivity. Therapeutic concepts for maladaptive diseases therefore need to reconsider BDNF replacement strategies in view of the present ﬁndings. Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells Nuclei stained with DAPI (blues). Scale bars: 5 µm. (C) Magniﬁed view from dashed frame in (A). Bdnf exon-VI-positive-YFP expression in mossy ﬁber terminals contacts overlaps with PV-positive (red) interneurons in the SL region of the CA3 and increases (80 dB SPL) or declines (120 dB SPL) with sound. Scale bars: 5 µm. (A–C) n = 6 animals/group. (D) Quantiﬁcation of CFP, YFP, and PV ﬂuorescence averaged over the SL and SP region of the CA3 region. Bdnf exon-IV-CFP ﬂuorescence intensity peaks in animals subjected to 80 dB SPL. Bdnf exon-VI-YFP and Bdnf exon-IV-CFP after 80 dB SPL sound exposure were also signiﬁcantly elevated when compared to 120 dB SPL-exposed animals. Data represented as mean ± SD. (E) Model depicting assumed locations of altered CFP, YFP, and PV expression after 80 dB SPL sound exposure. Mossy ﬁber terminals (yellow) terminate on dendrites of excitatory CA3 pyramidal cells (PC) and PV-positive inhibitory interneuron (IN, red). CFP is found in the perisomatic areas of CA3 PCs (blue). SO, stratum oriens. July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 13 Matt et al. Visualizing BDNF During Sound-Induced Plasticity Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin after sound exposure in the hippocampal CA1 region. (A) Low-magniﬁcation of n (PV) IR in BLEV reporter mice 2 weeks after exposure to control, 80, 100, or 120 dB SPL. Note the up-regulation of PV IR (arrows) in the SP region (Continued) 6 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin after sound exposure in the hippocampal CA1 region. (A) Low-magn in (PV) IR in BLEV reporter mice 2 weeks after exposure to control, 80, 100, or 120 dB SPL. Note the up-regulation of PV IR (arrows) in the SP FIGURE 6 \| Dynamic changes of Bdnf exon-IV and -VI transcripts and parvalbumin after sound exposure in the hippocampal CA1 region. (A) Low-magniﬁcation of parvalbumin (PV) IR in BLEV reporter mice 2 weeks after exposure to control, 80, 100, or 120 dB SPL. Note the up-regulation of PV IR (arrows) in the SP region (Continued) Frontiers in Molecular Neuroscience \| www.frontiersin.org July 2018 \| Volume 11 \| Article 260 14 Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 6 \| parallel to Bdnf exon-IV-CFP (open arrows) in the vascularized FH region. Scale bars. 100 µm. Acoustic Enrichment but Not Severe Acoustic Trauma Decreases Dendritic Inhibition of CA1 Pyramidal Cells We additionally have to consider that immediate GC-mediated activities operate within minutes to hours after exposure (de Kloet, 2014) preceding BDNF-mediated actions. Accordingly, dependence of BDNF on Arc adaptation processes is documented only for late but not early activities (Chowdhury et al., 2006; Nakayama et al., 2015; Carmichael and Henley, 2018; Epstein and Finkbeiner, 2018). Consequently, the activation of glutamate-receptors by injection of kainaic acid into BLEV mice increased activity-dependent Bdnf transcription within hours (Singer et al., submitted). As both eﬀects occur during a similar timeframe, it is likely that sound exposure may promote the activation of Bdnf promoters IV and VI in the brainstem (Figure 2) following immediate corticosterone- induced accentuation of auditory ﬁber responses. The failure of Bdnf exon-IV-CFP and VI-YFP upregulation in the olfactory bulb after acoustic enrichment indicates that BDNF transcripts are not elevated ubiquitously but in an auditory-speciﬁc manner. While it is well established that activation of the Bdnf exon-IV promoter occurs in response to elevated calcium (Ca2+) levels subsequently to neuronal activity changes (West et al., 2014), it is only modestly activated by neuronal activity (Timmusk et al., 1993; Aid et al., 2007). It may, however, be activated indirectly in response to the same stimulus, through the involvement of the cAMP response element binding-protein (CREB) (Bambah- Mukku et al., 2014). Moreover, as previously proposed, Bdnf exon-IV in platelets may respond to the activity of store-operated calcium channels (Chacón-Fernández et al., 2016) and may thus be inﬂuenced by neuronal activity, which is known to tightly regulate blood ﬂow (Hillman, 2014). Accordingly, a sound exposure-induced modiﬁcation of auditory nerve activity may correspond to the driving force of altered activation of BDNF sensory system. Using the BLEV reporter mouse, we will now be able to identify the neuronal and non-neuronal cell populations within the brainstem and hippocampus that guide this behaviorally relevant adaptation process. Frontiers in Molecular Neuroscience \| www.frontiersin.org Acoustic Enrichment Persistently Increases Responsiveness of the Auditory System Coinciding With Increased Excitability and Translation of Bdnf Exon-IV and exon-VI In the present study, diﬀerent levels of acoustic exposure induced persisting changes in sound-sensitivity that apparently correlated with changes of Bdnf exon-IV-CFP and VI-YFP levels in the brainstem and hippocampus. Considering the driving force for these transcript-speciﬁc BDNF changes we have to take into account that these external sound exposure conditions, here presented by (novel) experience of enriching, mild traumatic, or severe traumatic sound for 40 min, will not only activate the auditory path but will also drive an immediate glucocorticoid (GC)-mediated stress response associated to arousal or fear (De Kloet et al., 2005; de Kloet, 2014; Hermans et al., 2014; Myers et al., 2014). In line with this assumption, the cortisol level is elevated 30 min after sound exposure onwards (Campeau and Watson, 1997). Likewise, elevated cortisol levels inﬂuence IHC synapses contacting postsynaptic auditory ﬁbers as well as sound responses of aﬀerent auditory ﬁbers (Singer et al., 2013). Additionally, an July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org Frontiers in Molecular Neuroscience \| www.frontiersin.org 15 Visualizing BDNF During Sound-Induced Plasticity Matt et al. FIGURE 7 \| Summary of the location pattern of CFP (blue), YFP (yellow), neuronal (PV) and activity markers (Arc) in the hippocampus. Depicted is the situation in response to enriching or mild traumatic events (80/100 dB SPL) (A–C) or to severe traumatic/stressful events (120 dB SPL) (D–F). (A,D) Schematic drawing of the auditory pathway including hippocampal connections (red open arrow). (B,E) Schematic drawing of hippocampus with perforant path (PP). (C,D) Schematic drawing of dendritic (BS) and somatic (BC) inhibitory inputs onto a pyramidal cell (PC). CFP, YFP, VGLUT1 (A, light green), GluA2 (A, light orange), Arc (B red dendrites), PV (B red ﬁlled circles), and the δ subunit containing GABAA receptors in PCs (green dots in C) were shown to be mobilized after exposure to 80 or 100 dB SPL (A–C) but not to 120 dB SPL (D–F). (C,F) The cooperative transcript-speciﬁc BDNF release may drive a long-lasting reduction of dendritic (C, α1 subunit containing GABAA receptors, white small dots) and increase of perisomatic inhibition (C, δ subunit containing GABAA receptors, green small dots) of CA1 PC by PV expressing inhibitory interneurons (BS, BC) and in parallel possibly improve the recruitment of vascularization (B,E, blood vessels). Acoustic Enrichment Persistently Increases Responsiveness of the Auditory System Coinciding With Increased Excitability and Translation of Bdnf Exon-IV and exon-VI This central facilitating or adaptive responsiveness (A–C, solid red arrow indicating increased input activity) is severely hampered when a critical damage of peripheral auditory input occurs (D–F discontinuous red lined arrow). L symbol indicates increased output activity in 80 or 100 dB SPL exposed animals (C). Green dots in (C) represent δ subunit containing GABAA receptors in SP, green dots in (F) represent α1 subunit containing GABAA receptors in SR. AC, auditory cortex; AN, Auditory nerve; BC, basket cell; BS, bistratiﬁed cell; CA3-SC, excitatory input from CA3 Schaffer collaterals; CN, cochlear nucleus; DG, dentate gyrus; EC, entorhinal cortex; FH, ﬁssura hippocampalis; GC, granule cell; IC, inferior colliculus; MF, mossy ﬁbers; SC, Schaffer collaterale; SG, stratum granulare; SL, stratum lucidum; SM, stratum moleculare; SO, stratum oriens; SOC, superior olivary complex; SP, stratum pyramidale; SR, stratum radiatum; SU, subiculum. FIGURE 7 \| Summary of the location pattern of CFP (blue), YFP (yellow), neuronal (PV) and activity markers (Arc) in the hippocampus. Depicted is the situation in response to enriching or mild traumatic events (80/100 dB SPL) (A–C) or to severe traumatic/stressful events (120 dB SPL) (D–F). (A,D) Schematic drawing of the auditory pathway including hippocampal connections (red open arrow). (B,E) Schematic drawing of hippocampus with perforant path (PP). (C,D) Schematic drawing of dendritic (BS) and somatic (BC) inhibitory inputs onto a pyramidal cell (PC). CFP, YFP, VGLUT1 (A, light green), GluA2 (A, light orange), Arc (B red dendrites), PV (B red ﬁlled circles), and the δ subunit containing GABAA receptors in PCs (green dots in C) were shown to be mobilized after exposure to 80 or 100 dB SPL (A–C) but not to 120 dB SPL (D–F). (C,F) The cooperative transcript-speciﬁc BDNF release may drive a long-lasting reduction of dendritic (C, α1 subunit containing GABAA receptors, white small dots) and increase of perisomatic inhibition (C, δ subunit containing GABAA receptors, green small dots) of CA1 PC by PV expressing inhibitory interneurons (BS, BC) and in parallel possibly improve the recruitment of vascularization (B,E, blood vessels). This central facilitating or adaptive responsiveness (A–C, solid red arrow indicating increased input activity) is severely hampered when a critical damage of peripheral auditory input occurs (D–F discontinuous red lined arrow). L symbol indicates increased output activity in 80 or 100 dB SPL exposed animals (C). Increased Activity in the Auditory System Enhances Hippocampal BDNF Transcription and Synaptic Plasticity We also observed a trend for increased usage of Bdnf exon- IV and -VI in the hippocampus similar to the brainstem and accompanied by elevated levels of the AMPAR subunit GluA2, a marker for synaptic activity in the hippocampus (Tanaka et al., 2000), after acoustic enrichment, but not after severe acoustic trauma (Figure 3). Thus, we may assume that BDNF- modiﬁed auditory-speciﬁc excitability changes after enriching or traumatic sound exposure reach the hippocampus. Indeed, neuronal activity of the auditory pathway after sound exposure can propagate from auditory association cortices through the entorhinal cortex (EC) and via the performant path (PP) into the dorsal hippocampus (Munoz-Lopez et al., 2010). Activation of the dorsal hippocampus was shown after stress (Fanselow and Dong, 2010; Kirby et al., 2013) and environmental enrichment (Tanti et al., 2012), processes that also led to increased LTP (Korz and Frey, 2005) or to improved memory (Hullinger et al., 2015). Furthermore, like environmental enrichment (Weinberger, 2003; Chavez et al., 2009), acoustic enrichment improves frequency discrimination in the auditory cortex as well as the sensitivity to quiet sounds (Engineer et al., 2004; Cai et al., 2009; Bose et al., 2010; Centanni et al., 2013) by a mechanism that most likely involves projection pathways of the medial geniculate body (MGB) (Malmierca and Merchan, 2004). Fittingly, the present study shows acoustic enrichment to coincide with elevated LTP and improved performance in a hippocampus-dependent memory task (Figure 3). Moreover, these processes led to Bdnf exon-VI-YFP ﬂuorescence that was most prominently increased in the hippocampal CA3 region, while Bdnf exon-IV-CFP expression was highest in capillaries of the FH and in PC soma (Figure 4). Mossy ﬁber terminals in the CA3 area drive rapid generation and contextualization of episodic memories through elevated neurogenesis and feed-forward inhibition (Donato et al., 2013), for example in response to environmental enrichment. This process fails in post-traumatic stress disorders (Kheirbek et al., 2012; Zaletel et al., 2017). We thus may assume that unlike severe acoustic trauma, acoustic enrichment and mild acoustic trauma improve the generation and contextualization of memory traces because the exposure conditions modify auditory input in such a way that it activates Bdnf transcripts in the brainstem and alters auditory-speciﬁc ascending glutamatergic excitability. This constitutes an auditory-speciﬁc information ﬂow that conveys to the hippocampus through the ascending auditory pathway. Acoustic Enrichment Persistently Increases Responsiveness of the Auditory System Coinciding With Increased Excitability and Translation of Bdnf Exon-IV and exon-VI PV-positive interneurons are also targeted by BDNF-expressing mossy-ﬁbers and there is evidence that PV expressing basket-cell (BC) interneurons are activated by BDNF (Danzer and Mcnamara, 2004; Danzer et al., 2008). The changes in PV expression observed here could therefore be the direct result of excitability changes arising from the auditory input side. This would perfectly explain the all-encompassing correlation of events that diﬀered between sound-exposure paradigms. In addition, altered PV and BDNF expression patterns observed after sound exposure share characteristics of feed-forward inhibition. First, there is a correlation of Bdnf exon- VI-YFP levels in mossy ﬁbers and PV-positive neurons in the CA1 that are likely to correspond to BCs (Figure 6). BCs contact neighboring PCs through perisomatic δ subunit containing GABAA-receptor positive synapses (Klausberger et al., 2003; Glykys et al., 2008). Here we observed Bdnf exon-IV-CFP expressing PCs and PV-positive BCs that express δ subunit containing GABAA-receptors located nearby the PC soma (Figures 7C,F). Future studies might demonstrate recruitment of δ subunit containing GABAA receptors to Bdnf exon-VI- YFP positive synapses in the CA1 pyramidal layer as previously observed (Glykys et al., 2008). Furthermore, we found a reduction of PV and α1 subunit containing GABAA-receptor IR in the SR of animals exposed to acoustic enrichment (80 dB SPL) or mild acoustic trauma (100 dB SPL, Figures 6C,E; Supplementary Figures 5D,F). This may represent BC-mediated inhibition of PV-positive bistratiﬁed (BS) interneurons that target CA1 dendrites through α1 subunit containing GABAA-receptor expressing synapses (Klausberger and Somogyi, 2008; Willadt et al., 2013). The present ﬁndings in BLEV mice do not support a causal link of BDNF levels with PV-mediated feed-forward inhibition. The absence of Bdnf transcript mobilization in the tri-synaptic path together with unaltered PV expression and LTP after the reduction of auditory input by severe acoustic trauma (120 dB SPL, Figures 7D–F), however, indicates that activity- dependent BDNF expression in the hippocampal tri-synaptic path is a consequence of task-speciﬁc BDNF activities in lower auditory brainstem regions. This might be the initial cue to adapt the processing of auditory information within the circuits. It is challenging to consider that during memory-linked adaptation processes the previously shown mossy cell-mediated BDNF- dependent enhancement of dentate granule cell output to the CA3 region (Hashimotodani et al., 2017) might also be activated. The monitoring of these processes is now empowered in BLEV mice. Our ﬁndings thus provide a model for a general mechanism Acoustic Enrichment Persistently Increases Responsiveness of the Auditory System Coinciding With Increased Excitability and Translation of Bdnf Exon-IV and exon-VI Green dots in (C) represent δ subunit containing GABAA receptors in SP, green dots in (F) represent α1 subunit containing GABAA receptors in SR. AC, auditory cortex; AN, Auditory nerve; BC, basket cell; BS, bistratiﬁed cell; CA3-SC, excitatory input from CA3 Schaffer collaterals; CN, cochlear nucleus; DG, dentate gyrus; EC, entorhinal cortex; FH, ﬁssura hippocampalis; GC, granule cell; IC, inferior colliculus; MF, mossy ﬁbers; SC, Schaffer collaterale; SG, stratum granulare; SL, stratum lucidum; SM, stratum moleculare; SO, stratum oriens; SOC, superior olivary complex; SP, stratum pyramidale; SR, stratum radiatum; SU, subiculum. BDNF regulates VGLUT1 expression during development and hippocampal LTP (Melo et al., 2013), and is able to prevent VGLUT1 reduction in cognitive diseases (Anglada-Huguet et al., 2016). This suggests that Bdnf exon-IV-CFP and VI-YFP levels in the brainstem following enriched or severe acoustic trauma may directly drive increased or reduced expression of VGLUT1, respectively. This process needs to be regarded in the context of activity-dependent BDNF which induces strengthening of promoters in brainstem neurons, targeted by auditory nerves. In line with this, adaptations of Bdnf exon-IV-CFP and VI- YFP levels in the brainstem appear to correlate with increased expression of VGLUT1 (Figure 2), a speciﬁc presynaptic marker for auditory-speciﬁc synapses in the brainstem (Zhou et al., 2007). This indicates elevated numbers of active release sites, which were previously associated with greater spike ﬁdelity of auditory speciﬁc synapses (Ngodup et al., 2015). Interestingly, July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 16 Visualizing BDNF During Sound-Induced Plasticity Matt et al. synapses (Kellner et al., 2014) and acts as trans-synaptic messenger to control auditory-speciﬁc excitability. synapses (Kellner et al., 2014) and acts as trans-synaptic messenger to control auditory-speciﬁc excitability. synapses (Kellner et al., 2014) and acts as trans-synaptic messenger to control auditory-speciﬁc excitability. the cochlea in an auditory-speciﬁc manner. This model could explain the simultaneous changes in the expression of the excitatory marker GluA2 and BDNF in the hippocampus that interestingly coincided with the expression of PV, a marker for a wide array of inhibitory interneurons (Kepecs and Fishell, 2014). Particularly fast-spiking PV-positive interneurons serve a crucial function for microcircuit formation (Hu et al., 2014) and feed- forward inhibition (Donato et al., 2013) under control of BDNF (Waterhouse et al., 2012). Frontiers in Molecular Neuroscience \| www.frontiersin.org Increased Activity in the Auditory System Enhances Hippocampal BDNF Transcription and Synaptic Plasticity This assumption is supported by the failure to recruit Bdnf transcript activation in the brainstem and hippocampus following severe acoustic trauma that is also associated with a failure to restore sound sensitivity (ABR wave I and IV reduction, Figure 1). In summary, the changes in Bdnf exon-IV usage in hippocampal projecting neurons and Bdnf exon-VI in mossy In summary, the changes in Bdnf exon-IV usage in hippocampal projecting neurons and Bdnf exon-VI in mossy ﬁbers may be interpreted as a result of a sound-induced alteration of the driving force that spreads along the auditory path from Our ﬁndings thus provide a model for a general mechanism through which sound stimulation is linked to behaviorally relevant alterations in activity of neural networks. Furthermore, our data suggest that the precise nature of auditory experience July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 17 Visualizing BDNF During Sound-Induced Plasticity Matt et al. may sculpt synaptic traﬃc in the auditory brain and its connections, including regions such as the hippocampus that are critical for navigating the environment and mapping memories. We propose that a mechanism of this kind may be relevant to establish critical auditory periods in diﬀerent species including humans (Chen and Yuan, 2015) and may also play an important role in the pathophysiology of human neurodegenerative diseases. Hearing loss has been shown to be associated with an increased risk of dementia in epidemiological studies (Lin et al., 2017), however the nature of this linkage has not been deﬁned. Loss of a critical driving force for recruitment of activity-dependent BDNF expression in neuronal, glial and vascular cells is a plausible mechanism that could contribute to a non-adapting metabolic supply. This may cause accelerated regional brain (in particular, parahippocampal) atrophy (Lin et al., 2014) followed by chronic peripheral hearing impairment such as in older people developing Alzheimer’s disease. The ﬁndings suggest that possible neurodegenerative proteinopathies (Hardy et al., 2016) may not necessarily be overcome by systemic BDNF therapies. In contrast, sustained patterns of neural network activity may promote the spread of speciﬁc proteinopathies (“molecular nexopathies”) (Warren et al., 2013), and BDNF expression might be one mediator of such activity-dependent network degenerations in the setting of proteinopathies, as recently also demonstrated in Alzheimer’s disease (Hardy et al., 2017). ACKNOWLEDGMENTS Thank you to Michael Paolillo for reading the manuscript. SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnmol. 2018.00260/full#supplementary-material FUNDING This work was supported by the Deutsche Forschung sgemeinschaft DFG-Kni-316-10-1 (RP-M, WS, H-SG, MK, and TO); FOR 2060 project RU 713/3-2 (LR and CH); SPP 1608 RU 316/12-1 (PE, KR, H-SG); KN 316/12-1 (MM, UZ, and MK); SPP 1608 FR 1784/17-1 (EF and NM); the Brain and Behavior Research Foundation NARSAD Young Investigator Grant 20748 (LM), BFU2013-40944 (TS); DFG KFO134 (PR and AB). We acknowledge support by Deutsche Forschungsgemeinschaft and Open Access Publishing Fund of University of Tübingen. p g In conclusion, here we demonstrate for the ﬁrst time that translation of exon-IV and -VI derived BDNF is elevated after sound exposure conditions that induce long-lasting changes in sound-sensitivity correlating with increased hippocampal LTP. We verify the BLEV reporter mouse as a model to identify and examine those neurons, non-neuronal glial and capillary cells that in a task-speciﬁc and orchestrated way respond to those environmental changes that induce behavioral relevant adaptation processes. BLEV mice may thus be used to demonstrate that the synchronized activation of BDNF in neurons, glia, and capillaries provides the speciﬁc cues for GC-mediated metabolic support (de Kloet, 2014; Jeanneteau and Arango-Lievano, 2016) in the context of a speciﬁc sensory organ, a hypothesis that needs to be tested Increased Activity in the Auditory System Enhances Hippocampal BDNF Transcription and Synaptic Plasticity From the ﬁnding in the present study it is challenging to speculate that “auditory enrichment” (for example, via regular music listening) might also engage a BDNF-dependent mechanism, with intriguing implications for neuroprotective strategies. in more detail in future studies. To sum up, BLEV mice allow monitoring of which networks and which of their parts are activated and altered to accentuate behaviorally important auditory input (Berlau and Weinberger, 2008; Munoz- Lopez et al., 2010; Kraus and White-Schwoch, 2015). Our ﬁndings suggest a candidate mechanism whereby auditory experience may sculpt neural networks in the ascending auditory pathway and beyond. This in turn has potentially wide-reaching implications for understanding the role of auditory stimulation in promoting normal development of the human auditory brain and the contribution of auditory dysfunction to disease states, notably the neurodegenerative proteinopathies. AUTHOR CONTRIBUTIONS WS, PE, LM, RP-W, H-SG, TO, and MK: Conceptualization. WS, LM, AB, PE, MM, CH, EF, LR, and MK: Analysis. WS, LM, H- SG, AB, PE, MM, NM, and KR: Investigation. WS, RP-W, LM, PE, UZ, LR, TS, and MK: Writing. WS, PR, EF, TO, LR, and MK: Supervision. LM, WS, PR, UZ, LR, and MK: Review and Editing. Barde, Y. A., Edgar, D., and Thoenen, H. (1982). Puriﬁcation of a new neurotrophic factor from mammalian brain. EMBO J. 1, 549–553. REFERENCES T., and Kilgard, M. P. (2013). Cortical speech-evoked response patterns in multiple auditory ﬁelds are correlated with behavioral discrimination ability. J. Neurophysiol. 110, 177–189. doi: 10.1152/jn.00092.2013 Ernfors, P., Lee, K. F., and Jaenisch, R. (1994). Mice lacking brain-derived neurotrophic factor develop with sensory deﬁcits. Nature 368, 147–150. doi: 10.1038/368147a0 Chacón-Fernández, P., Säuberli, K., Colzani, M., Moreau, T., Ghevaert, C., and Barde, Y. A. (2016). Brain-derived neurotrophic factor in megakaryocytes. J. Biol. Chem. 291, 9872–9881. doi: 10.1074/jbc.M116.720029 Fanselow, M. S., and Dong, H. W. (2010). Are the dorsal and ventral hippocampus functionally distinct structures? Neuron 65, 7–19. doi: 10.1016/j.neuron.2009.11.031 Chavez, C. M., Mcgaugh, J. L., and Weinberger, N. M. (2009). The basolateral amygdala modulates speciﬁc sensory memory representations in the cerebral cortex. Neurobiol. Learn. Mem. 91, 382–392. doi: 10.1016/j.nlm.2008.10.010 Ferrini, F., and De Koninck, Y. (2013). Microglia control neuronal network excitability via BDNF signalling. Neural Plast. 2013, 429815. doi: 10.1155/2013/429815 Chen, Z., and Yuan, W. (2015). Central plasticity and dysfunction elicited by aural deprivation in the critical period. Front. Neural. Circuits 9, 26. doi: 10.3389/fncir.2015.00026 Glykys, J., Mann, E. O., and Mody, I. (2008). Which GABAA receptor subunits are necessary for tonic inhibition in the hippocampus? J. Neurosci. 28, 1421–1426. doi: 10.1523/JNEUROSCI.4751-07.2008 Chenaux, G., Matt, L., Hill, T. C., Kaur, I., Liu, X.-B., Kirk, L. M., et al. (2016). Loss of SynDIG1 reduces excitatory synapse maturation but not formation in vivo. eNeuro 3, ENEURO.0130-16.2016. doi: 10.1523/ENEURO.0130- 16.2016 Gray, J. D., Milner, T. A., and Mcewen, B. S. (2013). Dynamic plasticity: the role of glucocorticoids, brain-derived neurotrophic factor and other trophic factors. Neuroscience 239, 214–227. doi: 10.1016/j.neuroscience.2012.08.034 Chourbaji, S., Brandwein, C., Vogt, M. A., Dormann, C., Hellweg, R., and Gass, P. (2008). Nature vs. nurture: can enrichment rescue the behavioural phenotype of BDNF heterozygous mice? Behav. Brain Res. 192, 254–258. doi: 10.1016/j.bbr.2008.04.015 Hardy, C. J. D., Hwang, Y. T., Bond, R. L., Marshall, C. R., Ridha, B. H., Crutch, S. J., et al. (2017). Donepezil enhances understanding of degraded speech in Alzheimer’s disease. Ann. Clin. Transl. Neurol. 4, 835–840. doi: 10.1002/acn3.471 Hardy, C. J., Marshall, C. R., Golden, H. L., Clark, C. N., Mummery, C. J., Griﬃths, T. D., et al. (2016). Hearing and dementia. J. Neurol. 263, 2339–2354. doi: 10.1007/s00415-016-8208-y Chowdhury, S., Shepherd, J. D., Okuno, H., Lyford, G., Petralia, R. S., Plath, N., et al. (2006). Arc/Arg3.1 interacts with the endocytic machinery to regulate AMPA receptor traﬃcking. Neuron 52, 445–459. REFERENCES Synapse 64, 97–110. doi: 10.1002/syn.20710 Dieni, S., Matsumoto, T., Dekkers, M., Rauskolb, S., Ionescu, M. S., Deogracias, R., et al. (2012). BDNF and its pro-peptide are stored in presynaptic dense core vesicles in brain neurons. J. Cell Biol. 196, 775–788. doi: 10.1083/jcb.2012 01038 Bramham, C. R., and Messaoudi, E. (2005). BDNF function in adult synaptic plasticity: the synaptic consolidation hypothesis. Prog. Neurobiol. 76, 99–125. doi: 10.1016/j.pneurobio.2005.06.003 Donato, F., Rompani, S. B., and Caroni, P. (2013). Parvalbumin-expressing basket- cell network plasticity induced by experience regulates adult learning. Nature 504, 272–276. doi: 10.1038/nature12866 Bramham, C. R., Worley, P. F., Moore, M. J., and Guzowski, J. F. (2008). The immediate early gene arc/arg3.1: regulation, mechanisms, and function. J. Neurosci. 28, 11760–11767. doi: 10.1523/JNEUROSCI.3864-08.2008 Donovan, M. J., Lin, M. I., Wiegn, P., Ringstedt, T., Kraemer, R., Hahn, R., et al. (2000). Brain derived neurotrophic factor is an endothelial cell survival factor required for intramyocardial vessel stabilization. Development 127, 4531–4540. Cai, R., Guo, F., Zhang, J., Xu, J., Cui, Y., and Sun, X. (2009). Environmental enrichment improves behavioral performance and auditory spatial representation of primary auditory cortical neurons in rat. Neurobiol. Learn. Mem. 91, 366–376. doi: 10.1016/j.nlm.2009.01.005 Duguid, I., Branco, T., London, M., Chadderton, P., and Hausser, M. (2012). Tonic inhibition enhances ﬁdelity of sensory information transmission in the cerebellar cortex. J. Neurosci. 32, 11132–11143. doi: 10.1523/JNEUROSCI.0460-12.2012 Campeau, S., and Watson, S. J. (1997). Neuroendocrine and behavioral responses and brain pattern of c-fos induction associated with audiogenic stress. J. Neuroendocrinol. 9, 577–588. Edelmann, E., Lessmann, V., and Brigadski, T. (2014). Pre- and postsynaptic twists in BDNF secretion and action in synaptic plasticity. Neuropharmacology 76, 610–627. doi: 10.1016/j.neuropharm.2013.05.043 Carmichael, R. E., and Henley, J. M. (2018). Transcriptional and post-translational regulation of Arc in synaptic plasticity. Semin. Cell. Dev. Biol. 77, 3–9. doi: 10.1016/j.semcdb.2017.09.007 Engineer, N. D., Percaccio, C. R., Pandya, P. K., Moucha, R., Rathbun, D. L., and Kilgard, M. P. (2004). Environmental enrichment improves response strength, threshold, selectivity, and latency of auditory cortex neurons. J. Neurophysiol. 92, 73–82. doi: 10.1152/jn.00059.2004 j Castrén, E., and Rantamäki, T. (2010). The role of BDNF and its receptors in depression and antidepressant drug action: reactivation of developmental plasticity. Dev. Neurobiol. 70, 289–297. doi: 10.1002/dneu.20758 Epstein, I., and Finkbeiner, S. (2018). The arc of cognition: Signaling cascades regulating arc and implications for cognitive function and disease. Semin. Cell Dev. Biol. 77, 63–72. doi: 10.1016/j.semcdb.2017.09.023 Centanni, T. M., Engineer, C. REFERENCES Arango-Lievano, M., Lambert, W. M., Bath, K. G., Garabedian, M. J., Chao, M. V., and Jeanneteau, F. (2015). Neurotrophic-priming of glucocorticoid receptor signaling is essential for neuronal plasticity to stress and antidepressant treatment. Proc. Natl. Acad. Sci. U.S.A. 112, 15737–15742. doi: 10.1073/pnas.1509045112 Aid, T., Kazantseva, A., Piirsoo, M., Palm, K., and Timmusk, T. (2007). Mouse and rat BDNF gene structure and expression revisited. J. Neurosci. Res. 85, 525–535. doi: 10.1002/jnr.21139 Anglada-Huguet, M., Vidal-Sancho, L., Giralt, A., Garcia-Diaz Barriga, G., Xifro, X., and Alberch, J. (2016). Prostaglandin E2 EP2 activation reduces memory decline in R6/1 mouse model of Huntington’s disease by the induction of BDNF-dependent synaptic plasticity. Neurobiol. Dis. 95, 22–34. doi: 10.1016/j.nbd.2015. 09.001 Bambah-Mukku, D., Travaglia, A., Chen, D. Y., Pollonini, G., and Alberini, C. M. (2014). A positive autoregulatory BDNF feedback loop via C/EBPb mediates hippocampal memory consolidation. J. Neurosci. 34, 12547–12559. doi: 10.1523/JNEUROSCI.0324-14.2014 Barde, Y. A., Edgar, D., and Thoenen, H. (1982). Puriﬁcation of a new neurotrophic factor from mammalian brain. EMBO J. 1, 549–553. July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 18 Visualizing BDNF During Sound-Induced Plasticity Matt et al. regulation of both excitation and feedforward inhibition of CA3 pyramidal cells. J. Neurosci. 24, 11346–11355. doi: 10.1523/JNEUROSCI.3846-04.2004 Bath, K. G., Schilit, A., and Lee, F. S. (2013). Stress eﬀects on BDNF expression: eﬀects of age, sex, and form of stress. Neuroscience 239, 149–156. doi: 10.1016/j.neuroscience.2013.01.074 de Fiebre, N. C., Sumien, N., Forster, M. J., and De Fiebre, C. M. (2006). Spatial learning and psychomotor performance of C57BL/6 mice: age sensitivity and reliability of individual diﬀerences. Age (Dordr). 28, 235–253. doi: 10.1007/s11357-006-9027-3 Bausch, A. E., Dieter, R., Nann, Y., Hausmann, M., Meyerdierks, N., Kaczmarek, L. K., et al. (2015). The sodium-activated potassium channel Slack is required for optimal cognitive ﬂexibility in mice. Learn. Mem. 22, 323–335. doi: 10.1101/lm.037820.114 de Kloet, E. R. (2014). From receptor balance to rational glucocorticoid therapy. Endocrinology 155, 2754–2769. doi: 10.1210/en.2014-1048 Berlau, K. M., and Weinberger, N. M. (2008). Learning strategy determines auditory cortical plasticity. Neurobiol. Learn. Mem. 89, 153–166. doi: 10.1016/j.nlm.2007.07.004 De Kloet, E. R., Joëls M., and Holsboer, F. (2005). Stress and the brain: from adaptation to disease. Nat. Rev. Neurosci. 6, 463–475. doi: 10.1038/nrn1683 Bose, M., Muñoz-Llancao, P., Roychowdhury, S., Nichols, J. A., Jakkamsetti, V., Porter, B., et al. (2010). Eﬀect of the environment on the dendritic morphology of the rat auditory cortex. REFERENCES Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci. 24, 677–736. doi: 10.1146/annurev.neuro.24.1.677 Mallei, A., Baj, G., Ieraci, A., Corna, S., Musazzi, L., Lee, F. S., et al. (2015). Expression and dendritic traﬃcking of BDNF-6 splice variant are impaired in knock-in mice carrying human BDNF Val66Met polymorphism. Int. J. Neuropsychopharmacol. 18, pyv069. doi: 10.1093/ijnp/pyv069 Hullinger, R., O’riordan, K., and Burger, C. (2015). Environmental enrichment improves learning and memory and long-term potentiation in young adult rats through a mechanism requiring mGluR5 signaling and sustained activation of p70s6k. Neurobiol. Learn. Mem. 125, 126–134. doi: 10.1016/j.nlm.2015.08.006 Malmierca, M., and Merchan, M. (2004). “The auditory system,” in The Rat Nervous System, ed G. Paxinos (San Diego, CA: Academic Press), 997–1082. Marosi, K., and Mattson, M. P. (2014). BDNF mediates adaptive brain and body responses to energetic challenges. Trends Endocrinol. Metab. 25, 89–98. doi: 10.1016/j.tem.2013.10.006 Jeanneteau, F., and Arango-Lievano, M. (2016). Linking mitochondria to synapses: new insights for stress-related neuropsychiatric disorders. Neural Plast. 2016:3985063. doi: 10.1155/2016/3985063 Marty, S. (2000). Diﬀerences in the regulation of neuropeptide Y, somatostatin and parvalbumin levels in hippocampal interneurons by neuronal activity and BDNF. Prog. Brain Res. 128, 193–202. doi: 10.1016/S0079-6123(00)28017-5 Jonasson, Z. (2005). Meta-analysis of sex diﬀerences in rodent models of learning and memory: a review of behavioral and biological data. Neurosci. Biobehav. Rev. 28, 811–825. doi: 10.1016/j.neubiorev.2004.10.006 Matt, L., Michalakis, S., Hofmann, F., Hammelmann, V., Ludwig, A., Biel, M., et al. (2011). HCN2 channels in local inhibitory interneurons constrain LTP in the hippocampal direct perforant path. Cell. Mol. Life Sci. 68, 125–137. doi: 10.1007/s00018-010-0446-z Kellner, Y., Gödecke N., Dierkes, T., Thieme, N., Zagrebelsky, M., and Korte, M. (2014). The BDNF eﬀects on dendritic spines of mature hippocampal neurons depend on neuronal activity. Front. Synaptic Neurosci. 6:5. doi: 10.3389/fnsyn.2014.00005 Maynard, K. R., Hill, J. L., Calcaterra, N. E., Palko, M. E., Kardian, A., Paredes, D., et al. (2016). Functional role of BDNF production from unique promoters in aggression and serotonin signaling. Neuropsychopharmacology 41, 1943–1955. doi: 10.1038/npp.2015.349 Kepecs, A., and Fishell, G. (2014). Interneuron cell types are ﬁt to function. Nature 505, 318–326. doi: 10.1038/nature12983 Kheirbek, M. A., Klemenhagen, K. C., Sahay, A., and Hen, R. (2012). Neurogenesis and generalization: a new approach to stratify and treat anxiety disorders. Nat. Neurosci. 15, 1613–1620. doi: 10.1038/nn.3262 Melo, C. V., Mele, M., Curcio, M., Comprido, D., Silva, C. G., and Duarte, C. B. (2013). REFERENCES doi: 10.1016/j.neuron.2006.08.033 Hashimotodani, Y., Nasrallah, K., Jensen, K. R., Chávez, AE., Carrera, D., and Castillo, P. E. (2017). LTP at Hilar Mossy cell-dentate granule cell synapses modulates dentate gyrus output by increasing excitation/inhibition balance. Neuron 95, 928–943.e923. doi: 10.1016/j.neuron.2017.07.028 Chumak, T., Rüttiger, L., Lee, S. C., Campanelli, D., Zuccotti, A., Singer, W., et al. (2016). BDNF in lower brain parts modiﬁes auditory ﬁber activity to gain ﬁdelity but increases the risk for generation of central noise after injury. Mol. Neurobiol. 53, 5607–5627. doi: 10.1007/s12035-015-9474-x Hermans, E. J., Henckens, M. J., Joëls M., and Fernandez, G. (2014). Dynamic adaptation of large-scale brain networks in response to acute stressors. Trends Neurosci. 37, 304–314. doi: 10.1016/j.tins.2014.03.006 Danzer, S. C., Kotloski, R. J., Walter, C., Hughes, M., and Mcnamara, J. O. (2008). Altered morphology of hippocampal dentate granule cell presynaptic and postsynaptic terminals following conditional deletion of TrkB. Hippocampus 18, 668–678. doi: 10.1002/hipo.20426 Hill, J. L., Hardy, N. F., Jimenez, D. V., Maynard, K. R., Kardian, A. S., Pollock, C. J., et al. (2016). Loss of promoter IV-driven BDNF expression impacts oscillatory activity during sleep, sensory information processing and fear regulation. Transl. Psychiatry 6, e873. doi: 10.1038/tp.2016.153 Danzer, S. C., and Mcnamara, J. O. (2004). Localization of brain-derived neurotrophic factor to distinct terminals of mossy ﬁber axons implies July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 19 Visualizing BDNF During Sound-Induced Plasticity Matt et al. Hillman, E. M. (2014). Coupling mechanism and signiﬁcance of the BOLD signal: a status report. Annu. Rev. Neurosci. 37, 161–181. doi: 10.1146/annurev-neuro-071013-014111 Liston, C., Cichon, J. M., Jeanneteau, F., Jia, Z., Chao, M. V., and Gan, W. B. (2013). Circadian glucocorticoid oscillations promote learning-dependent synapse formation and maintenance. Nat. Neurosci. 16, 698–705. doi: 10.1038/n n.3387 Hong, E. J., Mccord, A. E., and Greenberg, M. E. (2008). A biological function for the neuronal activity-dependent component of Bdnf transcription in the development of cortical inhibition. Neuron 60, 610–624. doi: 10.1016/j.neuron.2008.09.024 Lu, B., Wang, K. H., and Nose, A. (2009). Molecular mechanisms underlying neural circuit formation. Curr. Opin. Neurobiol. 19, 162–167. doi: 10.1016/j.conb.2009.04.004 Hu, H., Gan, J., and Jonas, P. (2014). Interneurons. Fast-spiking, parvalbumin(+) GABAergic interneurons: from cellular design to microcircuit function. Science 345:1255263. doi: 10.1126/science.1255263 Lyons, M. R., and West, A. E. (2011). Mechanisms of speciﬁcity in neuronal activity-regulated gene transcription. Prog. Neurobiol. 94, 259–295. doi: 10.1016/j.pneurobio.2011.05.003 Huang, E. J., and Reichardt, L. F. (2001). REFERENCES BDNF regulates the expression and distribution of vesicular glutamate transporters in cultured hippocampal neurons. PLoS ONE 8:e53793. doi: 10.1371/journal.pone.0053793 Kirby, E. D., Muroy, S. E., Sun, W. G., Covarrubias, D., Leong, M. J., Barchas, L. A., et al. (2013). Acute stress enhances adult rat hippocampal neurogenesis and activation of newborn neurons via secreted astrocytic FGF2. Elife 2, e00362. doi: 10.7554/eLife.00362 Messaoudi, E., Kanhema, T., Soule, J., Tiron, A., Dagyte, G., Da Silva, B., et al. (2007). Sustained Arc/Arg3.1 synthesis controls long- term potentiation consolidation through regulation of local actin polymerization in the dentate gyrus in vivo. J. Neurosci. 27, 10445–10455. doi: 10.1523/JNEUROSCI.2883-07.2007 Klausberger, T., Magill, P. J., Marton, L. F., Roberts, J. D., Cobden, P. M., Buzsaki, G., et al. (2003). Brain-state- and cell-type-speciﬁc ﬁring of hippocampal interneurons in vivo. Nature 421, 844–848. doi: 10.1038/nature01374 Klausberger, T., and Somogyi, P. (2008). Neuronal diversity and temporal dynamics: the unity of hippocampal circuit operations. Science 321, 53–57. doi: 10.1126/science.1149381 Minichiello, L. (2009). TrkB signalling pathways in LTP and learning. Nat. Rev. Neurosci. 10, 850–860. doi: 10.1038/nrn2738 Mitre, M., Mariga, A., and Chao, M. V. (2017). Neurotrophin signalling: novel insights into mechanisms and pathophysiology. Clin. Sci. (Lond). 131, 13–23. doi: 10.1042/CS20160044 Knipper, M., Panford-Walsh, R., Singer, W., Rüttiger, L., and Zimmermann, U. (2015). Speciﬁc synaptopathies diversify brain responses and hearing disorders: you lose the gain from early life. Cell Tissue Res. 361, 77–93. doi: 10.1007/s00441-015-2168-x Miyamoto, N., Pham, L. D., Seo, J. H., Kim, K. W., Lo, E. H., and Arai, K. (2014). Crosstalk between cerebral endothelium and oligodendrocyte. Cell. Mol. Life Sci. 71, 1055–1066. doi: 10.1007/s00018-013-1488-9 Knipper, M., Van Dijk, P., Nunes, I., Rüttiger, L., and Zimmermann, U. (2013). Advances in the neurobiology of hearing disorders: recent developments regarding the basis of tinnitus and hyperacusis. Prog. Neurobiol. 111, 17–33. doi: 10.1016/j.pneurobio.2013.08.002 Munoz-Lopez, M. M., Mohedano-Moriano, A., and Insausti, R. (2010). Anatomical pathways for auditory memory in primates. Front. Neuroanat. 4:129. doi: 10.3389/fnana.2010.00129 Korz, V., and Frey, J. U. (2005). Bidirectional modulation of hippocampal long-term potentiation under stress and no-stress conditions in basolateral amygdala-lesioned and intact rats. J. Neurosci. 25, 7393–7400. doi: 10.1523/JNEUROSCI.0910-05.2005 Myers, B., Mcklveen, J. M., and Herman, J. P. (2014). Glucocorticoid actions on synapses, circuits, and behavior: implications for the energetics of stress. Front. Neuroendocrinol. 35, 180–196. doi: 10.1016/j.yfrne.2013.12.003 Nahmani, M., and Turrigiano, G. G. (2014). Adult cortical plasticity following injury: recapitulation of critical period mechanisms? Neuroscience 283, 4–16. REFERENCES Neurotrophins: transcription and translation. Handb. Exp. Pharmacol. 220, 67–100. doi: 10.1007/978-3-642-45106-5_4 Serra-Millàs, M. (2016). Are the changes in the peripheral brain-derived neurotrophic factor levels due to platelet activation? World J. Psychiatry 6, 84–101. doi: 10.5498/wjp.v6.i1.84 Willadt, S., Nenniger, M., and Vogt, K. E. (2013). Hippocampal feedforward inhibition focuses excitatory synaptic signals into distinct dendritic compartments. PLoS ONE 8:e80984. doi: 10.1371/journal.pone.0080984 Singer, W., Geisler, H. S., Panford-Walsh, R., and Knipper, M. (2016). Detection of excitatory and inhibitory synapses in the auditory system using ﬂuorescence immunohistochemistry and high-resolution ﬂuorescence microscopy. Methods Mol. Biol. 1427, 263–276. doi: 10.1007/978-1-4939-3615-1_15 Yamada, K., and Nabeshima, T. (2003). Brain-derived neurotrophic factor/TrkB signaling in memory processes. J. Pharmacol. Sci. 91, 267–270. doi: 10.1254/jphs.91.267 Singer, W., Kasini, K., Manthey, M., Eckert, P., Armbruster, P., Vogt, M. A., et al. (2018). The glucocorticoid antagonist mifepristone attenuates sound-induced long-term deﬁcits in auditory nerve response and central auditory processing in female rats. FASEB J. 32, 3005–3019. doi: 10.1096/fj.201701041RRR Zaletel, I., Filipovi´c, D, Puškaš, N. (2017). Hippocampal BDNF in physiological conditions and social isolation. Rev. Neurosci. 28, 675–692. doi: 10.1515/revneuro-2016-0072 Singer, W., Zuccotti, A., Jaumann, M., Lee, S. C., Panford-Walsh, R., Xiong, H., et al. (2013). Noise-induced inner hair cell ribbon loss disturbs central arc mobilization: a novel molecular paradigm for understanding tinnitus. Mol. Neurobiol. 47, 261–279. doi: 10.1007/s12035-012-8372-8 Zheng, K., An, J. J., Yang, F., Xu, W., Xu, Z. Q., Wu, J., et al. (2011). TrkB signaling in parvalbumin-positive interneurons is critical for gamma-band network synchronization in hippocampus. Proc. Natl. Acad. Sci. U.S.A. 108, 17201–17206. doi: 10.1073/pnas.1114241108 Tan, J., Rüttiger, L., Panford-Walsh, R., Singer, W., Schulze, H., Kilian, S. B., et al. (2007). Tinnitus behavior and hearing function correlate with the reciprocal expression patterns of BDNF and Arg3.1/arc in auditory neurons following acoustic trauma. Neuroscience 145, 715–726. doi: 10.1016/j.neuroscience.2006.11.067 Zhou, J., Nannapaneni, N., and Shore, S. (2007). Vessicular glutamate transporters 1 and 2 are diﬀerentially associated with auditory nerve and spinal trigeminal inputs to the cochlear nucleus. J. Comp. Neurol. 500, 777–787. doi: 10.1002/cne.21208 Tanaka, H., Grooms, S. Y., Bennett, M. V., and Zukin, R. S., (2000). The AMPAR subunit GluR2: still front and center-stage. Brain Res. 886, 190–207. doi: 10.1016/S0006-8993(00)02951-6 Zuccotti, A., Kuhn, S., Johnson, S. L., Franz, C., Singer, W., Hecker, D., et al. (2012). Lack of brain-derived neurotrophic factor hampers inner hair cell synapse physiology, but protects against noise-induced hearing loss. J. Neurosci. 32, 8545–8553. REFERENCES doi: 10.1101/gad.217902 Vorhees, C. V., and Williams, M. T. (2006). Morris water maze: procedures for assessing spatial and related forms of learning and memory. Nat. Protoc. 1, 848–858. doi: 10.1038/nprot.2006.116 Wall, M. J., and Corrêa, SAL. (2017). The mechanistic link between Arc/Arg3.1 expression and AMPA receptor endocytosis. Semin. Cell Dev. Biol. 77, 17–24. doi: 10.1016/j.semcdb.2017.09.005 Rauskolb, S., Zagrebelsky, M., Dreznjak, A., Deogracias, R., Matsumoto, T., Wiese, S., et al. (2010). Global deprivation of brain-derived neurotrophic factor in the CNS reveals an area-speciﬁc requirement for dendritic growth. J. Neurosci. 30, 1739–1749. doi: 10.1523/JNEUROSCI.5100-09.2010 Wardle, R. A., and Poo, M. M. (2003). Brain-derived neurotrophic factor modulation of GABAergic synapses by postsynaptic regulation of chloride transport. J. Neurosci. 23, 8722–8732. doi: 10.1523/JNEUROSCI.23-25-08722.2003 Rios, M., Fan, G., Fekete, C., Kelly, J., Bates, B., Kuehn, R., et al. (2001). Conditional deletion of brain-derived neurotrophic factor in the postnatal brain leads to obesity and hyperactivity. Mol. Endocrinol. 15, 1748–1757. doi: 10.1210/mend.15.10.0706 Warren, J. D., Rohrer, J. D., Schott, J. M., Fox, N. C., Hardy, J., and Rossor, M. N. (2013). Molecular nexopathies: a new paradigm of neurodegenerative disease. Trends Neurosci. 36, 561–569. doi: 10.1016/j.tins.2013.06.007 Rüttiger, L., Singer, W., Panford-Walsh, R., Matsumoto, M., Lee, S. C., Zuccotti, A., et al. (2013). The reduced cochlear output and the failure to adapt the central auditory response causes tinnitus in noise exposed rats. PLoS ONE 8:e57247. doi: 10.1371/journal.pone.0057247 Waterhouse, E. G., An, J. J., Oreﬁce, L. L., Baydyuk, M., Liao, G. Y., Zheng, K., et al. (2012). BDNF promotes diﬀerentiation and maturation of adult- born neurons through GABAergic transmission. J. Neurosci. 32, 14318–14330. doi: 10.1523/JNEUROSCI.0709-12.2012 Sakata, K., Jin, L., and Jha, S. (2010). Lack of promoter IV-driven BDNF transcription results in depression-like behavior. Genes Brain Behav. 9, 712–721. doi: 10.1111/j.1601-183X.2010.00605.x Waterhouse, E. G., and Xu, B. (2009). New insights into the role of brain-derived neurotrophic factor in synaptic plasticity. Mol. Cell. Neurosci. 42, 81–89. doi: 10.1016/j.mcn.2009.06.009 Sakata, K., Woo, N. H., Martinowich, K., Greene, J. S., Schloesser, R. J., Shen, L., et al. (2009). Critical role of promoter IV-driven BDNF transcription in GABAergic transmission and synaptic plasticity in the prefrontal cortex. Proc. Natl. Acad. Sci. U.S.A. 106, 5942–5947. doi: 10.1073/pnas.0811431106 Weinberger, N. M. (2003). The nucleus basalis and memory codes: auditory cortical plasticity and the induction of speciﬁc, associative behavioral memory. Neurobiol. Learn. Mem. 80, 268–284. doi: 10.1016/S1074-7427(03)00072-8 West, A. E., Pruunsild, P., and Timmusk, T. (2014). REFERENCES doi: 10.1016/j.neuroscience.2014.04.029 Kraus, N., and White-Schwoch, T. (2015). Unraveling the biology of auditory learning: a cognitive-sensorimotor-reward framework. Trends Cogn. Sci. 19, 642–654. doi: 10.1016/j.tics.2015.08.017 Nakayama, D., Iwata, H., Teshirogi, C., Ikegaya, Y., Matsuki, N., and Nomura, H. (2015). Long-delayed expression of the immediate early gene Arc/Arg3.1 reﬁnes neuronal circuits to perpetuate fear memory. J. Neurosci. 35, 819–830. doi: 10.1523/JNEUROSCI.2525-14.2015 Kuipers, S. D., Trentani, A., Tiron, A., Mao, X., Kuhl, D., and Bramham, C. R. (2016). BDNF-induced LTP is associated with rapid Arc/Arg3.1- dependent enhancement in adult hippocampal neurogenesis. Sci. Rep. 6:21222. doi: 10.1038/srep21222 Ngodup, T., Goetz, J. A., Mcguire, B. C., Sun, W., Lauer, A. M., and Xu-Friedman, M. A. (2015). Activity-dependent, homeostatic regulation of neurotransmitter release from auditory nerve ﬁbers. Proc. Natl. Acad. Sci. U.S.A. 112, 6479–6484. doi: 10.1073/pnas.1420885112 Lin, F. R., Ferrucci, L., An, Y., Goh, J. O., Doshi, J., Metter, E. J., et al. (2014). Association of hearing impairment with brain volume changes in older adults. Neuroimage 90, 84–92. doi: 10.1016/j.neuroimage.2013.12.059 Pariante, C. M. (2009). Risk factors for development of depression and psychosis. Glucocorticoid receptors and pituitary implications for treatment with antidepressant and glucocorticoids. Ann. N. Y. Acad. Sci. 1179, 144–152. doi: 10.1111/j.1749-6632.2009.04978.x Lin, V. Y., Chung, J., Callahan, B. L., Smith, L., Gritters, N., Chen, J. M., et al. (2017). Development of cognitive screening test for the severely hearing impaired: Hearing-impaired MoCA. Laryngoscope 127, S4–S11. doi: 10.1002/lary.26590 July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 20 Visualizing BDNF During Sound-Induced Plasticity Matt et al. Park, H., and Poo, M. M. (2013). Neurotrophin regulation of neural circuit development and function. Nat. Rev. Neurosci. 14, 7–23. doi: 10.1038/nrn3379 Vanevski, F., and Xu, B. (2013). Molecular and neural bases underlying roles of BDNF in the control of body weight. Front. Neurosci. 7:37. doi: 10.3389/fnins.2013.00037 Parkhurst, C. N., Yang, G., Ninan, I., Savas, J. N., Yates, J. R. III, et al. (2013). Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor. Cell 155, 1596–1609. doi: 10.1016/j.cell.2013.11.030 van Praag, H., Christie, B. R., Sejnowski, T. J., and Gage, F. H. (1999). Running enhances neurogenesis, learning, and long-term potentiation in mice. Proc. Natl. Acad. Sci. U.S.A. 96, 13427–13431. doi: 10.1073/pnas.96.23.13427 Postigo, A., Calella, A. M., Fritzsch, B., Knipper, M., Katz, D., Eilers, A., et al. (2002). Distinct requirements for TrkB and TrkC signaling in target innervation by sensory neurons. Genes Dev. 16, 633–645. REFERENCES doi: 10.1523/JNEUROSCI.1247-12.2012 Tanti, A., Rainer, Q., Minier, F., Surget, A., and Belzung, C. (2012). Diﬀerential environmental regulation of neurogenesis along the septo- temporal axis of the hippocampus. Neuropharmacology 63, 374–384. doi: 10.1016/j.neuropharm.2012.04.022 Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V., Saarma, M., et al. (1993). Multiple promoters direct tissue-speciﬁc expression of the rat BDNF gene. Neuron 10, 475–489. doi: 10.1016/0896-6273(93)90335-O Copyright © 2018 Matt, Eckert, Panford-Walsh, Geisler, Bausch, Manthey, Müller, Harasztosi, Rohbock, Ruth, Friauf, Ott, Zimmermann, Rüttiger, Schimmang, Knipper and Singer. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Copyright © 2018 Matt, Eckert, Panford-Walsh, Geisler, Bausch, Manthey, Müller, Harasztosi, Rohbock, Ruth, Friauf, Ott, Zimmermann, Rüttiger, Schimmang, Knipper and Singer. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Tuvikene, J., Pruunsild, P., Orav, E., Esvald, E. E., and Timmusk, T. (2016). AP-1 transcription factors mediate BDNF-positive feedback loop in cortical neurons. J. Neurosci. 36, 1290–1305. doi: 10.1523/JNEUROSCI.3360-15.2016 Vaghi, V., Polacchini, A., Baj, G., Pinheiro, V. L., Vicario, A., and Tongiorgi, E. (2014). Pharmacological proﬁle of brain-derived neurotrophic factor (BDNF) splice variant translation using a novel drug screening assay: a “quantitative code.” J. Biol. Chem. 289, 27702–27713. doi: 10.1074/jbc.M114.586719 July 2018 \| Volume 11 \| Article 260 Frontiers in Molecular Neuroscience \| www.frontiersin.org 21
https://openalex.org/W4361885555	https://figshare.com/articles/journal_contribution/Figure_S1_from_Immune_Activation_in_Early-Stage_Non_Small_Cell_Lung_Cancer_Patients_Receiving_Neoadjuvant_Chemotherapy_Plus_Ipilimumab/22462917/1/files/39914217.pdf	English	null	Figure S2 from Immune Activation in Early-Stage Non–Small Cell Lung Cancer Patients Receiving Neoadjuvant Chemotherapy Plus Ipilimumab	null	2,023	cc-by	594	Supplementary Figure 1. Supplementary Figure 1. Supplementary Figure 1. Spaghetti Plots of CD4+HLA-DR+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 5 10 15 20 25 % Spaghetti Plots of CD4+CTLA-4+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 10 20 30 40 50 % Spaghetti Plots of CD4+PD-1+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 20 % Spaghetti Plots of CD4+ICOS+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 10 20 30 % Spaghetti Plots of CD4-ICOS+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 5 10 15 20 25 30 % Spaghetti Plots of CD4-PD-1+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 20 25 % Spaghetti Plots of CD4-HLA-DR+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 20 40 60 % Spaghetti Plots of CD4-CTLA-4+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 % A. C. G. E. B. D. H. F. pp y g Spaghetti Plots of CD4-ICOS+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 5 10 15 20 25 30 % B. A. Spaghetti Plots of CD4+ICOS+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 10 20 30 % A. B. Supplementary Figure 1. Spaghetti Plots of CD4-HLA-DR+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 20 40 60 % D. C. Spaghetti Plots of CD4+HLA-DR+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 5 10 15 20 25 % C. E. Spaghetti Plots of CD4+CTLA-4+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 10 20 30 40 50 % E. F. Spaghetti Plots of CD4-CTLA-4+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 % Spaghetti Plots of CD4+PD-1+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 20 % G. Spaghetti Plots of CD4-PD-1+ by Visit Stratified by Overall Response Visit Response = PR Response = SD Response = PD 1 2 3 1 2 3 1 2 3 0 5 10 15 20 25 % G. H.
https://openalex.org/W3095802530	https://www.mdpi.com/2076-393X/8/4/659/pdf?version=1604580652	English	null	Potential of Anti-MUC1 Antibodies as a Targeted Therapy for Gastrointestinal Cancers	Vaccines	2,020	cc-by	13,534	Received: 30 September 2020; Accepted: 3 November 2020; Published: 5 November 2020 Abstract: Gastrointestinal cancers (GI) account for 26% of cancer incidences globally and 35% of all cancer-related deaths. The main challenge is to target cancer speciﬁc antigens. Mucins are heavily O-glycosylated proteins overexpressed in diﬀerent cancers. The transmembrane glycoprotein MUC1 is the most likeable target for antibodies, owing to its speciﬁc overexpression and aberrant glycosylation in many types of cancers. For the past 30 years, MUC1 has remained a possible diagnostic marker and therapeutic target. Despite initiation of numerous clinical trials, a comprehensively eﬀective therapy with clinical beneﬁt is yet to be achieved. However, the interest in MUC1 as a therapeutic target remains unaltered. For all translational studies, it is important to incorporate updated relevant research ﬁndings into therapeutic strategies. In this review we present an overview of the antibodies targeting MUC1 in GI cancers, their potential role in immunotherapy (i.e., antibody-drug and radioimmunoconjugates, CAR-T cells), and other novel therapeutic strategies. We also present our perspectives on how the mechanisms of action of diﬀerent anti-MUC1 antibodies can target speciﬁc hallmarks of cancer and therefore be utilized as a combination therapy for better clinical outcomes. Keywords: MUC1; immunotherapy; monoclonal antibody; gastrointestinal cancers; CAR-T cells Potential of Anti-MUC1 Antibodies as a Targeted Therapy for Gastrointestinal Cancers Mukulika Bose * and Pinku Mukherjee Department of Biological Sciences, University of North Carolina, Charlotte, NC 28223, USA; pmukherj@u * Correspondence: mbose@uncc.edu Review Potential of Anti-MUC1 Antibodies as a Targeted Therapy for Gastrointestinal Cancers Mukulika Bose * and Pinku Mukherjee Department of Biological Sciences, University of North Carolina, Charlotte, NC 28223, USA; pmukherj@uncc.edu * Correspondence: mbose@uncc.edu R i d 30 S t b 2020 A t d 3 N b 2020 P bli h d 5 N b 2020 1. Global Burden of GI Cancers Gastrointestinal (GI) cancers collectively refer to cancers of the esophagus and stomach (gastroesophageal cancers), the colon and rectum (colorectal cancers), pancreas, liver, gallbladder, small intestine, appendix, and anus. Following lung cancer (18.4%), colorectal cancer (9.2%), stomach cancer (8.2%), and liver cancer (8.2%) form the leading causes of cancer-related deaths worldwide [1]. According to the American Cancer Society (ACS) (www.cancer.org), gastrointestinal (GI) cancers have the highest incidence and are the second leading cause of cancer-related deaths in the United States. Esophageal cancer is the seventh most commonly diagnosed cancer and the 6th leading cause of cancer-related deaths worldwide [1]. It is often detected late and there are usually no early symptoms. The overall ﬁve-year survival rate for advanced esophageal cancer in the United States is about 15% [2]. Stomach cancer, or gastric cancer, is the ﬁfth most common cancer in the world and the second highest cause of cancer-related deaths globally [3]. Pancreatic Cancer is the twelfth most common cancer globally and the seventh leading cause of cancer-related deaths [1]. However, in the US it is the third leading cause of cancer-related deaths and is projected to become the second by the end of the year 2020. Most of the pancreatic tumors are detected at a very advanced stage thus making it a lethal disease. It has a dismal 5% 5-year survival rate globally, a mean life expectancy of <6 months, and a high degree of resistance to standard therapy. In the US the ﬁve-year survival rate is 9%, which is the lowest of all major cancers. Liver cancer is the sixth most commonly diagnosed cancer and the fourth leading cause of cancer-related deaths worldwide [1]. Colorectal cancer is the third most common cancer worldwide and the second leading cause of cancer mortality [1]. Vaccines 2020, 8, 659; doi:10.3390/vaccines8040659 www.mdpi.com/journal/vaccines 2 of 21 Vaccines 2020, 8, 659 Chemotherapy and radiation therapy alone or in combination with surgery remain the main modes of treatment so far. However, various immunotherapies are undergoing trials with monoclonal antibodies, combination therapies, CAR-T cell, dendritic cell therapies etc. In the last 40 years, the incidence and mortality of GI cancers have only increased without improvement in therapy. The main challenge is to target speciﬁc antigens that are not expressed in normal tissues. Mucins have always been shown to be key immunological players in various chronic and infectious diseases including cancer. 2.1. Structure of MUC1 Mucins are high molecular weight glycoproteins and their main function is to lubricate epithelial cell surfaces and protect them against invading pathogens [4]. Mucins are broadly divided into secretory gel-forming mucins (MUC2, MUC5AC, MUC5B, MUC6, MUC7 and MUC19, as protective barriers for underlying mucosal cells) and membrane-bound mucins (MUC1, MUC3A, MUC3B, MUC4, MUC12, MUC13, MUC15, MUC16, MUC17, and MUC20) that have a transmembrane, N-terminal extracellular domain (ECD), and a C-terminal cytoplasmic tail. Secretory gel-forming mucins work as protective barriers for underlying mucosal cells, while membrane-bound mucins also play a key role in cell signaling pathways and cellular interactions [4–6]. g g p y Mucin 1 or MUC1 (also known as episialin, PEM, EMA, H23Ag, MCA, and CA15-3) was the ﬁrst transmembrane mucin to be identiﬁed and structurally characterized [7–10]. MUC1 is a single pass type I transmembrane glycoprotein with a hyperglycosylated extracellular N- terminal domain that extends up to 200–500 nm from the cell surface [11,12]. Normally, MUC1 is expressed on the apical surface of glandular or luminal epithelial cells of almost all tissues including the mammary gland, stomach, lungs, esophagus, duodenum, pancreas, uterus, prostate, and the hematopoietic cells [13,14]. In healthy tissues, the extended hyperglycosylated branches of MUC1 create a physical barrier and prevent pathogenic access, thus protecting the underlying epithelia [15,16]. The extended sugar branches form a mucinous gel by oligomerization and protect the underlying epithelia from desiccation, pH changes, and invading microbes [17]. During translation, MUC1 is cleaved [18,19], and the extracellular domain with tandem repeats (25–100) is bound to the membrane by noncovalent interaction with the C-terminal domain of MUC1 (MUC1-CD) that consists of a short extracellular domain (ED), the transmembrane domain (TM) and the cytoplasmic domain (MUC1-CT). The MUC1 gene encodes a single polypeptide chain which is cleaved by auto-proteolysis process at a sea -urchin sperm protein enterokinase and agrin (SEA) domain to generate two peptide fragments and heterodimeric MUC1 [11,20]. The β subunit or MUC1-C contains a C-terminal cytoplasmic domain (MUC1-CT) with 69 amino acids, a hydrophobic transmembrane domain (TMD) with 28 amino acids and a short extracellular domain (ECD) with 58-amino acids that is noncovalently attached to the N-terminal extracellular domain (MUC1-N) or α subunit [21]. The cytoplasmic tail of MUC1 (MUC1-CT) aids in signal transduction [17,22]. Among diﬀerent types of glycosylation, O- and N-glycosylations dominate in MUC1 [23]. 1. Global Burden of GI Cancers In this review, we will provide a detailed overview of various immunotherapies developed against the mucin protein MUC1 in GI cancers including monoclonal antibodies, CAR-T cells and bi-speciﬁc antibodies that have successfully been through preclinical and clinical trials. We will also provide perspectives on how some of these antibodies target speciﬁc hallmarks of cancer so that they can be combined with other drugs for better outcomes in the clinic. 2.1. Structure of MUC1 The MUC1-N subunit in normal cells, consists of a heavily O-glycosylated- VNTR (variable number of tandem repeat) sequence of 20–21 amino acids (PDTRPAPGSTAPPAHGVTSA), which masks the peptide core and protects it from cleavage by proteolytic enzymes, and also prevents it from undergoing clathrin-mediated endocytosis [24]. The molecular weight of MUC1 can vary between 250–500 kDa based on the percentage of glycosylation (in the range of 50–90% of its molecular mass) and the number of tandem repeats [25]. N-glycosylation of MUC1 occurs at ﬁve potential sites, one in the ECD of MUC1-CD, and four in the degenerate repeat of MUC1-N [8]. N-glycosylation patterns are important Vaccines 2020, 8, 659 3 of 21 for MUC1 folding, sorting, apical expression and secretion, whereas O-glycosylation is crucial for its biological properties [26,27]. MUC1 glycosylation depends on the tissue of origin and is regulated by a large number of glycosyltransferases. O-glycosylation is initiated by adding N-acetyl-galactosamine (GalNAc) to the VNTR region highly rich in threonine (Thr) and serine (Ser) residues. Following that, a large family of up to 20 distinct polypeptide GalNAc transferases (ppGalNAc-Ts) form the initial O-linked GalNAcα-Ser/Thr structure (Tn antigen) in the endoplasmic reticulum (ER) and ER-Golgi compartments. This forms the initial O-linked GalNAcα-Ser/Thr structure (Tn antigen) [28]. Following the formation of Tn antigen, GalNAc residue can be further modiﬁed by various distinct glycosyltransferases and construct diﬀerent glycan structures of core 1 also known as T or TF (Thompson-Friedenreich) antigen (by addition of Gal residue) and core 3 (by adding GlcNAcβ1-3GalNAcα) and Sialyl-Tn antigen (STn, by addition of sialic acid residue). Glycosylation continues by extension and chain termination by the addition of carbohydrates such as sialic acid [28–30]. However, in cancer cells, MUC1 mostly displays hypoglycosylation of the core glycans, like sialyation of Tn and T antigens via sialyltransferase enzymes that lead to premature chain termination [30–34]. MUC1 expression has been shown to be up to 10 times higher in many human carcinomas than in normal tissues, which provides resistance to chemotherapy [34–36]. Therefore, antibodies against tumor associated MUC1 are more likely to bind to the antigen on the surface of tumor cells and not MUC1 on the surface of normal cells. This makes tMUC1 a top molecular target to both detect cancers as well as design antibodies against the altered glycopeptide epitopes in the TR domain. 2.1. Structure of MUC1 These antibodies are also used to design human T cells to target tMUC1, called Chimeric Antigen Receptor T-cells (CAR T cells) [37–39]. 2.2. Role in GI Tumors MUC1 is overexpressed and aberrantly glycosylated in most human epithelial cancers [40]. The aberrantly glycosylated MUC1 expressed on malignant cells, called the tumor associated MUC1 or tMUC1 renders usually inaccessible MUC1 epitopes open to detection. MUC1 has been a molecule of interest for immunotherapy for a long time. It is a highly overexpressed cell surface antigen and has altered glycosylation in tumors [41]. However, MUC1 has been shown to play a paradoxical role following infections, acting as an anti-inﬂammatory molecule in healthy cells and as a pro-inﬂammatory molecule in cancer cells [42]. In 2009, the National Cancer Institute (NCI) had ranked tMUC1 as the second most targetable antigen out of 75 for developing cancer vaccines [43]. g g p g MUC1 has been reported to play a role in tumorigenesis by inhibition of cell death and promotion of metastasis [44–46]. MUC1 induces signaling through its cytoplasmic domain (MUC1-CT) and binds to the EGFR family of growth factor tyrosine kinases and enhances signaling through ERK activation and cell proliferation [47]. MUC1-CT interacts with β-catenin, stabilizes it and co-activates Wnt signaling [48]. MUC1 overexpression and its interactions with p53 and FO × O3a transcription factor dampen drug-induced apoptosis and resist oxidative cell damage [49,50]. MUC1 also reduces pro-apoptotic signaling via the heat shock protein (HSP) 90, PI3K/Akt and Caspase-8 pathways [45,51,52]. An increase in depolarized MUC1 leads to the disruption of the normal cell-cell and cell-matrix adhesion and increase in cell-endothelial adhesion, allowing increased metastasis in preclinical models [53]. The hypoglycosylated tMUC1 has increased interaction with cell adhesion molecules ICAM-1 and E-selectin, both of which can improve cellular migration and vascular invasion [54]. MUC1 confers drug resistance in pancreatic ductal adenocarcinoma cells by upregulating multidrug resistance genes [55]. MUC1 has also been reported to increase metastasis through the induction of platelet-derived growth factor (PDGF-A) expression by hypoxia inducible factor (HIF)1-α [56] and leads to epithelial-to-mesenchymal transition in pancreatic cancer [57,58]. MUC1 has also been shown to regulate function of transforming growth factor-β (TGF-β) and switch it from a tumor suppressor to a tumor promoter in PDA cells [59,60]. MUC1 is a prognostic factor that marks poor outcome in gastric cancer patients [61]. Expression of MUC1 has also been reported to be signiﬁcantly correlated to metastasis in colorectal cancer [44]. 2.2. Role in GI Tumors 4 of 21 4 of 21 Vaccines 2020, 8, 659 Vaccines 2020, 8, x Overexpression in multiple epithelial tumors, expression all over the surface of a tumor cell due to loss of apicobasal polarity in cancer cells, thus making it accessible to antibodies and tumor-speciﬁc aberrant glycosylation with truncated carbohydrate antigens Tn and TF in the VNTR region are features that make MUC1 an attractive target for immunotherapy [37]. Various preclinical and clinical trials have been performed in GI cancers with antibodies against diﬀerent MUC1 domains (MUC1-N, SEA and MUC1-C), some of them targeting speciﬁc hallmarks of cancer (Figure 1). Overexpression in multiple epithelial tumors, expression all over the surface of a tumor cell due to loss of apicobasal polarity in cancer cells, thus making it accessible to antibodies and tumor-specific aberrant glycosylation with truncated carbohydrate antigens Tn and TF in the VNTR region are features that make MUC1 an attractive target for immunotherapy [37]. Various preclinical and clinical trials have been performed in GI cancers with antibodies against different MUC1 domains (MUC1- N, SEA and MUC1-C), some of them targeting specific hallmarks of cancer (Figure 1). Figure 1. A schematic diagram showing the different antibodies recognizing different domains of MUC1 and also the hallmarks of cancer that they target. The various domains of MUC1 are denoted with different colors, ED and VNTR in sea green, SEA domain in blue, transmembrane domain in orange, and CT in light green. Figure 1. A schematic diagram showing the diﬀerent antibodies recognizing diﬀerent domains of MUC1 and also the hallmarks of cancer that they target. The various domains of MUC1 are denoted with diﬀerent colors, ED and VNTR in sea green, SEA domain in blue, transmembrane domain in orange, and CT in light green. Figure 1. A schematic diagram showing the different antibodies recognizing different domains of MUC1 and also the hallmarks of cancer that they target. The various domains of MUC1 are denoted with different colors, ED and VNTR in sea green, SEA domain in blue, transmembrane domain in orange, and CT in light green. Figure 1. A schematic diagram showing the diﬀerent antibodies recognizing diﬀerent domains of MUC1 and also the hallmarks of cancer that they target. The various domains of MUC1 are denoted with diﬀerent colors, ED and VNTR in sea green, SEA domain in blue, transmembrane domain in orange, and CT in light green. 2.2. Role in GI Tumors The objective of this review is to highlight the recent advances made in the treatment of gastrointestinal cancers utilizing antibodies, immunoconjugates and antibody-derived molecular therapies against tMUC1. We have also provided perspectives on how different anti-MUC1 antibodies target different hallmarks of cancer and thus can be utilized as a combination therapy to have better clinical outcomes. The objective of this review is to highlight the recent advances made in the treatment of gastrointestinal cancers utilizing antibodies, immunoconjugates and antibody-derived molecular therapies against tMUC1. We have also provided perspectives on how diﬀerent anti-MUC1 antibodies target diﬀerent hallmarks of cancer and thus can be utilized as a combination therapy to have better clinical outcomes. 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials Antibody-based immunotherapy has been used for cancer treatment for the past two decades and is one of the most effective ways to treat hematological malignancies and solid tumors [62,63]. Monoclonal antibodies (mAbs) can be generated by immunizing immunocompetent mice with tumor antigens or tumor cell lysates, or synthetically engineered to bind to specific proteins on cancer cells [64,65]. The fundamental mechanism of therapeutic mAbs are to tag cancer cells for phagocytosis by macrophages or killing by NK or effector T-cells, block the downstream signaling of the target molecule, induce programmed cell death (or autophagy) in the antigen expressing cancer cell, and aid in targeted delivery of therapeutic agents to specifically destroy cancer cells [64 66] Antibody-based immunotherapy has been used for cancer treatment for the past two decades and is one of the most eﬀective ways to treat hematological malignancies and solid tumors [62,63]. Monoclonal antibodies (mAbs) can be generated by immunizing immunocompetent mice with tumor antigens or tumor cell lysates, or synthetically engineered to bind to speciﬁc proteins on cancer cells [64,65]. The fundamental mechanism of therapeutic mAbs are to tag cancer cells for phagocytosis by macrophages or killing by NK or eﬀector T-cells, block the downstream signaling of the target molecule, induce programmed cell death (or autophagy) in the antigen expressing cancer cell, and aid in targeted delivery of therapeutic agents to speciﬁcally destroy cancer cells [64–66]. aid in targeted delivery of therapeutic agents to specifically destroy cancer cells [64–66]. Many anti-MUC1 antibodies are in clinical trials or under pre-clinical or experimental studies. The anti-MUC1 antibody-based therapeutics developed against GI cancers that are in pre-clinical and clinical trials have been summarized in Tables 1 and 2 respectively Many anti-MUC1 antibodies are in clinical trials or under pre-clinical or experimental studies. The anti-MUC1 antibody-based therapeutics developed against GI cancers that are in pre-clinical and clinical trials have been summarized in Tables 1 and 2 respectively. 5 of 21 Vaccines 2020, 8, 659 Table 1. MUC1 antibodies under preclinical trials for GI cancers. 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials Antibody Epitope Original Antigen Treatment under Trial GI Cancer Type Year Reference KL-6 a sialylated sugar of Krebs von den Lugen-6 (KL-6) PDTRPAP sequence a sialylated sugar of Krebs von den Lugen-6 (KL-6) PDTRPAP sequence 99mTc labeled anti-KL-6/MUC1 Pancreatic Cancer 2008 [67–69] MY.1E12 sialyla2–3galactosylh1-3Nacetylgalactosaminide linked to a distinct threonine residue in the MUC1 tandem repeat HMFG 3-ICG-acyl-1,3-thiazolidine-2-thione labeled MY.1E12 Gastric Cancer 2008 [70–73] 5E5, 2D9 Tn or STn in the tandem repeat domain GalNAc-glycosylated MUC1 glycopeptide (VTSAPDTRPAPGSTAPPAHG) conjugated to KLH 5E5 MUC1-CAR-T cells Pancreatic Cancer 2016 2019 [74] hMUC1-1H7 extracellular domain of MUC1 C-terminal subunit (MUC1-C) recombinant human (rh) protein including extracellular region of MUC1- C (rhMUC1-EC192) obtained from MCF7 cells hMUC1-1H7 Pancreatic Cancer 2004 [75,76] TAB004 STAPPVHNV within the TR sequence Protein lysate from MUC1-expressing tumors that developed in a MUC1 transgenic mice (PDA mice) that expressed human MUC1 (1) TAB 004 (2) CAR-T cell therapy (3) Bispeciﬁc antibody with anti-CD3 Pancreatic Cancer 2008–2019 [77–82] Table 1. MUC1 antibodies under preclinical trials for GI cancers. 6 of 21 Vaccines 2020, 8, 659 Table 2. MUC1 antibodies under clinical trials for GI cancers. Antibody Epitope Original Antigen Treatment under Trial GI Cancer Type Clinical Trial Status Year Reference huC242 Sialyl-Lewis a epitope CanAg glycoprotein which is similar to MUC1 Human colorectal adenocarcinoma cell line COLO205 huC242-DM4 (1) Non-colorectal Cancer, Pancreatic Cancer (2) Locally Advanced and metastatic Stomach, Gastric and other GI cancers (1) Phase I completed (2) Phase II withdrawn 2006 2008 [83–88] huPAM4 Domain located between the amino terminus and start of the repeat domain of a MUC1 antigen (non- VNTR) and also react with MUC5AC Mucin puriﬁed from the xenografted RIP I human pancreatic carcinoma 111In-huPAM4 Pancreatic Cancer Phase I terminated 2006 [89–91] hPAM4 (Clivatuzumab) Domain located between the amino terminus and start of the repeat domain of a MUC1 antigen (non-VNTR) and also react with MUC5AC Mucin puriﬁed from the xenografted RIP I human pancreatic carcinoma (1) 90Y-hPAM4 (Clivatuzumab) (2) 90Y-hPAM4-Tetraxetan & Gemcitabine vs. 3.1.1. Antibodies Recognizing Non-Glycopeptide Epitope 3.1.1. Antibodies Recognizing Non-Glycopeptide Epitope Human milk fat globule 1 (HMFG1) is an IgG1 murine antibody with kappa light chain, recognizing PDTR epitope within the VNTR region of MUC1-ED. The humanized HMFG1 (AS1402, huHMFG1, Therex, BTH-1704, R-1550) was generated by transferring the complementarity determining regions (CDRs) of the murine HMFG1 onto selected human framework with the same aﬃnity to MUC1 [101,102]. To directly target MUC1 positive advanced pancreatic tumors and trigger neutrophil-mediated immune response, the binding capacity of this mAb in combination with a polysaccharide beta 1,3/1,6 glucan (derived from S.cerevisiae) as an immune stimulator with two drugs gemcitabine and Imprime PGG was evaluated [103]. The secondary objectives were to characterize the adverse eﬀects, time to progression, clinical response, progression-free and overall survival. However, this phase Ib trial (NCT02132403) was terminated due to drug recall. PAM4 is another lgG1 murine mAb, generated by immunizing mice with mucin puriﬁed from the xenografted RIP I human pancreatic carcinoma [91]. This mAb can recognize 85% of the pancreatic carcinomas and 50% of the colon carcinomas. However, it does not detect breast, ovarian, renal, prostate and liver cancers [90]. It has been reported that PAM4 is not related to the core epitopes of VNTR and that it also binds to other mucin proteins like MUC5AC [91,104]. In the preclinical studies, 131I- and 90Y-labeled PAM4, was shown to control pancreatic cancer with enhanced survival and clinical responses in pancreatic cancer patients [89,90]. In the phase I clinical trial, 131I-PAM4 IgG and 99mTc-PAM4 Fab′ showed the speciﬁc tumor localization in four out of ﬁve patients, therefore ensuring these are ideal candidates for further trials [90,105]. Humanized PAM4 (hPAM4, IMMU-107) also known as clivatuzumab was constructed and radiolabeled with Yttrium (90Y) and used for patients with stage III and IV of pancreatic cancer. In a phase I trial, it was shown that 90Y-Clivatuzumab tetraxetan was well tolerated with toxicity restricted to the bone marrow and manageable hematologic toxicity was seen at the maximal tolerated dose of 90Y. Tumor targeting was observed in most patients by using 111In-labeled antibody, and even with mucin antigen present in the serum, there were apparently no issues with the biodistribution or clearance of the antibody. All patients demonstrated disease progression at or after week eight, and some of them had stable target lesions at four weeks after treatment [92]. Hence, combination of chemotherapy and radioimmunotherapy agents was considered for future trials. 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials 3. Anti-MUC1 Antibodies in Preclinical and Clinical Trials Placebo & Gemcitabine (1) Pancreatic Cancer (2) Metastatic Pancreatic Cancer (1) Phases I and II completed (2) Phase III terminated 2008 2013 [92,93] SAR56665 8huDS6-DM4 O-linked glycans with α2,3-sialylated and β1,4-galactosylated termini in VNTR Human serous ovarian carcinoma SAR56665 8huDS6-DM4 Pancreas Phase II completed 2010 PankoMab-GEX™ (Gatipotuzumab) Epitope...PDTRP..., where T is O-glycosylated with GalNAca1- or a similar short, non-sialylated glycan such as Galb1-3GalNAca1-(core-1) Tumor MUC1 from a desialylated human breast cancer source (1) PankoMab-GEX™ (Gatipotuzumab) (2) Combination of Gatipotuzumab and anti-EGFR Tomuzotuximab (1) Pancreatic (2) Colorectal (1) Phase II, ongoing (2) Recruiting for Phase I 2010 2017 [94–97] PD-1 inhibitor armed with an anti-MUC 1 and anti- CD3 bispeciﬁc antibody Information unavailable Information unavailable PD-1 inhibitor armed with an anti-MUC 1 and anti- CD3 bispeciﬁc antibody Advanced Gastric, Colorectal, Pancreatic and Liver cancers Recruiting for Phase II 2018 [98] AR20.5 DTRPAP and DTnRPAP MUC1 from an ovarian cancer patient, derived from human ﬂuids and breast cancer cell MCF-7 culture medium (1) AR20.5 (2) Combination of mAb-AR20.5, anti-PD-L1 and Poly ICLC (1) Advanced adenocarcinoma (2) Pancreatic Cancer (1) Completed Phase I (2) Phase I/II ongoing 2004 2018 [24,99,100] Table 2. MUC1 antibodies under clinical trials for GI cancers. Original Antigen 7 of 21 7 of 21 Vaccines 2020, 8, 659 3.1. Monoclonal Antibodies 3.1.1. Antibodies Recognizing Non-Glycopeptide Epitope Phase I/II trials with 80 participants are ongoing (NCT00603863) to test whether diﬀerent doses of 90Y-hPAM4 in combination with gemcitabine are safe to give in patients with previously untreated pancreatic cancer. Clinical eﬃcacy of Y-clivatuzumab tetraxetan (DOTA) with or without low-dose gemcitabine (PANCRIT®-1) was assessed in a phase I/II/III trial with metastatic pancreatic cancer patients which appeared to be an active ﬁrst-line therapy for pancreatic cancer [93], but eventually, it was discontinued due to insuﬃcient improvement in overall survival in comparison to placebo [NCT01956812]. GP1.4 is an anti-MUC1 antibody that caused internalization of EGFR in pancreatic cancer cells. This inhibited ERK phosphorylation by EGF stimulation in a MUC1 dependent manner. Inhibition of ERK phosphorylation by GP1.4 resulted in the suppression of proliferation and migration of pancreatic cancer cells [106]. TAB004 is a murine IgG1 mAb that was initially developed by immunizing Balb/c mice with lysates from MUC1-expressing tumors that developed in a human tMUC1 bearing transgenic mouse [77]. TAB004 targets the epitope area with sequence STAPPVHNV present within the TR sequence (AA950-958) of hypoglycosylated tMUC1 [13,17,78,107]. TAB004 distinguishes between normal and tumor-associated forms of MUC1 solely based on the expression of hypo-glycosylated or aberrantly glycosylated MUC1. TAB004 alone or in conjugation with dye-doped mesoporous silica nanoparticles was used to detect breast cancer in vivo [80,108]. TAB004 was also shown to be a diagnostic marker for cancer stem cells and circulating MUC1 in mice and patients with pancreatic 8 of 21 Vaccines 2020, 8, 659 cancer [109]. TAB004 in combination with IL2 was shown to improve survival in PDA models by the following mechanisms: (1) reduction in tumor-induced immune regulation and (2) increasing recruitment of CD45+CD11b+ cells, thus increasing antibody-dependent-cellular-cytotoxicity or antibody-dependent-cellular-phagocytosis (ADCC/ADCP) [79]. It has also been reported that, the TAB004 antibody induces complement-independent growth inhibitory eﬀect on PDA cells and signiﬁcantly increases the anti-tumor eﬃcacy of chemotherapy drugs like 5-FU, Gemcitabine and Paclitaxel [81]. In another study, humanized TAB004 was conjugated to 111In and 225Ac-DOTA and this immunoconjugate not only could target the tumor speciﬁcally but also showed complete preclinical response in triple negative breast cancer [110]. MUC1-014E is another anti-MUC1 antibody raised against an intracellular nonrepeating 19-amino-acid sequence (RYVPPSSTDRSPYEKVSAG) of the MUC1-CT, using a synthetic peptide with the 7-amino-acid epitope (STDRSPY). MUC1-014E showed sharp and speciﬁc staining of carcinoma cells, but no staining in ﬁbroblasts, endothelial cells, and inﬂammatory cells. 3.1.1. Antibodies Recognizing Non-Glycopeptide Epitope High rates of positive immunohistochemical staining (97–100%) was found in 107 gastrectomy specimens compared with the other MUC1-related antibodies (MUC1-DF3, MUC1-Ab-5 and PAb anti-MUC1*1110-ecd). MUC1-014E also recognized isolated cancer cells of signet-ring cell carcinoma (sig) and non-solid type poorly diﬀerentiated stomach adenocarcinoma (por2). Therefore, this mAb could be used to detect cells in scirrhous gastric cancer [111]. g hMUC1-1H7 is an anti-hMUC1 murine mAb developed against a recombinant MUC1 obtained from the breast cancer cell MCF7. It signiﬁcantly reduced proliferation of breast cancer cells in which it is internalized and speciﬁcally localized in MUC1-expressing tumors in the xenograft mouse models. hMUC1-1H7 is speciﬁc for the extracellular domain of MUC1-CD and can bind to shed MUC1 as well [76]. It has also been reported that, G3 can inhibit EGF-mediated ERK phosphorylation and cyclin D1 expression, thus, inhibiting EGFR signaling pathways in pancreatic cancer models [75]. 3.1.2. Antibodies Recognizing Glycopeptide Epitopes 3.1.2. Antibodies Recognizing Glycopeptide Epitopes PankoMab is a murine IgG1, kappa light chain mAb recognizing tMUC1 glycopeptide. It has shown a reduced rate of binding to circulating tMUC1 and mononucleated cells in the serum of colon and pancreatic cancer patients [94]. There are various chimeric and humanized formats of PankoMab under clinical trials as suitable candidates for therapeutic and diagnostic applications [95]. PankoMab-GEX™(PMG) also known as Gatipotuzumab (previously known as PankoMab-GEX™), is a glyco-optimised mAb with many advantages. For example, it has higher tumor speciﬁcity and aﬃnity with an increased number of binding sites, reduced binding to shed MUC1 from colon and pancreatic carcinoma, no binding to peripheral blood mononucleated cells, stronger ADCC, and rapid internalization compared to other antibodies [95]. Its mechanisms of action include ADCC and ADCP. A phase I study in patients with tMUC1 positive advanced solid tumor showed that PMG was safe, well tolerated and showed promising anti-tumor activity [96]. The phase 2 study evaluated the eﬃcacy and safety of PMG’s maintenance therapy compared to placebo in patients with recurrent ovarian, fallopian tube or primary serous peritoneal cancer [97]. This randomized double blinded study reported that PMG failed to improve the time without disease recurrence when given as a single entity [97]. However, it showed a good safety proﬁle, hence, targeting tMUC1 by this antibody in combination with other standard chemotherapy or developing a bi-speciﬁc antibody to modulate the immune system holds promise to improve its anti-tumor eﬃcacy [97]. AR20.5 (BrevaRex) is a murine monoclonal antibody (IgG1) developed by immunizing mice with three diﬀerent sources including MUC1 derived from an ovarian cancer patient, human ﬂuids and MCF-7 cell culture medium. It reacts with six amino acids within the VNTR region (DTRPAP). However, addition of a single GalNAc enhanced the binding aﬃnity of AR20.5 to the MUC1 epitope [24]. AR20.5 forms a complex with circulating MUC1 and/or transmembrane MUC1 on tumor cells. This complex can be internalized by dendritic cells which facilitates eﬀective antigen-processing and cross-presentation of MUC1 to T cells, and leads to the activation of cytotoxic T cells to kill the tumor [112]. In the phase I trial 9 of 21 Vaccines 2020, 8, 659 of AR20.5 patients with advanced adenocarcinoma were treated, it induced MUC1-speciﬁc immune responses, did not have dose-limiting toxicity, and induced no hypersensitivity reactions. The 2-mg dose showed the strongest biological activity, and was evaluated in future trials [100]. 3.1.2. Antibodies Recognizing Glycopeptide Epitopes The combination of AR20.5, anti-PD-L1 antibody and PolyICLC rejected human MUC1 expressing tumors and provided a long-lasting, MUC1-speciﬁc cellular immune response, which when adoptively transferred to human MUC1 transgenic (MUC.Tg) mice, provided protection against tumor formation. CD8+ cells were found to be the eﬀectors for the MUC1-speciﬁc immune response generated by this combination. In the US, a phase I/II clinical trial is ongoing for pancreatic cancer by OncoVent Co., Ltd., with this combination [99]. The DS6 antibody is an IgG1 murine antibody recognizing the CA6 sialoglycotope of tMUC1 that is overexpressed in a variety of solid tumors, including ovarian, breast, cervical, pancreatic and lung cancers. DS6 detects a CA6 antigen that is diﬀerent from well-characterized tumor-associated antigens, such as MUC1, CA125 and the histo-blood group–related antigens sLea, sLex and sTn [113]. DS6 speciﬁcally binds to the tandem repeat domain of CA6-positive MUC1 based on the presence of mucin type O-linked glycans with α2,3-sialylated and β1,4-galactosylated termini [114]. Humanized DS6 (huDS6) antibody was conjugated to the cytotoxic maytansinoid derivative drug DM4 through a cleavable linker. The ADC was called SAR566658 and it showed antitumor eﬃcacy against CA6-positive human pancreas, cervix, bladder, and ovary in vivo tumor xenograft models, with a minimal eﬀective dose correlating with CA6 expression as well as better eﬃcacy than standard-of-care nontargeted tubulin binders. SAR566658 was used in a phase I clinical trial with 114 patients with refractory solid tumors. It showed a satisfactory safety proﬁle and antitumor activity. Tumor improvement was shown in 35–60% of patients at diﬀerent dosages of SAR566658 [115]. The monoclonal IgG1-kappa antibody C242 was developed by immunizing a mouse with human colorectal adenocarcinoma cell line COLO205. Humanized C242 (HuC242 or Cantuzumab) has the CA242 epitope and reacts with a novel glycoform of MUC1 also known as CanAg glycoprotein (cancer antigen) [83]. CanAg is very highly glycosylated, rich in fucose and sialic acid and Hx-CanAg (heavy subunit) is very similar to MUC1 in amino acid composition, but L-CanAg (light subunit) is diﬀerent. Deglycosylated H-CanAg can be recognized by the monoclonal antibodies SM-3 and HMFG-2 [84]. Also, due to its high expression in most pancreatic, biliary and colorectal cancers, CanAg is a potential candidate for mAb-based therapies. In a phase I trial, Cantuzumab was conjugated to an anti-microtubule agent mertansine (DM1) and diﬀerent doses were used to treat colon and rectum carcinomas or other malignancies with positive CanAg antigen as a single intravenous infusion. 3.2. Bispeciﬁc Antibodies for MUC1 Bispeciﬁc antibodies (bsAbs) can recognize two distinct epitopes or antigens simultaneously and therefore enhance the ability of immune cells to engage to tumor cells. Recently, MUC1 has been considered for designing bsAbs. MUC1-CD16-Bi antibody is a novel bispeciﬁc antibody generated via a Serine-Glycine linkage between single domain antibodies (VHH segments) against tMUC1, and CD16 presented on natural killer (NK) cells. The bsAb against MUC1 named MUC1-Bi-1 was humanized by grafting the CDRs of both segments to DP-47 V-segment. Both MUC1-Bi-1 and its humanized version speciﬁcally detected tMUC1 on several cancer cell lines (SKOV3, HT29, and LS174) and potentially introduced them to NK cells. These bsAbs had no binding aﬃnity and cytotoxicity to MUC1 negative CHO and HepG2 cells even in the presence of NK cells [118,119]. Diﬀerent types of bsAbs were constructed with binding aﬃnity to both tMUC1 and CD3 on T-cells. Fab’-S-NB fragments of OKT-3 mAb (anti-CD3) and Fab-SH fragments of MUSE11 mAb (anti-tMUC1) were used to generate the ﬁrst bsAb which increased the antitumor activity of CD3+ T-LAK cells. MUSE11 is a mouse IgG1 mAb developed against the ascites ﬂuid of gastric cancer patients. The epitope of this antibody could be within the amino acid sequence PDTRPAPG of tMUC1 [120]. MUC1 × CD3 BsAb was constructed with MUSE11 (anti tMUC1) and OKT-3 (anti-CD3), and MUC1 × CD2S BsAb was constructed with MUSE11 and 15E8 (anti-CD28) antibodies. The Fab’-SH from MUSE11 and Fab-S-NB of mouse IgG1 15E8 (anti-CD28) antibodies were used. These BsAbs showed growth inhibition of TFK-1 cancer cells and bile duct carcinoma in SCID mice [121]. The BsAbs (MUC1 × CD3 BsAb and MUC1 × CD28 BsAb) together exhibited 60% cytotoxicity in vitro, similar to that shown by BsAb (MUC1 × CD3) alone. Although reduction in tumor growth was limited, simultaneous administration of a combination of three bsAbs (M × 3, M × 28 and M × 2 bsAb) with peripheral blood mononuclear cells (PBMCs) or T-LAK cells in vitro showed higher cytotoxicity against MUC1-expressing bile duct carcinoma cells [121]. Mx3 diabody is a recombinant BsAb generated using the variable domains of two mAbs directed at eﬀector cells, one against CD3 (OKT-3, mouse IgG2a) and the other against CD28 (l5E8, mouse IgGl), and MUSE11 (mouse IgGl), directed at tMUC1 [122]. One chain consists of a variable heavy chain speciﬁc for MUC1 linked to a variable light chain speciﬁc for CD3 with a short polypeptide linker GlyGlyGlyGlySer (GGGGS). 3.1.2. Antibodies Recognizing Glycopeptide Epitopes Results showed that HuC242-DM1 is safe and well tolerated with eﬀective antitumor activity [85,86]. In another phase I trial, cantuzumab conjugated to potent cytotoxic maytansinoid drug ravtansine (DM4), called IMGN242, was found to be well tolerated in colorectal and pancreatic cancer patients at 168 mg/m2 dose. This provided a basis to perform phase II clinical studies [87]. The phase II trial was started in CanAg-expressing gastric cancer patients at a dose of 168 mg/m2. The data has been amended to diﬀerentiate the administered dose of IMGN242 based on the patient’s plasma CanAg levels [88]. KL-6 is a mouse IgG1 mAb that speciﬁcally recognizes a sialylated sugar of Krebs von den Lugen-6 (KL-6), which is considered a MUC1-derived glycoprotein antigen. The minimal antigenic epitope for binding of this antibody is PDTRPAP. It has been reported that anti-KL-6/MUC1 mAb increased aggregation of MUC1 glycoproteins at one pole of the cell, called capping of MUC1 on the surface and facilitated E-cadherin-mediated cell-cell interaction in breast cancer cell lines YMB-S and ZR-75-1S. Anti-KL-6 also enhanced the cytotoxic activity of lymphokine-activated killer (LAK) cells. The mechanism of action of this antibody is capping of MUC1 and restoring cell–cell adhesion by E-cadherin, which induces cell cycle arrest by upregulation of the cyclin-dependent kinase, p27 [116]. This also leads to increased accessibility for eﬀector cells to kill tumor cells [67–69]. 99mTc labeled anti-KL-6/MUC1 antibody was shown to be a tumor-speciﬁc radiotracer that detects pancreatic cancer in vivo, but no further information is available [117]. 10 of 21 Vaccines 2020, 8, 659 MY.1E12 is another murine anti-human MUC1 mAb that binds to MUC1 bearing sialylated O-linked oligosaccharides. MY.1E12 was generated by immunizing mice with HMFG. It can identify colon carcinoma tissue [70,71]. MY.1E12 speciﬁcally reacts to T structure (ST) attached to Thr8. The sialylation of the T structure (ST) enhances its reactivity with MUC1 [72]. ICG-N-hydroxysulfosuccinimide ester (ICG-sulfo-OSu) and 3-ICG-acyl-1,3-thiazolidine-2-thione (ICG-ATT) were developed as infrared ﬂuorescent-labeling reagents, and anti-human CEA antibody and FMY.1E12 were labelled with 3-ICG-acyl-1,3-thiazolidine-2-thione. This was shown to recognize the gastric cancer tissue specimens with a strong ﬂuorescent signal [73]. 5E5 and 2D9 are mouse IgG1k mAbs that were generated by immunization of wild-type Balb/c mice with GalNAc-glycosylated MUC1 glycopeptide (VTSAPDTRPAPGSTAPPAHG) conjugated to KLH. These antibodies exhibited high selectivity for MUC1 tandem repeat glycopeptides with Tn and STn O-glycans and showed preference for Tn-MUC1 glycoforms that had the highest O-glycan occupancy. 3.1.2. Antibodies Recognizing Glycopeptide Epitopes They can bind to MUC1 with Tn or STn in the GSTA sequence of tandem repeats but do not bind to the GSTA epitope carrying T [74]. 3.3. CAR-T Cells Targeting MUC1 TAB004 has been used to make a CAR-T cell construct, which has exhibited signiﬁcant cytotoxic activity against pancreatic cancer cells and reduced growth of orthotopic pancreatic tumors in a NOD-SCID mouse model [82]. Some PDA cells, for example CFPAC and HPAF II, were found to be resistant to the therapy and several genes were overexpressed in them such as indoleamine 2, 3-dioxygenases-1 (IDO1), cyclooxygenase 1 and 2 (CO × 1/2), and galectin-9 (Gal-9) [82]. This study showed that combining biological inhibitors of IDO1, CO × 1/2, and Gal-9 with the CAR-T cells resulted in signiﬁcant enhancement of CAR-T cell cytotoxicity against PDA cells. 5E5 mAb showed high speciﬁcity to breast cancer cells and tissue [37,123] and was used to develop MUC1 CAR-T cells. These CAR-T cells showed cytotoxicity against leukemia and pancreatic cancer cells and also enhanced survival of mice by eliminating the barriers for engagement of the endogenous immune system [38,125]. 3.2. Bispeciﬁc Antibodies for MUC1 The second chain has a variable light chain speciﬁc for MUC1 linked to a variable heavy chain speciﬁc for CD3. Therefore, Mx3 diabody can speciﬁcally bind to both MUC1 and CD3 positive LAK cells with a T cell phenotype (T-LAK). Mx3 diabody with T-LAK showed growth inhibition in about 98% of TFK-1 cells with an eﬀector:target ratio of 10 [122]. Mx3 was fused genetically to the mutated superantigen staphylococcal enterotoxin A (SEA) D227A to speciﬁcally target bile duct Vaccines 2020, 8, 659 11 of 21 carcinoma (BDC). This super-antigen fused diabody also showed the potential to inhibit the BDC cell line TFK-1 and reduce tumor size when compared to the Mx3 diabody alone [123]. carcinoma (BDC). This super-antigen fused diabody also showed the potential to inhibit the BDC cell line TFK-1 and reduce tumor size when compared to the Mx3 diabody alone [123]. A bsAb containing F(ab′)2/F(ab′) fragments with a functional chemical linker is the anti-MUC1/anti-Ga chelate. A mouse IgG1 12H12 mAb raised against a mouse glycosylated form of MUC1 called TAG-12 was combined to another mouse IgG3 anti-Ga chelate mAb. Prior to 3A10 F(ab′) coupling, the 12H12 F(ab′)2 fragment was labeled with 125I. This bispeciﬁc-mAb showed improved immunoscintigraphic tumor localization in breastcarcinoma bearing mice [124]. Another bsAb has been constructed with a novel PD-1 inhibitor-induced cytokine- induced killer cells (CIKs) armed with an anti-tMUC1 and anti- CD3 antibodies. This bsAb is currently under several phase II randomized clinical trials for advanced gastric, kidney, lung, breast, colorectal, pancreatic and liver cancers, but there is no further information available ([NCT03554395], [NCT03540199], [NCT03501056], [NCT03524261], [NCT03524274], [NCT03509298], and [NCT03484962]) [98]. 5. Concluding Remarks and Future Perspectives In spite of MUC1 being a top target, multiple trials with MUC1 antibodies and antibody-derived immunotherapies have failed to translate to the clinic. Most of the trials have been discontinued for not being suﬃciently eﬀective. There may be various reasons for the ineﬃciency of the antibodies. As of now, many anti-MUC1 antibodies have been developed against the highly immunogenic VNTR region of MUC1 α chain (MUC1-ED) [135]. After cleavage at the SEA domain, the MUC1-N is often shed from the surface of cells and released into the peripheral blood. The shed α subunits (MUC1-N) sequester anti-MUC1 antibodies against the VNTR region, preventing them from binding to the surface MUC1 [95]. To overcome this problem, antibodies against MUC1-CD could be used as a more eﬀective strategy. Shedding of MUC1-N increases its levels in the serum of patients with various cancers [136–138], thus, reducing the speciﬁcity and eﬀective binding of the antibodies to MUC1 on the tumor cells [136,139]. Therefore, serum levels of MUC1 in diﬀerent cancer patients need to be evaluated to ﬁnd an eﬀective dose of the antibodies [140]. In addition, bsAbs can be made by combining immune checkpoint inhibitors such as anti-PD1 and anti-PD-L1 antibodies with anti-MUC1 antibodies. This will increase engagement of the immune cells with the tumor. In recent years, antibodies are being designed against the other domains of MUC1 including SEA, extracellular, and intracellular MUC1-CT. Therefore, rational designing of antibodies and combination therapy strategies are important to achieve a good safety and eﬃcacy proﬁle against MUC1 expressing cancers. Antibodies to the non-glycopeptide part of the VNTR region have not been able to generate an eﬀective cellular or humoral immune response to tMUC1 [141]. Antibodies to MUC1 peptide also do not eﬀectively recognize MUC1-expressing tumor cells. However, antibodies raised against shortened glycopeptide structures with a simple T antigen (T, Galβ1-3GalNAc), sialyl Tn (NeuAcα2-6GalNAc) and Tn (GalNAc) elicit the strongest immune response against MUC1-expressing tumor cells [142]. This happens due to the speciﬁc presence of Tn and STn glycans on MUC1 expressing cancer cells, but not on normal epithelial cells and the blocked regions of the VNTR domain get exposed to recognition by antibodies, thereby, producing tumor-speciﬁc recognition sites. As evident, studying the glycosylation changes have led to the development of potentially eﬀective MUC1-based immunotherapy [143,144]. Some anti-MUC1 antibodies can recognize the MUC1 epitopes on both normal epithelial and tumor cells thus compromising the speciﬁcity [145]. 4. Molecular Interactions between MUC1 and Its Antibodies X-ray crystallography of antibody crystal structures [126] and NMR analysis of glycopeptides [127] are used to understand the biochemical interactions or molecular recognition between the antigen and antibody. The Tn antigen is one of the most important structural motifs of tMUC1 found widely in many diﬀerent aggressive carcinomas [128,129]. It has been shown by years of extensive eﬀort to develop antibodies targeting tMUC1 having the Tn antigen, that most anti-MUC1 antibodies do not directly bind to carbohydrates. However, the binding aﬃnities with the immunodominant MUC1 are shown to be signiﬁcantly increased by O-glycosylation in this area [130–132]. AR20.5 bound to the glycopeptide with stronger aﬃnity than the naked peptide. These observations led to the hypothesis that the antibody must speciﬁcally bind the carbohydrate as well as the peptide. X-ray crystallography of the structures of AR20.5 [24] and SM3 [133] in complex with both peptide and glycopeptide revealed that the carbohydrate did not have any speciﬁc polar contacts with the antibody. The high aﬃnity for the glycopeptide and the lack of speciﬁc binding contacts of AR20.5 suggest that glycosylation of MUC1 stabilizes an extended bioactive conformation of the peptide that is recognized by the antibody. Evidence suggests that glycosylation of the peptide alters the conformational equilibrium of the antigen, and this allows the antibody to select the correct conformation. Therefore, glycosylation of MUC1 is important for the generation of high aﬃnity therapeutic antibodies [24]. The anti-MUC1 KL-6 antibody distinguishes between the ST, Tn, and T antigens at the same O-glycosylation site independent of the modiﬁcations at other potential sites [68,131,132]. The NMR study suggests that KL6 mAb strictly recognizes the epitope from the extended trans conformation of a glycopeptide, which has been modiﬁed with the ST antigen. Detailed molecular recognition studies on MUC1 and anti-MUC1 antibodies and the use of synthetic glycopeptide library to develop a new class of antibodies targeting “dynamic glycopeptidic neoepitopes” with disease-relevant O-glycosylation in immunodominant mucin domains have been described recently [134]. 12 of 21 Vaccines 2020, 8, 659 The lack of carbohydrate-binding speciﬁcities in most anti-MUC1 mAbs is a huge challenge for the development of MUC1-based therapeutic antibodies. Antibodies binding to cancer-relevant glycopeptidic neoepitopes with higher speciﬁcities in carbohydrate recognition will be beneﬁcial in the development of anti-MUC1 mAbs as therapeutic and diagnostic agents in the clinical settings. 5. Concluding Remarks and Future Perspectives Conﬂicts of Interest: The authors declare no Conﬂict of Interest. 5. Concluding Remarks and Future Perspectives Also, heterogeneity of MUC1 expression levels, the glycosylation pattern and subcellular distribution contribute to reduced binding eﬃciency. The diﬀerent glycoforms may confer an evolutionary advantage on the tumor cells to be resistant against antibody-based therapies [145,146]. Therefore, a combination of antibodies that can detect many glycoforms of MUC1 can be considered for clinical trials. Anti-MUC1 antibodies directed against the SEA domain target the junction of MUC1 α and β subunits, which is composed of intact epitopes from both [147,148]. These anti-SEA domain antibodies have shown high aﬃnity and eﬀectivity compared to antibodies targeting the VNTR region [148]. The mechanism of action anti-MUC1 mAbs target one or more hallmarks of cancer. For example, some antibodies have been reported to show ADCC and ADCP, some others block anti-apoptotic mechanisms thus inducing cell death, also some antibodies reduce expression of pro-survival genes. Gatipotuzumab is a glycooptimized antibody developed by Glycotope’s GlycoExpress™platform that signiﬁcantly improved treatment outcome with mechanisms such ADCC, tumor cell phagocytosis and induction of apoptosis compared to non-glycooptimized biotherapeutics [97]. Other antibodies against MUC1 glycopeptide, such as 5E5 and 1B2, have been shown to be eﬀective as immunotherapy strategies because of their high speciﬁcity to tMUC1 and ability to induce ADCC [149]. Therefore, by utilizing the mechanism of action of an antibody, strategies could be developed to eliminate the tumor. 13 of 21 Vaccines 2020, 8, 659 However, a decrease in concentration of anti-MUC1 antibodies targeting the tumor and their poor internalization due to the extracellular MUC1-N barrier remain major hurdles. To overcome this, development of antibody fragments can be considered [143,150]. Also, a whole or fragmented antibody could be conjugated to potent drugs to target speciﬁc types of tumor cells. For example, Napabucasin, which is a STAT-3 inhibitor was under Phase III clinical trials for PDA but was discontinued due to futility [151]. However, it has been shown that high-MUC1 PDA cells are more sensitive toward the STAT-3 inhibitor Napabucasin [152]. Therefore, anti-MUC1 antibodies armed with Napabucasin may be a promising strategy to eliminate high-MUC1 tumors. Bispeciﬁc and trispeciﬁc antibodies armed with anti-PD-1, anti-MUC1 and anti-CD3 are new products under clinical trials [98]. Author Contributions: M.B. and P.M. conceived the idea and prepared the manuscript. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Funding: This research received no external funding. References 1. Bray, F.; Ferlay, J.; Soerjomataram, I.; Siegel, R.L.; Torre, L.A.; Jemal, A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA A Cancer J. Clin. 2018, 68, 394–424. [CrossRef] [PubMed] 1. Bray, F.; Ferlay, J.; Soerjomataram, I.; Siegel, R.L.; Torre, L.A.; Jemal, A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA A Cancer J. Clin. 2018, 68, 394–424. [CrossRef] [PubMed] 1. Bray, F.; Ferlay, J.; Soerjomataram, I.; Siegel, R.L.; Torre, L.A.; Jemal, A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA A Cancer J. Clin. 2018, 68, 394–424. [CrossRef] [PubMed] 2. Yang, S.; Wu, S.; Huang, Y.; Shao, Y.; Chen, X.Y.; Xian, L.; Zheng, J.; Wen, Y.; Chen, X.; Li, H. Screening for oesophageal cancer. Cochrane Database Syst. Rev. 2012, CD007883. [CrossRef] [PubMed] 2. Yang, S.; Wu, S.; Huang, Y.; Shao, Y.; Chen, X.Y.; Xian, L.; Zheng, J.; Wen, Y.; Chen, X.; Li, H. Screening for oesophageal cancer. Cochrane Database Syst. Rev. 2012, CD007883. [CrossRef] [PubMed] 3. Bjelakovic, G.; Nikolova, D.; Simonetti, R.G.; Gluud, C. Antioxidant supplements for preventing gastrointestinal cancers. Cochrane Database Syst. Rev. 2008. [CrossRef] [PubMed] 3. Bjelakovic, G.; Nikolova, D.; Simonetti, R.G.; Gluud, C. Antioxidant supplements for preventing gastrointestinal cancers. Cochrane Database Syst. Rev. 2008. [CrossRef] [PubMed] 4. Linden, S.; Sutton, P.; Karlsson, N.; Korolik, V.; McGuckin, M. Mucins in the mucosal barrier to infection. Mucosal Immunol. 2008, 1, 183–197. [CrossRef] 4. Linden, S.; Sutton, P.; Karlsson, N.; Korolik, V.; McGuckin, M. Mucins in the mucosal barrier to infection. Mucosal Immunol. 2008, 1, 183–197. [CrossRef] 5. Ostedgaard, L.S.; Moninger, T.O.; McMenimen, J.D.; Sawin, N.M.; Parker, C.P.; Thornell, I.M.; Powers, L.S.; Gansemer, N.D.; Bouzek, D.C.; Cook, D.P. Gel-forming mucins form distinct morphologic structures in airways. Proc. Natl. Acad. Sci. USA 2017, 114, 6842–6847. [CrossRef] 5. Ostedgaard, L.S.; Moninger, T.O.; McMenimen, J.D.; Sawin, N.M.; Parker, C.P.; Thornell, I.M.; Powers, L.S.; Gansemer, N.D.; Bouzek, D.C.; Cook, D.P. Gel-forming mucins form distinct morphologic structures in airways. Proc. Natl. Acad. Sci. USA 2017, 114, 6842–6847. [CrossRef] 6. Rachagani, S.; Torres, M.P.; Moniaux, N.; Batra, S.K. Current status of mucins in the diagnosis and therapy of cancer. Biofactors 2009, 35, 509–527. [CrossRef] [PubMed] 6. Rachagani, S.; Torres, M.P.; Moniaux, N.; Batra, S.K. Current status of mucins in the diagnosis and therapy of cancer. References [CrossRef] 20. Levitin, F.; Stern, O.; Weiss, M.; Gil-Henn, C.; Ziv, R.; Prokocimer, Z.; Smorodinsky, N.I.; Rubinstein, D.B.; Wreschner, D.H. The MUC1 SEA module is a self-cleaving domain. J. Biol. Chem. 2005, 280, 33374–33386. [CrossRef] 21. Parry, S.; Silverman, H.S.; McDermott, K.; Willis, A.; Hollingsworth, M.A.; Harris, A. Identiﬁcation of MUC1 proteolytic cleavage sites in vivo. Biochem. Biophys. Res. Commun. 2001, 283, 715–720. [CrossRef] 21. Parry, S.; Silverman, H.S.; McDermott, K.; Willis, A.; Hollingsworth, M.A.; Harris, A. Identiﬁcation of MUC1 proteolytic cleavage sites in vivo. Biochem. Biophys. Res. Commun. 2001, 283, 715–720. [CrossRef] 22. Kufe, D.W. MUC1-C oncoprotein as a target in breast cancer: Activation of signaling pathways and therapeutic h O 32 1 1 1 2. Kufe, D.W. MUC1-C oncoprotein as a target in breast cancer: Activation of signaling pathways and therape approaches. Oncogene 2013, 32, 1073–1081. 23. Bennett, E.P.; Mandel, U.; Clausen, H.; Gerken, T.A.; Fritz, T.A.; Tabak, L.A. Control of mucin-type O-glycosylation: A classiﬁcation of the polypeptide GalNAc-transferase gene family. Glycobiology 2012, 22, 736–756. [CrossRef] 24. Movahedin, M.; Brooks, T.M.; Supekar, N.T.; Gokanapudi, N.; Boons, G.-J.; Brooks, C.L. Glycosylation of MUC1 inﬂuences the binding of a therapeutic antibody by altering the conformational equilibrium of the antigen. Glycobiology 2017, 27, 677–687. [CrossRef] [PubMed] 25. Piyush, T.; Rhodes, J.M.; Yu, L.-G. MUC1 O-glycosylation contributes to anoikis resistance in epithelial cancer cells. Cell Death Discov. 2017, 3, 1–9. [CrossRef] 26. Parry, S.; Hanisch, F.G.; Leir, S.-H.; Sutton-Smith, M.; Morris, H.R.; Dell, A.; Harris, A. N-Glycosylation of the MUC1 mucin in epithelial cells and secretions. Glycobiology 2006, 16, 623–634. [CrossRef] 27. Tian, E.; Ten Hagen, K.G. Recent insights into the biological roles of mucin-type O-glycosylation. Glycoconj. J. 2009, 26, 325–334. [CrossRef] 28. Ju, T.; Wang, Y.; Aryal, R.P.; Lehoux, S.D.; Ding, X.; Kudelka, M.R.; Cutler, C.; Zeng, J.; Wang, J.; Sun, X. T n and sialyl-Tn antigens, aberrant O-glycomics as human disease markers. Proteom. Clin. Appl. 2013, 7, 618–631. [CrossRef] 29. Clausen, H.; Bennett, E.P. A family of UDP-GalNAc: Polypeptide N-acetylgalactosaminyl-transferases control the initiation of mucin-type O-linked glycosylation. Glycobiology 1996, 6, 635–646. [CrossRef] 30. Breloy, I.; Hanisch, F.-G. Functional Roles of O-Glycosylation. Molecules 2018, 3063. [CrossRef] 31. Brockhausen, I. Mucin-type O-glycans in human colon and breast cancer: Glycodynamics and functions. EMBO Rep. 2006, 7, 599–604. [CrossRef] [PubMed] 32. Hanson, R.L.; Hollingsworth, M.A. Functional consequences of diﬀerential O-glycosylation of MUC1, MUC4, and MUC16 (downstream eﬀects on signaling). Biomolecules 2016, 6, 34. References Biofactors 2009, 35, 509–527. [CrossRef] [PubMed] 7. Gendler, S.; Burchell, J.; Duhig, T.; Lamport, D.; White, R.; Parker, M.; Taylor-Papadimitriou, J. Cloning of partial cDNA encoding diﬀerentiation and tumor-associated mucin glycoproteins expressed by human mammary epithelium. Proc. Natl. Acad. Sci. USA 1987, 84, 6060–6064. [CrossRef] 8. Lan, M.S.; Batra, S.K.; Qi, W.-N.; Metzgar, R.S.; Hollingsworth, M.A. Cloning and sequencing of a human pancreatic tumor mucin cDNA. J. Biol. Chem. 1990, 265, 15294–15299. 9. Ligtenberg, M.; Vos, H.; Gennissen, A.; Hilkens, J. Episialin, a carcinoma-associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini. J. Biol. Chem. 1990, 265, 5573–5578. 10. Siddiqui, J.; Abe, M.; Hayes, D.; Shani, E.; Yunis, E.; Kufe, D. Isolation and sequencing of a cDNA coding for the human DF3 breast carcinoma-associated antigen. Proc. Natl. Acad. Sci. USA 1988, 85, 2320–2323. [CrossRef] [PubMed] 11. Hattrup, C.L.; Gendler, S.J. Structure and function of the cell surface (tethered) mucins. Annu. Rev. Physiol. 2008, 70, 431–457. [CrossRef] 12. Gendler, S.J.; Spicer, A. Epithelial mucin genes. Annu. Rev. Physiol. 1995, 57, 607–634. [CrossRef] 13. Gendler, S.J. MUC1, the renaissance molecule. J. Mammary Gland Biol. Neoplasia 2001, 6, 339–353. [C 14. Chang, J.-F.; Zhao, H.-L.; Phillips, J.; Greenburg, G. The epithelial mucin, MUC1, is expressed on resting T lymphocytes and can function as a negative regulator of T cell activation. Cell. Immunol. 2000, 201, 83–88. [CrossRef] 15. Yolken, R.H.; Peterson, J.A.; Vonderfecht, S.L.; Fouts, E.T.; Midthun, K.; Newburg, D.S. Human milk mucin inhibits rotavirus replication and prevents experimental gastroenteritis. J. Clin. Investig. 1992, 90, 1984–1991. [CrossRef] 14 of 21 Vaccines 2020, 8, 659 16. Schroten, H.; Hanisch, F.; Plogmann, R.; Hacker, J.; Uhlenbruck, G.; Nobis-Bosch, R.; Wahn, V. Inhibition of adhesion of S-ﬁmbriated Escherichia coli to buccal epithelial cells by human milk fat globule membrane components: A novel aspect of the protective function of mucins in the nonimmunoglobulin fraction. Infect. Immun. 1992, 60, 2893–2899. [CrossRef] [PubMed] 17. Kufe, D.W. Mucins in cancer: Function, prognosis and therapy. Nat. Rev. Cancer 2009, 9, 874–885. [CrossRef] [PubMed] 18. Baruch, A.; Hartmann, M.-L.; Yoeli, M.; Adereth, Y.; Greenstein, S.; Stadler, Y.; Skornik, Y.; Zaretsky, J.; Smorodinsky, N.I.; Keydar, I. The breast cancer-associated MUC1 gene generates both a receptor and its cognate binding protein. Cancer Res. 1999, 59, 1552–1561. 19. Stasyk, T.; Huber, L.A. Spatio-temporal parameters of endosomal signaling in cancer: Implications for new treatment options. J. Cell. Biochem. 2016, 117, 836–843. References [CrossRef] 33. Chandrasekaran, E.; Xue, J.; Xia, J.; Locke, R.D.; Patil, S.A.; Neelamegham, S.; Matta, K.L. Characterization of cancer associated mucin type O-glycans using the exchange sialylation properties of mammalian sialyltransferase ST3Gal-II. J. Proteome Res. 2012, 11, 2609–2618. [CrossRef] 34. Saeland, E.; Belo, A.I.; Mongera, S.; van Die, I.; Meijer, G.A.; van Kooyk, Y. Diﬀerential glycosylation of MUC1 and CEACAM5 between normal mucosa and tumour tissue of colon cancer patients. Int. J. Cancer 2012, 131, 117–128. [CrossRef] 35. Horm, T.M.; Schroeder, J.A. MUC1 and metastatic cancer: Expression, function and therapeutic targeting. Cell Adhes. Migr. 2013, 7, 187–198. [CrossRef] 36. Ghosh, S.K.; Pantazopoulos, P.; Medarova, Z.; Moore, A. Expression of underglycosylated MUC1 antigen in cancerous and adjacent normal breast tissues. Clin. Breast Cancer 2013, 13, 109–118. [CrossRef] [PubMed] 36. Ghosh, S.K.; Pantazopoulos, P.; Medarova, Z.; Moore, A. Expression of underglycosylated MUC1 antigen in cancerous and adjacent normal breast tissues. Clin. Breast Cancer 2013, 13, 109–118. [CrossRef] [PubMed] 37. Taylor-Papadimitriou, J.; Burchell, J.M.; Graham, R.; Beatson, R. Latest developments in MUC1 cancerous and adjacent normal breast tissues. Clin. Breast Cancer 2013, 13, 109–118. [CrossRef] [PubMed] 37. Taylor-Papadimitriou, J.; Burchell, J.M.; Graham, R.; Beatson, R. Latest developments in MUC1 immunotherapy. Biochem. Soc. Trans. 2018, 46, 659–668. [CrossRef] 37. Taylor-Papadimitriou, J.; Burchell, J.M.; Graham, R.; Beatson, R. Latest developments in MUC1 immunotherapy. Biochem. Soc. Trans. 2018, 46, 659–668. [CrossRef] 15 of 21 Vaccines 2020, 8, 659 15 of 21 38. Posey, A.D., Jr.; Schwab, R.D.; Boesteanu, A.C.; Steentoft, C.; Mandel, U.; Engels, B.; Stone, J.D.; Madsen, T.D.; Schreiber, K.; Haines, K.M. Engineered CAR T cells targeting the cancer-associated Tn-glycoform of the membrane mucin MUC1 control adenocarcinoma. Immunity 2016, 44, 1444–1454. [CrossRef] 9. Pourjafar, M.; Samadi, P.; Khoshinani, H.M.; Saidijam, M. Are mimotope vaccines a good alternativ monoclonal antibodies? Immunotherapy 2019, 11, 795–800. [CrossRef] 40. Lau, S.K.; Weiss, L.M.; Chu, P.G. Diﬀerential expression of MUC1, MUC2, and MUC5AC in carcinomas of various sites: An immunohistochemical study. Am. J. Clin. Pathol. 2004, 122, 61–69. [CrossRef] 41. Nath, S.; Mukherjee, P. MUC1: A multifaceted oncoprotein with a key role in cancer progression. Trends Mol. Med. 2014, 20, 332–342. [CrossRef] 42. Bose, M.; Mukherjee, P. Microbe–MUC1 Crosstalk in Cancer-Associated Infections. Trends Mol. Med. 2020, 26, 324–336. [CrossRef] [PubMed] 43. Cheever, M.A.; Allison, J.P.; Ferris, A.S.; Finn, O.J.; Hastings, B.M.; Hecht, T.T.; Mellman, I.; Prindiville, S.A.; Viner, J.L.; Weiner, L.M. References The prioritization of cancer antigens: A national cancer institute pilot project for the acceleration of translational research. Clin. Cancer Res. 2009, 15, 5323–5337. [CrossRef] 44. Zeng, Y.; Zhang, Q.; Zhang, Y.; Lu, M.; Liu, Y.; Zheng, T.; Feng, S.; Hao, M.; Shi, H. MUC1 predicts colorectal cancer metastasis: A systematic review and meta-analysis of case controlled studies. PLoS ONE 2015, 10, e0138049. [CrossRef] 45. Agata, N.; Ahmad, R.; Kawano, T.; Raina, D.; Kharbanda, S.; Kufe, D. MUC1 oncoprotein blocks death receptor–mediated apoptosis by inhibiting recruitment of caspase-8. Cancer Res. 2008, 68, 6136–6144. [CrossRef] 46. Yin, L.; Kosugi, M.; Kufe, D. Inhibition of the MUC1-C oncoprotein induces multiple myeloma cell death by down-regulating TIGAR expression and depleting NADPH. Blood J. Am. Soc. Hematol. 2012, 119, 810–816. [CrossRef] 47. Pochampalli, M.; El Bejjani, R.; Schroeder, J. MUC1 is a novel regulator of ErbB1 receptor traﬃcking. Oncogene 2007, 26, 1693–1701. [CrossRef] 48. Liu, X.; Caﬀrey, T.; Steele, M.; Mohr, A.; Singh, P.; Radhakrishnan, P.; Kelly, D.L.; Wen, Y.; Hollingsworth, M.A. MUC1 regulates cyclin D1 gene expression through p120 catenin and β-catenin. Oncogenesis 2014, 3, e107. [CrossRef] 49. Wei, X.; Xu, H.; Kufe, D. Human mucin 1 oncoprotein represses transcription of the p53 tumor suppressor gene. Cancer Res. 2007, 67, 1853–1858. [CrossRef] 50. Yin, L.; Huang, L.; Kufe, D. MUC1 oncoprotein activates the FOXO3a transcription factor in a survival response to oxidative stress. J. Biol. Chem. 2004, 279, 45721–45727. [CrossRef] 51. Ren, J.; Bharti, A.; Raina, D.; Chen, W.; Ahmad, R.; Kufe, D. MUC1 oncoprotein is targeted to mitochondria by heregulin-induced activation of c-Src and the molecular chaperone HSP90. Oncogene 2006, 25, 20–31. [CrossRef] [PubMed] 52. Raina, D.; Kharbanda, S.; Kufe, D. The MUC1 oncoprotein activates the anti-apoptotic phosphoinositide 3-kinase/Akt and Bcl-xL pathways in rat 3Y1 ﬁbroblasts. J. Biol. Chem. 2004, 279, 20607–20612. [CrossRef] 53. Satoh, S.; Hinoda, Y.; Hayashi, T.; Burdick, M.D.; Imai, K.; Hollingsworth, M.A. Enhancement of metastatic properties of pancreatic cancer cells by MUC1 gene encoding an anti-adhesion molecule. Int. J. Cancer 2000, 88, 507–518. [CrossRef] 54. Regimbald, L.H.; Pilarski, L.M.; Longenecker, B.M.; Reddish, M.A.; Zimmermann, G.; Hugh, J.C. The breast mucin MUC1 as a novel adhesion ligand for endothelial intercellular adhesion molecule 1 in breast cancer. Cancer Res. 1996, 56, 4244–4249. 55. Nath, S.; Daneshvar, K.; Roy, L.; Grover, P.; Kidiyoor, A.; Mosley, L.; Sahraei, M.; Mukherjee, P. References MUC1 induces drug resistance in pancreatic cancer cells via upregulation of multidrug resistance genes. Oncogenesis 2013, 2, e51. [CrossRef] 56. Singh, P.K.; Wen, Y.; Swanson, B.J.; Shanmugam, K.; Kazlauskas, A.; Cerny, R.L.; Gendler, S.J.; Hollingsworth, M.A. Platelet-derived growth factor receptor β-mediated phosphorylation of MUC1 enhances invasiveness in pancreatic adenocarcinoma cells. Cancer Res. 2007, 67, 5201–5210. [CrossRef] 57. Roy, L.D.; Sahraei, M.; Subramani, D.B.; Besmer, D.; Nath, S.; Tinder, T.L.; Bajaj, E.; Shanmugam, K.; Lee, Y.Y.; Hwang, S.I. MUC1 enhances invasiveness of pancreatic cancer cells by inducing epithelial to mesenchymal transition. Oncogene 2011, 30, 1449–1459. [CrossRef] [PubMed] Vaccines 2020, 8, 659 16 of 21 16 of 21 58. Sahraei, M.; Roy, L.D.; Curry, J.M.; Teresa, T.L.; Nath, S.; Besmer, D.; Kidiyoor, A.; Dalia, R.; Gendler, S.J.; Mukherjee, P. MUC1 regulates PDGFA expression during pancreatic cancer progression. Oncogene 2012, 31, 4935–4945. [CrossRef] 9. Grover, P.; Nath, S.; Nye, M.D.; Zhou, R.; Ahmad, M.; Mukherjee, P. SMAD4-independent activation of TG signaling by MUC1 in a human pancreatic cancer cell line. Oncotarget 2018, 9, 6897. [CrossRef] [PubMe 60. Grover, P.; Nath, S.; Bose, M.; Sanders, A.J.; Brouwer, C.; Fnu, N.; Zhou, R.; Yazdanifar, M.; Ahmed, M.; Wu, S.-T. Tumor-Associated MUC1 Regulates TGF-β Signaling and Function in Pancreatic Ductal Adenocarcinoma. bioRxiv 2020. [CrossRef] 61. Wang, X.-T.; Kong, F.-B.; Mai, W.; Li, L.; Pang, L.-M. MUC1 immunohistochemical expression as a prognostic factor in gastric cancer: Meta-analysis. Dis. Markers 2016, 9421571. [CrossRef] [PubMed] 62. Scott, A.M.; Allison, J.P.; Wolchok, J.D. Monoclonal antibodies in cancer therapy. Cancer Immun. Arch. 2012, 12, 14–22. 63. Weiner, L.M.; Murray, J.C.; Shuptrine, C.W. Antibody-based immunotherapy of cancer. Cell 2012, 148, 1081–1084. [CrossRef] 4. Redman, J.; Hill, E.; AlDeghaither, D.; Weiner, L. Mechanisms of action of therapeutic antibodies for ca Mol. Immunol. 2015, 67, 28–45. [CrossRef] 64. Redman, J.; Hill, E.; AlDeghaither, D.; Weiner, L. Mechanisms of action of therapeutic antibodies for cancer. Mol. Immunol. 2015, 67, 28–45. [CrossRef] 65. Shuptrine, C.W.; Surana, R.; Weiner, L.M. Monoclonal antibodies for the treatment of cancer. In Seminars in 5. Shuptrine, C.W.; Surana, R.; Weiner, L.M. Monoclonal antibodies for the treatment of cancer. In Semina Cancer Biology; Elsevier: Amsterdam, The Netherlands, 2012; pp. 3–13. 66. Samadi, P.; Saki, S.; Dermani, F.K.; Pourjafar, M.; Saidijam, M. Emerging ways to treat breast cancer: Will promises be met? Cell. Oncol. 2018, 41, 605–621. [CrossRef] 67. Doi, M.; Yokoyama, A.; Kondo, K.; Ohnishi, H.; Ishikawa, N.; Hattori, N.; Kohno, N. References Anti-tumor eﬀect of the anti-KL-6/MUC1 monoclonal antibody through exposure of surface molecules by MUC1 capping. Cancer Sci. 2006, 97, 420–429. [CrossRef] 68. Ohyabu, N.; Hinou, H.; Matsushita, T.; Izumi, R.; Shimizu, H.; Kawamoto, K.; Numata, Y.; Togame, H.; Takemoto, H.; Kondo, H. An essential epitope of anti-MUC1 monoclonal antibody KL-6 revealed by focused glycopeptide library. J. Am. Chem. Soc. 2009, 131, 17102–17109. [CrossRef] 69. Namba, M.; Hattori, N.; Hamada, H.; Yamaguchi, K.; Okamoto, Y.; Nakashima, T.; Masuda, T.; Sakamoto, S.; Horimasu, Y.; Miyamoto, S. Anti-KL-6/MUC1 monoclonal antibody reverses resistance to trastuzumab-mediated antibody-dependent cell-mediated cytotoxicity by capping MUC1. Cancer Lett. 2019, 442, 31–39. [CrossRef] [PubMed] 70. Yamamoto, M.; Bhavanandan, V.; Nakamori, S.; Irimura, T. A novel monoclonal antibody speciﬁc for sialylated MUC1 mucin. Jpn. J. Cancer Res. 1996, 87, 488–496. [CrossRef] 71. Suzuki, H.; Shoda, J.; Kawamoto, T.; Shinozaki, E.; Miyahara, N.; Hotta, S.; Iizuka, Y.; Nakahara, A.; Tanaka, N.; Yanaka, A. Expression of MUC1 recognized by monoclonal antibody MY. 1E12 is a useful biomarker for tumor aggressiveness of advanced colon carcinoma. Clin. Exp. Metastasis 2004, 21, 321–329. [CrossRef] 72. Yoshimura, Y.; Denda-Nagai, K.; Takahashi, Y.; Nagashima, I.; Shimizu, H.; Kishimoto, T.; Noji, M.; Shichino, S.; Chiba, Y.; Irimura, T. Products of Chemoenzymatic Synthesis Representing MUC1 Tandem Repeat Unit with T-, ST-or STn-antigen Revealed Distinct Speciﬁcities of Anti-MUC1 Antibodies. Sci. Rep. 2019, 9, 1–12. [CrossRef] Muguruma, N.; Ito, S. Labeled anti-mucin antibody detectable by infrared-ﬂuorescence endoscopy. Cancer Biomark. 2008, 4, 321–328. [CrossRef] 74. Tarp, M.A.; Sørensen, A.L.; Mandel, U.; Paulsen, H.; Burchell, J.; Taylor-Papadimitriou, J.; Clausen, H. Identiﬁcation of a novel cancer-speciﬁc immunodominant glycopeptide epitope in the MUC1 tandem repeat. Glycobiology 2007, 17, 197–209. [CrossRef] [PubMed] 75. Wu, G.; Maharjan, S.; Kim, D.; Kim, J.N.; Park, B.K.; Koh, H.; Moon, K.; Lee, Y.; Kwon, H.-J. A novel monoclonal antibody targets Mucin1 and attenuates growth in pancreatic cancer model. Int. J. Mol. Sci. 2018, 19, 2004. [CrossRef] 76. Wu, G.; Kim, D.; Kim, J.N.; Park, S.; Maharjan, S.; Koh, H.; Moon, K.; Lee, Y.; Kwon, H.-J. A Mucin1 C-terminal subunit-directed monoclonal antibody targets overexpressed Mucin1 in breast cancer. Theranostics 2018, 8, 78. [CrossRef] 17 of 21 Vaccines 2020, 8, 659 77. Tinder, T.L.; Subramani, D.B.; Basu, G.D.; Bradley, J.M.; Schettini, J.; Million, A.; Skaar, T.; Mukherjee, P. MUC1 enhances tumor progression and contributes toward immunosuppression in a mouse model of spontaneous pancreatic adenocarcinoma. J. Immunol. 2008, 181, 3116–3125. [CrossRef] 78. References Mukherjee, P.; Madsen, C.S.; Ginardi, A.R.; Tinder, T.L.; Jacobs, F.; Parker, J.; Agrawal, B.; Longenecker, B.M.; Gendler, S.J. Mucin 1-speciﬁc immunotherapy in a mouse model of spontaneous breast cancer. J. Immunother. 2003, 26, 47–62. [CrossRef] 79. Dréau, D.; Moore, L.J.; Wu, M.; Roy, L.D.; Dillion, L.; Porter, T.; Puri, R.; Momin, N.; Wittrup, K.D.; Mukherjee, P. Combining the speciﬁc anti-MUC1 antibody TAB004 and Lip-MSA-IL-2 limits pancreatic cancer progression in immune competent murine models of pancreatic ductal adenocarcinoma. Front. Oncol. 2019, 9, 330. [CrossRef] 80. Moore, L.J.; Roy, L.D.; Zhou, R.; Grover, P.; Wu, S.-T.; Curry, J.M.; Dillon, L.M.; Puri, P.M.; Yazdanifar, M.; Puri, R. Antibody-guided in vivo imaging for early detection of mammary gland tumors. Transl. Oncol. 2016, 9, 295–305. [CrossRef] 81. Bose, M.; Mukherjee, P. A Novel Antibody Blocks Anti-Apoptotic Activity of MUC1 in Pancreatic Cancer Cell Lines. Cancer Res. 2019, 79 (Suppl. 13), 2052. [CrossRef] 82. Yazdanifar, M.; Zhou, R.; Grover, P.; Williams, C.; Bose, M.; Moore, L.J.; Wu, S.-t.; Maher, J.; Dreau, D.; Mukherjee, P. Overcoming Immunological Resistance Enhances the Eﬃcacy of a Novel Anti-tMUC1-CAR T Cell Treatment against Pancreatic Ductal Adenocarcinoma. Cells 2019, 8, 1070. [CrossRef] 83. Calvete, J.A.; Newell, D.R.; Wright, A.F.; Rose, M.S. In vitro and in vivo antitumor activity of ZENECA ZD0490, a recombinant ricin A-chain immunotoxin for the treatment of colorectal cancer. Cancer Res. 1994, 54, 4684–4690. 84. Baeckström, D.; Hansson, G.C.; Nilsson, O.; Johansson, C.; Gendler, S.J.; Lindholm, L. Puriﬁcation and characterization of a membrane-bound and a secreted mucin-type glycoprotein carrying the carcinoma-associated sialyl-Lea epitope on distinct core proteins. J. Biol. Chem. 1991, 266, 21537–21547. 85. Tolcher, A.W.; Ochoa, L.; Hammond, L.A.; Patnaik, A.; Edwards, T.; Takimoto, C.; Smith, L.; de Bono, J.; Schwartz, G.; Mays, T. Cantuzumab mertansine, a maytansinoid immunoconjugate directed to the CanAg antigen: A phase, I.; pharmacokinetic, and biologic correlative study. J. Clin. Oncol. 2003, 21, 211–222. [CrossRef] 86. Rodon, J.; Garrison, M.; Hammond, L.A.; De Bono, J.; Smith, L.; Forero, L.; Hao, D.; Takimoto, C.; Lambert, J.M.; Pandite, L. Cantuzumab mertansine in a three-times a week schedule: A phase I and pharmacokinetic study. Cancer Chemother. Pharmacol. 2008, 62, 911–919. [CrossRef] 87. Mita, M.; Ricart, A.; Mita, A.; Patnaik, A.; Sarantopoulos, J.; Sankhala, K.; Fram, R.; Qin, A.; Watermill, J.; Tolcher, A. A phase I study of a CanAg-targeted immunoconjugate, huC242-DM4, in patients with Can Ag-expressing solid tumors. J. Clin. Oncol. 2007, 25 (Suppl. 18), 3062. References [CrossRef] 88. Goﬀ, L.; Papadopoulos, K.; Posey, J.; Phan, A.; Patnaik, A.; Miller, J.; Zildjian, S.; O’Leary, J.; Qin, A.; Tolcher, A. A phase II study of IMGN242 (huC242-DM4) in patients with CanAg-positive gastric or gastroesophageal (GE) junction cancer. J. Clin. Oncol. 2009, 27 (Suppl. 15), e15625. [CrossRef] 89. Gold, D.V.; Cardillo, T.; Vardi, Y.; Blumenthal, R. Radioimmunotherapy of experimental pancreatic cancer with 131I-labeled monoclonal antibody PAM4. Int. J. Cancer 1997, 71, 660–667. [CrossRef] 90. Cardillo, T.M.; Ying, Z.; Gold, D.V. Therapeutic advantage of 90yttrium-versus 131iodine-labeled PAM4 antibody in experimental pancreatic cancer. Clin. Cancer Res. 2001, 7, 3186–3192. [PubMed] 91. Gold, D.V.; Karanjawala, Z.; Modrak, D.E.; Goldenberg, D.M.; Hruban, R.H. PAM4-reactive MUC1 is a biomarker for early pancreatic adenocarcinoma. Clin. Cancer Res. 2007, 13, 7380–7387. [CrossRef] [PubMed] 92. Gulec, S.; Pennington, K.; Bruetman, D.; Garl, S.; Horne, H.; Gold, D.; Wegener, W.; Goldenberg, D. A phase-I study of 90Y-hPAM4 (humanized anti-MUC1 monoclonal antibody) in patients with unresectable and metastatic pancreatic cancer. J. Nucl. Med. 2007, 48 (Suppl. 2), 393p. 93. Picozzi, V.J.; Ramanathan, R.K.; Lowery, M.A.; Ocean, A.J.; Mitchel, E.P.; O’Neil, B.H.; Guarino, M.J.; Conkling, P.R.; Cohen, S.J.; Bahary, N. 90Y-clivatuzumab tetraxetan with or without low-dose gemcitabine: A phase Ib study in patients with metastatic pancreatic cancer after two or more prior therapies. Eur. J. Cancer 2015, 51, 1857–1864. [CrossRef] 94. Dian, D.; Janni, W.; Kuhn, C.; Mayr, D.; Karsten, U.; Mylonas, I.; Friese, K.; Jeschke, U. Evaluation of a novel anti-mucin 1 (MUC1) antibody (PankoMab) as a potential diagnostic tool in human ductal breast cancer; comparison with two established antibodies. Oncol. Res. Treat. 2009, 32, 238–244. [CrossRef] 18 of 21 18 of 21 Vaccines 2020, 8, 659 95. Danielczyk, A.; Stahn, R.; Faulstich, D.; Löﬄer, A.; Märten, A.; Karsten, U.; Goletz, S. PankoMab: A potent new generation anti-tumour MUC1 antibody. Cancer Immunol. Immunother. 2006, 55, 1337–1347. [CrossRef] 96. Fiedler, W.; DeDosso, S.; Cresta, S.; Weidmann, J.; Tessari, A.; Salzberg, M.; Dietrich, B.; Baumeister, H.; Goletz, S.; Gianni, L. A phase I study of PankoMab-GEX, a humanised glyco-optimised monoclonal antibody to a novel tumour-speciﬁc MUC1 glycopeptide epitope in patients with advanced carcinomas. Eur. J. Cancer 2016, 63, 55–63. [CrossRef] [PubMed] 97. Ledermann, J.; Sehouli, J.; Zurawski, B.; Raspagliesi, F.; De Giorgi, U.; Banerjee, S.; Arranz Arija, J.; Romeo Marin, M.; Lisyanskaya, A.; Póka, R. References LBA41A double-blind, placebo-controlled, randomized, phase 2 study to evaluate the eﬃcacy and safety of switch maintenance therapy with the anti-TA-MUC1 antibody PankoMab-GEX after chemotherapy in patients with recurrent epithelial ovarian carcinoma. Ann. Oncol. 2017, 28 (Suppl. 5), v605–v649. [CrossRef] 98. Runcie, K.; Budman, D.R.; John, V.; Seetharamu, N. Bi-speciﬁc and tri-speciﬁc antibodies-the next big thing in solid tumor therapeutics. Mol. Med. 2018, 24, 50. [CrossRef] 99. Mehla, K.; Tremayne, J.; Grunkemeyer, J.A.; O’Connell, K.A.; Steele, M.M.; Caﬀrey, T.C.; Zhu, X.; Yu, F.; Singh, P.K.; Schultes, B.C. Combination of mAb-AR20. 5, anti-PD-L1 and PolyICLC inhibits tumor progression and prolongs survival of MUC1. Tg mice challenged with pancreatic tumors. Cancer Immunol. Immunother. 2018, 67, 445–457. [CrossRef] 100. De Bono, J.; Rha, S.Y.; Stephenson, J.; Schultes, B.; Monroe, P.; Eckhardt, G.; Hammond, L.; Whiteside, T.; Nicodemus, C.; Cermak, J. Phase I trial of a murine antibody to MUC1 in patients with metastatic cancer: Evidence for the activation of humoral and cellular antitumor immunity. Ann. Oncol. 2004, 15, 1825–1833. [CrossRef] 101. Ibrahim, N.K.; Yariz, K.O.; Bondarenko, I.; Manikhas, A.; Semiglazov, V.; Alyasova, A.; Komisarenko, V.; Shparyk, Y.; Murray, J.L.; Jones, D. Randomized phase II trial of letrozole plus anti-MUC1 antibody AS1402 in hormone receptor–positive locally advanced or metastatic breast cancer. Clin. Cancer Res. 2011, 17, 6822–6830. [CrossRef] 102. Corraliza-Gorjón, I.; Somovilla-Crespo, B.; Santamaria, S.; Garcia-Sanz, J.A.; Kremer, L. New strategies using antibody combinations to increase cancer treatment eﬀectiveness. Front. Immunol. 2017, 8, 1804. [CrossRef] 103. Venepalli, N.K.; Gandhi, C.C.; Ozer, H.; Ho, D.; Lu, Y.; Xie, H.; Berg, S.A.; Chowdhery, R.A.; Gargano, M.A.; B A H Ph Ib t d f PGG b t l i bi ti ith ti MUC1 tib d (BTH1704) d 102. Corraliza-Gorjón, I.; Somovilla-Crespo, B.; Santamaria, S.; Garcia-Sanz, J.A.; Kremer, L. New strategies using antibody combinations to increase cancer treatment eﬀectiveness. Front. Immunol. 2017, 8, 1804. [CrossRef] 103. Venepalli, N.K.; Gandhi, C.C.; Ozer, H.; Ho, D.; Lu, Y.; Xie, H.; Berg, S.A.; Chowdhery, R.A.; Gargano, M.A.; Braun, A.H. Phase Ib study of PGG beta glucan in combination with anti-MUC1 antibody (BTH1704) and gemcitabine for the treatment of advanced pancreatic cancer. Am. Soc. Clin. Oncol. 2015, 33. [CrossRef] 104. Gold, D.V.; Newsome, G.; Liu, D.; Goldenberg, D.M. Mapping PAM4 (clivatuzumab), a monoclonal antibody in clinical trials for early detection and therapy of pancreatic ductal adenocarcinoma, to MUC5AC mucin. Mol. Cancer 2013, 12, 143. [CrossRef] 105. References Gold, D.V.; Cardillo, T.; Goldenberg, D.M.; Sharkey, R.M. Localization of pancreatic cancer with radiolabeled monoclonal antibody PAM4. Crit. Rev. Oncol./Hematol. 2001, 39, 147–154. [CrossRef] 106. Hisatsune, A.; Nakayama, H.; Kawasaki, M.; Horie, I.; Miyata, T.; Isohama, Y.; Kim, K.C.; Katsuki, H. Anti-MUC1 antibody inhibits EGF receptor signaling in cancer cells. Biochem. Biophys. Res. Commun. 2011, 405, 377–381. [CrossRef] [PubMed] 107. Hollingsworth, M.A.; Swanson, B.J. Mucins in cancer: Protection and control of the cell surface. Nat. Rev. Cancer 2004, 4, 45–60. [CrossRef] [PubMed] 108. Vivero-Escoto, J.L.; Jeﬀords, L.M.; Dréau, D.; Alvarez-Berrios, M.; Mukherjee, P. Mucin1 antibody-conjugated dye-doped mesoporous silica nanoparticles for breast cancer detection in vivo. In Colloidal Nanoparticles for Biomedical Applications XII: 2017; International Society for Optics and Photonics: San Francisco, CA, USA, 2017; p. 100780B. 109. Curry, J.M.; Thompson, K.J.; Rao, S.G.; Besmer, D.M.; Murphy, A.M.; Grdzelishvili, V.Z.; Ahrens, W.A.; McKillop, I.H.; Sindram, D.; Iannitti, D.A. The use of a novel MUC1 antibody to identify cancer stem cells and circulating MUC1 in mice and patients with pancreatic cancer. J. Surg. Oncol. 2013, 107, 713–722. [CrossRef] 110. Kelly, V.J.; Wu, S.-T.; Gottumukkala, V.; Coelho, R.; Palmer, K.; Nair, S.; Erick, T.; Puri, R.; Ilovich, O.; Mukherjee, P. Preclinical evaluation of an 111In/225Ac theranostic targeting transformed MUC1 for triple negative breast cancer. Theranostics 2020, 10, 6946. [CrossRef] 19 of 21 Vaccines 2020, 8, 659 19 of 21 111. Yonezawa, S.; Kitajima, S.; Higashi, M.; Osako, M.; Horinouchi, M.; Yokoyama, S.; Kitamoto, S.; Yamada, N.; Tamura, Y.; Shimizu, T. A novel anti-MUC1 antibody against the MUC1 cytoplasmic tail domain: Use in sensitive identiﬁcation of poorly diﬀerentiated cells in adenocarcinoma of the stomach. Gastric Cancer 2012, 15, 370–381. [CrossRef] 112. Qi, W.; Schultes, B.C.; Liu, D.; Kuzma, M.; Decker, W.; Madiyalakan, R. Characterization of an anti-MUC1 monoclonal antibody with potential as a cancer vaccine. Hybrid. Hybridomics 2001, 20, 313–324. [CrossRef] 113. Kearse, K.P.; Smith, N.L.; Semer, D.A.; Eagles, L.; Finley, J.L.; Kazmierczak, S.; Kovacs, C.J.; Rodriguez, A.A.; Kellogg-Wennerberg, A.E. Monoclonal antibody DS6 detects a tumor-associated sialoglycotope expressed on human serous ovarian carcinomas. Int. J. Cancer 2000, 88, 866–872. [CrossRef] 114. Trombe, M.; Caron, A.; Tellier, A.; Carrez, C.; Guérif, S.; Clavier, S.; Karst, N.; Saarinen, J.; Satomaa, T.; Pitkänen, V. Preclinical Activity of an Antibody Drug Conjugate Targeting Tumor Speciﬁcmuc1 Structural Peptide-Glycotope. Cancer Res. 2019, 235. [CrossRef] 115. References Gomez-Roca, C.A.; Boni, V.; Moreno, V.; Morris, J.C.; Delord, J.-P.; Calvo, E.; Papadopoulos, K.P.; Rixe, O.; Cohen, P.; Tellier, A. A phase I study of SAR566658, an anti CA6-antibody drug conjugate (ADC), in patients (Pts) with CA6-positive advanced solid tumors (STs) (NCT01156870). Am. Soc. Clin. Oncol. 2016, 34, 2511. [CrossRef] 116. St. Croix, B.; Sheehan, C.; Rak, J.W.; Flørenes, V.A.; Slingerland, J.M.; Kerbel, R.S. E-cadherin–dependent growth suppression is mediated by the cyclin-dependent kinase inhibitor p27KIP1. J. Cell Biol. 1998, 142, 557–571. [CrossRef] [PubMed] 117. Matsumura, K.; Niki, I.; Tian, H.; Takuma, M.; Hongo, N.; Matsumoto, S.; Mori, H. Radioimmunoscintigraphy of pancreatic cancer in tumor-bearing athymic nude mice using 99 m technetium-labeled anti-KL-6/MUC1 antibody. Radiat. Med. 2008, 26, 133–139. [CrossRef] 118. Li, Y.; Zhou, C.; Li, J.; Liu, J.; Lin, L.; Li, L.; Cao, D.; Li, Q.; Wang, Z. Single domain based bispeciﬁc antibody, Muc1-Bi-1, and its humanized form, Muc1-Bi-2, induce potent cancer cell killing in muc1 positive tumor cells. PLoS ONE 2018, 13, e0191024. [CrossRef] 119. Kodama, H.; Suzuki, M.; Katayose, Y.; Shinoda, M.; Sakurai, N.; Takemura, S.-I.; Yoshida, H.; Saeki, H.; Asano, R.; Ichiyama, M. Speciﬁc and eﬀective targeting cancer immunotherapy with a combination of three bispeciﬁc antibodies. Immunol. Lett. 2002, 81, 99–106. [CrossRef] 120. Hinoda, Y.; Nakagawa, N.; Ohe, Y.; Kakiuchi, H.; Tsujisaki, M.; Imai, K.; Yachi, A. Recognition of the polypeptide core of mucin by monoclonal antibody MUSE11 against an adenocarcinoma-associated antigen. Jpn. J. Cancer Res. 1990, 81, 1206–1209. [CrossRef] 121. Katayose, Y.; Kudo, T.; Suzuki, M.; Shinoda, M.; Saijyo, S.; Sakurai, N.; Saeki, H.; Fukuhara, K.; Imai, K.; Matsuno, S. MUC1-speciﬁc targeting immunotherapy with bispeciﬁc antibodies: Inhibition of xenografted human bile duct carcinoma growth. Cancer Res. 1996, 56, 4205–4212. 122. Takemura, S.-I.; Asano, R.; Tsumoto, K.; Ebara, S.; Sakurai, N.; Katayose, Y.; Kodama, H.; Yoshida, H.; Suzuki, M.; Imai, K. Construction of a diabody (small recombinant bispeciﬁc antibody) using a refolding system. Protein Eng. 2000, 13, 583–588. [CrossRef] 123. Takemura, S.-I.; Kudo, T.; Asano, R.; Suzuki, M.; Tsumoto, K.; Sakurai, N.; Katayose, Y.; Kodama, H.; Yoshida, H.; Ebara, S. A mutated superantigen SEA D227A fusion diabody speciﬁc to MUC1 and CD3 in targeted cancer immunotherapy for bile duct carcinoma. Cancer Immunol. Immunother. 2002, 51, 33–44. [CrossRef] 124. Schuhmacher, J.; Klivényi, G.; Kaul, S.; Henze, M.; Matys, R.; Hauser, H.; Clorius, J. Pretargeting of human mammary carcinoma xenografts with bispeciﬁc anti-MUC1/anti-Ga chelate antibodies and immunoscintigraphy with PET. References Nucl. Med. Biol. 2001, 28, 821–828. [CrossRef] 125. King, T.; Posey, A. Co-expression of an engineered cell-surface sialidase by CART cells improves anti-cancer activity of NK cells in solid tumors. Cytotherapy 2019, 21, S27. [CrossRef] 126. Brooks, C.L.; Schietinger, A.; Borisova, S.N.; Kufer, P.; Okon, M.; Hirama, T.; MacKenzie, C.R.; Wang, L.-X.; Schreiber, H.; Evans, S.V. Antibody recognition of a unique tumor-speciﬁc glycopeptide antigen. Proc. Natl. Acad. Sci. USA 2010, 107, 10056–10061. [CrossRef] 127. Borgert, A.; Heimburg-Molinaro, J.; Song, X.; Lasanajak, Y.; Ju, T.; Liu, M.; Thompson, P.; Ragupathi, G.; Barany, G.; Smith, D.F. Deciphering structural elements of mucin glycoprotein recognition. ACS Chem. Biol. 2012, 7, 1031–1039. [CrossRef] 20 of 21 Vaccines 2020, 8, 659 128. Springer, G.F. Immunoreactive T and Tn epitopes in cancer diagnosis, prognosis, and immunotherapy. J. Mol. Med. 1997, 75, 594–602. [CrossRef] [PubMed] 129. Kudelka, M.R.; Ju, T.; Heimburg-Molinaro, J.; Cummings, R.D. Simple sugars to complex disease—Mucin-type O-glycans in cancer. In Advances in Cancer Research; Elsevier: Amsterdam, The Netherlands, 2015; Volume 126, pp. 53–135. 130. Karsten, U.; Serttas, N.; Paulsen, H.; Danielczyk, A.; Goletz, S. Binding patterns of DTR-speciﬁc antibodies reveal a glycosylation-conditioned tumor-speciﬁc epitope of the epithelial mucin (MUC1). Glycobiology 2004, 14, 681–692. [CrossRef] 131. Matsushita, T.; Takada, W.; Igarashi, K.; Naruchi, K.; Miyoshi, R.; Garcia-Martin, F.; Amano, M.; Hinou, H.; Nishimura, S.-I. A straightforward protocol for the preparation of high performance microarray displaying synthetic MUC1 glycopeptides. Biochim. Biophys. Acta (BBA)-Gen. Subj. 2014, 1840, 1105–1116. [CrossRef] 132. Rangappa, S.; Artigas, G.; Miyoshi, R.; Yokoi, Y.; Hayakawa, S.; Garcia-Martin, F.; Hinou, H.; Nishimura, S.-I. Eﬀects of the multiple O-glycosylation states on antibody recognition of the immunodominant motif in MUC1 extracellular tandem repeats. MedChemComm 2016, 7, 1102–1122. [CrossRef] 133. Martínez-Sáez, N.; Castro-López, J.; Valero-González, J.; Madariaga, D.; Compañón, I.; Somovilla, V.J.; Salvadó, M.; Asensio, J.L.; Jiménez-Barbero, J.; Avenoza, A. Deciphering the Non-equivalence of serine and threonine O-glycosylation points: Implications for molecular recognition of the Tn antigen by an anti-MUC1 antibody. Angew. Chem. Int. Ed. 2015, 54, 9830–9834. [CrossRef] 134. Wakui, H.; Tanaka, Y.; Ose, T.; Matsumoto, I.; Kato, K.; Min, Y.; Tachibana, T.; Sato, M.; Naruchi, K.; Martin, F.G. A straightforward approach to antibodies recognising cancer speciﬁc glycopeptidic neoepitopes. Chem. Sci. 2020, 11, 4999–5006. [CrossRef] 135. Thie, H.; Toleikis, L.; Li, J.; von Wasielewski, R.; Bastert, G.; Schirrmann, T.; Esteves, I.T.; Behrens, C.K.; Fournes, B.; Fournier, N. Rise and fall of an anti-MUC1 speciﬁc antibody. 145. McGuckin, M.A.; Hurst, T.G.; Ward, B.G. Heterogeneity in production, secretion and glycosylation of MUC1 epithelial mucin by primary cultures of ovarian carcinoma. Int. J. Cancer 1995, 63, 412–418. [CrossRef] References PLoS ONE 2011, 6, e15921. [CrossRef] 136. Moreno, M.; Bontkes, H.J.; Scheper, R.J.; Kenemans, P.; Verheijen, R.H.; von Mensdorﬀ-Pouilly, S. High level of MUC1 in serum of ovarian and breast cancer patients inhibits huHMFG-1 dependent cell-mediated cytotoxicity (ADCC). Cancer Lett. 2007, 257, 47–55. [CrossRef] [PubMed] 137. Tang, Y.; Cui, X.; Xiao, H.; Qi, S.; Hu, X.; Yu, Q.; Shi, G.; Zhang, X.; Gu, J.; Yu, Y. Binding of circulating anti-MUC1 antibody and serum MUC1 antigen in stage IV breast cancer. Mol. Med. Rep. 2017, 15, 2659–2664. [CrossRef] [PubMed] 138. Treon, S.P.; Maimonis, P.; Bua, D.; Young, G.; Raje, N.; Mollick, J.; Chauhan, D.; Tai, Y.-T.; Hideshima, T.; Shima, Y. Elevated soluble MUC1 levels and decreased anti-MUC1 antibody levels in patients with multiple myeloma. Blood J. Am. Soc. Hematol. 2000, 96, 3147–3153. 139. Storr, S.J.; Royle, L.; Chapman, C.J.; Hamid, U.M.A.; Robertson, J.F.; Murray, A.; Dwek, R.A.; Rudd, P.M. The O-linked glycosylation of secretory/shed MUC1 from an advanced breast cancer patient’s serum. Glycobiology 2008, 18, 456–462. [CrossRef] 140. Pegram, M.; Borges, V.; Fuloria, J.; Ibrahim, N.; Shapiro, C.; Perez, E.; Wang, K.; Schaedeli Stark, F.; Yeon, C.; Courtenay-Luck, N. Phase I pharmacokinetics (PK) of humanized anti-MUC-1 antibody R1550. J. Clin. Oncol. 2006, 24 (Suppl. 18), 2533. [CrossRef] pp 141. Zhou, D.; Xu, L.; Huang, W.; Tonn, T. Epitopes of MUC1 tandem repeats in cancer as revealed by antibody crystallography: Toward glycopeptide signature-guided therapy. Molecules 2018, 23, 1326. [CrossRef] 142. Sørensen, A.L.; Reis, C.A.; Tarp, M.A.; Mandel, U.; Ramachandran, K.; Sankaranarayanan, V.; Schwientek, T.; Graham, R.; Taylor-Papadimitriou, J.; Hollingsworth, M.A. Chemoenzymatically synthesized multimeric Tn/STn MUC1 glycopeptides elicit cancer-speciﬁc anti-MUC1 antibody responses and override tolerance. Glycobiology 2006, 16, 96–107. [CrossRef] 143. Rivalland, G.; Loveland, B.; Mitchell, P. Update on Mucin-1 immunotherapy in cancer: A clinical perspective. Expert Opin. Biol. Ther. 2015, 15, 1773–1787. [CrossRef] 144. Singh, R.; Bandyopadhyay, D. MUC1: A target molecule for cancer therapy. Cancer Biol. Ther. 2007, 6, 481–486. [CrossRef] 145. McGuckin, M.A.; Hurst, T.G.; Ward, B.G. Heterogeneity in production, secretion and glycosylation of MUC1 epithelial mucin by primary cultures of ovarian carcinoma. Int. J. Cancer 1995, 63, 412–418. [CrossRef] 21 of 21 21 of 21 Vaccines 2020, 8, 659 146. Walsh, M.D.; Luckie, S.M.; Cummings, M.C.; Antalis, T.M.; McGuckin, M.A. Heterogeneity of MUC1 expression by human breast carcinoma cell lines in vivo and in vitro. Breast Cancer Res. Treat. 1999, 58, 253–264. [CrossRef] [PubMed] 147. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Zhao, Q.; Piyush, T.; Chen, C.; Hollingsworth, M.A.; Hilkens, J.; Rhodes, J.M.; Yu, L.-G. MUC1 extracellular domain confers resistance of epithelial cancer cells to anoikis. Cell Death Dis. 2014, 5, e1438. [CrossRef] [PubMed] 148. Pichinuk, E.; Benhar, I.; Jacobi, O.; Chalik, M.; Weiss, L.; Ziv, R.; Sympson, C.; Karwa, A.; Smorodinsky, N.I.; Rubinstein, D.B. Antibody targeting of cell-bound MUC1 SEA domain kills tumor cells. Cancer Res. 2012, 72, 3324–3336. [CrossRef] 149. Madsen, C.B.; Wandall, H.H.; Pedersen, A.E. Potential for novel MUC1 glycopeptide-speciﬁc antibody in passive cancer immunotherapy. Immunopharmacol. Immunotoxicol. 2013, 35, 649–652. [CrossRef] 150. Reilly, R.M.; Sandhu, J.; Alvarez-Diez, T.M.; Gallinger, S.; Kirsh, J.; Stern, H. Problems of delivery of monoclonal antibodies. Clin. Pharmacokinet. 1995, 28, 126–142. [CrossRef] [PubMed] 151. Sonbol, M.B.; Ahn, D.H.; Goldstein, D.; Okusaka, T.; Tabernero, J.; Macarulla, T.; Reni, M.; Li, C.-P.; O’Neil, B.; Van Cutsem, E. CanStem111P trial: A Phase III study of napabucasin plus nab-paclitaxel with gemcitabine. Future Oncol. 2019, 15, 1295–1302. [CrossRef] , , [ ] 152. Bose, M.; Vora, A.; Colleton, T.; Mukherjee, P. MUC1 Confers Sensitivity to STAT-3 Inhibitor Napabucasin in Pancreatic Ductal Adenocarcinoma Cells. Cancer Res. 2020, 1837. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4376958544	https://www.research-collection.ethz.ch/bitstream/20.500.11850/614707/2/EcologyandEvolution-2023-Mller-FindingfoodinachangingworldSmall%e2%80%90scaleforaginghabitatpreferencesofan.pdf	English	null	Finding food in a changing world: Small‐scale foraging habitat preferences of an insectivorous passerine in the Alps	Ecology and evolution	2,023	cc-by	13,115	ETH Library Author(s): Rights / license: Creative Commons Attribution 4.0 International Originally published in: Ecology and Evolution 13(5), https://doi.org/10.1002/ece3.10084 Originally published in: This page was generated automatically upon download from the ETH Zurich Research Collection. For more information, please consult the Terms of use. Received: 27 January 2023 \| Revised: 20 April 2023 \| Accepted: 27 April 2023 Received: 27 January 2023 \| Revised: 20 April 2023 \| Accepted: 27 April 2023 DOI: 10.1002/ece3.10084 DOI: 10.1002/ece3.10084 R E S E A R C H A R T I C L E R E S E A R C H A R T I C L E Finding food in a changing world: Small-scale foraging habitat preferences of an insectivorous passerine in the Alps Thomas M. Müller1,2 \| Christoph M. Meier1 \| Florian Knaus2 \| Pius Korner1 \| Barbara Helm1 \| Valentin Amrhein1,3 \| Yann Rime1,3 Thomas M. Müller1,2 \| Christoph M. Meier1 \| Florian Knaus2 \| Pius Korner1 \| Barbara Helm1 \| Valentin Amrhein1,3 \| Yann Rime1,3 1Swiss Ornithological Institute, Sempach, Switzerland 2Department of Environmental Systems Sciences, Swiss Federal Institute of Technology Zurich (ETH Zurich), Zurich, Switzerland 3Department of Environmental Sciences, Zoology, University of Basel, Basel, Switzerland Correspondence Thomas M. Müller, Swiss Ornithological Institute, Sempach 6204, Switzerland. Email: thomas_mueller@hotmail.com Funding information Eidgenössische Technische Hochschule Zürich; Swiss Ornithological Institute 1Swiss Ornithological Institute, Sempach, Switzerland 2Department of Environmental Systems Sciences, Swiss Federal Institute of Technology Zurich (ETH Zurich), Zurich, Switzerland 3Department of Environmental Sciences, Zoology, University of Basel, Basel, Switzerland Correspondence Thomas M. Müller, Swiss Ornithological Institute, Sempach 6204, Switzerland. Email: thomas_mueller@hotmail.com Funding information Eidgenössische Technische Hochschule Zürich; Swiss Ornithological Institute Abstract Abstract Organisms living in high-elevation habitats are usually habitat specialists who occupy a narrow ecological niche. To envision the response of alpine species to a changing environment, it is fundamental to understand their habitat preferences on multiple spatial and temporal scales. However, information on small-scale habitat use is still widely lacking. We investigated the foraging habitat preferences of the migratory northern wheatear Oenanthe oenanthe during the entire presence at a breeding site in the central Alps. We repeatedly observed 121 adult and juvenile individuals. We applied Bayesian logistic regression models to investigate which habitat character- istics influenced foraging habitat selection on a fine spatial scale, and how habitat use varied temporally. Throughout their presence on the breeding grounds, northern wheatears showed a consistent preference for a mosaic of stones and bare ground patches with slow-growing, short vegetation. The proximity of marmot burrows was preferred, whereas dense and low woody vegetation was avoided. After arrival at the breeding site, short vegetation, preferably close to the snow, was favored. The prefer- ence for open habitat patches that provide access to prey underlines the critical role of small-scale habitat heterogeneity for northern wheatears. The strong and consist- ent preference for a habitat that is under pressure from land-use and climate change suggests that this alpine bird species may be sensitive to habitat loss, leading to a potential range contraction. We highlight the need to conserve habitat diversity on a small spatial scale to ensure the long-term availability of suitable habitat for northern wheatears in the Alps. Correspondence K E Y W O R D S alpine birds, elevation, ground cover, habitat heterogeneity, insectivorous T A X O N O M Y C L A S S I F I C A T I O N Behavioural ecology, Biodiversity ecology, Community ecology, Ecosystem ecology, Global change ecology, Landscape ecology, Life history ecology, Phenology, Spatial ecology 2 of 16 \| 1 \| INTRODUCTION Food abundance and accessibility, however, are often promoted by different habitat characteristics and are temporally variable (Atkinson et al., 2004; Dussault et al., 2005; Fuller et al., 2007). Particularly, species with nar- row requirements, so-called specialists, are expected to be relatively sensitive to changes in food availability (McPeek, 1996). Typically, alpine species are often adapted to a short vegetation period, and they are restricted to a higher elevational range that is characterized by habitat heterogeneity on a finer scale, compared to lowland hab- itats (Cortés & Wheeler, 2018). For insectivorous alpine birds, prey abundance is driven by a stronger seasonality at high elevation (Pilar et al., 2020; Resano-Mayor et al., 2019). Arthropod abundance, di- versity, and species richness peak in early summer and then decrease to relatively low levels until autumn (Pilar et al., 2020). Consequently, the time window is limited for prey availability to match food demand for brood provisioning, for expensive maintenance such as molt, and for juvenile post-fledging establishment (Arlt & Pärt, 2008; Resano- Mayor et al., 2019; Tulp & Schekkerman, 2008). The ecological niche of a species is defined on multiple spatial and temporal scales (Mahon et al., 2016). Hence, to understand or pre- serve a species, it is necessary to identify its relevant habitat prefer- ences from large-scale distributions to small-scale habitat features. Here, we conducted an observational study on uniquely iden- tifiable individuals to determine the preferred foraging habitat of northern wheatears in their Alpine breeding range throughout their stay. We focused on the microhabitat at foraging locations and com- pared it with the available habitat at random locations within the territory. We investigated the role of vegetation height and ground cover composition in providing accessibility to prey. Foraging pref- erences may change throughout the annual cycle. Therefore, we considered the birds' entire presence at the breeding site, including during the pre-breeding and postbreeding periods. This also covers key processes such as molt and premigratory fuel deposition, as well as the high-risk phase of post-fledging establishment of juveniles. To determine the role of prey accessibility on Alpine breeding grounds, we examined the importance of small-scale heterogeneity in pro- viding suitable foraging habitat. Furthermore, we explored the role of grazing cattle and alpine marmots (Marmota marmota) in shaping habitat heterogeneity on a small scale. Alpine regions are more vulnerable to climate change than low- elevation areas (Brunetti et al., 2009). 2 of 16 \| 1 \| INTRODUCTION 2 of 16 MÜLLER et al. and shifting vegetation and prey phenology on multiple spatial and temporal scales (Jähnig et al., 2020; Sander et al., 2021, 2022). The species has a circumpolar distribution and overwinters in sub-Saharan Africa (Bairlein et al., 2012; Dunn et al., 2020; Meier et al., 2022; Rime et al., 2023). In Northern European lowland breed- ing sites, where seasonality is less strong compared to alpine hab- itats, northern wheatears favor open fields with short vegetation (Arlt et al., 2008; Arlt & Pärt, 2007; Paquet et al., 2019) and seem to be more limited by prey accessibility than by prey abundance (van Oosten et al., 2014). Unlike lowland breeding ranges, in Switzerland, the species is limited to high elevations above the tree line (Knaus et al., 2018). While in most parts of Europe, northern wheatear pop- ulations are declining, the Swiss Alpine population is stable overall while experiencing an upward shift in the elevational distribution (Hallman et al., 2022; Keller et al., 2020; Knaus et al., 2018). The population trend of the Alpine northern wheatear population points toward an increasingly important role of alpine habitats for the conservation of this species in central Europe (Knaus et al., 2018). This Alpine population faces spatial and temporal landscape dynam- ics that are different from those in the European lowland (Brunetti et al., 2009; Pilar et al., 2020). To examine the sensitivity of the spe- cies to current and future habitat changes and shifting vegetation phenology in the Alps, it is important to understand how the species interacts with the highly seasonal and variable habitat that the alpine ecosystem provides on a fine spatial and temporal scale. The ecological niche of a species is defined on multiple spatial and temporal scales (Mahon et al., 2016). Hence, to understand or pre- serve a species, it is necessary to identify its relevant habitat prefer- ences from large-scale distributions to small-scale habitat features. The availability of suitable foraging habitat plays a special role in the niche configuration and is crucial for survival and successful repro- duction. More specifically, food availability, comprised of food abun- dance and accessibility, is a major driver of foraging habitat selection that is influenced by habitat features on a fine scale (Arlettaz, 1999; Barras et al., 2020; Cody, 1985; Dussault et al., 2005). K E Y W O R D S K E Y W O R D S alpine birds, elevation, ground cover, habitat heterogeneity, insectivorous T A X O N O M Y C L A S S I F I C A T I O N Behavioural ecology, Biodiversity ecology, Community ecology, Ecosystem ecology, Global change ecology, Landscape ecology, Life history ecology, Phenology, Spatial ecology This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2023 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd. \| 1 of 16 www.ecolevol.org Ecology and Evolution. 2023;13:e10084. https://doi.org/10.1002/ece3.10084 2 of 16 \| 1 \| INTRODUCTION They experience adverse effects of rising temperatures, altered precipitation patterns, as well as advanced snowmelt and vegetation development that lead to an upward shift of the treeline (Gehrig-Fasel et al., 2007; Gobiet et al., 2014; Keller et al., 2005; Theurillat & Guisan, 2001). Moreover, land-use changes influence vegetation development in alpine areas through two opposed processes (Kulakowski et al., 2011): Low- intensity agricultural activities such as livestock grazing are being abandoned, leading to bush encroachment and ultimately to forest encroachment (Baur et al., 2006), while areas that are still managed tend to undergo agricultural intensification (Fischer et al., 2008). Land-use and climate change have fundamental effects on the major- ity of organisms across trophic levels, through either the loss of suit- able habitat or shifting vegetation phenology (Ferrarini et al., 2017; Hughes, 2000; Inouye, 2020; Keller et al., 2005). For migratory birds in particular, advanced vegetation phenology can lead to a potential phenological mismatch (Jones & Cresswell, 2010; Saino et al., 2011; Visser et al., 2004), because it reduces prey accessibility for ground- foraging species as a result of increased vegetation height and of ad- vances in the peak in arthropod abundance (Renner & Zohner, 2018; Tulp & Schekkerman, 2008). 2.1 \| Study area Our study area is located in Val Piora in the central Swiss Alps (46°33′N 8°42′E, Figure 1). It covers 6 km2 of mostly south- exposed slopes above the tree line, ranging from 1850 to 2200 m.a.s.l. and hosting more than 100 breeding pairs of north- ern wheatears. The habitat is characterized by heterogenous open As a long-distance migratory songbird, the northern wheatear (Oenanthe oenanthe) is affected by changing habitat characteristics 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library \| 3 of 16 MÜLLER et al. FI G U R E 1 The map (a) shows the position of all foraging points (green dots) recorded in the study area in Val Piora. Foraging habitat data were recorded for color-ringed northern wheatears (b) on a 1-m radius around foraging (presence) and pseudo-absence locations (c). Pseudo-absence locations were located randomly within 20–80 m and at a random angle (relative to true North) of each foraging location. background map: ©swisstopo, photos: ©Y. Rime. MÜLLER et al. 3 of 16 Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rule (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License FI G U R E 1 The map (a) shows the position of all foraging points (green dots) recorded in the study area in Val Piora. Foraging habitat data were recorded for color-ringed northern wheatears (b) on a 1-m radius around foraging (presence) and pseudo-absence locations (c). Pseudo-absence locations were located randomly within 20–80 m and at a random angle (relative to true North) of each foraging location. background map: ©swisstopo, photos: ©Y. Rime. 2.3 \| Statistical analysis We recorded the following set of habitat variables (Table 1 and Table S1) on a 1-m radius around foraging (presence) and pseudo- absence locations (Figure 1): ground cover estimates (percentage of live vegetation, dead vegetation, woody vegetation, bare ground, stones [granulometry >4 mm], and snow) and vegetation height. We calculated the vegetation height using the mean of three representa- tive measurements within the 1-m radius. Additionally, we estimated the distance to the closest marmot burrow and recorded cattle graz- ing activity, immediate cattle presence, and presence of cow dung within the 1-m radius. For each foraging and pseudo-absence loca- tion, we also computed the distance to the nest if it was found, and the normalized difference vegetation index (NDVI) and its rate of change between months. The distance to the nest was calculated based on the SwissALTI3D digital elevation model (swisstopo, 2018) in QGIS (QGIS Development Team, 2020). NDVI raster images for the study area were generated on Google Earth Engine (Gorelick et al., 2017) based on Sentinel-2 satellite images with a spatial res- olution of 10 m (ESA, 2015). After applying a cloud filter (<50% cloud area), the image with the clearest conditions for each month (April–September 2021) was manually selected, and the NDVI val- ues were extracted in R (R Core Team, 2021) using the extract func- tion from the package raster (Hijmans, 2021). To detect local shifts in We modeled the foraging habitat selection by comparing the re- corded variables between foraging (presence) and pseudo-absence locations using logistic regression models (logit-link function) with presence/absence as a binary outcome variable. In all models, the ground cover estimates, vegetation height, distance to marmot bur- row, NDVI, and its rate of change were included as fixed effects. To account for individual preferences and repeated observations of the same individual, we included the bird ID (color ring combination) and the point ID (unique number for each presence/pseudo-absence pair) as random effects (Korner-Nievergelt et al., 2015; Laird & Ware, 1982). All statistical analyses were conducted in R (R Core Team, 2021). Models were fitted in a Bayesian framework (Gelman et al., 2013; McElreath, 2016), using the brm function from the brms package (Bürkner, 2017). For each model, we ran four chains, each with 2000 iterations of which the first 1000 were discarded as the burn-in period (McElreath, 2016). greenness, the rate of NDVI change was computed as the difference between the NDVI values extracted from the images of the previous and the following month of the foraging event at each foraging and pseudo-absence location. To allow for a comparison between differ- ent habitat scales, we additionally recorded the same set of variables on a 2-m radius around the foraging and pseudo-absence locations. and three plastic color rings (Figure 1). Adults were caught using baited spring traps and cage-traps that were placed at the nest en- trance. Where they could be reached, chicks were ringed at their nest once they were 7 days old. We observed ringed individuals between May 12 and September 03, 2021, covering the arrival, incubation, feeding (period of food pro- visioning for chicks), and postbreeding stages. Northern wheatears arrive on their Alpine breeding grounds between the end of April and mid-May and depart for fall migration around mid-September (Glutz Von Blotzheim & Bauer, 2001; Meier et al., 2022; Rime et al., 2023; Sander et al., 2021). During this period, ringed birds were followed weekly from the distance, using binoculars and a scope, until the first foraging attempt of each observation. We recorded the exact location of the foraging event on a photograph taken through the scope. After the bird had left the foraging location, we immediately mapped the microhabitat on site and recorded the exact coordinates and information on the individual (color ring combination, sex, age, and nest ID) in QField (QGIS Development Team, 2020). To com- pare the foraging (presence) locations with locations that have not been chosen by the bird, we mapped the microhabitat at a nearby location within a randomly selected distance of 20–80 m to the for- aging location at a random angle (0°–360°) for each foraging event (Figure 1, Barbet-Massin et al., 2012; Johnson, 1980). This dis- tance range was selected to ensure that pseudo-absence locations were located within the territory of the observed bird (Glutz Von Blotzheim & Bauer, 2001). Adult birds usually remained within their territory throughout their entire stay in the study area, including for foraging activities (Rime et al., 2023). To make sure that each presence-pseudo-absence pair is independent, we moved on to the next territory after having recorded all ringed individuals sighted within their territory. As the birds' needs are expected to change during their pres- ence at the study site, we assigned three stages to each of the for- aging events on a per-breeding pair basis. The arrival and incubation stage lasts until the chicks hatch after an incubation period of 13– 15 days (Moreno, 1989a). This is followed by a feeding period that includes feeding chicks 13–15 days in the nest and feeding fledglings for 10 days out of the nest until they become largely independent (Glutz Von Blotzheim & Bauer, 2001; Moreno, 1984). The post- breeding period includes the remaining time until both adults and juveniles depart for fall migration (Arlt & Pärt, 2008). During this period, young wheatears must establish themselves, and both the adults and immatures undergo complete molt and deposit fuel for their long-distance migratory journey (Arlt & Pärt, 2008; Glutz Von Blotzheim & Bauer, 2001). In total, we recorded 620 foraging locations and an equal number of pseudo-absence locations (ntot = 1240) during the period of pres- ence of northern wheatears in the study area (Figure 1). We followed 121 ringed individuals (53 adult males, 47 adult females, and 21 juve- niles). Sixty-nine adults were returning individuals ringed in previous years, while 31 adults and 21 juveniles were newly ringed during the study period. We collected data for 193 foraging locations during the arrival and incubation stage, 193 during the feeding stage, and 182 during the postbreeding stage, of which 38 were from juveniles. nlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2.2 \| Study design grassland interspersed with rocks, boulders, debris fields, and re- mains of man-made rockpiles and stonewalls. Between July and September, the pastures are grazed in a rotational manner and the cattle are frequently moved, constituting a low-intensity grazing regime. The area is usually covered by snow between November and May. In the frame of a project on migration and ecology of northern wheatears, individuals have been ringed in the study area since 2010 (Meier et al., 2022; Rime et al., 2023; Schmaljohann et al., 2016). Each bird was ringed with a unique combination of one metal ring In the frame of a project on migration and ecology of northern wheatears, individuals have been ringed in the study area since 2010 (Meier et al., 2022; Rime et al., 2023; Schmaljohann et al., 2016). Each bird was ringed with a unique combination of one metal ring 4 of 16 MÜLLER et al. 2.3 \| Statistical analysis A prior sensitivity analysis (Figure S1) suggested that the model results were sufficiently robust to changing prior specification (Depaoli & van de Schoot, 2017; Link et al., 2002; Nicenboim et al., 2021). Hence, we chose uninforma- tive priors for our models (Berger, 2006; Kass & Wasserman, 1996; MÜLLER et al. Category Variables Description Unit Analysis Transformation General model (n = 542) Arrival and incubation (n = 193) Feeding (n = 190) Postbreeding (n = 182) Fixed effects Mean (range) (2.5%; 97.5%) Mean (range) (2.5%; 97.5%) Mean (range) (2.5%; 97.5%) Mean (range) (2.5%; 97.5%) Ground cover Dead vegetation Brown vegetation grown in the previous year % 10 (0–100) (0; 64.9) 25.9 (0–100) (0; 80) 5.3 (0–85) (0; 38.6) 3.6 (0–40) (0; 29.6) Polya, zb Woody vegetation Shrubs and other plants with a woody stem % 6.2 (0–100) (0; 70) 3.9 (0–93) (0; 40.8) 5.5 (0–100) (0; 67.6) 6.7 (0–100) (0; 76.5) Polya, zb Bare ground Bare ground with no vegetation cover % 17.3 (0–100) (0; 64.9) 20.9 (0–90) (0; 65) 17.7 (0–100) (0; 70) 12.7 (0–85) (0; 45.9) Polya, zb Stones Stones and rocks (granulometry >4 mm) % 12.1 (0–100) (0; 65) 9.6 (0–100) (0; 56.1) 10.2 (0–73) (0; 50) 14.3 (0–100) (0; 70) Polya, zb Snow cover Area covered by snow % Not included 15.5 (0:100) (0; 100) Not included Not included Polya, zb Vegetation height Mean height of three representative measurements cm 13.24 (0–94) (1; 36) 5.45 (0–55) (0; 18) 14.91 (0–94) (3; 36) 15.58 (0–55) (3; 38) zb Ecosystem engineers Distance to marmot burrow Distance to the closest marmot burrow (if within 100 m) m 7.93 (0–130) (0; 27) 8.75 (0–60) (0.63; 32.75) 8.88 (0–130) (0.48; 33.05) 6.86 (0–76) (1; 22.85) zb Vegetation Index NDVI Normalized difference vegetation index Index 0.65 (0–0.92) (0.1; 0.87) 0.38 (0–0.88) (0; 0.81) 0.68 (0–0.92) (0.15; 0.88) 0.72 (0.28–0.9) (0.52; 0.87) Polya, zb Rate of NDVI change Change in vegetation greenness Index change rate 0.18 (−0.36 to 0.83) (−0.22; 0.69) 0.4 (−0.05 to 0.86) (0; 0.74) 0.2 (−0.32 to 0.82) (−0.08; 0.73) −0.02 (−0.36 to 0.58) (−0.27; 0.3) Polya, zb Random effects No. of levels No. of levels No. of levels No. of levels Point info Point ID Unique identifier for each point pair 542 193 190 182 Bird info Bird ID Color ring combination 119 57 72 95 Note: Snow cover was only included in the arrival and incubation model. 3 \| RESULTS Our models revealed a positive effect of short vegetation and bare ground on the foraging probability (presence vs. pseudo-absence) of northern wheatears after the snow has melted, while the habi- tat characteristics changed as the season advanced. The most com- mon ground cover type at foraging and pseudo-absence locations was live vegetation with a mean ± SD of 54.3% ± 30.0%, followed by bare ground (17.3% ± 17.0%), stones (12.1% ± 17.8%), dead vegetation (10.0% ± 17.3%), and woody vegetation (6.2% ± 16.9; Table 1). Snow was only present during the arrival and incubation period (15.5% ± 33.6%; Table 1). Characteristic seasonal developments were observed with decreasing snow, bare ground, and dead vegetation covers, while live vegetation increased as the season advanced (Figure S5). Due to the strong seasonality in the study area, snow can only be expected at the beginning of the season. As a result, we only used snow cover in the arrival and incubation model. Furthermore, snow cover may lead to biased relative estimates for the other ground covers. Therefore, all locations containing snow (n = 156) were re- moved from the general model. Whenever foraging locations had to be removed, the corresponding pseudo-absence location was dis- carded as well. Because ground cover variables always added up to 100%, they could not all be included in the models. Therefore, the main ground cover component, live vegetation, was not used in the models. Visual data exploration did not suggest differences in the topographic variables between foraging and pseudo-absence loca- tions, which can be explained by the small distance between them (Figure S3). Therefore, topographic variables were not included in statistical models. Similarly, grazing variables were discarded, as they always fell into the same category due to the small distance between foraging and corresponding pseudo-absence locations. We did not apply any further model selection steps, and no interactions were considered. The birds' foraging and pseudo-absence locations had a similar av- erage vegetation cover in May and June. However, pseudo-absence locations rose to a higher level of live vegetation before stabilizing at the beginning of June. After that, mean cover of live vegetation remained higher at pseudo-absence locations compared to foraging locations until the end of the study period. Nevertheless, the general seasonal patterns followed the same trend in foraging and pseudo- absence locations (Figure S5). Zhou et al., 2014). For the intercept and the group-level vari- ances (bird ID and foraging ID), we chose default student-t priors (훽∼Student(3, 0, 2.5)) and determined a normal prior distribution for the population-level effects (훽∼Normal(0, 100)). package (Gräler et al., 2016). In addition, we calculated the area under the curve (AUC) and visually evaluated the goodness of fit (Figure S4) by comparing the fitted values with the data (Korner- Nievergelt et al., 2015). For each model, we calculated the con- ditional and marginal Nakagawa's R2 (Nakagawa et al., 2017; Nakagawa & Schielzeth, 2013) using the performance package (Lüdecke et al., 2021). Prior to modeling, numeric variables were z-transformed (mean = 0, SD = 1). As we expected nonlinear relationships, we in- cluded the first two orthogonal polynomials of the ground cover variables and the vegetation index variables in the models using the poly function. We checked for collinearity between covariates by calculating the Spearman's correlation coefficient and did not detect strong collinearity among explanatory variables (all \|rs\| < .7). To quantify the effect of each predictor on the foraging prob- ability (probability of presence), we present effect plots for each predictor (Korner-Nievergelt et al., 2015). To do so, for each draw from the posterior distribution, we calculated the regression line over the range of the variable that is shown in the effect plot. From these regression lines, we used the median as a point estimate re- gression line and the 2.5% and 97.5% quantiles as 95% credible in- terval (CrI; Korner-Nievergelt et al., 2015). When showing the effect of a ground cover variable across its range, the remaining area was divided among the other ground cover variables (including live veg- etation) proportional to their mean proportions across all locations (and snow cover was set to zero). This was done due to the unit-sum constraint of ground cover variables. Data, code, and supplementary material used in this study are available under the DOI: 10.5281 at https://doi.org/10.5281/zenodo.7805040 (Müller et al., 2023). Observations in the field suggested potential differences in for- aging habitat preferences between adult and juvenile birds. To de- tect differential preferences of northern wheatears that are related to their age class (adult, juvenile) or sex (female, male), we applied principal component analysis (PCA) using the variables summa- rized in Table 1. PCA were generated with the ggbiplot R package (Vu, 2011) but did not reveal relevant differences between age classes or sexes (Figure S2). As a result, age and sex class were not included in the models. To detect stage-dependent differences in foraging habitat pref- erences during the study period, we analyzed each of the three stages in a separate model, in addition to a general model including the data from the entire study period. To compare foraging habitat preferences across different scales, we also fitted each of the four models with the data collected on the 2-m radius around the forag- ing and pseudo-absence locations. 2.3 \| Statistical analysis For each model and variable, the mean per 1-m radius plot, range, 2.5% quantile, and 97.5% quantile are given, and the applied data transformation is provided. Point ID and bird ID were included as random effects for which the number of levels is given for each model. aThe first two orthogonal polynomials were included. bCentered to 0 and scaled to 1 SD. 6 of 16 MÜLLER et al. 3 \| RESULTS Overall, vegetation height at foraging and pseudo-absence points had a mean ± SD of 13.24 cm ± 9.46 cm and increased throughout the study period. In accordance with the changing ground cover composition and vegetation development, the mean NDVI value was 0.65 ± 0.19 and increased throughout the season. The mean rate of NDVI change was 0.18 ± 0.24, indicating an increase in vegetation greenness from May to July until it started to decline in August (Figure S5). We verified model convergence based on Gelman–Rubin convergence diagnostics and visually confirmed convergence using “trace” plots (MCMC plots; Depaoli & van de Schoot, 2017; Rizzo, 2008). We checked for autocorrelation within the MCMC chains using the mcmc_plot function from the bayesplot package (Gabry & Mahr, 2021). Additionally, we checked for spatial auto- correlation using bubble plots and semivariograms from the gstat 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable C 6 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; O 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/te 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for r MÜLLER et al. 0457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commo \| FI G U R E 2 Predictions from logistic regression models showing the average effect (solid line) of each ground cover variable (labeled on the righthand side of the plots) on the foraging probability (presence vs. pseudoabsence) for the general model (whole study period; first column) and each period separately (other columns) within 1 m of the foraging (presence = 1) and pseudo-absence (0) locations. “Live vegetation” was not used as a predictor in the model but it is a derived parameter from the other ground cover parameters and is given here because all ground covers add up to 100%. 3 \| RESULTS The colored areas represent the 95% Bayesian credible intervals and the gray dots show the raw data. 58, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creativ FI G U R E 2 Predictions from logistic regression models showing the average effect (solid line) of each ground cover variable (labeled on the righthand side of the plots) on the foraging probability (presence vs. pseudoabsence) for the general model (whole study period; first column) and each period separately (other columns) within 1 m of the foraging (presence = 1) and pseudo-absence (0) locations. “Live vegetation” was not used as a predictor in the model but it is a derived parameter from the other ground cover parameters and is given here because all ground covers add up to 100%. The colored areas represent the 95% Bayesian credible intervals and the gray dots show the raw data. 8 of 16 MÜLLER et al. TA B LE 2 Summary of the output of the general (all-season) model, arrival and incubation, feeding (food provisioning for chicks), and postbreeding model using the 1-m data. 4 \| DISCUSSION During this period, the mean vegetation height at foraging points was 10.3 cm (±7.3 cm) and 16.2 cm (±10.4 cm) at pseudo-absence points. Bare ground was positively related to the foraging probability, especially during the feeding period (Table 2, Figure 2). However, no effect of bare ground was found for the arrival and incubation stage when short vegetation and melting snow patches prevailed. Stone cover had a positive effect: during the feeding period, only the linear effect was well supported by the data, whereas in all other models, a maximum probability of foraging was observed at an intermediate (20%–70%) stone cover (Table 2, Figure 2). Locations with low stone cover (<15%) were less likely to be chosen for foraging (Figure 2). In the general model, woody vegetation showed a maximum at a low woody vegetation cover (Figure 2). Especially while feeding, woody vegetation had a nega- tive effect on the foraging probability (Table 2). Locations with more than 40% woody vegetation were never used as foraging locations in the feeding period (Figure 2). In the postbreeding period, woody vegetation only had a weak negative effect (Table 2). Dead vegeta- tion did not play an important role during any of the periods and only showed a weak negative trend in the general as well as the arrival and incubation models (Table 2, Figure 2). During the arrival and in- cubation period, snow cover showed a strong quadratic effect, indi- cating a high foraging probability at low to intermediate snow cover levels (Table 2, Figure 2). Locations with more than 60% snow were avoided (Figure 2). Foraging attempts were never observed directly on snow, even when it still covered a large part of the study area. Nonetheless, our NDVI results indicate that vegetation produc- tivity is an important component of the foraging microhabitat. This result must be interpreted in the context of larger-scale effects. The minimal spatial resolution of sentinel-2 satellite data is 10 m, which means that the available information summarizes a larger area than the sampling locations, informing on the productivity in the habi- tat matrix around the foraging location. Even though patches with bare ground and stones were preferred on a small scale, they lay within the territories in the study area where heterogeneous and productive grassland is the dominating habitat type. 4 \| DISCUSSION 4 Based on AUC values as well as marginal and conditional R2, the all-season model (AUC 0.87, R2 marginal .59, R2 conditional .6), the arrival and incubation model (AUC 0.83, R2 marginal .78, R2 con- ditional .78), the feeding model (food provisioning for chicks, AUC 0.92, R2 marginal .88, R2 conditional .89), and the postbreeding model (AUC 0.92, R2 marginal .74, R2 conditional .74), all performed well. The difference between the marginal and the conditional R2 was consistently small, indicating a small effect of the random fac- tors (i.e., individual and local preferences). Our study highlights the importance of small-scale characteristics in the foraging preferences of a long-distance migrant breeding in high-alpine habitats. Accessibility to the ground and habitat heter- ogeneity determined, on a very fine scale, whether a location was chosen for foraging. Habitat structure and ground cover composition changed as the season advanced, but northern wheatears generally showed similar habitat preferences throughout their presence in the study area. Interestingly, the habitat preferences were consistent between females and males as well as between adults and juveniles. We found a specific preference for open patches, interspersed with stones within vegetated areas, where prey abundance is expected to be higher (Morris, 2000). This underpins that a diverse habitat is nec- essary to sustain food availability for northern wheatears through- out their stay on the Alpine breeding grounds. Preferred foraging habitat in the study area was composed of multiple types of ground cover. Especially the presence of bare ground patches seemed im- portant, which allow birds to detect and access prey more easily than in the surrounding vegetation (Schaub et al., 2010; Vickery & Arlettaz, 2012). In particular, bare ground plays a crucial role during food provisioning for chicks, when food demand is enhanced and vegetation is growing fast (Moreno, 1989b). Rocks and boulders may have played a similar role, as they served as perching positions, al- lowing the birds to detect prey more easily. Particularly in the post- breeding period, stones may also have hosted an increased amount of prey, as we have repeatedly observed birds picking ants and other prey items from boulders or directly from anthills located in rocky areas; this was not the case earlier in the season. Vegetation height had a strong negative linear effect on the for- aging probability (Figure 2), with the effect being strongest while feeding (Table 2). 9 of 16 9 of 16 3 \| RESULTS Variables General model Arrival and incubation Feeding Postbreeding Estimate CrI Estimate CrI Estimate CrI Estimate CrI 2.5% 97.5% 2.5% 97.5% 2.5% 97.5% 2.5% 97.5% Bird ID (Intercept) 0.13 0.01 0.35 0.19 0.01 0.54 0.35 0.02 0.94 0.22 0.01 0.64 Point ID (Intercept) 0.09 0.00 0.25 0.14 0.01 0.41 0.22 0.01 0.64 0.21 0.01 0.62 Dead vegetation −0.04 −0.24 0.15 −0.25 −0.64 0.13 0.14 −0.20 0.46 −0.24 −0.62 0.10 Dead vegetation2 0.21 0.06 0.37 0.14 −0.15 0.43 0.16 −0.16 0.45 −0.05 −0.39 0.27 Woody vegetation −0.20 −0.50 0.07 −0.54 −1.41 0.09 −2.74 −5.43 −0.70 −0.11 −0.54 0.28 Woody vegetation2 −0.33 −0.61 −0.09 −0.67 −1.46 −0.10 −1.96 −3.61 −0.62 −0.01 −0.42 0.36 Bare ground 0.81 0.62 1.01 0.09 −0.27 0.44 1.26 0.82 1.78 1.03 0.63 1.50 Bare ground2 −0.18 −0.36 0.00 −0.05 −0.36 0.27 −0.27 −0.67 0.20 −0.22 −0.56 0.18 Stones 1.15 0.96 1.37 0.41 0.10 0.73 1.58 1.10 2.18 1.80 1.31 2.35 Stones2 −0.49 −0.67 −0.29 −0.45 −0.73 −0.17 −0.37 −0.79 0.13 −0.95 −1.34 −0.52 Snow −2.92 −4.54 −1.72 Snow2 −1.18 −1.84 −0.66 Vegetation height −1.04 −1.29 −0.80 −1.10 −1.58 −0.65 −1.59 −2.18 −1.05 −1.21 −1.69 −0.78 Distance to marmot burrow −0.42 −0.64 −0.21 −0.37 −0.71 −0.05 −0.44 −0.92 0.01 −0.68 −1.18 −0.20 NDVI 0.49 0.29 0.70 0.08 −0.31 0.47 0.71 0.31 1.16 0.50 0.07 0.96 NDVI2 0.01 −0.16 0.18 0.03 −0.30 0.36 0.26 −0.11 0.68 0.02 −0.34 0.36 Rate of NDVI change −0.29 −0.49 −0.10 −0.18 −0.49 0.13 −0.62 −1.10 −0.18 −0.40 −0.82 −0.01 Rate of NDVI change2 0.03 −0.14 0.19 −0.18 −0.51 0.14 −0.30 −0.72 0.09 0.27 −0.13 0.66 Note: Foraging versus pseudo-absence points were modeled using a logistic regression with logit-link function. Bird ID and point ID are random effects, while all other variables are fixed effects. Given are the estimate and lower (2.5%) and upper (97.5%) limits of the 95% credible interval. Estimates where the 95% credible interval does not contain zero are highlighted in bold. library.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License MÜLLER et al. 4 \| DISCUSSION On the one hand, this result implies that northern wheatears selected produc- tive areas for foraging that offer high arthropod abundance and diversity (Morris, 2000), which increases with vegetation height (Atkinson et al., 2004). On the other hand, high vegetation decreases visibility and access to the ground (Atkinson et al., 2004; Vickery & Arlettaz, 2012) and reduces the probability of a foraging attempt being successful (Dennis et al., 2008). As a ground-foraging insec- tivore, the northern wheatear requires visibility of and access to the ground for foraging (Arlt & Pärt, 2007; van Oosten et al., 2014). The preference for short vegetation on a fine scale suggests that prey accessibility is more limiting for successful foraging than prey abundance. This result is consistent with findings from study sites in the lowland of Northern Europe. In the Netherlands, where prey abundance remains stable throughout the breeding season, Increasing distance to the closest marmot burrow had a negative effect throughout the study period (Table 2, Figure 3), indicating a preference for foraging locations close to burrows (Figure 3). This effect was strongest at the postbreeding stage (Table 2). Except for the arrival and incubation stage, NDVI had a positive linear effect in each model, being strongest during the feeding period (Table 2, Figure 3). Even though vegetation was greening and growing fast during arrival and incubation (Figure S5), the rate of NDVI change had no strong effect on the foraging probability at that stage (Table 2), but it had a negative effect in the other models (Table 2, Figure 3). In the postbreeding stage, the rate of NDVI change had a slightly negative linear effect (Table 2, Figure 3). We ran all models based on a 2-m-radius with very similar results (Table S2, Figure S6): Although some effects were stronger on the smaller scale, the general patterns were the same (Table 2, Table S2, Figure S6). 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Onl 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles a MÜLLER et al. 10 of 16 10 of 16 \| MÜLLER et al. 4 \| DISCUSSION FI G U R E 3 Model predictions from logistic regression models showing the average effect (solid line) of vegetation height, distance to marmot burrow, NDVI, and the rate of NDVI change (labeled on the righthand side of the plots) on the foraging probability (presence vs. pseudo-absence) for the general model (whole study period; first column) and each period separately (other columns) within 1 m of the foraging (presence = 1) and pseudo-absence (0) locations. The colored areas represent the 95% Bayesian credible intervals and the gray dots show the raw data. FI G U R E 3 Model predictions from logistic regression models showing the average effect (solid line) of vegetation height, distance to marmot burrow, NDVI, and the rate of NDVI change (labeled on the righthand side of the plots) on the foraging probability (presence vs. pseudo-absence) for the general model (whole study period; first column) and each period separately (other columns) within 1 m of the foraging (presence = 1) and pseudo-absence (0) locations. The colored areas represent the 95% Bayesian credible intervals and the gray dots show the raw data. Even though woody vegetation reduces ground accessibility and was usually avoided, it played a specific role later in the season when it provided berries as an additional food source, explaining the observed weaker avoidance of this habitat type in the post- breeding season. We then observed northern wheatears foraging on Vaccinium myrtillus and Daphne mezereum berries. Coloring of the feces confirmed the consumption of berries (García-Rodríguez et al., 2022). Many insectivorous birds become frugivorous when their main food source becomes scarce (Bairlein, 2003; Fry, 1992). Berries are important sources of nutrients that may enhance molt and are crucial for migration (Bairlein, 2003; Eeva et al., 2018). Berries are therefore actively chosen, while including northern wheatears are more limited by prey accessibility than by prey abundance, as they preferentially forage in short grass (van Oosten et al., 2014). Similarly, northern wheatear populations had improved growth rates in short-vegetation habitats compared to tall field layers in Swedish farmland (Arlt et al., 2008), where the pres- ence of short vegetation is a major driver of population growth (Arlt et al., 2008; Paquet et al., 2019) and an important clue for habitat se- lection (Arlt & Pärt, 2007). This could be explained by a sparser and generally lower vegetation in more mineral-based high-elevation habitats. This is in line with the preference for more productive grasslands with an intermediate ground cover of stones and rock in our study area. At our study site, most northern wheatears remained in their territories throughout their presence (Rime et al., 2023). Other ground-dwelling insectivo- rous birds seem to be less capable of finding suitable habitat in their breeding territories as the season advances. For example, white- winged snowfinches (Montifringilla nivalis) rely on Tipulidae larvae at the retreating snow front (Brambilla et al., 2017; Resano-Mayor et al., 2019), a food resource used by adult northern wheatears only in the pre-breeding period, while ring ouzels (Turdus torquatus alpestris) rely mainly on earthworms and perform diel and seasonal ltit di l t t t k it bl f i h bit t i For northern wheatears, an upward shift in elevation has been ob- served in Switzerland, with the increase above 2400 m being higher than the loss at lower elevation, resulting in a stable or increasing general population trend (Hallman et al., 2022; Knaus et al., 2018). However, such an ongoing shift could lead to range contraction in the future (Dirnböck et al., 2003; Jähnig et al., 2020). Furthermore, winters in the Alps tend to become shorter, and spring greening-up is expected to advance earlier (Asam et al., 2018; Chamberlain & Pearce-Higgins, 2013; Gobiet et al., 2014). The resulting rise of the tree line, the increasing bush encroachment, and higher vegeta- tion density are threatening the availability of accessible foraging habitat also for the northern wheatear (Ceresa et al., 2021; Jähnig et al., 2020). Land-use change enhances population threats even further (Kulakowski et al., 2011): Agricultural intensification leads to landscape homogenization (Benton et al., 2003) and to higher nutrient levels, accelerating vegetation development and altering species composition (Dirnböck et al., 2003; Fischer et al., 2008). Even more apparent in alpine regions, pastoral abandonment leads to an increased vegetation height and eventually to shrub and for- est encroachment (Gehrig-Fasel et al., 2007; Kulakowski et al., 2011; Laiolo et al., 2004). On the other hand, low-intensity grazing of cat- tle positively influences ground-foraging birds (Atkinson et al., 2004; Laiolo et al., 2004; Vickery et al., 1999) and maintains suitable forag- ing habitat for the northern wheatear (Maron & Lill, 2005). Furthermore, habitat heterogeneity maintains food availability temporally by allowing diverse vegetation phenology to coexist and supply sufficient food throughout the season (Benton et al., 2003; Hovick et al., 2015; Vickery & Arlettaz, 2012). This is important because the habitat characteristics in the study area were strongly influenced by seasonal changes, while the species' forag- ing habitat preferences remained similar. The availability of suitable foraging habitat mainly depended on the progress of spring green- ing-up. In 2021, the area experienced a late and cold spring. When the birds arrived in the breeding region in May, most of their territo- ries were still covered by snow. During the melting period, the edges of snow fields played an important role, providing accessible habitat with high prey abundance (Barras et al., 2020; Brambilla et al., 2017; Leingärtner et al., 2014; Resano-Mayor et al., 2019). Once vege- tation growth increased and ground accessibility declined, habitat heterogeneity and the availability of open habitat patches became crucial in providing suitable foraging habitat. Similar results were found at a larger scale in Swedish farmland habitats, where fields with low vegetation became increasingly important for northern wheatears and positively influenced reproductive success later in the season (Arlt & Pärt, 2007). Sander et al. (2022) showed that nest survival of northern wheatears benefitted from a higher vegetation at another site in the Alps with a broader elevational gradient. This could be explained by a sparser and generally lower vegetation in more mineral-based high-elevation habitats. This is in line with the preference for more productive grasslands with an intermediate ground cover of stones and rock in our study area. At our study site, most northern wheatears remained in their territories throughout their presence (Rime et al., 2023). Other ground-dwelling insectivo- rous birds seem to be less capable of finding suitable habitat in their breeding territories as the season advances. For example, white- winged snowfinches (Montifringilla nivalis) rely on Tipulidae larvae at the retreating snow front (Brambilla et al., 2017; Resano-Mayor et al., 2019), a food resource used by adult northern wheatears only in the pre-breeding period, while ring ouzels (Turdus torquatus alpestris) rely mainly on earthworms and perform diel and seasonal altitudinal movements to track suitable foraging habitat as spring The preference for highly accessible patches within more produc- tive areas has been described for a variety of ground-foraging insec- tivorous farmland birds (Atkinson et al., 2004; Martinez et al., 2010; Schaub et al., 2010; Tagmann-Ioset et al., 2012; Vickery et al., 1999; Vickery & Arlettaz, 2012; Weisshaupt et al., 2011) as well as for al- pine specialists (Barras et al., 2020; Brambilla et al., 2017; Resano- Mayor et al., 2019). Food abundance for insectivores is higher in heterogeneous habitat (Cole et al., 2010), and fine-scale habitat di- versity provides accessible patches within species-rich landscapes that support high food abundance (Atkinson et al., 2004; Vickery & Arlettaz, 2012). Furthermore, habitat heterogeneity maintains food availability temporally by allowing diverse vegetation phenology to coexist and supply sufficient food throughout the season (Benton et al., 2003; Hovick et al., 2015; Vickery & Arlettaz, 2012). This is important because the habitat characteristics in the study area were strongly influenced by seasonal changes, while the species' forag- ing habitat preferences remained similar. The availability of suitable foraging habitat mainly depended on the progress of spring green- ing-up. In 2021, the area experienced a late and cold spring. When the birds arrived in the breeding region in May, most of their territo- ries were still covered by snow. During the melting period, the edges of snow fields played an important role, providing accessible habitat with high prey abundance (Barras et al., 2020; Brambilla et al., 2017; Leingärtner et al., 2014; Resano-Mayor et al., 2019). Once vege- tation growth increased and ground accessibility declined, habitat heterogeneity and the availability of open habitat patches became crucial in providing suitable foraging habitat. Similar results were found at a larger scale in Swedish farmland habitats, where fields with low vegetation became increasingly important for northern wheatears and positively influenced reproductive success later in the season (Arlt & Pärt, 2007). Sander et al. (2022) showed that nest survival of northern wheatears benefitted from a higher vegetation at another site in the Alps with a broader elevational gradient. 4 \| DISCUSSION The preference for short vegetation has been consistently described for lowland bird communities (Atkinson et al., 2004; Rime et al., 2020; Vickery & Arlettaz, 2012) as well as for other insectivorous alpine birds (Barras et al., 2020; Brambilla et al., 2017; Resano-Mayor et al., 2019). MÜLLER et al. 11 of 16 berries in an insectivorous diet most likely also reduces foraging energy expenditure and further supports fattening for migration (Lindström, 2003). Nevertheless, northern wheatears still pre- ferred open habitat in the postbreeding period, suggesting a suffi- cient abundance of arthropods (Beck et al., 2010; Pilar et al., 2020; Resano-Mayor et al., 2019). advances (Barras et al., 2020, 2021). Similarly, water pipits (Anthus spinoletta) perform within-season movements to avoid dense and high grassland as vegetation growth progresses (Ceresa et al., 2020). advances (Barras et al., 2020, 2021). Similarly, water pipits (Anthus spinoletta) perform within-season movements to avoid dense and high grassland as vegetation growth progresses (Ceresa et al., 2020). Due to this strong dependence of northern wheatears on the small-scale habitat mosaic that maintains suitable foraging habitat, the species is likely sensitive to climate and land-use change (Scridel et al., 2018; Theurillat & Guisan, 2001). In most parts of Europe, northern wheatear populations are declining, while the Alpine pop- ulations are stable overall (Gideon et al., 2014; Hallman et al., 2022; Issa & Muller, 2015; Keller et al., 2020; Knaus et al., 2018). Northern wheatears might be less vulnerable to climate change than other high-elevation specialists as long as micro-habitat heterogeneity is maintained. The rock ptarmigan (Lagopus muta) and the white- winged snowfinch (Montifringilla nivalis) for instance show a decrease in all but the uppermost part of their distributional range where populations remain stable (Issa & Muller, 2015; Keller et al., 2020; Knaus et al., 2018). The population trends suggest that these spe- cies are limited in their ability to find suitable habitat, even at higher altitudes. The preference for highly accessible patches within more produc- tive areas has been described for a variety of ground-foraging insec- tivorous farmland birds (Atkinson et al., 2004; Martinez et al., 2010; Schaub et al., 2010; Tagmann-Ioset et al., 2012; Vickery et al., 1999; Vickery & Arlettaz, 2012; Weisshaupt et al., 2011) as well as for al- pine specialists (Barras et al., 2020; Brambilla et al., 2017; Resano- Mayor et al., 2019). Food abundance for insectivores is higher in heterogeneous habitat (Cole et al., 2010), and fine-scale habitat di- versity provides accessible patches within species-rich landscapes that support high food abundance (Atkinson et al., 2004; Vickery & Arlettaz, 2012). (equal); writing – original draft (supporting); writing – review and editing (equal). Thomas M. Müller: Conceptualization (supporting); data curation (lead); formal analysis (lead); investigation (lead); meth- odology (equal); project administration (equal); software (equal); validation (equal); visualization (lead); writing – original draft (lead); writing – review and editing (lead). height stopped increasing and leveled off. Additionally, northern wheatears showed a preference for patches with stable vegeta- tion dynamics that ensure long-term habitat heterogeneity (Hovick et al., 2015; Vickery & Arlettaz, 2012). Furthermore, the foraging habitat of northern wheatears was probably positively influenced by alpine marmots, as northern wheatears where often foraging close to their burrows. Despite field observations suggesting marmots as potential nest predators, benefits of association with marmots seem to persist. Marmots maintain structural heterogeneity and accessible habitat by creating patches of bare ground, keeping the vegetation short, and potentially improving arthropod abundance and species richness (Ballová et al., 2019; Buyandelger et al., 2021; Buyandelger & Otgonbayar, 2022; Davidson et al., 2012). ORCID Thomas M. Müller https://orcid.org/0000-0001-6199-5011 Christoph M. Meier https://orcid.org/0000-0001-9584-2339 Florian Knaus https://orcid.org/0000-0003-3919-4730 Pius Korner https://orcid.org/0000-0002-7837-610X Barbara Helm https://orcid.org/0000-0002-6648-1463 Valentin Amrhein https://orcid.org/0000-0001-5173-4571 AUTHOR CONTRIBUTIONS Christoph M. Meier: Conceptualization (supporting); formal analy- sis (supporting); funding acquisition (equal); methodology (equal); project administration (equal); resources (equal); supervision (equal); writing – original draft (supporting); writing – review and editing (equal). Florian Knaus: Conceptualization (equal); formal analysis (supporting); funding acquisition (equal); methodology (equal); pro- ject administration (equal); resources (equal); supervision (equal); writing – original draft (supporting); writing – review and editing (equal). Pius Korner: Data curation (supporting); formal analysis (equal); methodology (supporting); software (equal); visualization (supporting); writing – original draft (supporting); writing – review and editing (equal). Barbara Helm: Funding acquisition (equal); pro- ject administration (supporting); resources (supporting); supervision (supporting); writing – original draft (supporting); writing – review and editing (equal). Valentin Amrhein: Supervision (supporting); vali- dation (supporting); writing – review and editing (equal). Yann Rime: Conceptualization (lead); data curation (supporting); formal analy- sis (equal); funding acquisition (lead); investigation (supporting); methodology (lead); project administration (lead); resources (equal); software (equal); supervision (lead); validation (equal); visualization CONFLICT OF INTEREST STATEMENT The authors declare that they have no conflict of interest. ACKNOWLEDGMENTS We thank Henri Descombes and Gilles Hauser for their assistance in the field and Fenna von Hirschheydt, Lena Wiest, and all the other field assistants who ringed birds for this project in previous years. Further, we also thank Fränzi Korner-Nievergelt, Jaime Resano- Mayor, Jérôme Guélat, Arnaud Barras, Dominik Hagist, and Felix Liechti for their valuable advice and support. Even though resource availability and habitat characteristics change temporally within a season, the foraging habitat preferences of northern wheatears remained similar at the study site. Northern wheatears depend on the availability of suitable foraging habitat within the same territory for the entire presence at the breeding site, even after the chicks are fully independent. Within an ecosystem that is characterized by spatiotemporal dynamics that are different to those in lowland habitats, Alpine northern wheatears inhabit an ecological niche that features a mosaic of accessible patches within vegetated areas that provide high prey abundance. Due to pressures from climate and land-use change on alpine ecosystems, this hab- itat is fragile and northern wheatears may be sensitive to habitat loss and range contraction. Our study emphasizes the importance of the Alpine breeding area for northern wheatears. It underlines the necessity to maintain and preserve the spatiotemporal availability of structural diversity and small-scale habitat heterogeneity that is critical in providing suitable foraging habitat for northern wheatears in the Alps in the long term. DATA AVAILABILITY STATEMENT The data and codes used in this study are deposited on Zenodo under the DOI: 10.5281 at https://doi.org/10.5281/zenodo.7805040 (Müller et al., 2023). FUNDING INFORMATION This study was funded by the Swiss Ornithological Institute (Vogelwarte Sempach) and the Swiss Federal Institute of Technology Zurich (ETH Zurich). It is im- portant to note that the positive effects of grazing on grassland bird communities are associated with low-intensity grazing, as applied in our study area, whereas high-intensity grazing can negatively affect them (Brambilla et al., 2020; Garcia-Pausas et al., 2017). Even though we were not able to quantify the effects of grazing with our method focusing on small-scale habitat parameters, low-intensity grazing is an important driver of landscape dynamics (Laiolo et al., 2004; Yoshihara et al., 2010). In areas that were grazed, vegetation height was lower and more heterogenous, and the growing dynamic was dis- rupted. With the onset of grazing in the study area, mean vegetation rary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licens 12 of 16 MÜLLER et al. org/10.1111/j.1365-2664.2007.01378.x Depaoli, S., & van de Schoot, R. (2017). Improving transparency and rep- lication in Bayesian statistics: The WAMBS-checklist. Psychological Methods, 22(2), 240–261. https://doi.org/10.1037/met0000065 Benton, T. G., Vickery, J. A., & Wilson, J. D. (2003). Farmland biodiversity: Is habitat heterogeneity the key? Trends in Ecology & Evolution, 18(4), 182–188. https://doi.org/10.1016/S0169-5347(03)00011-9 Dirnböck, T., Dullinger, S., & Grabherr, G. (2003). A regional im- pact assessment of climate and land-use change on alpine veg- etation. Journal of Biogeography, 30(3), 401–417. https://doi. org/10.1046/j.1365-2699.2003.00839.x Berger, J. (2006). The case for objective Bayesian analysis. Bayesian Analysis, 1(3), 385–402. https://doi.org/10.1214/06-BA115 Dunn, E. H., Hussell, D. J. T., Kren, J., & Zoerb, A. C. (2020). Northern wheatear (Oenanthe oenanthe). In Birds of the World. Cornell Lab of Ornithology. https://doi.org/10.2173/bow.norwhe.02 Brambilla, M., Cortesi, M., Capelli, F., Chamberlain, D., Pedrini, P., & Rubolini, D. (2017). Foraging habitat selection by alpine white- winged Snowfinches Montifringilla nivalis during the nestling rear- ing period. Journal of Ornithology, 158(1), 277–286. https://doi. org/10.1007/s10336-016-1392-9 Dussault, C., Ouellet, J. P., Courtois, R., Huot, J., Breton, L., & Jolicoeur, H. (2005). Linking moose habitat selection to limiting factors. Ecography, 28(5), 619–628. https://doi. org/10.1111/j.2005.0906-7590.04263.x Brambilla, M., Gustin, M., Cento, M., Ilahiane, L., & Celada, C. (2020). Habitat, climate, topography and management differently affect occurrence in declining avian species: Implications for conservation in changing environments. Science of the Total Environment, 742, 140663. https://doi.org/10.1016/j.scitotenv.2020.140663 Eeva, T., Holmström, H., Espín, S., Sánchez-Virosta, P., & Klemola, T. (2018). Leaves, berries and herbivorous larvae of bilberry Vaccinium myrtillus as sources of metals in food chains at a Cu-Ni smelter site. Chemosphere, 210, 859–866. https://doi.org/10.1016/j.chemo sphere.2018.07.099 Brunetti, M., Lentini, G., Maugeri, M., Nanni, T., Auer, I., Böhm, R., & Schöner, W. (2009). Climate variability and change in the greater alpine region over the last two centuries based on multi-variable analysis. International Journal of Climatology, 29(15), 2197–2225. https://doi.org/10.1002/joc.1857 ESA. (2015). ESA's Optical High-Resolution Mission for GMES Operational Services. Ferrarini, A., Alatalo, J. M., & Gustin, M. (2017). Climate change will se- riously impact bird species dwelling above the treeline: A prospec- tive study for the Italian Alps. Science of the Total Environment, 590– 591, 686–694. https://doi.org/10.1016/j.scitotenv.2017.03.027 Bürkner, P.-C. (2017). Brms: An R package for Bayesian multilevel models using Stan. Journal of Statistical Software, 80(1), 1–28. https://doi. org/10.18637/jss.v080.i01 Fischer, M., Rudmann-Maurer, K., Weyand, A., & Stöcklin, J. (2008). 13 of 16 13 of 16 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (htt Ceresa, F., Brambilla, M., Monrós, J. S., Rizzolli, F., & Kranebitter, P. (2020). Within-season movements of alpine songbird distributions are driven by fine-scale environmental characteristics. Scientific Reports, 10(1), 5747. https://doi.org/10.1038/s41598-020-62661-0 Bairlein, F. (2003). Nutritional strategies in migratory birds. In P. Berthold, E. Gwinner, & E. Sonnenschein (Eds.), Avian migration (pp. 321– 332). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3- 662-05957-9_22 Ceresa, F., Kranebitter, P., S Monrós, J., Rizzolli, F., & Brambilla, M. (2021). Disentangling direct and indirect effects of local temperature on abundance of mountain birds and implications for understanding global change impacts. PeerJ, 9, e12560. https://doi.org/10.7717/ peerj.12560 Bairlein, F., Norris, D. R., Nagel, R., Bulte, M., Voigt, C. C., Fox, J. W., Hussell, D. J. T., & Schmaljohann, H. (2012). Cross-hemisphere mi- gration of a 25 g songbird. Biology Letters, 8(4), 505–507. https:// doi.org/10.1098/rsbl.2011.1223 Chamberlain, D., & Pearce-Higgins, J. (2013). Impacts of climate change on upland birds: Complex interactions, compensatory mechanisms and the need for long-term data. IBIS, 155(3), 451–455. https://doi. org/10.1111/ibi.12070 Ballová, Z., Pekárik, L., Píš, V., & Šibík, J. (2019). How much do ecosystem engineers contribute to landscape evolution? A case study on Tatra marmots. Catena, 182(June), 104121. https://doi.org/10.1016/j. catena.2019.104121 Cody, M. L. (1985). Habitat selection in birds. Academic Press. Barbet-Massin, M., Jiguet, F., Albert, C. H., & Thuiller, W. (2012). Selecting pseudo-absences for species distribution models: How, where and how many? Methods in Ecology and Evolution, 3(2), 327–338. https:// doi.org/10.1111/j.2041-210X.2011.00172.x Cole, L. J., Pollock, M. L., Robertson, D., Holland, J. P., McCracken, D. I., & Harrison, W. (2010). The influence of fine-scale habitat heteroge- neity on invertebrate assemblage structure in upland semi-natural grassland. Agriculture, Ecosystems and Environment, 136(1–2), 69– 80. https://doi.org/10.1016/j.agee.2009.11.010 Barras, A. G., Liechti, F., & Arlettaz, R. (2021). Seasonal and daily move- ment patterns of an alpine passerine suggest high flexibility in re- lation to environmental conditions. Journal of Avian Biology, 52(12), 1–10. https://doi.org/10.1111/jav.02860 80. https://doi.org/10.1016/j.agee.2009.11.010 Cortés, A. J., & Wheeler, J. A. (2018). The environmental heterogeneity of mountains at a fine scale in a changing world. In C. Hoorn, A. Perrigo, & A. Antonelli (Eds.), Mountains, climate and biodiversity (p. 544). Wiley. 1–10. https://doi.org/10.1111/jav.02860 1–10. https://doi.org/10.1111/jav.02860 Barras, A. G., Marti, S., Ettlin, S., Vignali, S., Resano-Mayor, J., Braunisch, V., & Arlettaz, R. (2020). The importance of seasonal environmental factors in the foraging habitat selection of alpine ring ouzels Turdus torquatus alpestris. IBIS, 162(2), 505–519. https://doi.org/10.1111/ ibi.12764 Davidson, A. D., Detling, J. K., & Brown, J. H. (2012). Ecological roles and conservation challenges of social, burrowing, herbivorous mammals in the world's grasslands. Frontiers in Ecology and the Environment, 10(9), 477–486. https://doi.org/10.1890/110054 Baur, P., Bebi, P., Gellrich, M., & Rutherford, G. (2006). ‘Wasalp. Waldausdehnung im Schweizer Alpenraum’, Schlussbericht. NFP 48 ‘Landschaften und Lebensräume der Alpen’, Hrsg. WSL, p. 65 S. Dennis, P., Skartveit, J., McCracken, D. I., Pakeman, R. J., Beaton, K., Kunaver, A., & Evans, D. M. (2008). The effects of livestock grazing on foliar arthropods associated with bird diet in upland grasslands of Scotland. Journal of Applied Ecology, 45(1), 279–287. https://doi. org/10 1111/j 1365-2664 2007 01378 x Beck, J., Altermatt, F., Hagmann, R., & Lang, S. (2010). Seasonality in the altitude-diversity pattern of alpine moths. Basic and Applied Ecology, 11(8), 714–722. https://doi.org/10.1016/j. baae.2010.08.009 REFERENCES Arlettaz, R. (1999). Habitat selection as a major resource partitioning mechanism between the two sympatric sibling bat species Myotis myotis and Myotis blythii. Journal of Animal Ecology, 68(3), 460–471. https://doi.org/10.1046/j.1365-2656.1999.00293.x Arlt, D., Forslund, P., Jeppsson, T., & Pärt, T. (2008). Habitat-specific pop- ulation growth of a farmland bird. PLoS One, 3(8), e3006. https:// doi.org/10.1371/journal.pone.0003006 Arlt, D., & Pärt, T. (2007). Nonideal breeding habitat selection: A mis- match between preference and fitness. Ecology, 88(3), 792–801. https://doi.org/10.1890/06-0574 Arlt, D., & Pärt, T. (2008). Post-breeding information gathering and breeding territory shifts in northern wheatears. Journal of Animal Ecology, 77(2), 211–219. https://doi.org/10.1111/ j.1365-2656.2007.01329.x Asam, S., Callegari, M., Matiu, M., Fiore, G., de Gregorio, L., Jacob, A., Menzel, A., Zebisch, M., & Notarnicola, C. (2018). Relationship be- tween spatiotemporal variations of climate, snow cover and plant phenology over the Alps-an earth observation-based analysis. Remote Sensing, 10(11), 1757. https://doi.org/10.3390/rs10111757 MÜLLER et al. Atkinson, P. W., Buckingham, D., & Morris, A. J. (2004). What fac- tors determine where invertebrate-feeding birds forage in dry agricultural grasslands? IBIS, 146, 99–107. https://doi. org/10.1111/j.1474-919X.2004.00346.x Ecological Engineering, 18, 239–245. https://doi.org/10.1007/s1135 5-022-00494-x org/10.1111/j.1365-2664.2007.01378.x Agricultural land use and biodiversity in the Alps: How cultural tradition and socioeconomically motivated changes are shaping grassland biodiversity in the Swiss Alps. Mountain Research and Development, 28(2), 148–155. https://doi.org/10.1659/mrd.0964 Buyandelger, S., Enkhbayar, T., Otgonbayar, B., Zulbayar, M., & Bayartogtokh, B. (2021). Ecosystem engineering effects of Mongolian marmots (Marmota sibirica) on terrestrial arthropod communities. Mongolian Journal of Biological Sciences, 19(2), 17–30. https://doi.org/10.22353/mjbs.2021.19.11 Fry, C. H. (1992). The Moreau ecological overview. The ecology and con- servation of Palaearctic-African migrants. Papers from a conference, Buyandelger, S., & Otgonbayar, B. (2022). Mongolian marmot burrow influences an occupancy of isabelline wheatear. Landscape and 14 of 16 MÜLLER et al. 4 of 16 \| Norwich, 1991 (pp. 3–6) https://doi.org/10.1111/j.1474-919x.1992. tb04726.x Johnson, D. H. (1980). The comparison of usage and availability mea- surements for evaluating resource preference. Ecology, 61(1), 65– 71. https://doi.org/10.2307/1937156 Fuller, A. K., Harrison, D. J., & Vashon, J. H. (2007). Winter habitat se- lection by Canada lynx in Maine: Prey abundance or accessibil- ity? Journal of Wildlife Management, 71(6), 1980–1986. https://doi. org/10.2193/2006-288 Jones, T., & Cresswell, W. (2010). The phenology mismatch hypothe- sis: Are declines of migrant birds linked to uneven global climate change? Journal of Animal Ecology, 79(1), 98–108. https://doi. org/10.1111/j.1365-2656.2009.01610.x Jones, T., & Cresswell, W. (2010). The phenology mismatch hypothe- sis: Are declines of migrant birds linked to uneven global climate change? Journal of Animal Ecology, 79(1), 98–108. https://doi. org/10.1111/j.1365-2656.2009.01610.x Gabry, J., & Mahr, T. (2021). bayesplot: Plotting for Bayesian models. Kass, R. E., & Wasserman, L. (1996). The selection of prior distri- butions by formal rules. Journal of the American Statistical Association, 91(435), 1343–1370. https://doi.org/10.1080/01621 459.1996.10477003 Garcia-Pausas, J., Romanyà, J., Montané, F., Rios, A. I., Taull, M., Rovira, P., & Casals, P. (2017). Are soil carbon stocks in mountain grasslands compromised by land-use changes? In High Mountain conservation in a changing world, advances in global change research 62 (p. 413). Springer Open. Keller, F., Goyette, S., & Beniston, M. (2005). Sensitivity analysis of snow cover to climate change scenarios and their impact on plant habi- tats in alpine terrain. Climatic Change, 72(3), 299–319. https://doi. org/10.1007/s10584-005-5360-2 García-Rodríguez, A., Albrecht, J., Farwig, N., Frydryszak, D., Parres, A., Schabo, D. G., & Selva, N. (2022). Functional complementarity of seed dispersal services provided by birds and mammals in an al- pine ecosystem. Journal of Ecology, 110(1), 232–247. https://doi. CRC. https://doi.org/10.1201/b16018 Gideon, K., Grüneberg, C., Mitschke, A., Sudfeldt, C., Eikhorst, W., Fischer, S., Flade, M., Frick, S., Geiersberger, I., Koop, B., & Kramer, M. (2014). Atlas Deutscher Brutvogelarten. Atlas of German breeding birds. Stiftung Vogel monitoring Deutschland und Dachverband Deutscher Avifaunisten. Korner-Nievergelt, F., Roth, T., von Felten, S., Guélat, J., Almasi, B., & Korner-Nievergelt, P. (2015). Bayesian data analysis in ecology using linear models with R, BUGS, and Stan. Elsevier. Glutz Von Blotzheim, U. N., & Bauer, K. M. (2001). Passeriformes (2. Teil): Turdidae – Schmätzer und Verwandte: Erithracinae, Handbuch der Vögel Mitteleuropas. Kulakowski, D., Bebi, P., & Rixen, C. (2011). The interacting effects of land use change, climate change and suppression of natural disturbances on landscape forest structure in the swiss Alps. Oikos, 120(2), 216– 225. https://doi.org/10.1111/j.1600-0706.2010.18726.x Gobiet, A., Kotlarski, S., Beniston, M., Heinrich, G., Rajczak, J., & Stoffel, M. (2014). 21st century climate change in the European Alps-a re- view. Science of the Total Environment, 493, 1138–1151. https://doi. org/10.1016/j.scitotenv.2013.07.050 225. https://doi.org/10.1111/j.1600-0706.2010.18726.x Laiolo, P., Dondero, F., Ciliento, E., & Rolando, A. (2004). Consequences of pastoral abandonment for the structure and diversity of the al- pine avifauna. Journal of Applied Ecology, 41(2), 294–304. https:// doi.org/10.1111/j.0021-8901.2004.00893.x Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017). Google earth engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment, 202, 18–27. https:// doi.org/10.1016/j.rse.2017.06.031 Laird, N. M., & Ware, J. H. (1982). Random-effects models for longitudi- nal data. Biometrics, 38(4), 963–974. Gräler, B., Pebesma, E., & Heuvelink, G. (2016). Spatio-temporal in- terpolation using gstat. The R Journal, 8, 204–218. https://doi. org/10.32614/rj-2016-014 Leingärtner, A., Krauss, J., & Steffan-Dewenter, I. (2014). Elevation and experimental snowmelt manipulation affect emergence phenology and abundance of soil-hibernating arthropods. Ecological Entomology, 39(4), 412–418. https://doi.org/10.1111/ een.12112 Hallman, T. A., Guélat, J., Antoniazza, S., Kéry, M., & Sattler, T. (2022). Rapid elevational shifts of Switzerland's avifauna and associated species traits. Ecosphere, 13(8), 1–16. https://doi.org/10.1002/ ecs2.4194 Lindström, Å. (2003). Fuel deposition rates in migrating birds: Causes, constraints and consequences. In Avian migration (pp. 307–320). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3- 662-05957-9_21 Hijmans, R. J. (2021). raster: Geographic Data Analysis and Modeling. https://cran.r-project.org/package=raster Hovick, T. J., Elmore, R. D., Fuhlendorf, S. D., Engle, D. M., & Hamilton, R. G. (2015). Spatial heterogeneity increases diversity and stability in grassland bird communities. Ecological Applications, 25(3), 662–672. https://doi.org/10.1890/14-1067.1 Link, W. A., Cam, E., Nichols, J. D., & Cooch, E. G. (2002). org/10.1111/j.1365-2664.2007.01378.x org/10.1111/1365-2745.13799 Keller, V., Herrando, S., Voříšek, P., Rodriguez-Franch, M., Kipson, M., Milanesi, P., Martí, D., Anton, M., Klvanová, A., Kalyakin, M. V., Bauer, H.-G., & Foppen, R. P. B. (2020). European breeding bird Atlas 2: Distribution, abundance and change. In European breed- ing bird Atlas 2. Distribution, abundance and change (pp. 86–87). Lynx Edicions. https://www.researchgate.net/publication/34657 3850_European_Breeding_Bird_Atlas_2_Distribution_Abund ance_and_Change Gehrig-Fasel, J., Guisan, A., & Zimmermann, N. E. (2007). Tree line shifts in the Swiss Alps: Climate change or land abandonment? Journal of Vegetation Science, 18(4), 571. https://doi.org/10.1658/1100- 9233(2007)18[571:tlsits]2.0.co;2 Gelman, A., Carlin, J. B., Stern, H. S., Dunson, D. B., Vehtari, A., & Rubin, D. B. (2013). Bayesian data analysis (3rd ed.). Chapman and Hall/ CRC. https://doi.org/10.1201/b16018 Knaus, P., Antoniazza, S., Wechsler, S., Guélat, J., Kéry, M., Strebel, N., & Sattler, T. (2018). Schweizer Brutvogelatlas 2013–2016. Verbreitung und Bestandsentwicklung der Vögel in der Schweiz und im Fürstentum Liechtenstein. Schweizerische Vogelwarte. http://www.worldcat. org/oclc/41298543 , Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (ht Rime, Y., Luisier, C., Arlettaz, R., & Jacot, A. (2020). Landscape hetero- geneity and management practices drive habitat preferences of wintering and breeding birds in intensively-managed fruit-tree plantations. Agriculture, Ecosystems & Environment, 295(2), 106890. https://doi.org/10.1016/j.agee.2020.106890 McElreath, R. (2016). Statistical rethinking: A Bayesian course with examples in R and Stan. Chapman and Hall/CRC. https://doi. org/10.1201/9781315372495 McPeek, M. A. (1996). Trade-offs, food web structure, and the coexis- tence of habitat specialists and generalists. The American Naturalist, 148, S124–S138. https://doi.org/10.1086/285906 Rime, Y., Nussbaumer, R., Briedis, M., Sander, M. M., Chamberlain, D., Amrhein, V., Helm, B., Liechti, F., & Meier, C. M. (2023). Multi- sensor geolocators unveil global and local movements in an alpine- breeding long-distance migrant. Movement Ecology, 11(1), 19. https://doi.org/10.1186/s40462-023-00381-6 Meier, C. M., Rime, Y., Lisovski, S., Buchmann, M., & Liechti, F. (2022). Locally adapted migration strategies? Comparing routes and tim- ing of northern wheatears from alpine and lowland European populations. Journal of Avian Biology, 2022(8), 1–11. https://doi. org/10.1111/jav.02932 Rizzo, M. L. (2008). Statistical computing with R. Chapman and Hall/CRC. Saino, N., Ambrosini, R., Rubolini, D., von Hardenberg, J., Provenzale, A., Hüppop, K., Hüppop, O., Lehikoinen, A., Lehikoinen, E., Rainio, K., Romano, M., & Sokolov, L. (2011). Climate warming, ecologi- cal mismatch at arrival and population decline in migratory birds. Proceedings of the Royal Society B: Biological Sciences, 278(1707), 835–842. https://doi.org/10.1098/rspb.2010.1778 Moreno, J. (1984). Parental care of fledged young, division of labor, and the development of foraging techniques in the northern wheatear (Oenanthe L.). The Auk, 101(4), 741–752. https://doi. org/10.2307/4086901 Moreno, J. (1989a). The breeding biology of the wheatear Oenanthe in South Sweden during three contrasting years. Journal of Ornithology, 130, 321–334. https://doi.org/10.1007/BF01644745 Sander, M. M., Chamberlain, D., Mermillon, C., Alba, R., Jähnig, S., Rosselli, D., Meier, C. M., & Lisovski, S. (2021). Early breeding con- ditions followed by reduced breeding success despite timely arrival in an alpine migratory songbird. Frontiers in Ecology and Evolution, 9(2), 240–261. https://doi.org/10.3389/fevo.2021.676506 Moreno, J. (1989b). Variation in daily energy expenditure in nesting northern wheatears (Oenanthe oenanthe). The Auk, 106(1), 18–25. https://doi.org/10.2307/4087752 Sander, M. M., Jähnig, S., Lisovski, S., Mermillon, C., Alba, R., Rosselli, D., & Chamberlain, D. (2022). High nest failure but better nestling qual- ity for early breeders in an alpine population of northern wheatear (Oenanthe oenanthe). IBIS, 165, 125–141. https://doi.org/10.1111/ ibi.13133 Morris, M. G. (2000). The effects of structure and its dynamics on the ecology and conservation of arthropods in British grasslands. , Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (ht Biological Conservation, 95(2), 129–142. https://doi.org/10.1016/ S0006-3207(00)00028-8 Müller, T. M., Rime, Y., Meier, C. M., Helm, B., Korner, P., Amrhein, V., & Knaus, F. (2023). Foraging Habitat Preferences of Northern Wheatears breeding in Val Piora (Version 1). https://doi. org/10.5281/zenodo.7805040 Schaub, M., Martinez, N., Tagmann-Ioset, A., Weisshaupt, N., Maurer, M. L., Reichlin, T. S., Abadi, F., Zbinden, N., Jenni, L., & Arlettaz, R. (2010). Patches of bare ground as a staple commodity for de- clining ground-foraging insectivorous farmland birds. PLoS One. Edited by D. M. Evans, 5(10), e13115. https://doi.org/10.1371/ journal.pone.0013115 Nakagawa, S., Johnson, P. C. D., & Schielzeth, H. (2017). The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. Journal of the Royal Society, Interface, 14, 20170213. https://doi. org/10.1098/rsif.2017.0213 Schmaljohann, H., Meier, C., Arlt, D., Bairlein, F., van Oosten, H., Morbey, Y. E., Åkesson, S., Buchmann, M., Chernetsov, N., Desaever, R., Elliott, J., Hellström, M., Liechti, F., López, A., Middleton, J., Ottosson, U., Pärt, T., Spina, F., & Eikenaar, C. (2016). Proximate causes of avian protandry differ between subspecies with con- trasting migration challenges. Behavioral Ecology, 27(1), 321–331. https://doi.org/10.1093/beheco/arv160 Nakagawa, S., & Schielzeth, H. (2013). A general and simple method for obtaining R2 from generalized linear mixed-effects mod- els. Methods in Ecology and Evolution, 4, 133–142. https://doi. org/10.1111/j.2041-210x.2012.00261.x Nicenboim, B., Schad, D., & Vasishth, S. (2021). An introduction to Bayesian data analysis for cognitive science. GitHub. https://vasishth. github.io/bayescogsci/book/ Scridel, D., Brambilla, M., Martin, K., Lehikoinen, A., Iemma, A., Matteo, A., Jähnig, S., Caprio, E., Bogliani, G., Pedrini, P., Rolando, A., Arlettaz, R., & Chamberlain, D. (2018). A review and meta-analysis of the ef- fects of climate change on Holarctic mountain and upland bird pop- ulations. IBIS, 160(3), 489–515. https://doi.org/10.1111/ibi.12585 Paquet, M., Arlt, D., Knape, J., Low, M., Forslund, P., & Pärt, T. (2019). Quantifying the links between land use and population growth rate in a declining farmland bird. Ecology and Evolution, 9(2), 868–879. https://doi.org/10.1002/ece3.4766 swisstopo. (2018). SwissALTI3D, Das hoch aufgelöste Terrainmodell der Schweiz. Pilar, F. C., Loïc, M., Emmanuel, D., & Sergio, R. (2020). Seasonal changes in arthropod diversity patterns along an alpine elevation gradient. Ecological Entomology, 45(5), 1035–1043. https://doi.org/10.1111/ een.12881 Tagmann-Ioset, A., Schaub, M., Reichlin, T. S., Weisshaupt, N., & Arlettaz, R. (2012). Bare ground as a crucial habitat feature for a rare terres- trially foraging farmland bird of Central Europe. Acta Oecologica, 39, 25–32. https://doi.org/10.1016/j.actao.2011.11.003 QGIS Development Team. (2020). 20457758, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. Se MÜLLER et al. 15 of 16 CRC. https://doi.org/10.1201/b16018 Of Bugs and birds: Markov chain Monte Carlo for hierarchical modeling in wild- life research. Journal of Wildlife Management, 66(2), 277–291. Hughes, L. (2000). Biological consequences of global warming: Is the sig- nal already apparent? Trends in Ecology & Evolution, 15(2), 56–61. https://doi.org/10.1016/S0169-5347(99)01764-4 Lüdecke, D., Ben-Shachar, M., Patil, I., Waggoner, P., & Makowski, D. (2021). Performance: An R package for assessment, comparison and testing of statistical models. Journal of Open Source Software, 6(60), 3139. https://doi.org/10.21105/joss.03139 Inouye, D. W. (2020). Effects of climate change on alpine plants and their pollinators. Annals of the New York Academy of Sciences, 1469(1), 26– 37. https://doi.org/10.1111/nyas.14104 Mahon, C. L., Holloway, G., Sólymos, P., Cumming, S. G., Bayne, E. M., Schmiegelow, F. K. A., & Song, S. J. (2016). Community structure and niche characteristics of upland and lowland western boreal birds at multiple spatial scales. Forest Ecology and Management, 361, 99–116. https://doi.org/10.1016/J.FORECO.2015.11.007 Issa, N., & Muller, Y. (2015). Atlas des oiseaux de France métropolitaine. Nidification et présence hivernale. Delachaux et Niestlé. Jähnig, S., Sander, M. M., Caprio, E., Rosselli, D., Rolando, A., & Chamberlain, D. (2020). Microclimate affects the distribution of grassland birds, but not forest birds, in an alpine environment. Journal of Ornithology, 161(3), 677–689. https://doi.org/10.1007/ s10336-020-01778-5 Maron, M., & Lill, A. (2005). The influence of livestock grazing and weed invasion on habitat use by birds in grassy woodland remnants. Biological Conservation, 124(4), 439–450. https://doi.org/10.1016/j. biocon.2005.02.002 Martinez, N., Jenni, L., Wyss, E., & Zbinden, N. (2010). Habitat struc- ture versus food abundance: The importance of sparse vegeta- tion for the common redstart Phoenicurus phoenicurus. Journal of Ornithology, 151(2), 297–307. https://doi.org/10.1007/s1033 6-009-0455-6 selection for a high-alpine passerine during breeding: Implications for species persistence in the face of climate change. Biodiversity and Conservation, 28(10), 2669–2685. https://doi.org/10.1007/ s10531-019-01786-9 Zhou, R., Sivaganesan, S., & Longla, M. (2014). An objective Bayesian es- timation of parameters in a log-binomial model. Journal of Statistical Planning and Inference, 146, 113–121. https://doi.org/10.1016/j. jspi.2013.09.006 Vickery, J. A., & Arlettaz, R. (2012). The importance of habitat hetero- geneity at multiple scales for birds in European agricultural land- scapes. In R. J. Fuller (Ed.), Birds and habitat: Relationships in chang- ing landscapes (pp. 177–204). Cambridge University Press. Vickery, J. A. Tallowin, J. R. B., Feber, R. E., Atkinson, P. W., Asteraki, E. J., Fuller, R. J., & Brown, V. K. (1999) Changes in lowland grassland management: Implications for invertebrates & birds. British Trust for Ornithology. , Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10084 by Eth Zürich Eth-Bibliothek, Wiley Online Library on [02/06/2023]. See the Terms and Conditions (ht Quantum GIS geographic information system, Open Source Geospatial Foundation Project. Theurillat, J. P., & Guisan, A. (2001). Potential impact of climate change on vegetation in the European alps: A review. Climatic Change, 50(1–2), 77–109. https://doi.org/10.1023/A:1010632015572 R Core Team. (2021). R Core Team 2021, R: A Language and Environment for Statistical Computing. Renner, S. S., & Zohner, C. M. (2018). Climate change and phenologi- cal mismatch in trophic interactions among plants, insects, and vertebrates. Annual Review of Ecology, Evolution, and Systematics, 49, 165–182. https://doi.org/10.1146/annurev-ecolsys-11061 7-062535 Tulp, I., & Schekkerman, H. (2008). Has prey availability for Arctic birds advanced with climate change ? Hindcasting the abundance of tun- dra arthropods using weather and seasonal variation. Arctic, 61(1), 48–60. van Oosten, H. H., van den Burg, A. B., Versluijs, R., & Siepel, H. (2014). Habitat selection of brood-rearing northern wheatears Oenanthe oenanthe and their invertebrate prey. Ardea, 102(1), 61–69. https:// doi.org/10.5253/078.102.0111 Resano-Mayor, J., Korner-Nievergelt, F., Vignali, S., Horrenberger, N., Barras, A. G., Braunisch, V., Pernollet, C. A., & Arlettaz, R. (2019). Snow cover phenology is the main driver of foraging habitat 16 of 16 MÜLLER et al. SUPPORTING INFORMATION Additional supporting information can be found online in the Supporting Information section at the end of this article. Visser, M. E., Both, C., & Lambrechts, M. M. (2004). Global climate change leads to mistimed avian reproduction. Advances in Ecological Research, 35, 89–110. https://doi.org/10.1016/S0065 -2504(04)35005-1 Vu, V. Q. (2011). ggbiplot: A ggplot2 based biplot. http://github.com/vqv/ ggbiplot How to cite this article: Müller, T. M., Meier, C. M., Knaus, F., Korner, P., Helm, B., Amrhein, V., & Rime, Y. (2023). Finding food in a changing world: Small-scale foraging habitat preferences of an insectivorous passerine in the Alps. Ecology and Evolution, 13, e10084. https://doi.org/10.1002/ ece3.10084 How to cite this article: Müller, T. M., Meier, C. M., Knaus, F., Korner, P., Helm, B., Amrhein, V., & Rime, Y. (2023). Finding food in a changing world: Small-scale foraging habitat preferences of an insectivorous passerine in the Alps. Ecology and Evolution, 13, e10084. https://doi.org/10.1002/ ece3.10084 How to cite this article: Müller, T. M., Meier, C. M., Knaus, F., Korner, P., Helm, B., Amrhein, V., & Rime, Y. (2023). Finding food in a changing world: Small-scale foraging habitat preferences of an insectivorous passerine in the Alps. Ecology and Evolution, 13, e10084. https://doi.org/10.1002/ ece3.10084 Weisshaupt, N., Arlettaz, R., Reichlin, T. S., Tagmann-Ioset, A., & Schaub, M. (2011). Habitat selection by foraging wrynecks Jynx torquilla during the breeding season: Identifying the optimal habitat pro- file. Bird Study, 58(2), 111–119. https://doi.org/10.1080/00063 657.2011.556183 Yoshihara, Y., Okuro, T., Buuveibaatar, B., Undarmaa, J., & Takeuchi, K. (2010). Complementary effects of disturbance by livestock and marmots on the spatial heterogeneity of vegetation and soil in a Mongolian steppe ecosystem. Agriculture, Ecosystems and Environment, 135(1–2), 155–159. https://doi.org/10.1016/j. agee.2009.09.009
https://openalex.org/W3203489540	https://www.e3s-conferences.org/articles/e3sconf/pdf/2021/81/e3sconf_rubis2021_03006.pdf	English	null	Progress of Rubber Breeding Program to Support Agroforestry System in Indonesia	E3S web of conferences	2,021	cc-by	4,267	Progress of Rubber Breeding Program to Support Agroforestry System in Indonesia Fetrina Oktavia 1,, Sahuri1 and Dwi Shinta Agustina 1 Fetrina Oktavia 1,, Sahuri1 and Dwi Shinta Agustina 1 1 Indonesian Rubber Research Institute Jl. Raya Palembang – Pangkalan Balai Km 29, Sembawa, Banyuasin, South Sumatra 1 Indonesian Rubber Research Institute Jl. Raya Palembang – Pangkalan Balai Km 29, Sembawa, Banyuasin, South Sumatra Abstract. The use of superior rubber planting materials is one of the most important components of technology to support the cultivation and sustainability of the natural rubber industry. The effect of the genetic components of planting materials to the rubber productivity can reach 60%, and the rest is the influence of agro-climatic conditions. The aim of the rubber plant breeding program is to obtain the new superior rubber clones that have a high latex yielding potential and good agronomic characters. The fluctuations of natural rubber price and climate changes also influence the direction and objectives of the rubber plant breeding program. To deal with the conditions, it is important to provide the rubber agroforestry technology by through intercropping of rubber with various other crops. The article will provide the information about progress of rubber breeding program in Indonesia and it is role in supporting agroforestry system. Several of new superior rubber clones have been released by IRRI, and some of these clones such as IRR 112 and IRR 118 had been planting with rice, corn, and other crops by through rubber agroforestry system. The system was estimated be able to maintain latex yielding potential of clones as well as farmers' income can be improved. E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 https://doi.org/10.1051/e3sconf/202130503006 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). 1. Introduction Rubber (Hevea brasiliensis) is one of the commodities that has the most important role for the Indonesian economy, both as a source of employment opportunity and income for farmers as well as a source of foreign exchange. Until 2018, the total area of rubber plantations in Indonesia was recorded at 3.67 million hectares, consisting of 88.13% smallholder plantations, 5.16% private plantations and 6.7% state plantations with total production reaching 3.63 million tons. Based on this information, it is known that the largest contribution to Indonesian natural rubber production was obtained from smallholder plantations. However, if we pay attention to the productivity of smallholder rubber plantations which just 1153 kg ha -1 per year, it is still low and below the productivity of state and private rubber plantations, 1529 kg ha -1 per year and 1549 kg ha -1 per year respectively, as well as under of the average productivity of Indonesian rubber productivity about 1205 kg ha -1 per year [1]. Corresponding author: fetrina_oktavia@yahoo.com Corresponding author: fetrina_oktavia@yahoo.com © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 https://doi.org/10.1051/e3sconf/202130503006 According to observations in the fields, it is known that the low productivity is mainly due to the fact that there are still many old rubber plants that are less productivity (2.6%) as well as low of adoption and implementation of rubber cultivation technology and management [2]. Limitation of capital, land and labour were the main cause of the low of technology adoption in Indonesia [3]. Other factors contributed to exacerbating the condition of rubber productivity in Indonesian recently was the outbreak of a new disease named by Pestalotipsis leaf fall disease that attacked rubber plantations in 2018. This condition also affects the productivity of Indonesian rubber plantations which results in a significant decrease in farmers' income. Various efforts must be done to increase the productivity of rubber plantations, especially smallholder rubber plantations which dominate the rubber plantations area in Indonesia. Several important factors that influence to productivity of rubber plantations are the use of high-yielding superior planting materials with a balanced type and age of clone composition as well as planting of clones in the appropriate agroecosystem, application of right cultivation techniques which includes land clearing, fertilization with appropriate dosage, frequency and method of application, prevention and control of disease, application of exploitation systems according to clone physiological characteristics, control of tapping panel dryness, and replanting of non-productive rubber plantation [4,5]. In addition, the change in the rubber cultivation system from monoculture to intercropping during the immature phase is also a solution in dealing with the problem of fluctuating rubber prices, especially related to farmers' income in smallholder plantations [6,7]. Productivity of rubber plantation will influence by two main factors. The most factors is genetic, the kind of clone that used as a planting material which will give effect about 60% to growth and productivity of rubber plant, and remains is environment effect such as rainfall, elevation, soil and management of rubber plantation. Using of of the superior rubber clones will increase the rubber productivity. This paper will report the progress of rubber breeding program to obtain the superior rubber clone to assist increasing of rubber productivity and using of the superior clones to support agroforestry systems in Indonesia. 2.1 Purpose, direction and strategy of the rubber breeding program in Indonesia The rubber breeding and selection activities in Indonesia began in 1910 and continued until present. At the first, the aim of rubber breeding program just to produce the high latex yielding clone, but along with the development of environmental conditions and climate change, there were a change not only focused on latex yielding, but also to obtain the plants that have a good agronomic character. Some of the agronomic characters that had been taken into consideration include fast and robust growth, resistant to biotic stress especially to major diseases, tolerant to abiotic stress such as drought and tapping panel dryness, have a good quality wood as well as have a good canopy architectural typology [8]. 2.2 Genetic material The success and progress of the rubber breeding program to increase rubber productivity depends on the availability of a wide variety of genetic material sources. Generally the rubber breeding still use a widely cultivated clones that come from Wickham population. Based on the genetic analysis it is known that these clones have a narrow genetic basis [9, 10, 11, 12]. 2 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 https://doi.org/10.1051/e3sconf/202130503006 This is estimated due to the clones came from the same area in Brazil. Another problem is rubber plants are inbreeding depression, so steadily using of Wickham's clones in crossing programs can lead to a decrease in the genetic value of important agronomic characters such as a latex production and thickness, increasing risk of disease epidemics and pest attacks as well as reducing the adaptability of plants to environmental changes [13]. The problem can be overcome by using the new germplasm resulted exploration of IRRDB (International Rubber Research Development Board) in three districts of Brazil, namely Rondonia, Acre and Matto Grosso in 1981. The genotypes were resulted from exploration called Amazonian or IRRDB 1981 population. According to growth and yield testing, it is known that the germplasm had fast growth the potential of latex yielding was below the Wickham clones[14]. This becomes a limitation to being the germplasm directly into production clones, but had the potential to be a source of genetic material in crossing for a good agronomic characters [15]. 2.3 Rubber breeding stage There are some stages in rubber breeding program to produce the new superior clones consists i.e hand pollination, selection of F1 progenies, small scall clone trial (SSCT), large scale clone trial (LSCT) and multilocation trial for adaptation selection of clone in different agroclimate. The program need a long time, begin from pollination up to the clone ready to release about 25-30 years which each step take a different time as showed in scheme of standard operating procedure of rubber breeding (figure 1). Some of constrains faced in the breeding process especially related to limitation of genetic material availability as a source of genes for the desired agronomic character, low low of success rate of crossing which was less than 3.2% [16], limitation of testing area, large of cost for every testing stage as well as the length of time required for a clone to be released as a new superior clone to public, it is about 25-30 years [17]. Fig. 1. Scheme of standard operating procedure of rubber breeding Regarding to type of planting material used and the increasing of planting material productivity, the progress of rubber breeding program in Indonesia can be clustered into five generations as listed in Table 1. There were a plenty high increasing of rubber productivity from only about 500 kg ha-1 per year to about 2500 kg ha-1 per year by change the using of rubber planting material from seedling of selected high-yielding mother trees become directed crossing from selected parents clones. Continuously efforts to increase the rubber productivity up to reach the maximum potential of the latex production about 7000 - 8000 kg ha-1 per year [18]. Fig. 1. Scheme of standard operating procedure of rubber breeding Fig. 1. Scheme of standard operating procedure of rubber breeding Regarding to type of planting material used and the increasing of planting material productivity, the progress of rubber breeding program in Indonesia can be clustered into five generations as listed in Table 1. There were a plenty high increasing of rubber productivity from only about 500 kg ha-1 per year to about 2500 kg ha-1 per year by change the using of rubber planting material from seedling of selected high-yielding mother trees become directed crossing from selected parents clones. Continuously efforts to increase the rubber productivity up to reach the maximum potential of the latex production about 7000 - 8000 kg ha-1 per year [18]. 2.3 Rubber breeding stage 3 3 https://doi.org/10.1051/e3sconf/202130503006 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 Table 1. Milestone of rubber planting material in Indonesia Generation Period Planting Material Clone Productivity (kg ha-1 per year) I 1910-1935 Selected seedling Selected seedling < 500 II 1953-1960 Propagation of selected seedling Tjir I, Tjir 16, GT 1, LCB 479, LCB 1320, PR 107 500 - 1000 III 1960 - 1985 Propagation clone which obtained from hand pollination AVROS 2037, BPM 1, PR 228, PR 255, PR 261, PB 5/51, RRIM 600 1000 - 1500 IV 1985 - 2010 Propagation clone which obtained from hand pollination BPM 24, BPM 107, BPM 109, PB 260, PB 330, PB 340, RRIC 100 1500 - 2000 V 2010 - present Propagation clone which obtained from hand pollination Latex Clone IRR 220, IRR 118, IRR 112, IRR 104, BPM 24, PB 260, PB 330 and PB 340 Latex Timber Clone IRR 230, IRR 5, IRR 39, IRR 42, IRR 119 and RRIC 100 Root Stock Clone AVROS 2037, GT 1, PB 26, PB 330, BPM 24 and RRIC 100 2000 - 2500 Table 1. Milestone of rubber planting material in Indonesia 2.4 Inovation technology of superior rubber clone IRR series Several studies showed that there are no universal clones or plant varieties, which have the same performance at different locations. The use of superior rubber clones as planting material can increase the productivity of rubber plantations. Type of clones will give an effect about 60% to growth and productivity of rubber plant, while the rest is environmental factors. There are two important environmental factors that influence the performance of a rubber clone, namely the agroclimate conditions such as soil, elevation and rainfall, as well as the management of cultivation such as planting systems, maintenance, fertilization, disease control and harvesting system. According to the results of selection in multilocation which have a different agroclimate condition, IRRI had been released some of superior rubber clones that have a good adaptatation in all of these location. The clones were named with IRR (Indonesian Rubber Research), and untill present IRRI has been release IRR series 00,100 and 200. The newest clones that released to public are IRR 112, IRR 118, IRR 220 and IRR 230. These clones are a latex yielding and a good agronomic characters such as a robust and fast growth, fast open tapping time, resistant to leaf disease, responsive to stimulants and have a good of canopy architecture typologi. Table 2 shows the latex yielding potential and agronomic characters of the each clone. 4 https://doi.org/10.1051/e3sconf/202130503006 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 Table 2. Characteristics of Indonesian superior rubber clones Characteristic IRR 112 IRR 118 IRR 220 IRR 230 Production potential (kg ha-1 per year) 2452 2200 2487 2008 Open tapping (year) 4 4.5 4 4 Resistance to leaf fall disease Resistant Moderate Resistant Resistant Response to Stimulant High Moderate Low Moderat Adaptation to Agroclimate All Dry climate All All Type of Clone Latex clone Latex clone Latex clone Latex-timber clone Suitable of latex Concentrate latex SIR 3WF SIR 3CV, SIR 3L and concentrate latex SIR 3CV, SIR 3L, RSS Table 2. Characteristics of Indonesian superior rubber clones Beside released clones, IRRI have been prepared the new superior clones which currently in the evaluation process in various agroclimate conditions. The prospective superior clones called IRR 300 and 400 series which are expected have the potential of production above the previous released IRR series clones. Both series of clones are expected to be released in the near future. 2.4 Inovation technology of superior rubber clone IRR series Figure 2 shows the increased productivity of IRR 300 and 400 series compare to IRR 00 100 and 200 series [8,17]. Fig. 2. Progress of increasing of latex productivity of IRR 00 – 400 series of rubber clone on 5 years tapping 0 500 1000 1500 2000 2500 3000 3500 4000 IRR 5 IRR 118 IRR 220 IRR 309 IRR 425 Productivity (kg h-1 per year) Rubber Clone Fig. 2. Progress of increasing of latex productivity of IRR 00 – 400 series of rubber clone on 5 years tapping 3. Utilization and role of superior clones in rubber agroforestry system There are efforts to increase the income and welfare of rubber farmers through increasing the productivity of rubber plantations through some approaches, including the provision the technology of superior rubber clones as a planting material and the use of intercropping plants. The change of the planting system from monoculture to agroforestry system is expected to be able to increase the land productivity. The main aims of RAS was to increase and to maintain the sustainability of farmers' income, wherein a s long as the rubber plant on 5 5 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 https://doi.org/10.1051/e3sconf/202130503006 immature periode, they can still get the income from the intercrops. Based on this concept, usually the farmers choose the annual crops such as a rice, chilie, corn and soybeans, which can faster produce than perennial crops. In addition, RAS is estimated can reduce the cost of maintaining the rubber plantation, increase the soil fertility, avoid the soil erosion and maintain the biodiversity compared to the monoculture planting systems [19, 20, 21]. immature periode, they can still get the income from the intercrops. Based on this concept, usually the farmers choose the annual crops such as a rice, chilie, corn and soybeans, which can faster produce than perennial crops. In addition, RAS is estimated can reduce the cost of maintaining the rubber plantation, increase the soil fertility, avoid the soil erosion and maintain the biodiversity compared to the monoculture planting systems [19, 20, 21]. y p p g y [ ] Development of RAS in Indonesia has been started since rubber cultivated by smallholders, but have not applied good agriculture practical. In order to increase the rubber productivity of smallholder, so since 1994 the World Agroforestry Center (ICRAF) in collaboration with the Center de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD) France and Indonesian Rubber Research Institute launched the SRAP (Smallholder Rubber Agroforestry) project. The networking established on farm trials in several central areas of smallholder rubber plantations in Indonesia such as South Sumatra, Jambi, West Sumatra, and West Kalimantan. The primary aim of the project was to study some of option to improved the RAS under management of the farmer that was being monitored by professionals and researcher. There were three types of distinct RAS technology that developed in the project. RAS-1 that used unselected rubber seddling were replaced by recommended clones. Stage 1. Non clonal planting material In the stage, smallholders still used rubber seedling as a planting material which planting together with the perennial crops such as Acacia sp, Shorea sp as well as fruits tree. In the other side, the smallholders were not adopted GAP to manage and maintain their plantation so that the plantations often look like a jungle rubber. The condition caused the low of rubber productivity in the smallholder just about 500 kg ha-1 per year. 3. Utilization and role of superior clones in rubber agroforestry system In RAS-2 rubber are planted with other perennial crops. RAS-3 is designed for rubber with fast growing and multipurpose of tree species that can be harvested before rubber trees reach tapping size. In the RAS-3 also using the cover crops such as Mucuna to suppress the growth of Imperata [22] The RAS system are developed to get the optimal gain. Based on the kind of planting material that be used and adoption of good agriculture practical (GAP) in the rubber plantation, so RAS in Indonesia can be classified into four stage: Stage 2. Clonal planting material and non GAP The generation II and III of rubber had started to be used in smallholder rubber plantations, however, there were not a good maintenance such as fertilizer, weeding and protection of disease infection this case. Some of G-II clones that had been widely adopted by smallholders was GT 1 and LCB 1230 and G-III such as BPM 1 and RRIM 600. The rubber plants were planted together with various perenial crops such as Acacia sp and Shorea sp as well as seasonal crops such as upland rice and maize. The use of superior rubber clones without being followed by GAP will cause the latex potential of clone is not being maximally achieved. In the condition non application GAP, the potential of the clones G-II and G-III will be decrease under 1000 kg ha-1 per year. Stage 3. Clonal planting material and GAP There is a significantly increasing of rubber productivity in the third stage wherein the G- III and G-IV clones such as BPM 24, PB 260 and RRIC 100 as well as application of GAP under monitoring of research institute were implemented in the smallholder rubber plantations. The latex production in the stage could be reached 1500 kg ha-1 per year [23]. Stage 4. New Superior rubber clone planting material and GAP The differences of stages 3 and 4 can be seen from the using of the clones and improved of planting system. In the stage, the smallholder have been adopted the new superior rubber 6 6 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 https://doi.org/10.1051/e3sconf/202130503006 clones that recommended by Indonesian Rubber Research Institute. The clones have a good latex potential and a good agronomic character such as resistant to rubber main disease, high response to stimulant and good canopy architecture. Some of new superior clones that adopted in RAS are IRR 112 and IRR 118 which have a latex potential reach 2500 kg ha-1 per year. There is a implementation of GAP and improving of planting system such as a double row spacing for intercropping with annual crops at this stage. The information showed the importance of the role of rubber planting materia l in effort to increase the rubber productivity in RAS. The progress of productivity some of rubber clones that be used in stage 1 to 4 in the RAS in Indonesia, starting from seedling as a planting material in stage I up to utilize the new superior clone in stage 4 can be seen in Table 3. Eventhough the productivity has not been maximally achieved according to the potential of clone which it can reach up to 8000 kg ha-1 per year [18], there has been a significant increasing of latex production between the planting material be used. Sahuri [24] stated that intercropping crops give the positive effect to girth of rubber plant in the RAS. This is due to the return of intercropping plant biomass compost to the soil, so that the soil structure becomes more loose and rich of nutrients [25]. The intercropping had no effect on latex yield per tree per tapping but yield per hectare was greater in the intercropping than monoculture due to the number of trees that could be tapped was significantly higher. Table 3. Productivity of rubber planting material on the 3rd year of tapping in RAS in Indonesia Table 3. Productivity of rubber planting material on the 3rd year of tapping in RAS in Indonesia Description Seedling BPM 1 RRIM 600 PB 260 RRIC 100 IRR 112 Yield g/tree/tapping 15 28 27 32 29 30 Nb. tappings/year 120 134 136 135 134 113 Nb. 4. Conclusions 4. Conclusions 1. It is important to use the superior rubber clone as a planting material to increase the rubber productivity 2. RAS can increase the land rubber productivity, so as to increase the income and welfare of rubber farmers 3. According to experiments showed that RAS significant increase the growth of rubber tree, reduces unproductive plant phases, and had no effect on latex yield per tree per tapping. 4. Utilization of superior rubber clone is one of the important key to success of RAS 3. According to experiments showed that RAS significant increase the growth of rubber tree, reduces unproductive plant phases, and had no effect on latex yield per tree per tapping. pp g 4. Utilization of superior rubber clone is one of the important key to success of RAS Stage 4. New Superior rubber clone planting material and GAP trees/ha 339 371 375 422 371 550 Yield kg/ha/year 518 1430 1442 1794 1508 1864 Source : Wibawa et al, 2008; Sahuri, 2019 References 1. Directorate General of Estate Crops. Rubber Tree Crops estate Statistic of Indonesia. Directorat General of estate Crops, Jakarta (2019) 2. D. I. S. Simamora, J. Yusri, N. Dewi. J. Online Riau University Faculty of Agriculture. 4 (2):1–12 (2017). 3. D. Iskandar. J. Sains and Technology Indonesia. 13(3):165-170 (2011). 4. Boerhendhy, K. Amypalupy. J. Agriculture Res & Dev. 30(1): 23-30 (2011). 5. Junaidi. Perspective. 19(1):17-28. (2020) 6. Sahuri, M.J. Rosyid. Indonesian Bulletin of natural Rubber. 34(2):77-88 (2015). 1. Directorate General of Estate Crops. Rubber Tree Crops estate Statistic of Indonesia. Directorat General of estate Crops, Jakarta (2019) 1. Directorate General of Estate Crops. Rubber Tree Crops estate Statistic of Indonesia. Directorat General of estate Crops, Jakarta (2019) 2. D. I. S. Simamora, J. Yusri, N. Dewi. J. Online Riau University Faculty of Agriculture. 4 (2):1–12 (2017). 3. D. Iskandar. J. Sains and Technology Indonesia. 13(3):165-170 (2011). 4. Boerhendhy, K. Amypalupy. J. Agriculture Res & Dev. 30(1): 23-30 (2011). y, yp py g ( ) ( ) 5. Junaidi. Perspective. 19(1):17-28. (2020) 5. Junaidi. Perspective. 19(1):17-28. (2020) p ( ) ( ) 6. Sahuri, M.J. Rosyid. Indonesian Bulletin of natural Rubber. 34(2):77-88 (2015). 7 7 https://doi.org/10.1051/e3sconf/202130503006 E3S Web of Conferences 305, 03006 (2021) RUBIS 2021 7. Sahuri, A. Nurcahyo, I.S. Nugraha. Indonesian Bulletin of natural Rubber.35(2):107- 120 (2016). 8. Daslin, S. Woelan, M. Lasminingsih. H. Hadi. Progress of rubber breeding and selection in Indonesia, in Proceedings of National Rubber Breeding Seminar, 4-6 Auhust 2009, Batam, Indonesia (2009). ( ) 9. N. Lekawipat, K. Teerawatanasuk, M. Rodier-Goud, M. Seguin, A. Vanavichit, T. Toojinda, S. Tragoonrung. J. Rubber Res. 6(1):36-47 (2003). 10. L.R.L. Gouvêa, L.B. Rubiano, A.F. Chioratto, M.I. Zucchi, P.D.S Gonçalves. Gen Mol Bio. 33:308-318 (2010) 11. F. Oktavia, M. Lasminingsih, Kuswanhadi. Hayati J. Biosci. 18(1):27-32 12. F. Oktavia, Kuswanhadi, D. Dinarty, Widodo, Sudarsono. Agrivita J. Agriculture Sci. 39(3): 239-251 (2017) ( ) ( ) 13. U.V. Lopes, J.R.B. Marques. Agrotrópica. 27(1):33-44 (2015). 14. Daslin. The selected of genotypes of IRRDB rubber germplasm for a latex timber producer. Prospect and development or rubber timber. Indonesian Rubber Research Institute: 75-84 (2009) 15. K. Mydin, C. Narayanan, T. Abraham. Incorporation of the 1981 IRRDB wild amazonian germplasm in Hevea breeding in India. IRRI-IRRDB Rubber Plant Breeding Seminar. Medan, Indonesia. (2012). 16. S. Woelan. Indonesian J. Natural Rubber. 24(1):17-31 (2006). 17. S.A. Pasaribu. M.R. Darojat, E. Bukit. 25. W. Pansak. Kasetsart J. 49:785-794 (2015). References Technolgy assembly of superior rubber clone with cost component. Talenta Confrence series. 1:42-46 (2018). 18. Azis. Introducing research result into practice. Research Management RRIM Kuala Lumpur (1998). 19. G. Wibawa, L. Joshi, M.V. Noordwijk, E. penot. In Proceeding of workshop : land use after the tsunami: Supporting Education, Research and Development in the Aceh Region. Syah Kuala University, Banda Aceh, Indonesia November 4-6 November 2008. Indonesia (2008). ( ) 20. C. Pye-Smith. Rubber Agroforestry. World Agroforestry Centre (2013). 21. W. Sukmawati, Y. Arkeman, S. Maarif. Industrial techniq J. 58-64 (2014). 22. E. Penot. From Shifting Cultivation to Sustainable Jungle Rubber: A History of Innovations in Indonesia. Chapter 48 of the book Voices from the Forest Integrating Indigenous Knowledge into Sustainable Upland Farming. Malcolm Cairns, editor. Browse Book. 880 p. (2006). 23. Budi, G. Wibawa, Ilahang, R. Akiefnawati, E. Penot, Janudianto. Manual for rubber agroforestry system -RAS. World Agroforestry Centre (2008). 24. Sahuri. J. Agriculture R & D. 38(1):23-34 (2019). 25. W. Pansak. Kasetsart J. 49:785-794 (2015). 8 8
https://openalex.org/W3209307076	https://link.springer.com/content/pdf/10.1007/s11273-021-09841-5.pdf	English	null	Advancing salt marsh restoration for coastal resilience: a learning exchange	Wetlands ecology and management	2,021	cc-by	9,323	Advancing salt marsh restoration for coastal resilience: a learning exchange Nicole Maher . Camilo Salazar . Alexa Fournier Received: 12 March 2021 / Accepted: 5 October 2021 / Published online: 28 October 2021 The Author(s) 2021 about project ideas and contributed to project designs in early development stages. Collaborating while engaged in on-the ground projects enabled partici- pants to implement lessons learned in real time. Field trips to restoration sites at different stages of devel- opment allowed a greater and more ﬂuid exchange of ideas and practical implementation advice. Practition- ers leveraged resources and developed new collabo- rations. Lessons learned and shared through this faster and more ﬂexible forum will inform the design, implementation, and monitoring of restoration pro- jects across the region and improve overall marsh health and resilience in the face of climate change. Learning exchanges like this should be used more frequently to improve the efﬁciency and effectiveness of coastal restoration particularly when there is a windfall of cash and a short window of opportunity such as with post-disaster federal spending. Abstract A multidisciplinary group of salt marsh professionals from Maine to Virginia participated in a collaborative learning exchange to improve restora- tion for the overall health and resilience of coastal wetlands. This was an unprecedented forum through which participants representing different geographies, backgrounds, and roles in salt marsh management were able to share and learn from one another to develop the best available restoration methods for on- the-ground projects that address multiple beneﬁts. By including mosquito control agencies, restoration prac- titioners, regulatory agencies, academic researchers, and conservation organizations in the learning exchange, we developed an understanding and accep- tance of different approaches. Regulators learned Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/ s11273-021-09841-5. Keywords Learning exchange Salt marsh restoration Coastal resilience Minga Field trips Climate change Keywords Learning exchange Salt marsh restoration Coastal resilience Minga Field trips Climate change N. Maher (&) The Nature Conservancy in New York, Cold Spring Harbor, NY, USA e-mail: nmaher@tnc.org C. Salazar Suffolk County Department of Economic Development and Planning, Suffolk County, NY, USA Wetlands Ecol Manage (2022) 30:1033–1047 https://doi.org/10.1007/s11273-021-09841-5 (0123456789().,-volV)( 01234567 89().,-volV) ORIGINAL PAPER A. Fournier Division of Marine Resources, New York State Department of Environmental Conservation, Kings Park, NY, USA Introduction Tidal wetlands are extremely valuable for the multiple beneﬁts they provide to both natural and human communities (UNEP 2006), but they are facing grave 12 3 3 Wetlands Ecol Manage (2022) 30:1033–1047 1034 risks: a large fraction has been lost to date (e.g. Basso et al. 2015; CEA 2015) and climate and habitat projections suggest further losses in the future (e.g. NYSERDA 2018) unless we engage in restoration at scale. This urgent situation calls for restoration approaches that are more effective and efﬁcient than the project by project model and traditional learning methods (e.g. annual conference presentations, peer- reviewed publications, individual projects conducted and evaluated locally). Collaborative Learning exchanges (e.g. Paolisso et al. 2019; Feurt 2008) provide measurable beneﬁts that are more responsive to the needs of diverse stakeholders and extend beneﬁts beyond the reach of individual projects. A restoration effort after Hurricane Sandy provides a good example of a successful regional collaborative learning exchange. After the storm roared up the east coast of the United States in October 2012, previously scarce federal dollars for restoration ﬂowed for projects across the states most heavily impacted by the storm. The US Fish and Wildlife Service (USFWS) initiated 14 coastal marsh restoration projects on federal lands (Babson et al. 2020). The U.S. Depart- ment of the Interior’s (DOI) Hurricane Sandy Coastal Resiliency Competitive Grant Program (HSCR) administered by the National Fish and Wildlife Foundation (NFWF), supported 15 additional marsh restoration projects led by non-federal partners includ- ing state and local governments, tribes, nonproﬁts, and universities (Dwyer et al. n.d.). All Sandy-funded projects were to be initiated and completed within four to ﬁve short years. Phragmites australis, and promote natural biological control of mosquito larvae. All Suffolk County restoration sites implemented through this project are part of the South Shore Estuary Reserve: Suffolk County Gardiner Park West (71 acres), Suffolk County Gardiner Park East (26 acres), West Sayville Marsh (113 acres), and Timber Point Marsh (51 acres). ( ), ( ) In this paper, we share the Methods of our salt marsh restoration learning exchange. We describe four types of Beneﬁts of the learning exchange: Achieving Multiple Beneﬁts, Engaging Regulators, Learning in Real Time, Practical Knowledge Exchange in the Field. Introduction We highlight two types of Outcomes: speciﬁc outcomes for implementation of the Suffolk County restoration projects (Suffolk County Project Imple- mentation), and more general outcomes for building social capital among participants (Professional Rela- tionships and New Collaborations). We put this learning exchange in the context of the state of the world (Considerations and Questions for the Future). Finally, we stress the urgent need for successful restoration and encourage more ﬂexible and expedient restoration learning collaborations like this one. We believe they are essential to generate the transforma- tion necessary to meet the needs of our changing planet (Discussions and Conclusions). Methods of the learning exchange In addition to active restoration, Suffolk County’s project included a learning exchange among salt marsh restoration experts across the Sandy-impacted region (from VA-ME) (Maher 2018). This learning exchange, assembled and led by The Nature Conser- vancy in NY (TNC), was called the Regional Tech- nical Workgroup (RTW). The RTW was a forum for practitioners to discuss the best available restoration methods and share lessons learned to improve the success of coastal wetland restoration projects both within Suffolk County and across the larger region. A collaboration among the Suffolk County Depart- ment of Economic Development and Planning, Suf- folk County Parks, and Suffolk County Vector Control in NY, jointly referred to as Suffolk County, received one of these HSCR NFWF grants to implement Integrated Marsh Management (IMM) as described in Rochlin et al. (2012). The funds were awarded to promote coastal resilience by restoring multiple func- tions and environmental beneﬁts to over 200 acres of tidal marsh on the south shore of Suffolk County and build the capacity to eventually rehabilitate 1500 acres. The County’s multifaceted IMM approach strives to improve tidal hydrology, remediate the historic mosquito grid ditch network, enhance pro- cesses for marsh accretion and resilience to sea level rise, promote healthy native vegetation, provide natural control of the invasive common reed, When we initiated this learning exchange, we were unaware of the established Collaborative Learning structure described by Daniels and Walker (2001) and demonstrated by Feurt (2008). We inadvertently used some of the principles they describe. Future efforts will beneﬁt from following their established protocols more closely. 123 123 1035 Wetlands Ecol Manage (2022) 30:1033–1047 Approximately 50 participants from the Sandy- impacted region comprised a multidisciplinary group of salt marsh professionals including the Principal Investigators of the post-Sandy funded marsh restora- tion projects, restoration practitioners, mosquito con- trol specialists, natural resource managers, regulators, academic researchers, and conservation scientists. Notably, the learning exchange also included agency representatives from Federal, state, county, and local municipalities with expertise in conducting and eval- uating restoration projects. Active participation grew and shrank over the course of the 2-year timeline as individuals’ availability changed and as their expertise and interest intersected with the objectives of each scheduled meeting and ﬁeld trip. was because at different sites project leads prioritized different measures of success. Methods of the learning exchange Participants were open and generous in sharing lessons learned from across the region at sites that were similar and others that were quite dissimilar. Hearing perspectives from different regions helped advance our thinking about the marshes in our own locales. We distributed presentations, meeting notes, meet- ing recordings, referenced papers and reports, ﬁeld trip notes, and photographs after each meeting and stored relevant information on a shared online folder for easy access by participants. Members used this online folder as a repository for other resources that they wanted to share. Discussions focused primarily on two types of restoration approaches: restoring hydrology and restoring elevation. A ﬁnal report (Maher 2018) captured the motivation of the learning exchange and bulleted lists of lessons learned by category. The learning exchange was conducted through both traditional in-person and remote meetings as well as ﬁeld trips to restoration sites over the course of 2 years. In-person meetings were held at the start (fall 2016) and end (fall 2018), two webinar meetings were held in the middle (winter 2017 and summer 2018). Ten days of ﬁeld trips scheduled throughout the project timeline visited a total of 20 marsh sites across the region. Beneﬁts of the learning exchange The group is standing in a patch of typical ‘‘high marsh’’ grasses (Spartina patens and Distichlis spicata) that is soggy with shallow standing water at ffolk County’s Gardiner County Park East toration (11/18/18). Changes can happen ne growing season. The group is standing high marsh’’ grasses (Spartina patens and t is soggy with shallow standing water at low tide from poor drainage. Notice the Robert Moses Causeway bridge in the background as a marker. A photo of this same spot two growing seasons after restoration is shown in Fig. 7. (N. Maher @TNC) low tide from poor drainage. Notice the Robert Moses Causeway bridge in the background as a marker. A photo of this same spot two growing seasons after restoration is shown in Fig. 7. (N. Maher @TNC) low tide from poor drainage. Notice the Robert Moses Causeway bridge in the background as a marker. A photo of this same spot two growing seasons after restoration is shown in Fig. 7. (N. Maher @TNC) Fig. 1 Field trip to Suffolk County’s Gardiner County Park East restoration site pre-restoration (11/18/18). Changes can happen quickly—even over one growing season. The group is standing in a patch of typical ‘‘high marsh’’ grasses (Spartina patens and Distichlis spicata) that is soggy with shallow standing water at There is growing recognition of the value of integrating science and management activities to achieve multiple beneﬁts (e.g. Dale and Huslman 1990; Dale and Knight 2008, 2012; Wolfe 2005). Dale and Knight (2008) speciﬁcally call this out as a knowledge gap that requires interdisciplinary collab- oration. Similarly, Wolfe (2005) identiﬁes communi- cation among researchers, mosquito control managers, and regulators as a missing link. Furthermore, Dale and Hulsman (1990) make it clear that the multidis- ciplinary approach necessary to manage complex coastal ecosystems is often beyond the resources of individual mosquito control districts’ capacity to achieve. Our learning exchange provided an opportu- nity for a multidisciplinary group to collaborate and ﬁll these critical gaps to identify the best available restoration methods to meet multi-stakeholder needs and address multiple beneﬁts (e.g. mosquito control, coastal resilience, ﬂood risk reduction, habitat protec- tion for threatened species, long term habitat sustainability). grid ditching and chemical control was destructive to marsh ecosystems (Hulsman et al. 1989; Wolfe 2005; Dale and Knight 2008). A growing appreciation for the value of marsh ecosystems inspired development of alternatives to these earlier destructive methods (Hulsman et al. Beneﬁts of the learning exchange Incorporating a learning exchange into an on-the- ground restoration project provided an efﬁcient plat- form for the continuous discussion of best practices and an opportunity to implement those recommenda- tions in real time. Some of the beneﬁts and lessons learned through this experience are described below under the headings: Achieving Multiple Beneﬁts, Engaging Regulators, Learning in Real Time, and Practical Knowledge Exchange in the Field. Participants contributed their experience and shared their projects for review during the meetings and ﬁeld trips. Projects included those in early stages of development, currently underway, and various ages post completion. Projects funded under the post-Sandy funding streams were emphasized. Suffolk County shared their initial restoration designs based on their holistic IMM approach for rigorous evaluation and peer review through the learning exchange to ensure implementation of the best available methods to achieve multiple beneﬁts and coastal resilience. Arguably, the most productive meetings were designed as ‘‘ﬁeld trips’’ at restoration project sites where conversations were more free form. Field trips were hosted at the Suffolk County sites pre-restoration (Fig. 1) and at partner sites both pre-restoration and at different stages post-restoration (Online Resource 1). Additional photographs of ﬁeld visits to the Suffolk County sites post-restoration are also available in Online Resource 1. Achieving multiple beneﬁts: exploring runnels as an example Restoration projects are typically designed locally to meet the goals of whichever agency is leading the effort. An avian conservation group may prioritize restoration of habitat for obligate salt marsh nesting birds; a sporting organization for speciﬁc species of waterfowl; a municipality for ﬂood risk reduction for a coastal community; a vector control agency for mosquito control. Restoration can also be optimized to promote long term resilience by facilitating marsh migration. However, when designed purposefully, restoration of coastal marshes can achieve multiple beneﬁts for wildlife, coastal communities, and coastal economies (Rochlin et al. 2012; Wigand et al. 2015; Wolfe 1996, 2005). During meetings, we shared our different perspec- tives, sometimes without directly addressing inherent disagreements. We never tried to reach consensus or informed consent. Perhaps this was because of the acknowledgement that every site is unique; perhaps it 12 23 Wetlands Ecol Manage (2022) 30:1033–1047 1036 Fig. 1 Field trip to Suffolk County’s Gardiner County Park East restoration site pre-restoration (11/18/18). Changes can happen quickly—even over one growing season. 123 Beneﬁts of the learning exchange 1989; Wolfe 2005; Meredith and Lesser 2007). More recent approaches include drain- ing water from the surface of the marsh (source reduction) and enhancing water ﬂow and ﬁsh access to depressions on the marsh surface that are ideal breeding grounds for salt marsh mosquitoes. This allows native killiﬁsh species to consume larvae before they become ﬂying, biting, adults (biological control). Runnels are a tool intentionally developed by mosquito control practitioners in Australia to facilitate biological control of mosquitoes while maintaining or at least minimizing impacts to natural marsh function (Hulsman et al. 1989). Runnels, as codiﬁed in Australia, are shallow ‘‘spoon-shaped’’ channels that are at least three times as wide as they are deep positioned to drain standing water from low spots on the marsh surface and prevent larval development in those areas (Hulsman et al. 1989). Mosquito control agencies offered unique experi- ences and expertise in this learning exchange, includ- ing long term familiarity with the sites they manage through different weather patterns and seasons. Although their goals are more holistic now, histori- cally their primary goal in saltmarsh management was to reduce the production of mosquitoes that are vectors of disease. This single-minded approach including In contrast to Australia, the US uses a more informal deﬁnition and approach to runnels and there are morphological differences between the runnels implemented for mosquito control in the US and Australia. In the US, runnels for mosquito control 123 123 Wetlands Ecol Manage (2022) 30:1033–1047 1037 (positioned for either source reduction or biological control) are always shallower than drainage ditches or creeks, but they are not always wider than they are deep. For example, in Suffolk County, NY, runnels are designed to serve the needs of the microtidal regime (tidal amplitude on the south shore of Suffolk County is only 21–33 cm). Accordingly, the dimensions of runnels here average 30.5 cm wide by 30.5 cm deep. They can be dug by hand with a shovel (23 cm blade), or mechanically using an excavator with a 70 cm bucket rotated 45, a 30.5 cm bucket, or a blade (Fig. 2). Most runnels are connected to tidal channels and some are connected to permanent micropools that offer ﬁsh habitat on the marsh surface. Figure 3 shows an example of a micropool and runnel installed at Suffolk County Gardiner Park West site less than a year after implementation. Beneﬁts of the learning exchange The average dimensions of micropools as implemented in these Suffolk County projects are 3 m 9 3 m 9 0.6 m. In Delaware, USA, Meredith et al. (1985) and Meredith and Lesser (2007) describe shallow channels called ‘‘sill ditches’’ to remove water from shallow surface water larval habitats and ‘‘radial ditches’’ to provide ﬁsh access between permanent ponds and outlying larval habitats. These shallow sill and radial ditches are sometimes informally called ‘‘runnels’’ in the US even though they do not ﬁt the intent or deﬁnition of runnels as created in Australia. Runnels are a good opportunity for collaboration among diverse stakeholders and to achieve multiple beneﬁts, but there are some terminology hurdles to overcome. Independently, salt marsh ecologists in excavator (30.5 cm bucket), and excavator with a blade. Average runnel dimensions 30.5 9 30.5 cm. (J. Montesano, SC Vector Control) Fig. 2 Runnel implementation at Suffolk County’s Gardiner County Park West restoration site: runnels were built by shovel (* 23 cm blade), excavator (70 cm bucket, rotated 45), excavator (30.5 cm bucket), and excavator with a blade. Average runnel dimensions 30.5 9 30.5 cm. (J. Montesano, SC Vector Control) excavator (30.5 cm bucket), and excavator with a blade. Average runnel dimensions 30.5 9 30.5 cm. (J. Montesano, SC Vector Control) Fig. 2 Runnel implementation at Suffolk County’s Gardiner County Park West restoration site: runnels were built by shovel (* 23 cm blade), excavator (70 cm bucket, rotated 45), excavator (30.5 cm bucket), and excavator with a blade. Average runnel dimensions 30.5 9 30.5 cm. (J. Montesano, SC Vector Control) 12 3 Wetlands Ecol Manage (2022) 30:1033–1047 1038 Fig. 3 Micropool and runnel in upper part of high marsh habitat near the invasive Phragmites line at Suffolk County Gardiner Park West less than one-year post-restoration (03/08/21). Dimensions of runnels installed here are 30.5 9 30.5 cm; dimensions of micropools are 3 m 9 3 m 9 0.6 m. (N. Maher @TNC) Fig. 3 Micropool and runnel in upper part of high marsh habitat near the invasive Phragmites line at Suffolk County Gardiner Park West less than one-year post-restoration (03/08/21). Dimensions of runnels installed here are 30.5 9 30.5 cm; dimensions of micropools are 3 m 9 3 m 9 0.6 m. (N. Maher @TNC) runnels was for source reduction or biological control of mosquito larvae (water and fauna). This compares to the later marsh ecology application which focused on promoting revegetation (water and ﬂora). Recently, the mosquito control use of runnels also includes the purpose of draining shallow, improperly impounded water created by parallel ditch levees to allow ﬂushing and vegetative recovery (R. Wolfe, pers. com.), similar to the intentions of marsh ecologists. For these reasons, being explicit about our intentions for runnels and how they implemented is important. There is a valuable opportunity to engage across these comple- mentary disciplines (mosquito control and marsh ecology) to develop minimally invasive restoration approaches that achieve multiple beneﬁts. Each group has a lot to gain from the other to inform effective restoration and resource management. By engaging representatives of these different disciplines simulta- neously, and sharing our diverse experiences, chal- lenges, and lessons learned we can appreciate each others’ perspectives and tools in the ﬁeld. However, there is not a universal understanding for the purpose and implementation of what we both call runnels. This remains an example of where we have more to learn from each other as we learn how to manage marshes to achieve multiple beneﬁts and how marshes respond to these restoration techniques. Fig. 3 Micropool and runnel in upper part of high marsh habitat near the invasive Phragmites line at Suffolk County Gardiner Park West less than one-year post-restoration (03/08/21). Dimensions of runnels installed here are 30.5 9 30.5 cm; dimensions of micropools are 3 m 9 3 m 9 0.6 m. (N. Maher @TNC) Dale and Knight (2008) emphasize that mosquito and wetland management communities need to communicate and collaborate to simultaneously protect both wetland and human health. We argue that multidisciplinary learning exchanges, like we demonstrate here, are more expedient and potentially more effective than attend- ing each others’ professional meetings to glean similar New England, USA also started using shallow drainage channels to connect areas of improperly impounded shallow water to the tide (Ferguson 2016). The primary marsh ecology application is to drain improperly impounded water off the marsh surface at low tide and drop the water table slightly to relieve root zone ﬂooding so the salt marsh plants can grow vigorously, build elevation, and halt the marsh subsi- dence trajectory. Without knowing that the mosquito community already used the term, they referred to these shallow drainage features as runnels. Runnels as implemented for marsh restoration in Rhode Island (RI) can be small (15 cm wide by 25 cm deep) or large (30 9 30 cm) and can be implemented either by hand or machinery (W. Ferguson pers. com.). See Online Resource 1 for photos of runnels as implemented in RI. Runnels are now one of the many inexpensive and innovative tools at the forefront of marsh restoration science to restore more natural hydrology, support vigorous vegetation growth, and promote increased vertical marsh elevation growth in response to sea level rise (Wigand et al. 2015; Adamowicz et al. 2020). This example of semantic discord on runnels reﬂects the communication challenges present when previously discussing Open Marsh Water Manage- ment (OMWM) (see Wolfe 2005). There is a super- ﬁcial similarity between these two types of shallow drainage channels both called runnels. Although they appear similar in many ways, the motivations to implement them, the ways that they are implemented in the ﬁeld, and the expectations of what they will achieve are sometimes (but not always) meaningfully different. The original mosquito control application of 123 123 Wetlands Ecol Manage (2022) 30:1033–1047 1039 knowledge, open lines of communication, and develop an understanding of our complimentary marsh man- agement and its critical role in climate resilience. efforts and maximize their likelihood of success. It was a tangible opportunity for interdisciplinary capac- ity building, providing opportunities for local project managers and regulatory staff to engage with restora- tion experts and research scientists. Practical knowledge exchange in the ﬁeld Field trips to restoration sites at different stages of implementation were exceptionally productive. It was invaluable to see the sites ﬁrsthand, develop a shared perspective, compare site conditions, ask questions, and discuss alternatives while assessing a site in real time. These onsite meetings allowed a more ﬂuid exchange of ideas including implementation advice and operational details that never make it into written reports, publications, or presentations. These crucial details play a large role in executing successful projects on the ground. In this informal setting, participants freely discussed insights gained, pros and cons of different types of equipment, challenges faced when working with contractors, unanticipated modiﬁcations necessitated by site conditions, adaptive management tactics, and other lessons learned. Field visits were a safe space to share lessons learned from failures and false starts without judgement by regula- tors or funders. We all recognized that there is no agreed upon guidance for deciding which treatments Although it may feel like they are in opposing positions during the permitting process, environmen- tal regulators and restoration practitioners seek a common outcome: a healthy, productive, and sustain- able ecosystem. Both sides come with inherent advantages. Regulators see a large number of projects over time and can often predict where issues may arise based on past experience. Practitioners, meanwhile, may be more knowledgeable about new restoration techniques and promising experimental methods. This learning exchange gave regulators a chance to see novel restoration techniques in the ﬁeld beyond their home states, broadening their understanding of current restoration science while increasing their comfort in the feasibility of these methods. Engaging regulators: building a mutually beneﬁcial relationship Engaging regulators: building a mutually beneﬁcial relationship The regulatory approval process for any wetland restoration project is a long and complicated endeavor. There are environmental protection regulations that must be followed under federal, state, and local governments, and grant-funded projects may have additional requirements for compliance. For these reasons, there can be friction between those hoping to implement restoration work and the regulatory author- ities they perceive to be standing in their way. This learning exchange provided a platform for ongoing, open communication between restoration practitioners and regulatory staff. Regulators were able to con- tribute to design ideas during early project develop- ment rather than reacting to permit requests. It was crucial to have regulators on the front lines and actively engaged so they were free to contribute insights and design improvements before there had been investment in an alternative project approach. Regulatory staff also learned from practitioners why a particular strategy was more likely to succeed, or why a commonly held permitting standard might not be appropriate for a particular project. The learning exchange also connected practitioners across a larger geography than is common for development of individual projects. Participants rep- resented active projects from Maine to Virginia which brought geographically diverse experiences and per- spectives. As a result, when they shared lessons learned from their individual projects, those lessons inﬂuenced multiple projects and practitioners from across the entire region. This approach also provided the opportunity for all involved to stay current on the latest thinking in restoration science. Discussions included critical evaluation of restoration site selection, design, implementation, monitoring, and adaptive management. Mosquito control special- ists, as well as local regulators and managers, bring valuable experience to these conversations but are frequently unable to participate in national scientiﬁc and restoration conferences that require travel across state lines. Local learning exchanges can be a critical opportunity for them to engage with their peers in other sectors. Suffolk County project implementation Engagement with the learning exchange and observa- tion of restoration sites with similar tidal regimes and restoration objectives informed strategic revision of Suffolk County’s preliminary restoration designs. Many of the proposed recommendations were incor- porated. This engagement reduced the risk of unan- ticipated results, promoted the efﬁcient use of staff and funding resources, and smoothed the permitting pro- cess because the permit application was improved by real time, vigorous peer-review. Figure 4 shows a side-by-side comparison of the initial and ﬁnal restoration designs for Suffolk County Gardiner Park East, as an illustration of the changes made across all Suffolk County sites implemented as part of their HSCR NFWF grant. All ﬁeld trips occurred before Suffolk County ﬁnalized and implemented its restoration plans, and the ﬁeld visits helped to reﬁne many of the designs’ technical details. The ﬁeld trip to a collection of restoration sites in Rhode Island (RI) led by Wenley Ferguson, Director of Habitat Restoration with Save the Bay (https://www.savebay.org) was especially enlightening. Suffolk County was considering if run- nels were the best way to address improperly impounded water on the marsh surface and if so, how and where to implement them. Practitioners in RI are restoration leaders in ﬁguring out when, where, and exactly how to implement minimally invasive yet effective runnels to address improperly impounded water causing marsh subsidence (Ferguson 2016; Wigand et al. 2015). The ﬁrst modiﬁcation to Suffolk County’s initial designs addressed the proposal for ‘‘naturalizing’’ or adding sinuosity to some linear mosquito ditches by cutting and repositioning sections of marsh peat along them to restore curvature. The primary motivation was to improve natural drainage and sediment deposition over the marsh while maintaining open channels for tidal exchange. A secondary aesthetic motivation was to remove artiﬁcial straight lines and restore sinuosity that more closely resembled natural tidal creeks. However, learning exchange participants thought the disturbance caused by ditch ‘‘naturalizing’’, in partic- ular cutting and moving high quality peat and vegetation to add curves, would cause more harm than beneﬁt. Regulators from the New York State Department of Environmental Conservation (NYS- DEC) also objected to the intensive ditch reshaping originally proposed due to the possibility that the disturbance would cause contaminants within the sediments to be resuspended and reintroduced to biota. Learning in real time This learning exchange provided real time peer review for all participants to improve their own restoration 12 3 3 Wetlands Ecol Manage (2022) 30:1033–1047 1040 are best for which sites because marsh restoration is not prescriptive; we need to experiment, measure, learn and share. Outcomes Tangible outcomes of the learning exchange included improved restoration and monitoring plans for projects initiated by Suffolk County and the professional relationships formed and strengthened among the participants across the region. These different types of outcomes are described below in greater detail. Meetings conducted on restoration sites enabled participants to understand the full context of those projects. Although there are many similarities among degraded salt marshes (often resulting from wide- spread historic practices such as grid ditching) each site is unique and warrants careful consideration of local conditions. Habitat heterogeneity necessitates customized restoration designs to meet site speciﬁc needs and stakeholder objectives. Field trips provided insight into local conditions and a more realistic view of potential restoration practices. Suffolk County project implementation Instead, the Suffolk County team added sinu- osity to the system with the purposeful placement of runnels (typically 30.5 cm 9 30.5 cm) through low Rhode Island presents a shining example of how conservation and mosquito control practitioners can work collaboratively to achieve multiple goals in their projects (W. Ferguson pers. com.). By visiting restoration sites in RI with similar site characteristics to those in Suffolk County (microtidal marshes with low sediment supply), we observed marsh response to runnel treatments of different ages implemented by Save the Bay, RI Department of Environmental Management Mosquito Abatement Coordination, and USFWS. We also learned how the experienced RI team evaluated a site in the planning stages by visiting one of their pre-restoration sites and engaging with them on the different considerations and alterna- tives under consideration for the strategic restoration plan. This helped Suffolk County better strategize the implementation of runnels on their own sites. See further discussion below. 12 123 Wetlands Ecol Manage (2022) 30:1033–1047 1041 Fig. 4 Side by side maps of preliminary (left) and revised/ﬁnal (right) restoration plans for Suffolk County’s Gardiner Park East in Bayshore, NY. Meaningful changes incorporated from the learning exchange are displayed here. The ditches to be maintained as the primary tidal creeks were not ‘‘naturalized’’ to add sinuosity by cutting and repositioning peat. This was important to minimize disturbance and especially to avoid cutting healthy high marsh peat. The extraneous ditches were ﬁlled with coir logs and natural peat from the site, but importantly, the slightly elevated levees along the sides of these ditches were not scraped down for use as ﬁll. Instead, those areas were left to provide some elevation refuge for obligate salt marsh nesting birds. Fewer runnels were incorporated into the ﬁnal designs, but those that were used were more strategically placed and used to add sinuosity. Multiple runnels drain into the same primary ditch to keep water ﬂowing in and out The concept of ‘‘ditch remediation’’ is to rebuild the peat that was once there by creating the conditions for ﬁbrous organic material, sediment, and living roots to accumulate in the linear ditch (Burdick et al. 2020). When material reaches the appropriate elevation, plants naturally colonize through seedling or colonial growth. The approach demonstrated by Burdick et al. Suffolk County project implementation Because material for peat building was scarce at these sites, Suffolk County used intact coir logs and some natural peat excavated from runnel and micro-pool construction on site to ﬁll ditches to the elevation of the marsh platform but did not scrape down material from the elevated ditch levees as originally proposed. Over two growing seasons since implementation, the coir logs have not caused the disruptions described by other learning exchange participants. A photograph of a ditch ﬁlled in this manner after 2 growing seasons is included in Online Resource 1. One possible explanation for the different response observed by Suffolk County is that the coir logs used in these projects were exposed to the elements for over 6 months and up to a year before deployment. This weathering may have made the coir material more stable when positioned, buried, and staked down in the ditches. A third modiﬁcation to the initial designs was the way runnels were strategically used to achieve three speciﬁc goals: to restore tidal hydrology and vegeta- tion growth, to promote marsh migration and invasive species control, and to address mosquito abatement issues in breeding hot spots. Consistent with the marsh ecology use of the term, runnels were deliberately placed to restore drainage of shallow improperly impounded water that was preventing vegetation growth and elevation accretion on the marsh platform. Figure 5 shows surface hydrology before and after runnel implementation. Consistent with the mosquito control use of the term, runnels, and a small number of ﬁsh refuge micropools were created near the upper edges of the marsh where Suffolk County will address all three goals simultaneously (Fig. 3). Figure 6 shows the improved surface hydrology and vegetation recov- ery of the same location pictured in Fig. 1 two growing seasons post-restoration. Suffolk County is actively monitoring the function of these runnels (currently 1–2 years post implementation) to determine how well they are meeting their restoration and manage- ment goals. It is important to note that there was disagreement about the use of coir logs as ﬁll. Many participants shared experiences with buoyant coir logs shifting out of position and causing complications on sites. One suggestion from the group was to open the coir logs and use the inner ﬁbrous material to line the ditches in a manner similar to how Burdick et al. Suffolk County project implementation (2020) relies on harvesting salt hay from the sur- rounding marsh to place in ditches as a ﬁlter to trap material and start the peat building process. However, in marshes without this resource, managers may need to ﬁnd substitutes for locally harvested salt hay. areas and focused on strategic restoration of only those linear ditches with poor tidal connectivity. This allowed the County to avoid the extensive toxicity testing required by NYSDEC when moving or disturbing larger volumes of sediments. A second improvement to the initial designs was partial modiﬁcation of how the non-functioning linear mosquito ditches were retired. The team attempted to identify and restore natural ‘tidesheds’ in the project design so each system had a single main channel or primary ditch to accommodate the incoming and outgoing tide. Supplementary ditches were retired. 12 3 3 Wetlands Ecol Manage (2022) 30:1033–1047 1042 Preliminary designs for Suffolk County proposed utilizing a combination of coconut ﬁber logs (coir logs) and material scraped from miniature levees along the sides of the linear ditches to ﬁll the space up to the marsh surface. that leaving ditches partially ﬁlled could create new mosquito breeding habitat in these microtidal marshes. In previous IMM projects, Suffolk County maintained that retired ditches had to be ﬁlled ﬂush with the surface of the marsh to prevent creation of new larval mosquito habitat (Rochlin et al. 2012). The learning exchange advised experimentation with ﬁlling non-functioning ditches, provided there was sufﬁcient drainage in the marsh, but it was best not to dig up stable marsh peat to do so, particularly along the slightly elevated ditch levees. They advised preserving elevation along the ditch edges to provide multiple habitat beneﬁts, most notably for obligate salt marsh breeding birds such as the Saltmarsh Sparrow (Ammodramus caudacutus) known to inhabit these marshes on the south shore of Suffolk County (eBird.org). Saltmarsh Sparrows use the peat mound refugia for nesting and protection from high tides and return to the same marshes to breed in successive years, often nesting within a few meters of their previously successful nests (Benvenuti et al. 2018). Maintaining these refugia in place is important. Suffolk County recognized the ecological and eco- nomic value of maintaining intact peat and vegetation along these miniature levees. Leaving them intact preserved habitat and kept total excavation volume under the contaminant testing limits, thus reducing project costs. Suffolk County project implementation (2020) used mowed salt hayThe recommendation was to leave the coir ﬁber material exposed (not buried with peat) so that water could ﬁlter through and trap sediments to start the peat rebuilding process. Professional relationships and new collaborations The City of New York Department of Parks and Recreation (NYC Parks), The Nature Conservancy in New Jersey and the Partnership for the Delaware Estuary (PDE), now consult one another regarding techniques, materials and monitoring approaches for their respective living shoreline projects capitalizing on one another’s resources and skills (Joshua Moody, Christopher Haight, and Adrianna Zito-Livingston pers. com.). Joshua Moody, Restoration Program Manager at the Partnership for the Delaware Estuary shared: ‘‘Having an open dialogue across regions regarding project development, implementation, and outcomes is a critical facet of restoration science, and these relationships are foundational in evaluating the transferability of speciﬁc goal-based restoration tech- niques.’’ Additionally, Ducks Unlimited had not previously partnered with NYC Parks but now they have leveraged resources on what might be the ﬁrst ever U.S. Fish and Wildlife Service $1 Million Standard North American Wetland Conservation Act grant for southern New York including restoration In the time since our learning exchange concluded, many groups have expressed interest in convening learning exchanges across other geographies and areas of expertise and especially incorporating ﬁeld trips into those learning exchanges. For example, USFWS and other partners recently convened conservation stakeholders throughout New Jersey (NJ) and Dela- ware (DE) to form a Salt Marsh Working Group. Communication and sharing lessons learned were top interests and purposes for the group. During initial working group meetings, this learning exchange was held up as a highly successful model that the NJ-DE group is hopeful to repeat, once interstate travel is again feasible (Adrianna Zito-Livingston and Mitch Hartley pers. com.). Working groups like this one are forming in several other locations in the northeast as well. Professional relationships and new collaborations Professional relationships and new collaborations In addition to the restoration outcomes on the ground, other valuable outcomes of the learning exchange include the productive partnerships formed and strengthened through the process (Fig. 7). Practition- ers across the region are now a ready resource to one The learning exchange participants familiar with ditch remediation cautioned that the process of peat building was important and that it was not possible to skip to the endpoint by just putting peat on the top of coir ﬁlling. However, Suffolk County had concerns 123 123 123 Wetlands Ecol Manage (2022) 30:1033–1047 1043 Fig. 5 Surface water before and after runnels (30.5 9 30.5 cm) are cut on the Suffolk County Gardiner Park West site. (J. Montesano, SC Vector Control) Fig. 5 Surface water before and after runnels (30.5 9 30.5 cm) are cut on the Suffolk County Gardiner Park West site. (J. Montesano, SC Vector Control) ore and after runnels (30.5 9 30.5 cm) are cut on the Suffolk County Gardiner Park West site. (J. Montesano, Fig. 5 Surface water before and after runnels (30.5 9 30.5 cm) are cut on the Suffolk County Gardiner Park West site. (J. Montesano, SC Vector Control) Fig. 6 Suffolk County Gardiner County Park East restoration site two growing seasons post-restoration (03/ 08/21). This is the same location as featured in Fig. 1 pre-restoration. Notice the Robert Moses Causeway bridge in the background as a marker. There is some Spartina alterniﬂora, but the spot is dominated by Salicorna spp. (20 cm tall) indicating signs of recovery. (N. Maher @TNC) 123 Wetlands Ecol Manage (2022) 30:1033–1047 1044 Fig. 7 Field visit to Cape May NWR in Cape May Court House, NJ. New introductions of partners. Jim Feaga, Ducks Unlimited meeting and shaking hands with Beth Watson, Academy of Natural Sciences of Drexel University (ANS). Ducks Unlimited ield visit to Cape May NWR in Cape May Court House, introductions of partners. Jim Feaga, Ducks Unlimited and shaking hands with Beth Watson, Academy of Sciences of Drexel University (ANS). Ducks Unlimited had not previously engaged with ANS but following the learning exchange they connected on a number of projects including one at the John Heinz NWR in Philadelphia, PA. (N. Maher @TNC) had not previously engaged with ANS but following the learning exchange they connected on a number of projects including one at the John Heinz NWR in Philadelphia, PA. (N. Professional relationships and new collaborations Maher @TNC) Fig. 7 Field visit to Cape May NWR in Cape May Court House, NJ. New introductions of partners. Jim Feaga, Ducks Unlimited meeting and shaking hands with Beth Watson, Academy of Natural Sciences of Drexel University (ANS). Ducks Unlimited had not previously engaged with ANS but following the learning exchange they connected on a number of projects including one at the John Heinz NWR in Philadelphia, PA. (N. Maher @TNC) sites from NYC and the eastern end of Suffolk County (Jim Feaga pers. com.). another and have the capacity to leverage resources for greater restoration and management outcomes. Although we didn’t measure these valuable outcomes the way that Paolisso et al. (2019) did so skillfully, we have anecdotal evidence to this effect. There are too many examples of connections formed or strengthened through this learning exchange to be inclusive, but two examples are included here for illustrative purposes. The City of New York Department of Parks and Recreation (NYC Parks), The Nature Conservancy in New Jersey and the Partnership for the Delaware Estuary (PDE), now consult one another regarding techniques, materials and monitoring approaches for their respective living shoreline projects capitalizing on one another’s resources and skills (Joshua Moody, Christopher Haight, and Adrianna Zito-Livingston pers. com.). Joshua Moody, Restoration Program Manager at the Partnership for the Delaware Estuary shared: ‘‘Having an open dialogue across regions regarding project development, implementation, and outcomes is a critical facet of restoration science, and these relationships are foundational in evaluating the transferability of speciﬁc goal-based restoration tech- niques.’’ Additionally, Ducks Unlimited had not previously partnered with NYC Parks but now they have leveraged resources on what might be the ﬁrst ever U.S. Fish and Wildlife Service $1 Million Standard North American Wetland Conservation Act grant for southern New York including restoration another and have the capacity to leverage resources for greater restoration and management outcomes. Although we didn’t measure these valuable outcomes the way that Paolisso et al. (2019) did so skillfully, we have anecdotal evidence to this effect. There are too many examples of connections formed or strengthened through this learning exchange to be inclusive, but two examples are included here for illustrative purposes. Discussion and conclusions experiences and best practices to build momentum and scale up ecosystem restoration to meet the needs of a changing climate. Suffolk County had the foresight to include a regional interdisciplinary learning exchange (the RTW) as part of their HSCR NFWF grant. They recognized the immense value in learning from experts and projects across the region and wanted to make the most of the opportunity to implement restoration on 200? acres across Suffolk County by ensuring implementation of the best available methods. Importantly, they were open to hearing new ideas and having their assump- tions challenged so the projects implemented had the greatest chance of success. g g Furthermore, there is renewed immediacy to restore Atlantic Coast salt marshes. The Saltmarsh Sparrow, an obligate salt marsh nesting bird endemic to this area, was recently recognized by the U.S. Fish and Wildlife Service as an ‘‘At Risk Species’’. While not placed on the list of federally threatened or endangered species, At Risk status indicates that federal and state agencies are placing special emphasis on conserving the sparrow’s numbers. Without such intervention, population numbers are predicted to hit severely low levels by 2030 (Field et al. 2017). Since ﬂooding of low elevation nests is one of the primary causes of decline, there is increasing emphasis to restore degraded marshes (Hartley and Weldon 2020). The best opportunity to do so is during the next 10 years while we are experiencing smaller tidal amplitudes in the current phase of the 18.6-year Metonic Cycle (Adamowicz et al. 2020). After that, the cycle switches to higher tidal ranges that will peak by about 2035. This learning exchange was an unprecedented assembly of marsh restoration practitioners across the Hurricane Sandy-impacted region. Because restoration science is still a developing and interdis- ciplinary ﬁeld, this experience has identiﬁed the beneﬁts of such platforms at the local and regional scale. Among the many take-away points from this exchange, we learned that diverse stakeholder groups, including mosquito control agencies, bring valuable experience to these conversations and should be included in marsh restoration and management dis- cussions. Engaging regulators provided a platform for open communication and simultaneous learning for restoration practitioners and regulatory staff that both improved project design and streamlined the permit- ting process. It was also clear that real time learning and capacity building at the local level are paramount for getting restoration to scale. Considerations and questions for the future The United Nations General Assembly declared 2021–2030 the ‘‘Decade on Ecosystem Restoration.’’ This call to action recognizes the need to accelerate restoration of degraded ecosystems across the globe to get the world on track for a sustainable future. The declaration speciﬁcally calls out the importance of capacity building and cooperation including sharing The United Nations General Assembly declared 2021–2030 the ‘‘Decade on Ecosystem Restoration.’’ 12 123 Wetlands Ecol Manage (2022) 30:1033–1047 1045 Discussion and conclusions Conversations that occurred on ﬁeld trips were especially valuable for conveying practical knowledge. Not all the recom- mendations from the learning exchange were incor- porated into the projects implemented by Suffolk County, but the project designs were greatly improved by the recommendations that were adopted. But how do we organize restoration efforts on a scale large enough to make a difference? And how do we provide a quality assurance/quality control aspect to the endeavor? Building on the example of the learning paradigm described here, Adamowicz et al. (2020), describe a two-level approach—local and regional. The Salt Marsh Adaptation and Resiliency Teams (SMARTeams) model provides three support teams at the regional level: Design Review, Technical Support, and Training, Outreach, and Education. These teams assist local Field Teams—partnerships of landowners, NGOs, researchers, and others—who implement actual projects. Field Teams build within themselves economies of scale, enabling projects larger than any single partner could do alone. That economy of scale is used again in the regional level teams—providing services that could not be ﬁnanced by individual Field Teams. Lessons learned are also conveyed through the SMARTeams in a comprehen- sive and rapid method—keeping all projects through- out the region connected in continual learning and improving. The effort proposed here (building capacity for efﬁcient climate adaptation) is like a minga, a Quechua word from the Andean region which means equally beneﬁting collective work for all its partici- pants. We propose that this type of engagement is a valuable model for our ﬁeld. Even though it was difﬁcult for participants to add another activity to their already full schedules, (especially up against the ambitious accelerated timelines instituted by funding agencies) the beneﬁts of engagement were greater than the costs of doing so. 123 123 1046 Wetlands Ecol Manage (2022) 30:1033–1047 We believe that this learning exchange has already had long-lasting effects on the Sandy-impacted north- east because it promoted better science and more successful individual restoration projects, connected practitioners from across the region, fostered new relationships, brought in new perspectives, and expanded the ‘toolkit’ for restoration project man- agers. We anticipate that successful learning exchanges like this will encourage additional regional collaborations and improve future salt marsh restora- tion projects. These beneﬁts will provide better information to policy makers and funders to inform decisions about conservation, restoration, and man- agement of salt marshes in the face of climate change. Consent for publication All authors read and approved the ﬁnal manuscript. Open Access This article is licensed under a Creative Com- mons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any med- ium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Discussion and conclusions We still have much to learn from one another about how our marshes respond to restoration. The future of our salt marshes depends on our collaborations. 43006. However, the grant did not cover the time of the co- authors to write this article. N. Maher is grateful for the generous support of the William E. & Maude S. Pritchard Charitable Trust. Camilo Salazar’s funding comes from The Water Quality Unit, Department of Economic Development and Planning, Suffolk County, Water Quality Unit, Department of Economic Development and Planning. Data availability The ﬁnal report and ﬁeld trip notes from the learning exchange described here are available from the corresponding author and Suffolk County on request. Declarations Conﬂict of interest During the time of the learning exchange described in this manuscript, author, N. Maher, as an employee of The Nature Conservancy received support from Suffolk County (NY) as a sub-award from their U.S. Department of the Interior’s (DOI) Hurricane Sandy Coastal Resiliency Competi- tive Grant Program (HSCR) administered by the National Fish and Wildlife Foundation (NFWF). Author, C. Salazar is an employee of Suffolk County (NY) and managed the execution of the NFWF grant that supported this learning exchange and the restoration projects implemented. Author, A. Fournier has no conﬂicts to disclose. Acknowledgements This article was inspired by the Regional Technical Workgroup of saltmarsh restoration practitioners funded by NFWF grant funded project: Wetland Restoration in Suffolk County (NY) project ID 43006. N. Maher is grateful for the generous support of the William E. & Maude S. Pritchard Charitable Trust. The ﬁndings and conclusions in this article are those of the author(s) and do not necessarily represent the views of Suffolk County. Dominick Ninivaggi, former Superintendent at Suffolk County (NY) Vector Control, played an instrumental role in steering Suffolk County’s mosquito program towards more holistic marsh restoration and was particularly supportive of creating an advisory group with a broad range of professionals to implement successful restoration projects. Thank you to all the RTW participants who leaned in to make this a valuable experience. Thank you to Susan Adamowicz for her review and contribution on SMARTeams. Thank you to Joseph Montesano for contributing site photos. Ilia Rochlin provided encouragement and valuable suggestions during the process. Conversations with participants including Adam Starke, Adrianna Zito-Livingston, Elizabeth Watson, Jim Feaga, Roger Wolfe, Wenley Ferguson, Joshua Moody, and Mitch Hartley and editing by Kristin France and Anna Bartlett enhanced the manuscript. Constructive comments from the WEM editor, Eric Wolanski, and 2 anonymous reviewers greatly improved the manuscript. Consent for publication All authors read and approved the ﬁnal manuscript. g doi.org/10.1007/s12237-019-00656-5 Meredith WH and Lesser CR (2007) Open marsh water man- agement in Delaware: 1979–2007. New Jersey mosquito control association. In: Proceedings 94th annual meeting, pp 55–69 Cameron Engineering and Associates (CEA) (2015) Long Island tidal wetland trends analysis. Prepared for the New England Interstate Water Pollution Control Commission, p 207, http://www.dec.ny.gov/lands/5113.html New York State Energy Research and Development Authority (NYSERDA) (2018) Integrating SLAMM results and stakeholder priorities to deﬁne marsh adaptation strategies, NYSERDA Report Number 18–04. Prepared by Warren Pinnacle Consulting, Inc. Waitsﬁeld VT. nyserda.ny.gov/ publications Dale PER, Hulsman K (1990) A critical-review of salt-marsh management methods for mosquito-control. Crit Rev Aquat Sci 3:281–311 Dale PER, Knight JM (2008) Wetlands and mosquitoes: a review. Wetl Ecol Manag 16:255–276. https://doi.org/10. 1007/s11273-008-9098-2 Paolisso M, Prell C, Johnson KJ, Needelman B, Khan IMP, Hubacek K (2019) Enhancing socio-ecological resilience in coastal regions through collaborative science, knowl- edge exchange and social networks: a case study of the Deal Island Peninsula, USA. Socio Ecol Pract Res 1:109–123. https://doi.org/10.1007/s42532-019-00010-w Dale PER, Knight JM (2012) Managing mosquitoes without destroying wetlands: an eastern Australian approach. Wetl Ecol Manag 20:233–242. https://doi.org/10.1007/s11273- 012-9262-6 Daniels SE, Walker GB (2001) Working through environmental conﬂict: the collaborative learning approach. Praeger Publishers, Westport Rochlin I, James-Pirri MJ, Adamowicz SC, Wolfe RJ, Capo- tosto P, Dempsey ME, Iwanejko T, Ninivaggi DV (2012) Integrated Marsh Management (IMM): a new perspective on mosquito control and best management practices for salt marsh restoration. Wetl Ecol Manag 20:219–232. https:// doi.org/10.1007/s11273-012-9251-9 Dwyer L, Bassow A, Flynn C (no date) Hurricane Sandy Coastal Resiliency Competitive Grant Program. https://www.nfwf. org/programs/hurricane-sandy-coastal-resiliency- competitive-grant-program Accessed 2 Feb 2021 UN Environ. Prog. (2006) Marine and coastal ecosystems and human well-being: a synthesis report based on the ﬁndings of the millennium ecosystem assessment. UNEP, Nairobi, p 76 Ferguson W (2016) A matter of millimeters: helping our salt marsh survive sea level rise. Tides Mag save Bay 48(1):10–11 Feurt, C (2008) Collaborative Learning Guide for Ecosystem Management. Wells National Estuarine Research Reserve. Wells, Maine, p 20. https://www.wellsreserve.org/sup/ downloads/collaborative_learning_guide.pdf Accessed May 2021 Wigand C, Ardito T, Chaffee C, Ferguson W, Paton S, Raposa K, Vandemoer C, Watson E (2015) A climate change adaptation strategy for management of coastal marsh sys- tems. Estuar Coasts 40:682–693. https://doi.org/10.1007/ s12237-015-0003-y Field CR, Bayard TS, Gjerdrum C, Hill JM, Meiman S, Elphick CS (2017) High-resolution tide projections reveal extinc- tion threshold in response to sea-level rise. References Author contributions All authors contributed to the paper conception and design. Figure 4 was created by Camilo Salazar. The ﬁrst draft of the manuscript was written by Nicole Maher and all authors commented on and contributed to later versions of the manuscript. Adamowicz SC, Wilson G, Burdick DM, Ferguson W, Hopping R (2020) Farmers in the marsh: lessons from history and case studies for the future. Wetl Sci Pract 37(3):182–195 Babson AL, Bennett RO, Adamowicz SC, Stevens S (2020) Coastal impacts, recovery, and resilience post-hurricane sandy in the Northeastern US. Estuar Coasts 43(7):1603–1609 Funding Support for the learning exchange described in this article was through a U.S. Department of the Interior (DOI) Hurricane Sandy Coastal Resiliency Competitive Grant Program (HSCR) administered by the National Fish and Wildlife Foundation (NFWF) to Suffolk County (NY): Wetland Restoration in Suffolk County (NY) project ID Basso GK, O’Brien K, Albino Hegeman M, and V. O’Neill V (2015) Status and trends of wetlands in the Long Island Sound Area: 130 year assessment. U.S. Department of the Interior, Fish and Wildlife Service, p 35 12 123 Wetlands Ecol Manage (2022) 30:1033–1047 1047 support of Suffolk County’s National Fish and Wildlife Foundation (‘‘NFWF’’) Sandy resiliency grant: Coastal Resiliency via Integrated Wetland Management, p 22 Benvenuti B, Walsh J, O’Brien KM, Kovach AI (2018) Plas- ticity in nesting adaptations of a tidal marsh endemic bird. Ecol Evol 8:10780–10793. https://doi.org/10.1002/ece3. 4528 Meredith WH, Saveikis DE, Stachecki CJ (1985) Guidelines for ‘‘open marsh water management’’ in Delaware’s salt marshes—objectives, system designs, and installation procedures. Wetlands 5:119–133 Burdick DM, Moore GE, Adamowicz SC, Wilson GM, Peter CR (2020) Mitigating the legacy effects of ditching in a New England salt marsh. Estuar Coasts 43:1672–1679. https:// doi.org/10.1007/s12237-019-00656-5 g doi.org/10.1007/s12237-019-00656-5 Glob Change Biol 23(5):2058–2070 Wolfe RJ (1996) Effects of open marsh water management on selected tidal marsh resources: a review. J Am Mosq Control Assoc 12:701–712 Hartley MJ, Weldon AJ (eds) (2020) Saltmarsh Sparrow Con- servation Plan. Atlantic Coast Joint Venture, acjv.org/ saltmarshsparrowplan https://www.acjv.org/documents/ SALS_plan_ﬁnal.pdf Accessed 2 Feb 2021 Wolfe RJ (2005) Open marsh water management: a review of system designs and installation guidelines for mosquito control and integration in wetland habitat management. New Jersey Mosquito Control Association. In: Proceedings ninety second annual meeting Hulsman K, Dale PER, Kay BH (1989) The runneling method of habitat modiﬁcation—an environment-focused tool for salt-marsh mosquito management. J Am Mosq Control Assoc 5:226–234 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Maher NP (2018) Saltmarsh Restoration Regional Technical Workgroup (RTW): Final Report. Submitted to Suffolk County Dept. Economic Development and Planning in 123
https://openalex.org/W3091971690	https://ueaeprints.uea.ac.uk/id/eprint/77445/1/Published_Version.pdf	English	null	Fireside tales: understanding experiences of previous eruptions and factors influencing the decision to evacuate from activity of Volcán de Fuego	Volcanica	2,020	cc-by	19,892	Abstract Volcán de Fuego (Guatemala) is an active stratovolcano capable of large (VEI ~ě2) explosive eruptions like that of 3rd June 2018, which triggered pyroclastic ﬂows that devastated the community of San Miguel Los Lotes and caused hundreds of fatalities and severe long-term socio-economic impacts. Future volcanic risk mitigation eﬀorts are likely to involve temporary evacuation of local communities, the success of which requires co-operation between locals, scientists, and decision-makers. However, locals’ experiences of eruptive activity, and how these experiences inﬂuence their responses to evacuation, have not been studied in detail. In 2019 we conducted an investigation of these themes through qualitative research methods involving semi-structured interviews that focussed on direct experience as opposed to volcanic risk perception. We found substantial diﬀerences between scientists’ and locals’ observations of Fuego’s activity. Furthermore, a clear disparity emerged between communities on Fuego’s west and east ﬂanks in terms of direct prior experience of eruptions and communication with authorities. These ﬁndings have serious implications for future evacuation eﬀorts at Fuego and at other highly populated active volcanoes. Fireside tales: understanding experiences of previous eruptions among other factors that inﬂuence the decision to evacuate from eruptive activity of Volcán de Fuego. Ailsa K. Naismith∗α, M. Teresa Armijosβ, Edgar Antonio Barrios Escobarγ, William Chignaδ, I. Matthew Watsonα αSchool of Earth Sciences, Wills Memorial Building, University Bristol, Queens Road, Bristol, BS8 1RJ, UK. βSchool of International Development, University of East Anglia, Norwich, Norfolk, NR4 7TJ, UK. γInstituto Nacional de Sismología, Vulcanología, Meteorología, e Hidrología (INSIVUMEH), Ediﬁcio Central, 7ª. Avenida 14-57, Zona 13, Ciudad de Guatemala, Guatemala. δCoordinadora Nacional para la Reducción de Desastres (CONRED), Avenida Hincapié 21–72, Zona 13, Ciudad de Guatemala, Guatemala. Ailsa K. Naismith∗α, M. Teresa Armijosβ, Edgar Antonio Barrios Escobarγ, William Chignaδ, I. Matthew Watsonα αSchool of Earth Sciences, Wills Memorial Building, University Bristol, Queens Road, Bristol, BS8 1RJ, UK. βSchool of International Development, University of East Anglia, Norwich, Norfolk, NR4 7TJ, UK. γInstituto Nacional de Sismología, Vulcanología, Meteorología, e Hidrología (INSIVUMEH), Ediﬁcio Central, 7ª. Avenida 14-57, Zona 13, Ciudad de Guatemala, Guatemala. δCoordinadora Nacional para la Reducción de Desastres (CONRED), Avenida Hincapié 21–72, Zona 13, Ciudad de Guatemala, Guatemala. Resumen El Volcán de Fuego (Guatemala) es capaz de erupciones catastróﬁcas como la del 3 de junio de 2018, cuando ﬂujos piroclásticos destruyeron la comunidad de San Miguel Los Lotes y causaron numerosos muertos e impactos socio- economicos severos en el largo plazo. Al futuro, la mitigación del riesgo volcánico implicará la evacuación de las comunidades locales, cuyo éxito requiere la cooperación entre autoridades y la población. Sin embargo, se sabe poco sobre cuales son las experiencias de actividad eruptiva de estos grupos, y cómo estas experiencias afectan sus respuestas a la evacuación. En 2019 realizamos una investigación de estos temas y encontramos diferencias signiﬁ- cativas entre las observaciones cientíﬁcas y de la población de la actividad de Fuego. Además, surgió una disparidad entre las comunidades en diferentes ﬂancos del volcán en términos de la experiencia vivida de las erupciones an- teriores y de la comunicación con autoridades. Estos resultados tienen implicaciones para los futuros esfuerzos de evacuación en Fuego y en volcanes análogos. Keywords: Pyroclastic flows; Self-evacuation; Trust; Knowledge; Risk; Volcán de Fueg ∗Corresponding author: ailsanaismith@gmail.com RESEARCH ARTICLE RESEARCH ARTICLE RESEARCH ARTICLE RESEARCH ARTICLE RESEARCH ARTICLE RESEARCH ARTICLE Fireside tales: understanding experiences of previous eruptions among other factors that inﬂuence the decision to evacuate from eruptive activity of Volcán de Fuego. Ailsa K. Naismith∗α, M. Teresa Armijosβ, Edgar Antonio Barrios Escobarγ, William Chignaδ, I. Matthew Watsonα αSchool of Earth Sciences, Wills Memorial Building, University Bristol, Queens Road, Bristol, BS8 1RJ, UK. βSchool of International Development, University of East Anglia, Norwich, Norfolk, NR4 7TJ, UK. γInstituto Nacional de Sismología, Vulcanología, Meteorología, e Hidrología (INSIVUMEH), Ediﬁcio Central, 7ª. Avenida 14-57, Zona 13, Ciudad de Guatemala, Guatemala. δCoordinadora Nacional para la Reducción de Desastres (CONRED), Avenida Hincapié 21–72, Zona 13, Ciudad de Guatemala, Guatemala. O V L O V NI O V O V L O V NI O V O V L O V NI O V RESEARCH ARTICLE RESEARCH ARTICLE RESEARCH ARTICLE 1 Introduction cially missing, although independent estimates suggest that up to 2,900 people were killed [Associated Press 2018]. In addition, an estimated 5,000 people lost their homes [Noticias ONU 2018]. Scientists informed on ac- tivity continuously from 06:30 a.m., while authorities persistently tried to remove people from high-risk ar- eas. However, national media later highlighted the dis- connect between these authorities supposedly fulﬁlling their responsibilities and the high death toll. In partic- On 3rd June 2018, an explosive paroxysmal eruption of Volcán de Fuego (“Chi’gag” in Kaqchikel Maya), an active stratovolcano in southern Guatemala, generated a series of pyroclastic ﬂows that descended Barranca Las Lajas and buried the community of San Miguel Los Lotes. 332 people have been reported as oﬃ- Fireside tales: Volcán de Fuego Naismith et al., 2020 ular, media focussed on the diﬀerent fates of geograph- ically close communities: why did the private golf re- sort of La Reunión successfully evacuate, yet Los Lotes, two kilometres further south, suﬀer such extensive hu- man loss [Tobar 2018]?. This question relates to the larger issue of the ability and willingness of commu- nities to evacuate from eruptive crisis. By investigating the diﬀerent ways in which people experience Fuego’s eruptive activity, and the factors that inﬂuence evac- uation, this paper provides possible explanations and future actions to prevent these situations from happen- ing again. It highlights the importance of understand- ing local residents’ priorities, interests and decision- making processes when managing volcanic risk. volcano encounter. Our ﬁndings show that although authorities’ experiences of Fuego’s activity are similar to changes seen in remote sensing data, locals see the same events in an entirely diﬀerent way. Local residents are highly aware of Fuego’s activity and knowledgeable of most volcanic hazards. However, since Fuego’s reac- tivation in 1999, the only eruptions they clearly remem- ber and identify are those that required a community- wide response which interrupted day-to-day life. This paper ﬁnds that the root causes of risk identiﬁed at other volcanoes (e.g. security concerns, maintaining livelihoods) are also present at Fuego. This paper also shows that a component of volcanic risk particular to Fuego is the disparity between communities on its west and east ﬂanks in terms of direct experience of previ- ous eruptions and communication with INSIVUMEH and CONRED. 1 Introduction Through reference to volcanic risk per- ception and evacuation literature, this paper conﬁrms that direct experience of eruptions is only one of many factors informing response to eruptive crisis at Fuego. For local residents, many competing factors (including existing socio-economic pressures and speciﬁc impacts associated with evacuation) create conditions that im- pede evacuation. At Fuego, the current policy places the majority of the responsibility for evacuation on lo- cals, ignoring the implications of these competing fac- tors. We argue that both the great variability in expe- riences of eruptive activity (both between authorities and locals, and between locals in diﬀerent communi- ties) and the social pressures aﬀecting locals have im- plications for volcanic risk at Fuego. These act par- allel and sometimes in opposition to any potential in- crease in local risk awareness and may have severe con- sequences for the success of future evacuations. Pyroclastic ﬂows are frequently produced by erup- tive activity of Volcán de Fuego [Naismith et al. 2019]. However, the estimated 15.1 million cubic meters of py- roclastic ﬂow material deposited in Las Lajas on 3rd June [Albino et al. 2020] was exceptionally large for a single eruption. It was more than double the aver- age volume of pyroclastic ﬂows registered since 1999 [Ferres and Escobar-Wolf 2018]. Nevertheless, erup- tions producing smaller pyroclastic ﬂow volumes have repeatedly triggered evacuation (e.g. September 2012, May 2017, November 2018). The high velocity of py- roclastic ﬂows means that evacuation is the only pro- cedure that eﬀectively prevents loss of life. However, prior evacuation is a complex and costly procedure that involves signiﬁcant resources from national authorities. The additional social and economic pressures aﬀecting members of communities like Los Lotes suggests that their compliance with evacuation may be even more diﬃcult. Yet, as this paper reports, authorities believe that locals have the capacity and responsibility to rec- ognize changes in volcanic activity and to self-evacuate when volcanic risk becomes intolerable. It is precisely this diﬀerence in opinion that continues to generate risk for the people living near Volcán de Fuego. This paper argues that understanding diﬀerences in experiences of previous eruptions and in volcanic risk tolerance be- tween locals and authorities is critical to eﬀective vol- canic risk mitigation (including evacuation). It does so through an exploration of the diverse coping strategies and important factors that inﬂuence peoples’ decision- making in the face of volcanic crisis. 2.1 Developments in volcanic risk literature Apparent underestimation of risk has been observed in commu- nities near Katla, Iceland [Jóhannesdóttir and Gísladót- tir 2010], and Ruapehu, New Zealand [Johnston et al. 1999]. This paper contributes to the debate by simul- taneously presenting perspectives of locals, volcanolo- gists, and oﬃcials in a single environment. tive in reducing volcanic risk to the most vulnerable [Gaillard and Mercer 2013]. Storytelling is an aspect of local knowledge that may be particularly important for volcanic risk mitiga- tion. Many disparate populations have used oral tra- dition to comprehend the trauma of a volcanic erup- tion [Cashman and Cronin 2008]. Although telling stories to understand volcanic eruption occurs in both pre-literature and literate societies, this method has largely been neglected in modern volcanic hazard miti- gation strategies [Cronin and Cashman 2016]. Fortu- nately, this is changing. Storytelling through digital ﬁlm has been used to sustain cultural memories of vol- canic eruptions, with inspirational outcomes [Hicks et al. 2017]. The power of storytelling for building re- silience to natural hazards in the Global South is in- creasingly recognized [Van Loon et al. 2020]. Story- telling for future disaster prevention is recognized in research disciplines beyond natural hazards, such as technical safety [Hayes 2018]. In this discipline, the responsibility for incorporating storytelling into disas- ter management strategies lies with professional safety managers [Hayes 2018]. Our paper includes stories told by local residents around Fuego to illustrate how story- telling may contain powerful truths about volcanic risk mitigation. The view that locals underestimate volcanic risk can lead to the mistaken belief that locals are deﬁcient in knowledge or have miscalculated their priorities. Communities aﬀected by natural hazards often develop “coping cultures” to adapt to their environment [e.g. Bankoﬀ2007]. Conversely, academic knowledge of risk is not authoritative, although this group is often cred- ited with an ‘accurate perception’ of the risk [Christie et al. 2015]. This is illustrated by a recent review of per- ception and social behaviour dealing with various natu- ral hazards including ﬂoods, earthquakes, and volcanic eruptions [Wachinger et al. 2013]. The review found no consistent inﬂuence of multiple personal factors (in- cluding age, gender, and level of education) on individ- ual risk perception, despite widespread academic be- lief that these factors are inﬂuential [Wachinger et al. 2013]. 2.1 Developments in volcanic risk literature Gilbert White’s 1942 work on ﬂood hazard in the US is an early study of human response to natural hazards. White [1942] argued that response is formed of ad- justment to one’s environment and implementing prac- tices to minimize loss. In the 1970s, concern that re- searchers were focussed on extreme natural events at the expense of root social causes of risk [e.g. Hewitt 1983] drove studies that focussed on vulnerability, or the existing socio-economic conditions that render an individual susceptible to disaster [Blaikie et al. 2014]. Loss of livelihood was identiﬁed as a key facet of vul- nerability, both in more economically developed coun- tries [e.g. Dibben 1999] and less developed [e.g. Lane et al. 2003]. At Volcán Tungurahua in Ecuador, ef- forts to create livelihood alternatives outside areas of high volcanic risk have evolved with adaptive forms of risk management. Here, local residents beneﬁt from the greater security of such alternative livelihoods and This paper presents ﬁndings from studies conducted in 2018 and 2019 that explicitly compare (1) how local people experience recent activity of Fuego; (2) how members of Guatemalan authorities, INSIVUMEH (Instituto Nacional de Sismología, Vulcanología, Meteo- rología e Hidrología) and CONRED (Coordinadora Na- cional para la Reducción de Desastres), experience the same and; (3) the potential implications of these diﬀer- ences for volcanic risk and its mitigation at Fuego. We deliberately chose to study direct experience of erup- tive activity as opposed to volcanic risk perception. The term “volcanic risk perception” implicitly assumes that volcanic risk is the only risk that local residents of a Presses universitaires de rasbourg Page 206 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica take collective decisions to temporarily evacuate, thus minimizing the disruptive eﬀects of forced evacuation [Armijos and Few 2015]. Local residents in other coun- tries have also engaged with temporary evacuation, for instance at Mt Merapi, Indonesia [Andreastuti et al. 2019]. This demonstrates that physical and social drivers of volcanic risk are not of equal priority in lo- cal peoples’ response to volcanic activity, as local peo- ple respond to socio-economic pressures before adjust- ing to hazards [Dibben 2008; Gaillard 2008]. From the perspective of some volcanologists and disaster risk of- ﬁcials, however, locals consistently appear to underes- timate volcanic risk [Donovan et al. 2014]. 2.1 Developments in volcanic risk literature The only deﬁnitive drivers of volcanic risk per- ception were (1) communication and trust between au- thorities and locals, and (2) direct previous experience of hazards. Furthermore, these drivers are themselves volatile: as volcanic eruptions rarely develop consis- tently, they will variably aﬀect surrounding popula- tions. Thus risk will vary even between neighbouring communities around the same volcano [Donovan et al. 2012]. An emerging area of volcanic risk research is showing how diﬀerent stakeholders focus on diﬀerent periods of eruptive activity. Dove [2008] explored local and gov- ernment perspectives of Mt Merapi’s activity to argue that not only ‘risk perception’ but even the conceptual- isation of risk itself varies. At Mt Merapi, locals contex- tualized changes in volcanic behaviour within their fo- cus on long periods of calm, while authorities focussed on Mt Merapi in times of crisis and thus separated it from daily life. While it is uncontroversial to state that a volcano demands more attention from authori- ties during an eruption, this diﬀerence in focus between stakeholders and consequent implications for volcanic risk and its mitigation has been little explored in other countries. In contrast to most complementary literature, this paper explicitly studies “direct experience (of previous eruptive activity)” as opposed to “volcanic risk percep- tion”. We believe that focussing on the latter isolates volcanic risk as the only risk people face in a volcani- cally active environment. Instead, “volcanic risk per- ception . . . is one form of risk perception balanced with other forms of perception including risks to livelihood and cultural heritage” [Gaillard 2008]. We hope that by focussing on how diﬀerent people experience erup- tions, we can contribute towards more complete under- standing of responses to volcanic activity of Fuego. Local knowledge has the advantage of coming from direct experience of activity [van Manen 2014]. Recent research shows the importance of including local peo- ples’ experience in managing volcanic risk, including in decision-making during crisis. Recognition of the ﬂaws in a traditional linear approach to communicat- ing risk [Donovan et al. 2014], successful integration of local and academic knowledge for participatory risk mitigation [Cronin et al. 2004], and proof of situations where locals have hazard knowledge equivalent to sci- entists [Gaillard 2008] have all highlighted the valuable contributions that local knowledge can make to under- standing volcanic risk. Conversely, a failure to integrate local and institutional knowledge often proves ineﬀec- 2.2 Factors aﬀecting evacuation Before making such a judgement, they should seek ﬁrst to understand tem- poral and spatial changes in social, political, and eco- nomic factors, as well as changes in volcanic hazard and responses to risk, all of which may encourage re- turn [Few et al. 2017]. Responses to risk are related to local peoples’ priorities, which themselves are often closely linked to the existing social and economic pres- sures that place individuals at risk. Pressures that en- courage evacuees to return while risk is still high can be summarized as “push” (e.g. poor shelter conditions) or “pull” (e.g. concern for livestock) factors [Barclay et al. 2019]. These pressures, as they express a desire to act against further impoverishment, may interfere with an otherwise apparently more logical desire to protect life. Although eruptions from Fuego have frequently trig- gered evacuation and disrupted the lives and liveli- hoods of local residents, few studies explore the link between volcanic activity and evacuation at this vol- cano. Early literature studied risk through the lens of human and agricultural vulnerability to volcanic haz- ards [Bonis and Salazar 1973]. Although Fuego had not caused signiﬁcant damage to surrounding populations, the authors presciently detail possible future losses, as- tutely observing that “the human problems faced by the geologist on the site not only will be repeated, but may be increased manifold in the future.” [Bonis and Salazar 1973, p84]. The next similar study, conducted among communities on Fuego’s south-west ﬂanks, dis- covered high awareness of volcanic risk coupled with widespread normalisation of Fuego’s behaviour among locals, and increased risk awareness with age [Graves 2007]. These ﬁndings were conﬁrmed by León-Ramírez Carné [2012], although this study found that awareness did not translate to a willingness to evacuate. The most recent study of volcanic risk at Fuego was conducted by Escobar-Wolf [2013] to consider situational and percep- tion variables in decision-making during volcanic cri- sis. This work consisted of a pilot study in 2009 involv- ing 38 individuals that informed a quantitative survey conducted in 2010 with 155 individuals in 8 commu- nities around Fuego [Escobar-Wolf 2013]. These stud- ies showed that locals frequently faced the decision of whether to evacuate or not from an eruption. Factors aﬀecting the decision to evacuate included fear of loot- ing and poor shelter conditions. 2.2 Factors aﬀecting evacuation An ex- ample of eﬀective collaboration between stakeholder groups comes from Tungurahua, where trust between local vigías (watchmen) and scientists permits eﬀective risk communication and evacuation processes [Armijos et al. 2017]. In this case, trust has evolved together with improved shelter conditions, evacuation routes, and re- sources together with possibilities for locals to maintain livelihoods inside and outside of the risk zone. in risk mitigation [Andreastuti et al. 2019]. An ex- ample of eﬀective collaboration between stakeholder groups comes from Tungurahua, where trust between local vigías (watchmen) and scientists permits eﬀective risk communication and evacuation processes [Armijos et al. 2017]. In this case, trust has evolved together with improved shelter conditions, evacuation routes, and re- sources together with possibilities for locals to maintain livelihoods inside and outside of the risk zone. to evacuate is often diﬃcult to make because all choices may have negative consequences. An individ- ual may decide to reduce personal risk when an erup- tion reaches its climax. This decision may involve evac- uation, particularly if the hazards associated with the eruption are impossible to manage from that individ- ual’s current situation. But what factors inﬂuence the decision to evacuate, and which are inconsequential? Recent literature suggests that risk awareness appears not to be a primary factor. While direct experience of hazards may increase risk perception, it does not neces- sarily lead to better preparedness [Johnston et al. 1999]. Wachinger et al. [2013] attribute this weak link be- tween risk awareness and preparedness to three poten- tial causes: ﬁrst, experience and motivation (e.g. an in- dividual understands the risk but perceives that bene- ﬁts outweigh risk); second, trust and responsibility (e.g. individual understands the risk but transfers responsi- bility elsewhere); third, personal ability (e.g. individual understands the risk but does not have resources to af- fect situation). Often the three causes can intersect. For example, at Montserrat, peoples’ premature return to the exclusion zone was driven by factors varying from economic hardship to a lack of shared thresholds of tol- erable risk [Barclay et al. 2008]. To outsiders, behaviour such as returning to an exclusion zone may seem illog- ical, as it increases threat to life. 2.2 Factors aﬀecting evacuation The most eﬀective action to mitigate risk to life from most volcanic hazards is evacuation. The decision Presses universitaires de rasbourg Page 207 Page 207 Naismith et al., 2020 Fireside tales: Volcán de Fuego to evacuate is often diﬃcult to make because all choices may have negative consequences. An individ- ual may decide to reduce personal risk when an erup- tion reaches its climax. This decision may involve evac- uation, particularly if the hazards associated with the eruption are impossible to manage from that individ- ual’s current situation. But what factors inﬂuence the decision to evacuate, and which are inconsequential? Recent literature suggests that risk awareness appears not to be a primary factor. While direct experience of hazards may increase risk perception, it does not neces- sarily lead to better preparedness [Johnston et al. 1999]. Wachinger et al. [2013] attribute this weak link be- tween risk awareness and preparedness to three poten- tial causes: ﬁrst, experience and motivation (e.g. an in- dividual understands the risk but perceives that bene- ﬁts outweigh risk); second, trust and responsibility (e.g. individual understands the risk but transfers responsi- bility elsewhere); third, personal ability (e.g. individual understands the risk but does not have resources to af- fect situation). Often the three causes can intersect. For example, at Montserrat, peoples’ premature return to the exclusion zone was driven by factors varying from economic hardship to a lack of shared thresholds of tol- erable risk [Barclay et al. 2008]. To outsiders, behaviour such as returning to an exclusion zone may seem illog- ical, as it increases threat to life. Before making such a judgement, they should seek ﬁrst to understand tem- poral and spatial changes in social, political, and eco- nomic factors, as well as changes in volcanic hazard and responses to risk, all of which may encourage re- turn [Few et al. 2017]. Responses to risk are related to local peoples’ priorities, which themselves are often closely linked to the existing social and economic pres- sures that place individuals at risk. Pressures that en- courage evacuees to return while risk is still high can be summarized as “push” (e.g. poor shelter conditions) or “pull” (e.g. concern for livestock) factors [Barclay et al. 2019]. These pressures, as they express a desire to act against further impoverishment, may interfere with an otherwise apparently more logical desire to protect life. in risk mitigation [Andreastuti et al. 2019]. 2.3 Providing context: eruptive history of Volcán de Fuego and associated stakeholder groups Volcanic risk mitigation at Fuego is managed through a network of institutions and the community. Table 1 deﬁnes acronyms of several institutions in this net- work. Figure 1 shows how these institutions com- municate between themselves and with the public. INSIVUMEH was founded in 1976 after the 1976 Guatemala earthquake. The institution is responsible for monitoring geophysical phenomena and advising the government and private sector on natural hazards. INSIVUMEH monitor volcanic activity through a geo- physical monitoring network managed from a central oﬃce in Guatemala City, aided by visual observations via two observatories in the communities of Panimaché Uno (OVFGO1) and Sangre de Cristo (OVFGO2)∗. The observatories are staﬀed by observers also resident in those communities. CONRED was founded in 1996 to reduce the impacts of disasters on Guatemalan society and to co-ordinate relief eﬀorts. CONRED carries out training in hazard awareness and preparatory actions in communities around Fuego. This is achieved primar- ily through a subsidiary oﬃce, UPV (Unidad para Pre- vención en Volcanes) in Antigua Guatemala which orga- nizes voluntary community groups known as COLRE- Des (Coordinadora Local para la Reducción de Desas- tres) in local communities. Communication between UPV and a COLRED is maintained via in-person vis- its, WhatsApp, and radio. Radio UPV is the network of community radio bases. As of April 2019, UPV had ra- dio bases installed in 28 communities and two private farms (ﬁncas) around Fuego. Each community radio base is housed by a radio operator who also belongs to that community’s COLRED. Participation of COLRED Figure 1: Schematic of diﬀerent institutions and com- munities aﬀected by activity of Volcán de Fuego show- ing pathways of communication between them. Dia- gram co-created by G. Chigna, W. Chigna, and A. Nai- smith. Figure 1: Schematic of diﬀerent institutions and com- munities aﬀected by activity of Volcán de Fuego show- ing pathways of communication between them. Dia- gram co-created by G. Chigna, W. Chigna, and A. Nai- smith. INSIVUMEH prepare bulletin reports that are pub- lished on their website (www.insivumeh.gob.gt) and on social media, and delivered to CONRED. Published re- ports are further disseminated through WhatsApp and CONRED’s radio network. The pathways of commu- nication between institutions and the public regard- ing eruptive activity of Fuego are illustrated in Fig- ure 1. 2.3 Providing context: eruptive history of Volcán de Fuego and associated stakeholder groups members in WhatsApp chats is highly variable due to inconsistent phone signal and costs of mobile data pre- venting local peoples’ access to the conversation. Within the last 450 years, Fuego has produced at least 60 summit eruptions that usually manifest as short- lived explosive events producing pyroclastic ﬂows and tephra fall [Rose et al. 2008]. A sub-Plinian eruptive sequence in 1974 was followed by two decades of qui- escence. Since its reactivation in May 1999, Fuego’s be- haviour has been characterized by persistent Strombo- lian activity and lava ﬂows. Fuego entered a new erup- tive regime in January 2015 associated with more fre- quent explosive paroxysmal eruptions that form part of a repeated cycle of activity [Naismith et al. 2019]. These recent explosive paroxysmal eruptions (“paroxysms”) consist of three stages where the central, most intense phase consists of 24–48 hours of intense activity involv- ing a sustained eruptive plume, continuous explosions, and occasional pyroclastic ﬂows. A full description of this activity is given in Naismith et al. [2019]. We use “paroxysm” in this paper to describe this activity that is characteristic of Fuego. Policy is a primary factor determining the success of evacuation from volcanic crisis. As of June 2020, a policy of auto-evacuación (self-evacuation) is being actively promoted by CONRED at Volcán de Fuego. This policy requires active involvement of a commu- nity in decision and responsive action. The deci- sion to evacuate a community from activity of Fuego should be made in agreement between a community’s COLRED and its local council or COCODE (Consejos Comunitarios de Desarrollo). Furthermore, in the self- evacuation policy, a community is supposed to man- age the initial stages of evacuation including gathering family members, moving to a pre-deﬁned safe point, and beginning to leave a community on foot or by ve- hicle if necessary. Communication would ideally be maintained with UPV throughout evacuation. Theo- retically, a community which initiates its own evac- uation would ﬁnd a secondary response co-ordinated by UPV involving temporary evacuation shelters and transport from the safe point to the shelters. In re- ality, several factors prevent this policy from work- ing as it should; these factors are explored in Sec- tion 5. A full description of the roles of CONRED and INSIVUMEH can be found in the National Response Plan on CONRED’s website (https://www.conred.gob. gt/site/Plan-Nacional-de-Respuesta). ∗After the events of 3rd June 2018, Sangre de Cristo was evacu- ated and its observer relocated to Panimaché Uno. By the end of my ﬁeldwork in April 2019, this observer was still in Panimaché Uno and working at OVFGO1. 2.2 Factors aﬀecting evacuation This study made an enormous contribution towards understanding of pop- ulation demographics around Fuego, including a sum- mary of the complex origin and development of ru- ral communities from plantations subsequently trans- formed by resettlement policies following the civil war years (1960–1996). While not explored in depth, such information contextualized the (un)willingness of local residents to evacuate their homes due to Fuego’s be- haviour. Additionally, this study determined quanti- tatively that willingness to evacuate from future erup- tive crisis would be inﬂuenced by the conditions under which evacuation took place. However, this work did not explicitly compare direct experiences of activity be- tween local residents and authorities. Our paper thus contributes to the debate on eruption and evacuation at Fuego through study of comparative experiences to un- derstand the conditions under which evacuation does or does not take place Local actions labelled as “illogical” may instead be driven by misunderstandings arising from poor com- munication between stakeholder groups that lead to disagreement regarding the nature of the risk and a disincentive to evacuate (e.g. in the reoccupation in the town of Baños near Volcán Tungurahua described by Lane et al. [2003]). In addition to breakdowns in communication, diﬃculties in evacuation management may occur because of peoples’ resistance to leaving an area of high risk [Mei et al. 2013], driven by factors such as place attachment and security fears. Mei et al. [2013] identiﬁed ﬁve interrelated factors negatively af- fecting successful evacuation, including uncertainty in forecasting eruption and resistance associated with eco- nomic factors. Conversely, good communication and a shared understanding between diﬀerent stakehold- ers may result in a shared commitment to participate Presses universitaires de rasbourg Page 208 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica 2.3 Providing context: eruptive history of Volcán de Fuego and associated stakeholder groups Presses universitaires de rasbourg 3 Methods In order to study experiences of previous eruptive ac- tivity and factors aﬀecting evacuation, we chose qual- itative data collection methods because of the ﬂex- ible and exploratory approach to research they af- forded, and because they allowed in-depth understand- ing of the motivations and interactions between diﬀer- ent stakeholder groups. This paper is a case study, a “detailed study of a single class of phenomena” [Fly- vbjerg 2006] which is valuable because of its depth of focus. Our study is informed by previous similar case studies such as Jóhannesdóttir and Gísladóttir [2010] and Stone et al. [2014]. We chose in-depth interviews as our main study method to facilitate discussion with participants about their experiences, similar to Jóhan- nesdóttir and Gísladóttir [2010]. As mentioned in Sec- tion 2.2, the most recent study capturing local expe- riences of eruptive activity at Volcán de Fuego was in 2007. Therefore, interviews provided an opportunity for better understanding of “[a] phenomenon about which little is yet known . . . to gain more in-depth information that may be diﬃcult to convey quantita- tively” [Hoepﬂ1997]. Our paper is published from the results of the lead author’s thesis, which includes theo- retical assumptions and philosophical stance involved in her research [Naismith 2020]. In total, 37 interviews were completed and audio- recorded, of which 32 were with local residents of the slopes of Volcán de Fuego and ﬁve with oﬃcials from INSIVUMEH and CONRED. Interviews were held in nine communities and a golf resort near the volcano (Figure 2). INSIVUMEH and CONRED staﬀwere in- terviewed in Antigua, Alotenango, and Guatemala City. To preserve conﬁdentiality these people are referred to as “Oﬃcial” when quoted. Table 2 gives demographic data of local residents who gave recorded interviews in 2019. Interviews were held in Spanish. Before an interview, permission to record it was sought from participant(s). The interviewer also explained to the participant(s) the interview’s purpose and how the data would be used. The interviewer clariﬁed to participants that any infor- mation they provided would be treated conﬁdentially. To redress some peoples’ wariness of appearing igno- rant when discussing a volcano with a volcanologist, the interviewer explicitly stated that she wished to hear stories of Fuego from those who had knowledge that she did not. Interviews involved the use of questionnaires involv- ing open questions categorized by theme†; neverthe- less, interviews could meander to explore interesting deviations. †The questionnaires used in this study have been included as Sup- plementary Material in English and Spanish. 2.3 Providing context: eruptive history of Volcán de Fuego and associated stakeholder groups While in theory the roles of INSIVUMEH and CONRED are distinct, there is no single piece of doc- umentation that clearly separates their responsibilities, Presses universitaires de rasbourg Presses universitaires de rasbourg Page 209 Fireside tales: Volcán de Fuego Naismith et al., 2020 2020 Table 1 – List of acronyms used in this text. All acronyms relate to institutions involved in managing volcanic risk at Fuego. Table 1 – List of acronyms used in this text. All acronyms relate to institutions involved in managing volcanic risk at Fuego. e 1 – List of acronyms used in this text. All acronyms relate to institutions involved in managing volca at Fuego. Acronym Institution ALFA INSIVUMEH’s centre of communications for information dissemination CTE CONRED’s centre of transmissions of emergencies COCODE Community Development Council COLRED Local Co-ordinator for Disaster Reduction DGAC Civil Aviation Authority OVFGO1 Observatory One of Volcán de Fuego OVFGO2 Observatory Two of Volcán de Fuego INSIVUMEH National Institute of Seismology, Volcanology, Meteorology, and Hydrology SE-CONRED Executive Secretary of CONRED UPV Union for Prevention in Volcanoes RADIO UPV Network of community radio bases managed by UPV proved by the University of Bristol Ethics Committee. All participants in the 2018 project were already known to the interviewer and were approached individually. People interviewed in 2019 were recruited through a mixture of purposive sampling [Palinkas et al. 2015] and ‘snowball’ sampling [Atkinson and Flint 2004]. In- terview data was supplemented by participant obser- vation, a non-intrusive data collection method involv- ing observing and participating in community activi- ties [Atkinson and Flint 2004]. and in practice the institutions’ eﬀorts frequently over- lap. This confusion has implications for personal and institutional responsibility for volcanic risk mitigation at Fuego. ‡NVivo software works in Spanish. 3 Methods Questionnaire design was guided by Bird We collected our ﬁndings in two studies: ﬁrst, a pi- lot study involving interviews with INSIVUMEH and CONRED staﬀin 2018; second, a study of experiences of local people, supplemented by those of INSIVUMEH and CONRED staﬀ, in 2019. Both projects were ap- Presses universitaires de rasbourg Page 210 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica Figure 2: Map of locations visited at Fuego in 2019. Blue circles represent communities on west ﬂanks: LC (La Colonia), LY (Los Yucales), MO (Morelia), PD (Panimaché Dos), PU (Panimaché Uno). Red diamonds represent communities further east: AL (Alotenango), CC (Chu-chu), CL (Ceilán), LR (La Reunión), SO (San Andrés Os- una). White squares represent INSIVUMEH observatories: OF1 (Panimaché Uno) and OF2 (Sangre de Cristo, permanently evacuated since June 2018). Large yellow triangle represents summit of Fuego (3763 m). Barrancas are labelled. Map data: Google Earth [2020]. Figure 2: Map of locations visited at Fuego in 2019. Blue circles represent communities on west ﬂanks: LC (La Colonia), LY (Los Yucales), MO (Morelia), PD (Panimaché Dos), PU (Panimaché Uno). Red diamonds represent communities further east: AL (Alotenango), CC (Chu-chu), CL (Ceilán), LR (La Reunión), SO (San Andrés Os- una). White squares represent INSIVUMEH observatories: OF1 (Panimaché Uno) and OF2 (Sangre de Cristo, permanently evacuated since June 2018). Large yellow triangle represents summit of Fuego (3763 m). Barrancas are labelled. Map data: Google Earth [2020]. Table 2 – Demographic data of local people who gave recorded interviews during 2019 study. People under 18 years old were not invited to participate due to additional ethical approval requirements. Initials for locations are same as for Figure 2. Note that number of participants in this table (n = 35) does not match number of recorded interviews with locals (n = 32) because three interviews contained two participants. Table 2 – Demographic data of local people who gave recorded interviews during 2019 study. People under 18 years old were not invited to participate due to additional ethical approval requirements. Initials for locations are same as for Figure 2. Note that number of participants in this table (n = 35) does not match number of recorded interviews with locals (n = 32) because three interviews contained two participants. Presses universitaires de rasbourg 3 Methods Characteristic Sex Age Male Female 18–29 30–39 40–49 50–59 60+ Count 14 21 5 6 10 8 6 Characteristic Location (Figure 2) AL CL CC LC LR LY MO PD PU SO Count 1 3 1 1 1 6 6 2 11 3 Characteristic Sex Age Male Female 18–29 30–39 40–49 50–59 60+ Count 14 21 5 6 10 8 6 Characteristic Location (Figure 2) AL CL CC LC LR LY MO PD PU SO Count 1 3 1 1 1 6 6 2 11 3 [2009]. Interviews were recorded with an Olympus WS-853 digital voice recorder. After interview, notes were added to a master Excel spreadsheet and inter- views manually transcribed into written Spanish. The transcripts were loaded into NVivo software‡ and sub- jected to thematic analysis to generate “codes”, or units of meaning in transcript data that had potential ana- lytical signiﬁcance. We analysed our material induc- tively, allowing for themes to emerge, and “building from particulars to general themes” [Cresswell 2014]. Our inductive approach to thematic analysis was in- formed by Pistrang and Barker [2012]. To distinguish the most signiﬁcant themes, we performed iterative Page 211 Fireside tales: Volcán de Fuego Naismith et al., 2020 coding, where inductively derived codes were applied to transcripts and ﬁeld notes, which were then re-read and analysed to generated additional codes [Ritchie 2003]. Codes included single words (e.g. ‘evacuación’ or ‘evacuation’) and phrases referencing local experiences (e.g. ‘Oct 1974’). This process allowed us to identify the themes that form the results of this paper, which include: a disparity between communities on Fuego’s east and west ﬂanks in terms of experience of previous eruptions and communication with INSIVUMEH and CONRED; eﬀects of experiences of 20th-century erup- tions on peoples’ responses towards recent activity of Fuego; and vast diﬀerences in the way locals and non- locals focussed attention on eruptive activity at Fuego. Quotations illustrating these themes have been trans- lated from Spanish and included in Section 4 along with the community location (west or east ﬂanks) to further validate the results. For readers of monochrome version of this paper, west communities are represented by circles in Figure 2, communities further east by dia- monds. but are seen by locals as periods of unrest only, describ- ing these periods as “thunder” or “rumbling”. Local people often connected descriptions of eruptive activ- ity with a sense of having become accustomed to the activity. Resident of Panimaché Dos (West): Sometimes we think it’s only going to make rum- bles again, or ash fall, or a bit of arena (sand) will fall, and then it’ll pass. So, that has allowed some of us to stay here. We are already used to the rum- bles and we have this already as an experience— that the rumbles don’t scare us any longer, nor do the ﬂares of ﬁre that appear each night. When we asked local people about previous eruptive activity of Fuego, they classiﬁed only the largest parox- ysms since 1999 as “eruptions”. Locals did not distin- guish between Fuego’s activity pre- and post-2015 and this was consistent across age and location. By contrast, INSIVUMEH and CONRED staﬀobserved a change in the frequency of paroxysms at Fuego similar to satellite observations: Oﬃcial 1: And then we arrived in 2015, when the eruptions were very frequent, almost every 20 days. 15, 20 days, there were eruptions with pyroclastic ﬂows, eﬀusive . . . in 2015 there were 15 eruptions, in 2016, 16 eruptions. In 2017 the number had dropped already, 9 in the year . . . including a very large eruption in May 2017 . . . So we are waiting, see, to see what will happen. Because just as it changed in 2015, there could be another change so the activity decreases again, to how it was before. But it’s still uncertain, how can we know what will happen? 4 Results Results are divided thematically into the following sub- sections: 4.1 How diﬀerent people experience volcanic activity; 4.1.1 Diﬀerences in focus; 4.1.2 Signiﬁcance of previous large (VEI ě 2) eruptions; 4.2 Factors af- fecting evacuation; 4.2.1 Trust between stakeholder groups; 4.2.2 Responsibility for decision-making and self-evacuation policy. Section 5 (Discussion) also fol- lows this structure. 3 Methods This was true of activity they experienced both before and after June 2018. The resident below speaks of activity after June: 4.1.1 Diﬀerences in focus We found substantial diﬀerences between how locals and authorities experience eruptive activity of Fuego. Locals observe that Fuego’s activity has increased since 1999; however, they report that they have neither ob- served nor experienced the sudden increase in explo- sive paroxysmal eruptions since 2015. When asked how many eruptions had occurred within the past ﬁve years, locals frequently estimated fewer than ten. Fur- thermore, when asked about a speciﬁc eruption, locals frequently replied that it did not happen. For example, in March 2019 a resident of Panimaché Uno stated that the only eruptions in the last two years were those of June and November 2018§. Nevertheless, local people see Fuego as persistently active, with ﬂares of activity set within periods of calm. Smaller eruptive events in 2015–2018 are reported as paroxysms by INSIVUMEH Figure 3 illustrates the diﬀerence between observa- tions by local residents and by satellite of Fuego’s ac- tivity in 1999–2019. Oﬃcial descriptions of activity closely mirror satellite observations, particularly since 2015, as quoted above. All eruptions illustrated by or- ange arrows resulted in widespread evacuations of local communities [Escobar-Wolf 2013]. Figure 3 contrasts these observations to show that both local and scientiﬁc observations are valid experiences of previous eruptive activity of Fuego. Our interpretations of this ﬁgure ap- pear in Section 5.1.1. 4.1.2 Signiﬁcance of previous large (VEI ě 2) erup- tions 4.1.2 Signiﬁcance of previous large (VEI ě 2) erup- tions How important is past experience in determining present views of risk at Fuego? Locals remember two major eruptions in the 20th century: one sometime in 1966, and the VEI-3 eruption of October 1974. The §INSIVUMEH recognized four explosive paroxysmal eruptions in this period: those beginning 31st January, 3rd June, 12th October, and 18th November 2018. No paroxysms occurred between January 2019 and June 2020. Presses universitaires de rasbourg Presses universitaires de rasbourg Page 212 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica Figure 3: Contrasting local experiences of eruptions with satellite remote sensing observations. Orange dashed lines represent local experiences of major eruptions 1999–2019, while blue timeseries represents satellite-based thermal timeseries data derived from MIROVA night-time values of Fuego activity 2000–2018 (adapted from Naismith et al. [2019]). Each one of the eruptions highlighted by locals is associated with largest MIROVA value for that eruption, where possible. Figure 3: Contrasting local experiences of eruptions with satellite remote sensing observations. 4.1.1 Diﬀerences in focus Orange dashed lines represent local experiences of major eruptions 1999–2019, while blue timeseries represents satellite-based thermal timeseries data derived from MIROVA night-time values of Fuego activity 2000–2018 (adapted from Naismith et al. [2019]). Each one of the eruptions highlighted by locals is associated with largest MIROVA value for that eruption, where possible. precise date of the 1966 eruption was uncertain, but all testimonies in 2019 agreed it was extremely powerful, denoting either Feb 7 It was not possible to meet anyone on the east of Fuego who had directly experienced either the 1966 or the 1974 eruption. However, several residents of Ceilán shared their grandparents’ experiences of 1974, talking of fall of rocks three inches in diameter, and a descending darkness that caused the blinded birds to ﬂy into trees in confusion. The fact that these sto- ries have been passed down through generations shows that large eruptions form a key part of eastern local residents’ experiences of Fuego’s activity. However, these experiences were of a short, intense eruption. Communities on Fuego’s eastern ﬂanks did not in gen- eral evacuate [Escobar-Wolf 2013]. Descriptions further west emphasise the signiﬁcant impacts that followed the 1974 eruption. Evacuations lasted weeks or even months, and as with 1966 the impacts of 1974 were ex- tensive death of native forest and wildlife and soil dam- age that has left a lasting legacy for those that draw a livelihood from agriculture in these communities. Even in 2019 the impacts of these eruptions remained: while locals on Fuego’s east ﬂanks stated, “the land here is very good land”, western locals agreed that to ﬁnd good land one had to dig. A resident of Morelia explained, “the volcano brings more poverty . . . one has to make a hole to [one metre] depth, to reach the good soil.” Resident of Morelia (West): So, we were playing marbles . . . when it made the rumble, the thunder, like a bombshell. And it rose like a mushroom, a rising bubble, and it spread. And at that moment the rays of the sun were ob- scured, and we remained in darkness. Only 15 minutes from the start of the explosion, 15 min- utes and we were blind . . . it was totally de- stroyed. The houses fell, the rivers ran dry . . . and the arena (sand) fell, fell, fell, and lots of lightning, a lot of friction in the atmosphere, the clouds. . . . and any trees which were still half alive were killed by the lightning. And all of the vegetation died. All of it. All locals interviewed in 2019 who had experienced the 1966 eruption also lived on the west ﬂanks of Fuego, in the communities of Morelia, Panimaché Uno, and Panimaché Dos (Figure 2). When these resi- dents described the eruption, they focussed on its long- lasting impacts, in particular its devastating eﬀect on agriculture and associated livelihoods: Experiences highlighted by western locals in 2019 are consistent with existing literature on the 1974 erup- tion, including isopach maps showing principal tephra deposition towards the SW (Figure 4). Locals use the term piedra (“stone”) to describe tephra fall in 1974. Size cannot be quantiﬁed precisely through descrip- tion, but the use of piedra alludes to the eruption’s severity. This is also true of the term arena (“sand”), which accumulated to a much greater depth in west- ern communities (Table 3: compare 3 inches (7.6 cm) in Ceilán with 50–100 cm in Los Yucales). These terms have been left untranslated in following quotations to Resident of Morelia (West): The sad thing is that my generation . . . until 2000, from 66, until 2000, lost any agriculture. Because now no-one could farm. No-one could harvest anything because nothing would grow. And we tried it, those of us who were already farmers, we tried. It didn’t work. And those who were born in a land which is not for farming . . . they did not learn anything. Here nothing was produced and the easiest thing is to leave. Presses universitaires de rasbourg Page 213 Naismith et al., 2020 Fireside tales: Volcán de Fuego Figure 4: Local descriptions of tephra fall from the 1974 eruption. Symbols describe local communi- ties: pink star (Ceilán), blue diamond (Los Yucales), green square (Panimaché Dos), purple cross (Pani- maché Uno), orange circle (San Andres Osuna). Large yellow circle is summit of Fuego. Arrows represent wind directions of the October 14th 1974 eruption and show predominant wind direction towards WSW. Lines represent altitudes: dashed (4000 m altitude at 215°), dashed (10000 m at 225°), dot-dash (7000 m at 285°). Wind directions from Figure 4 of Rose et al. [2008]. Lo- cals’ verbatim descriptions of tephra fall are found in Table 3. Map data: Google Earth [2020]. Table 3 – Community descriptions of tephra fall associ- ated with the October 1974 eruptive episode. Table 3 – Community descriptions of tephra fall associ- ated with the October 1974 eruptive episode. with the October 1974 eruptive episode. Resident of Morelia (West): Community Symbol Distance from summit (km, direction) Side Description Description (English) Ceilán star 8.5 S E “tres pulgadas de arena” “three inches of sand” Los Yucales diamond 12.3 SW W “llegó un metro de arena”; “cayó piedras de 3 cm”; “50 cm de arena” “a metre of sand came”; “rocks of 3 cm fell”; “50 cm of sand” Panimaché Dos square 8.2 SW W “dos metros de arena”; “cayó piedras de 10 cm”; “Casi topaba la arena, y había que entrar agachados para el corredor”; “había subido casi un metro de arena” “two metres of sand”; “10 cm rocks fell”; “the sand almost overcame us, and one had to enter the corridor bent over”; “almost a metre of sand had piled up” Panimaché Uno cross 7.8 SW W “piedras”; “arena” “rocks”; “sand” San Andres Osuna circle 11.90 S E “cayó arena”; “muchas piedras pero piedras pequeñas” “sand fell”; “rocks fell, but they were small rocks” P i it i d b Figure 4: Local descriptions of tephra fall from the 1974 eruption. Symbols describe local communi- ties: pink star (Ceilán), blue diamond (Los Yucales), green square (Panimaché Dos), purple cross (Pani- maché Uno), orange circle (San Andres Osuna). Large yellow circle is summit of Fuego. Arrows represent wind directions of the October 14th 1974 eruption and show predominant wind direction towards WSW. Lines represent altitudes: dashed (4000 m altitude at 215°), dashed (10000 m at 225°), dot-dash (7000 m at 285°). Wind directions from Figure 4 of Rose et al. [2008]. Lo- cals’ verbatim descriptions of tephra fall are found in Table 3. Map data: Google Earth [2020]. preserve accuracy. Many locals we interviewed in 2019 stated that the 3rd June 2018 eruption was the largest eruption of Fuego they had seen. We reported in Section 4.1.1 that many locals had become accustomed to activity both before and after June 2018. The eruption was described as singularly powerful by all on Fuego’s eastern ﬂanks, who had less prior experience of severe eruptive im- pacts: Resident of Ceilán (East): [The eruption] began in the morning. And it was completely cloudless in the morning. But the er- ror . . . was that this [activity] was already typ- ical for us, we didn’t give it much signiﬁcance. Because we were already used to seeing this . . . it’s common. It’s going to throw out lava again. It’s ordinary. It’s already throwing out ash. Ev- eryone who has their crops up there, towards the main square, ash is going to rain on them. Nor- mal. Resident of Ceilán (East): [The eruption] began in the morning. And it was completely cloudless in the morning. But the er- ror . . . was that this [activity] was already typ- ical for us, we didn’t give it much signiﬁcance. Because we were already used to seeing this . . . it’s common. It’s going to throw out lava again. It’s ordinary. It’s already throwing out ash. Ev- eryone who has their crops up there, towards the main square, ash is going to rain on them. Nor- mal. This resident then described the pyroclastic ﬂows generated by the eruption. They do not use the term “pyroclastic ﬂow”, but from its characteristics and Presses universitaires de rasbourg Page 214 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica 4.2 Factors aﬀecting evacuation through triangulation with other sources we can deﬁni- tively conclude that this was the hazard observed. The lack of a speciﬁc term to describe pyroclastic ﬂows was common among local people: Resident of Morelia (West): Resident of Panimaché Uno (West): Resident: So that is the fear that we have. But I tell people, “You are all afraid of what the volcano is doing now, this is nothing!”, I tell them. Resident: So that is the fear that we have. But I tell people, “You are all afraid of what the volcano is doing now, this is nothing!”, I tell them. Interviewer: And how did you know? What was the alert, the information? Interviewer: “This is nothing”? So – lately, this is nothing? Resident: Because it came from INSIVUMEH up there. They advised us. They gave the alarm that we had to leave, so we got together in order to leave. They alert us in case we have to leave. If they don’t advise anything, we do not leave. If the volcano is erupting, and they don’t advise us, we don’t leave. While they – we wait for their voice in order to leave. Resident: Yes. But these that are – these erup- tions are nothing compared to what I have lived, I tell them. Believe me that I, the day of that erup- tion, when everything went dark, a lot of arena fell. After that came another eruption. In which the arena was not ﬁne sand so much as ﬁne rocks. Older locals considered the 2018 eruption minor compared to 1966/1974, therefore did not evacuate in 2018. Another theme in older locals’ responses was in how they determined eruptive severity. Several stated that their response to an eruption is determined by the type of ash fall: In contrast to the good relationship between locals and INSIVUMEH/observers, western locals’ interac- tions with CONRED were less positive: Resident of Panimaché Uno (West): 4.2.1 Trust between stakeholder groups Communities around Fuego in 2019 had inconsistent levels of trust in authorities, with local opinion vary- ing between complete faith in authority and total self- reliance. This variability was most clearly expressed in a striking diﬀerence between communities on Fuego’s west and east ﬂanks in terms of their communica- tion with INSIVUMEH and CONRED. The presence of INSIVUMEH-owned observatories in the western communities of Panimaché Uno (OVFGO1) and San- gre de Cristo (OVFGO2, until June 2018) (Figure 2), and the regular visits of scientists from Guatemala City, builds locals’ conﬁdence in INSIVUMEH. Western lo- cals appreciate the presence of the observatories and INSIVUMEH’s role in issuing information during vol- canic crisis. These factors have generated so much trust between locals and observers that some locals believe INSIVUMEH carries the responsibility for prepared- ness for an eruptive crisis. Many western locals, when asked about their decision to evacuate or not, stated that they could not distinguish when Fuego was becom- ing dangerous, and that they relied on advice from “up there” (OVFGO1 in Panimaché Uno) to tell them when to evacuate: At eleven, twelve [o’clock], we began to see that it started to throw out more, as if it were smoke. We said, “Ah! What a humazón [cloud of smoke] it’s throwing out towards the other side!”. But we didn’t know the name of the material it was erupting. And from then on, from midday on- wards, we saw that it started to throw out mate- rial this side too. . . . But we saw that all this side of the volcano, towards the side with Los Lotes, we saw it as smoke, that is how it came. As if it were an enveloping ball. As if it were a ball of gas. It was buried like that. Many people in western communities agreed that June 2018 was one of the largest eruptions they had seen. However, a notable exception was among older people in these communities. To them, the 2018 erup- tion was smaller in scale and impact than those of 1966 and/or 1974. They remarked on the diﬀerence in ash fall: arena and piedra in 1966 and 1974, and ceniza (ash) in 2018. Here a resident explains the diﬀerent severities of eruption they experienced: Resident of Panimaché Uno (West): COLREDes on Fuego’s east ﬂanks include vol- unteer participants who perform similar roles to ob- servers at Panimaché Uno, acting as knowledgeable in- dividuals who inform others of changes in eruptive be- haviour and co-ordinating community response when Fuego is more active: Oﬃcial 4: They demand of us the answer, “Why did La Re- unión leave?” – and it is true, they evacuated in time. “And why not the communities?”. They demand of us the answer, “Why did La Re- unión leave?” – and it is true, they evacuated in time. “And why not the communities?”. All oﬃcials we interviewed spoke of the diﬃculty of successful evacuation of communities from eruptive crisis of Fuego. The oﬃcial quoted above spoke of their experience in encouraging evacuation at Los Lotes: I – as I said, I work here in the COLRED for my community. [That time] I noticed the wailing. People were saying, “And when will we leave? What is going to happen to us?”. And I said . . . I have an example. One day – about three days after the tragedy, I was here living through diﬃ- culty. Take it from me, I had never lived through something like that, here on the side of the – and a woman called, “[Resident’s name]! Over there you can see the lava! What should we do?”. And I – with my training, I told her, “Well, let us leave, over there are my associates.” When we were going through Los Lotes, we were going through warning them, with a siren and ev- erything. I did not see anyone come out. No one, no one. That is, no one expected that . . . maybe they imagined that . . . somehow, they could have escaped, if [the hazard] had come down the road. But they never imagined that it was going to come out from behind them. As explained in Section 2.3, the decision to evac- uate is co-ordinated between stakeholder groups at Fuego. Oﬃcial responsibility for calling an evacua- tion at Fuego requires approval either from a commu- nity’s COCODE or agreed between the community’s COCODE and COLRED. In a crisis, self-evacuation policy would have these groups convene and agree to temporarily evacuate their community. Self-evacuation also makes local residents responsible for communicat- ing their evacuation to, and requesting resources from, CONRED (via UPV). Resident of Panimaché Uno (West): Self-evacuation was being pro- moted both before and after 3rd June: While western locals’ trust in INSIVUMEH and its observers is evident, the level of conﬁdence of east- ern locals in their COLRED was less clear. The organ- isation and size of a COLRED is highly variable be- tween communities and subject to rapid, unexpected change: in the nine weeks’ duration of the 2019 study, at least two COLREDes were restructured. Motivation appeared higher among COLRED volunteers who host a community’s radio, many of whom stated that this role was a source of pride. However, the politics of owning a radio are complicated. In several commu- nities, ownership disputes have led to a breakdown in communication, either within the COLRED or between the COLRED and UPV. Happily, this diﬃcult situation appears to be improving. In 2020, communication be- tween 27 communities is regularly maintained, with reports of eruptive activity occurring four times daily (W. Chigna, personal communication). Participation in COLRED is a voluntary, unpaid role, and in 2019 many locals expressed that they were disincentivized to par- ticipate. This was partly due to a lack of recognition of the role from their community and partly due to incon- sistent support from UPV. In 2019 we found that sev- eral COLREDes had infrequent contact with UPV and received little support from this entity that supposedly acts as advisor to, and co-ordinator of, COLREDes. One local said eloquently: “Why should we continue, when no side supports us?”. 2018: Interviewer: If an eruption like 1974 were to occur again, what would be a ‘success’ for CONRED? Oﬃcial 2: Success would be that everyone evacu- ates without us coming. That people call us to say they have begun evacuating, so that we can co- ordinate transport, how to catch buses and reach the shelters. Not that they call us, “Look, what shall we do, come here and get us”, but that the decision is theirs and they evacuate. Interviewer: So that they make the decision – and communicate with you, and you are the ones that give support. Oﬃcial 2: Mm-hm. It is quite diﬃcult. Oﬃcial 2: Mm-hm. It is quite diﬃcult. 4.2.2 Responsibility for decision-making and self- evacuation policy 4.2.2 Responsibility for decision-making and self- evacuation policy 2019: Resident of Panimaché Uno (West): [CONRED] supposedly their work is to look out for the communities, give talks, so that disasters don’t happen. Right? But always . . . CONRED never come to give talks. You’ve seen now that CONRED said, “One week on this side of Fuego, one week on the other” – and see, four weeks and they haven’t come. They are always like this . . . they always fall short. Resident of Los Yucales (West): I stayed and set to thinking. I told them, “No”. I told them, “No. Here . . . don’t be afraid, my chil- dren, because this is ceniza. When arena falls, yes. When that happens, we must leave.” This resident elaborated that their decision to evacu- ate would be inﬂuenced by ash fall. If it was ﬁne ceniza, the eruption was not severe enough to require evacua- tion. When fall material resembled the arena that fell on western communities in 1966 and 1974, evacuation would be necessary. Residents of eastern communities were far less fa- miliar with INSIVUMEH. They had neither an obser- vatory nor familiar faces with which they could asso- ciate INSIVUMEH’s work. Instead they knew CONRED Presses universitaires de rasbourg Page 215 Fireside tales: Volcán de Fuego Naismith et al., 2020 found that understanding a community’s (lack of) will- ingness to evacuate was still a central challenge for of- ﬁcials at Fuego: found that understanding a community’s (lack of) will- ingness to evacuate was still a central challenge for of- ﬁcials at Fuego: through its sub-department UPV and their frequent visits to local COLREDes. Figure 1 gives an overview of how these groups communicate during activity of Fuego. COLREDes on Fuego’s east ﬂanks include vol- unteer participants who perform similar roles to ob- servers at Panimaché Uno, acting as knowledgeable in- dividuals who inform others of changes in eruptive be- haviour and co-ordinating community response when Fuego is more active: through its sub-department UPV and their frequent visits to local COLREDes. Figure 1 gives an overview of how these groups communicate during activity of Fuego. Oﬃcial 4: It is true that many people stated that they were newly aware of what Fuego could do to damage them: Resident of San Andres Osuna (East): Such a beautiful view, but today we know the ability, the capacity that this volcano has to de- stroy, don’t we? Resident of San Andres Osuna (East): La Reunión staﬀ1 (East): But that one of Febru- ary 2018, the ﬁrst of February I think it was, it scared me a lot when I left the oﬃce here and I jumped to see the volcano, and I saw that this cloud, as if it had come above us – then it scared me. But then nothing happened, right? It was only fear! Resident of San Andres Osuna (East): Such a beautiful view, but today we know the ability, the capacity that this volcano has to de- stroy, don’t we? However, other locals did not appear to have in- creased risk perception. The resident from Panimaché Dos quoted in Section 4.1.1 represented the views of many locals who in 2019 had readjusted to Fuego’s ac- tivity. Older western locals also expressed a diﬃdence towards the scale of the 2018 eruption that contrasts with the quoted statement of Oﬃcial 4 above. De- spite these instances of diﬀering perspectives, our cod- ing did show that, after the 3rd June eruption, many lo- cals would be prepared to evacuate when the situation demanded it: Interviewer: A-ha, yes. So you stayed here? Staﬀ: No. We evacuated, that time we evacuated . . . and then on the next day we returned. Staﬀ: No. We evacuated, that time we evacuated . . . and then on the next day we returned. Interviewer: Ah. And in previous years when there were eruptions, . . . was there a warning to evacuate or not? Staﬀ: Yes, they informed us. Yes. They always warned us that we had to evacuate. Resident of Panimaché Uno (West): We saw how [Los Lotes] was left. Well, we are afraid, having seen everything that happened and waiting here. Better to leave . . . leave and not wait any more. We saw how [Los Lotes] was left. Well, we are afraid, having seen everything that happened and waiting here. Better to leave . . . leave and not wait any more. From interviews we determined several factors facil- itating a culture of preventative evacuation at La Re- unión. Oﬃcial 4: We began this paper by highlighting the diﬀerent fates of San Miguel Los Lotes and La Reunión on 3rd June. This relates to a larger issue of the willingness of com- munities to evacuate from eruptive crisis. In 2019 we Ultimately, I think things have changed. But not everything that we would have liked. What I mean is that possibly there is better risk percep- tion. They know what it is, that there is a very Presses universitaires de rasbourg Presses universitaires de rasbourg Page 216 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica serious risk. But . . . but I don’t think that the speciﬁc ways have been focussed on. For example, in the communities, people wait and see, in any situation in which it is worth evacuating, we have to come and remove them. Everyone. We have to send trucks, we have to send vehicles, transporta- tion. Just like what happened on November 16. People did not leave on their own. ... a better example is, "My house is catching ﬁre. I don’t ex- pect ﬁreﬁghters to come to take me out. I’ll leave, I ...". That is, the ﬁrst response comes – from the individual level, family level, right, and as a com- munity, it is important. But I understand that it is also diﬃcult. separate interviews with locals. Locals employed this phrase to express uncertainty regarding evacuation. We were able to directly interview staﬀat La Re- unión golf resort. This gave us direct insight into fac- tors aﬀecting the resort’s decision to evacuate in June 2018. From interviews with La Reunión staﬀand IN- SIVUMEH/CONRED staﬀwho had worked with them, we found that La Reunión had implemented a cul- ture of preventative self-evacuation where guests and staﬀevacuated when an explosive paroxysmal erup- tion of Fuego developed. The crucial threshold for self- evacuation was the descent of pyroclastic ﬂows below a certain height in visible barrancas [conﬁrmed in Tobar 2018]. This threshold was not decided by La Reunión staﬀbut by CONRED, who had regular communication with La Reunión management. La Reunión had evacu- ated several times before June 2018, most notably in the paroxysmal eruption of 31st January–1st February 2018: While oﬃcials in 2019 stated that locals have better risk perception, this belief was not fully borne out by our interviews with locas. Oﬃcial 4: Its position as a private resort allows manage- ment to force guests’ departure; it enjoys good commu- nication with both INSIVUMEH and CONRED; and re- sources such as 4ˆ4 vehicles allow rapid escape. With all these advantages, staﬀargued that evacuation was still not simple: We found strong consensus that it was better to leave immediately when Fuego began displaying signs of un- rest rather than waiting for further increase in activity. Consensus was particularly strong in Panimaché Uno, where nine of the 11 people we interviewed stated this in some form. We suggest this relates to the strong rela- tionship between locals and observers in OVFGO1 (see Section 4.2.1). However, while locals outside of Pani- maché Uno also expressed desire to evacuate promptly, this desire was consistently tied to concerns about risks involved in evacuation. Risks mentioned included leav- ing houses unattended (thus vulnerable to looting), livestock welfare, and evacuation shelter conditions— or whether such shelters existed. A particular phrase, “a donde ir” (“where to go”) was recorded in nine 5.1.1 Diﬀerences in focus In Section 4.1.1 we presented evidence for observa- tional diﬀerences of Fuego’s activity by local residents and by satellite (Figure 3). Observations by oﬃcials closely matched satellite data. What causes these dif- ferences? One reason is that oﬃcials have information that is unavailable to locals. INSIVUMEH’s geophysi- cal monitoring network detects changes in seismic ac- tivity at Fuego. These can be correlated with other in- formation to which INSIVUMEH has access, like satel- lite imagery such as NASA’s MODIS and LandSat plat- forms. While INSIVUMEH and CONRED staﬀ’s experi- ences of recent activity at Fuego correlate with satellite observations, this is likely because of coincidence with visual observations of Fuego’s summit eruptive activ- ity from OVFGO1 [Lyons et al. 2010; Naismith et al. 2019]. Visual observations from OVFGO1 (Figure 2) generally match satellite data because thermal anoma- lies detectable by satellite are also visually distinctive (e.g. incandescent ﬁre fountaining). INSIVUMEH com- bine visual observations with other monitoring tools because frequent cloud cover frustrates observations of Fuego from surrounding communities. We reason that the fewer sources of information on Fuego that locals can access partly accounts for the diﬀerence in experi- ences between locals and oﬃcials. We advocate here for a recognition of diﬀerent points of view, rather than a resolution. Local people should not have to adapt to resolve this diﬀerence in focus on Fuego’s activity (and consequently of its risk). The view of some scientists that local knowledge of vol- canic risk as insuﬃcient [e.g. Donovan et al. 2014] does not acknowledge that diﬀerent points of view can co- exist and be valid. As Dove [2008, p 336] states: “au- thority views of risk are themselves inevitably socially constructed and thus contingent in value and eﬃcacy”. Diﬀerent views of volcanic risk stemming from diﬀer- ent focuses on activity can explain conﬂicting responses to past eruptive crises at Fuego, and if unresolved these conﬂicts may be repeated in future. For example, dur- ing the November 2018 paroxysm, despite CONRED issuing a red alert and sending vehicles to aid evac- uation, many locals refused to leave their homes [RT 2018]. In this case, risk was viewed as suﬃciently high to require evacuation by authorities but not by locals. Accordingly, authorities considered that locals were underestimating risk, and conﬂict arose as the groups diﬀered in their response (or lack of) to the perceived risk. La Reunión staﬀ1 (East): What is complicated is that not everyone leaves. Some leave immediately, others we have to insist to them that they leave . . . they didn’t leave, be- cause always, “Why should we leave if nothing is going to happen”, right? We discuss these ﬁndings and their implications in Sec- tion 5.2.2. Presses universitaires de rasbourg Page 217 Fireside tales: Volcán de Fuego Naismith et al., 2020 5 Discussion did, and therefore did not require communities to take responsive action. This response is the critical factor which preserves an eruption in local memory at Fuego. Notably, the isolated events that locals remember (e.g. September 2012, June 2018) were all associated with disruption of daily community life. This is substanti- ated by Figure 3: the orange arrows indicate both erup- tions that caused evacuations, and eruptions that locals frequently described. Figure 3: shows the great dif- ference in local and scientiﬁc foci on Fuego’s activity. Crucial is that both views are partial: years of eruptive activity are not acknowledged by locals, while several eruptions that locals consider signiﬁcant barely regis- ter in satellite data. For example, the eruption of 13th September 2012 provoked the evacuation of thousands of people but produced a peak thermal radiance of 1612.27 MW, smaller than many events occurring since onset of the new eruptive regime in 2015 [Naismith et al. 2019]. By contrasting local and scientiﬁc obser- vations in Figure 3, we are not trying to undermine the utility of scientiﬁc observations for understanding eruptive activity. Instead, we argue that both sets of ob- servations are valid and need to be recognized in terms of what matters to diﬀerent stakeholder groups around Fuego. Appreciating the diﬀerence between locals’ and scientists’ views is critical for eﬀective future risk mit- igation at Fuego, if these groups wish to collaborate to protect life and assets from eruptive activity. Through- out volcanic risk literature there is evidence that shared views of risk between stakeholders contributes to more eﬀective risk mitigation procedures, for instance at Vol- cán Tungurahua in Ecuador [Armijos et al. 2017], at Sinabung and Kelud volcanoes in Indonesia [Andreas- tuti et al. 2019], and at Tristan de Cunha [Hicks et al. 2014]. We have divided the discussion into subsections that follow the structure of Section 4. Findings from Sec- tion 4 are discussed and we interpret them with refer- ence to literature presented in Section 2. We also in- clude the implications of our interpretations. 5.1 How diﬀerent people experience volcanic activity Presses universitaires de rasbourg 5.1.1 Diﬀerences in focus An inverse diﬀerence in views of risk A second reason for this diﬀerence could be the ‘nor- malisation bias’ encountered in other literature [e.g. Haynes et al. 2008]. Normalisation bias may impel lo- cals to expect only the experienced, desensitising them to changing risks; this has previously been documented at Fuego [Graves 2007]. Our ﬁndings from interviews with locals in 2019 suggest many were accustomed to Fuego’s eruptive phenomena (Section 4.1.1) Excepting uncommonly large eruptions such as September 2012, relative changes in Fuego’s activity do not register with locals. In contrast, INSIVUMEH staﬀand CONRED staﬀface the volcano rather than use it as their back- drop. Their point of observation is diﬀerent. Although we conclude that locals do normalise Fuego’s activity, we argue that this only partly resolves the diﬀerence between locals’ and authorities’ experi- ences of recent eruptive activity. Quotations from lo- cals in Section 4.1.2 are rich descriptions that show how vividly they remember previous eruptions. We propose an alternative explanation: while any eruption since 1999 could have evolved into a larger event, few Presses universitaires de rasbourg Page 218 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica has also recently occurred at Fuego, when on 6th June 2018 residents of communities in Escuintla observed behavioural changes of Fuego that they interpreted as a reactivation [G. Chigna, personal communication]. Many people evacuated and requested help from au- thorities that was not given. On this occasion, locals recognized a risk from Fuego that authorities did not acknowledge. Locals then desired a reaction that au- thorities did not feel responsible for. On both occa- sions, diﬀerences between stakeholder judgements of risk caused diﬀerent responses and resulted in conﬂict. We believe that acknowledging multiple points of view would avoid such conﬂicts in the event of future erup- tive crisis at Fuego. ing this change with diﬀerent eruptive severities indi- cates some understanding that fall of larger tephra is associated with a more energetic eruption producing a strong eruptive plume capable of more widespread tephra dispersal and representing greater hazard. This ﬁnding demonstrates an impressive understanding of volcanic hazard. Nevertheless, we assert that evalua- tion of tephra size to decide response to Fuego’s erup- tions has potentially dangerous implications. Firstly, making the decision to evacuate when arena falls may already be too late. Secondly, this evaluation does not acknowledge other independently-occurring hazards, such as lahar or pyroclastic ﬂow, that could put locals at risk. 5.1.2 Signiﬁcance of previous large (VEI ě 2) erup- tions In Section 4.1.2, we reported our ﬁndings on local ex- periences of eruptions in 1966, 1974, and 2018. Older locals in western communities describe the long legacy of 1966 and 1974 eruptions on agriculture. They de- scribed the 2018 eruption as lesser. In contrast, people in eastern communities described the 2018 eruption as larger than any they or their predecessors had experi- enced. Some of this disparity is due to fewer older in- terviewees in the east. However, the disparity seems genuine. Locals in Ceilán shared parents’ stories of an acute event without long-lasting impacts. Diﬀerences in impacts are clear in Figure 4 and Table 3 (compare “three inches of sand” that fell in Ceilán with “two me- tres of sand” in Panimaché Dos). Given that commu- nities around Fuego are separated by only a few kilo- metres, Figure 4 illustrates that previous direct expe- riences of volcanic activity can vary dramatically be- tween even adjacent communities. This diﬀerence mat- ters in inﬂuencing the decision to evacuate [Wachinger et al. 2013]. We also found that older western locals frequently compared 1966/1974 with 2018 in terms of severity (Section 4.1.2). That older locals experienced the 2018 eruption as relatively minor is key in evalu- ating volcanic risk at Fuego because of how these lo- cals interpret the risk of an eruption, and subsequently their vulnerability and response. We conclude from their failure to evacuate in 2018 that locals would not do so until a future eruption is comparable to 1974. This has problematic implications. 5.1.1 Diﬀerences in focus In 2019, people who reported distinguishing between arena and ceniza in determining response to eruption were located in Panimaché Uno, Los Yucales, and Ceilán (Figure 2). These communities all have evacuation routes that cross barrancas which could be cut oﬀby lahars. Panimaché Uno also lies close to Bar- ranca Taniluyá, a principal route for pyroclastic ﬂow descent. As such, using only arena/ceniza to determine response to eruption may provide an incomplete assess- ment of hazards. Ameliorating volcanic hazard assess- ment is not solely the responsibility of local people, but should depend also on better information and its eﬀec- tive communication from INSIVUMEH and CONRED. 5.1.2 Signiﬁcance of previous large (VEI ě 2) erup- tions Presses universitaires de rasbourg 5.2.1 Trust between stakeholder groups Placing such trust in INSIVUMEH has ambiguous connotations. It may disincentivize lo- cals to take personal measures to mitigate risk. On the other hand, it is essential in ensuring advice from IN- SIVUMEH will be considered during a crisis. This ad- vice may decisively change local peoples’ response to volcanic risk, either by the content of the advice itself, or more likely because the high levels of conﬁdence in observers automatically renders the advice impor- tant, as seen with the vigías of Tungurahua [Stone et al. 2014]. However, this trust is complicated. Whether it is the vigía network and IG-EPN of Tungurahua, or the OVFGO1 observers and INSIVUMEH in western communities of Fuego, scientists do not have a man- date to call an evacuation. As such, if INSIVUMEH ad- vises locals to evacuate due to activity of Fuego, it is not in an oﬃcial context. This occurred in September 2012, when a widespread evacuation of communities was led by the observers of OFVGO1. Unoﬃcially, sev- eral people in 2019 stated that this event was the im- petus for CONRED’s current policy of self-evacuation. However, it cannot be assumed that in future eruptive crises, other communities could carry out such an evac- uation: for example, no communities other than Pani- maché Uno have trained observers already resident. In- deed, the reliance on INSIVUMEH and observers for an informal evacuation alarm may lead to a conﬂict be- tween local’s and CONRED’s judgements of risk and subsequent response (as discussed in Section 5.1.1). In addition, the lack of conﬁdence in CONRED displayed by some western locals may dissuade them from com- plying with CONRED’s advice during future eruptive crises. There are many other conditions that compli- cate the relationship and trust between authorities and residents of western communities, and more research is required to understand the situation better. However, our ﬁndings indicate that in the speciﬁc case of Pani- maché Uno, recognition of diﬀerent views and a closer relationship between residents and authorities has had a positive inﬂuence on local residents’ willingness to evacuate. Trust in authorities among residents of eastern com- munities was less clear than in the west From inter- of COLREDes themselves expressed a lack of support from authorities. There are some parallels to be drawn between COLREDes and the vigía network at Tungu- rahua. 5.2.1 Trust between stakeholder groups The latter has ﬂourished in part because vol- unteers feel they are playing a critical role in provid- ing early warning of volcanic activity and contribut- ing to risk mitigation [Stone et al. 2014]. Contrary to the “trust and responsibility” clause cited in the previ- ous paragraph, a lack of trust in authorities has multi- ple eﬀects on local peoples’ perspective and behaviour, including an increased tendency to underestimate risk and a reduction in willingness to take preparatory ac- tions against risk from natural hazards [Wachinger et al. 2013]. Therefore, we promote support of existing COLREDes in eastern communities at Fuego, as in the absence of a permanent observatory a voluntary net- work may be the best line of communication between locals and authorities. We found interesting dynamics in being a COLRED radio operator: on the one hand, it was an important source of pride, but on the other, it could inspire envy. In the comparable vigía commu- nication network at Tungurahua, a radio is used as a shared resource that provides an important informal communication pathway [Stone et al. 2014]. A practical way to support COLREDes could be to maintain the ra- dio network with good batteries, ongoing training, and encouraging commitment in participating in reports. rahua between locals, vigías, and scientists from Quito [Few et al. 2017; Stone et al. 2014]. We found at Fuego that western locals were so trusting that they were willing to outsource decisions on evacuation to INSIVUMEH and the observers of OVFGO1, similar to the “trust and responsibility” cause cited by Wachinger et al. [2013] for the disconnect between risk perception and preparedness. Placing such trust in INSIVUMEH has ambiguous connotations. It may disincentivize lo- cals to take personal measures to mitigate risk. On the other hand, it is essential in ensuring advice from IN- SIVUMEH will be considered during a crisis. This ad- vice may decisively change local peoples’ response to volcanic risk, either by the content of the advice itself, or more likely because the high levels of conﬁdence in observers automatically renders the advice impor- tant, as seen with the vigías of Tungurahua [Stone et al. 2014]. However, this trust is complicated. Whether it is the vigía network and IG-EPN of Tungurahua, or the OVFGO1 observers and INSIVUMEH in western communities of Fuego, scientists do not have a man- date to call an evacuation. 5.2.1 Trust between stakeholder groups As such, if INSIVUMEH ad- vises locals to evacuate due to activity of Fuego, it is not in an oﬃcial context. This occurred in September 2012, when a widespread evacuation of communities was led by the observers of OFVGO1. Unoﬃcially, sev- eral people in 2019 stated that this event was the im- petus for CONRED’s current policy of self-evacuation. However, it cannot be assumed that in future eruptive crises, other communities could carry out such an evac- uation: for example, no communities other than Pani- maché Uno have trained observers already resident. In- deed, the reliance on INSIVUMEH and observers for an informal evacuation alarm may lead to a conﬂict be- tween local’s and CONRED’s judgements of risk and subsequent response (as discussed in Section 5.1.1). In addition, the lack of conﬁdence in CONRED displayed by some western locals may dissuade them from com- plying with CONRED’s advice during future eruptive crises. There are many other conditions that compli- cate the relationship and trust between authorities and residents of western communities, and more research is required to understand the situation better. However, our ﬁndings indicate that in the speciﬁc case of Pani- maché Uno, recognition of diﬀerent views and a closer relationship between residents and authorities has had a positive inﬂuence on local residents’ willingness to evacuate. 5.2.1 Trust between stakeholder groups Evacuation is a powerful tool when seeking to protect life during a volcanic eruption. However, it is compli- cated and often costly, thus often invoked only in crisis. The diﬀerent fates of people in La Reunión and in Los Lotes on 3rd June 2018 has been heavily scrutinized, often with the perception that wealth, and consequent access to resources and information, was the author of their destinies. This is debatable because the preven- tative evacuation that protected lives at La Reunión on 3rd June also has precedence among rural communities with many fewer resources. Sangre de Cristo, for in- stance, has developed a practice of preventative evacu- ation since the eruption of 7th August 2007, similar to the self-evacuations at Tungurahua reported in Armi- jos et al. [2017]. The ability to evacuate or not must therefore be determined by something more sophisti- cated than simply being rich, just as vulnerability is often correlated to, but not synonymous with, poverty [Blaikie et al. 2014]. Local residents’ descriptions of eruptive activity have been documented at Volcán Tungurahua, and the obser- vations interpreted in terms of diﬀerent eruptive pro- cesses [Armijos et al. 2017]. At Volcán de Fuego, we found a similar phenomenon where residents use qual- itative assessment of tephra size as a method of eval- uating eruptive severity. Furthermore, they include this assessment in their decision to evacuate. This speaks of a remarkable awareness of volcanic hazard among locals: their identiﬁcation of changes in tephra fall is a clever form of monitoring. Furthermore, link- The importance of trust in authorities is crucial in taking preventative action during crisis [Wachinger et al. 2013]. We found that local trust in authori- ties around Fuego in 2019 was inconsistent (see Sec- tion 4.1.2). Western locals had good trust in IN- SIVUMEH due to the presence of community-based observatories and regular visits from scientists in Guatemala City. A similar conﬁdence occurs at Tungu- Presses universitaires de rasbourg Page 219 Fireside tales: Volcán de Fuego Naismith et al., 2020 rahua between locals, vigías, and scientists from Quito [Few et al. 2017; Stone et al. 2014]. We found at Fuego that western locals were so trusting that they were willing to outsource decisions on evacuation to INSIVUMEH and the observers of OVFGO1, similar to the “trust and responsibility” cause cited by Wachinger et al. [2013] for the disconnect between risk perception and preparedness. 5.2.2 Responsibility for decision-making and self- evacuation policy The diﬀerences between local and author- ity views of responsibility for decision-making, and im- plications for success of self-evacuation policy, are dis- cussed in more depth below. planning and community co-operation are appropriate, we argue these eﬀorts should be in support of, and not in place of, their own greater capacity and oﬃcial re- sponsibility. Authors such as Haynes et al. [2008] have shown that local residents face signiﬁcant barriers responding to volcanic activity that are deeply connected to the root causes of risk. Better evacuation systems and methods of communication may serve to improve this. However, these improvements should be primarily the responsi- bility of authorities, who have more resources available than local residents to achieve this. The self-evacuation policy contains assumptions about local knowledge, resources, and experiences that conﬂict with quotations from local people. Assump- tions of knowledge are implicit in quotations by oﬃ- cials in 2018 and 2019 (see Section 4.2.2 and Oﬃcial 1, end of this section). Oﬃcial 1 considers that recent di- rect experience of eruptive activity has increased local residents’ risk awareness (supported by literature, e.g. Johnston et al. [1999]), but also that this awareness will motivate future willingness to evacuate. However, Sec- tion 4.2.1 showed that only nine months after 3rd June, many local people appear again to normalise Fuego’s behaviour. This suggests that in future eruptions local residents might not have the knowledge to distinguish normal eruptive activity from an eruptive crisis that re- quires them to evacuate. The oﬃcial quoted in 2018 references local peoples’ uncertainty in the face of vol- canic activity (“What shall we do?”) and presents this as an undesirable situation illustrating failure of the self-evacuation policy. This may be true, but it does not follow that the failure is due to local uncertainty. One must ask why local people don’t have access to infor- mation about Fuego. Escobar-Wolf [2013] showed that many local people (66 % of respondents in 2010, 101 people) considered themselves insuﬃciently informed of Fuego, agreeing that “you don’t know when the vol- cano will become dangerous, and you need someone else with more knowledge to tell you when you should evacuate”. Aside from a distinction between arena and ceniza, local people in 2019 were similarly uncertain, despite their recent experiences of June and November 2018. 5.2.2 Responsibility for decision-making and self- evacuation policy For example, the Ceilán resident who described pyroclastic ﬂows in Section 4.1.2 was a member of the community COLRED, but had never seen pyroclastic ﬂows before 3rd June. This resident may not feel conﬁ- dent to advise others in the community about appropri- ate action to take – which for authorities would be evac- uation. While knowledge gained by direct experience of natural hazards is important, combining such knowl- edge with training and access to scientiﬁc information is vital to developing local ability to cope with persis- tent eruptive activity and facility in decision-making regarding evacuation [Few et al. 2017; Mei et al. 2013]. Our interpretations of the self-evacuation policy’s as- sumptions about knowledge and resources are sup- ported by analysis of a CONRED infographic promot- ing self-evacuation (Figure 5). First, Step 1 assumes that every family can create a Family Response Plan, which requires internet access, literacy (both written and computational), and value judgements of priori- ties that themselves require knowledge. Many locals around Fuego lack some or all of these requirements and would not be able to create such a plan. Steps 2–3 assume that locals will (a) be able to access infor- mation through media such as radio; (b) be able to in- terpret the information received and incorporate it into their decision-making. However, there have been sev- eral major eruptions of Fuego in which communication pathways were damaged, including 3rd June 2018. Fur- thermore, the lack of a shared vocabulary between lo- cals and oﬃcials at Fuego (e.g. locals do not have a speciﬁc word to describe pyroclastic ﬂows – see Sec- tion 4.1.2) suggests that these groups do not speak the same language, and therefore interpretation of any re- ceived information will be diﬃcult. Development of a shared vocabulary is critically important in eﬀective risk mitigation at analogue volcanoes [Armijos et al. 2017]. It appears that a shared understanding of tol- erable risk does not exist at Fuego. Moreover, whose responsibility is it that this information is (a) accessible to, and (b) interpretable by locals? It is certainly not that of locals themselves. Finally, Step 4 uses ambigu- ous language (“begin self-evacuation”) that implicitly assumes both a willingness and ability to evacuate. 5.2.2 Responsibility for decision-making and self- evacuation policy 5.2.2 Responsibility for decision-making and self- evacuation policy Supporting COLREDes is also important for the success of the self-evacuation policy at Fuego in future erup- tive crises. Set-up of a community COLRED is not triv- ial. UPV is responsible for volunteer recruitment and training. However, once founded, a COLRED is sup- posed to act as a separate entity. Members of a com- munity COLRED are supposed to act as a source of in- formation for the community and advise the COCODE on whether to evacuate during volcanic crisis. Based on this structure of responsibilities and the quotations from oﬃcials in Section 4.2.2, the presence of COLRE- Des appears to be an attempt to transfer more responsi- bility for volcanic risk preparedness to local people by CONRED. However, it is uncertain whether a newly- founded COLRED and its members have suﬃcient knowledge and training to perform equivalent work to INSIVUMEH in advising a COCODE in times of cri- sis. While Section 4.2.1 illustrates that locals have good knowledge of Fuego’s eruptive hazards drawn from di- rect experience, this does not translate to an ability to distinguish when such hazards have reached a critical level requiring evacuation. Furthermore, although we interpret the creation of COLREDes as a transfer of re- sponsibility for decision-making from CONRED to lo- cal residents, it is unclear whether residents accept this responsibility. In fact, there is no clear evidence for a lessening in locals’ belief that CONRED are responsible Trust in authorities among residents of eastern com- munities was less clear than in the west. From inter- views with local residents who are voluntary COLRED members, we found these residents considered their COLRED a trusted source of information in the com- munity (similar to INSIVUMEH’s observers at OVFGO1 in Panimaché Uno). However, while the trust between western locals and observers was ratiﬁed by locals’ words, we could not conﬁrm the extent of trust between eastern locals and COLRED members. Some members Presses universitaires de rasbourg Page 220 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica for decision-making and initial response during erup- tive crisis. The diﬀerences between local and author- ity views of responsibility for decision-making, and im- plications for success of self-evacuation policy, are dis- cussed in more depth below. for decision-making and initial response during erup- tive crisis. 5.2.2 Responsibility for decision-making and self- evacuation policy We conclude that this judgement among older people means they would decide to evacuate only when it is too late to do so. their environments and make adjustments to minimize loss. Supported by more recent literature as well as this study’s ﬁndings, we suggest this approach has certain ﬂaws. We found that local residents describe their di- rect experiences of previous eruptive activity in a man- ner that shows that what they see is diﬀerent from what oﬃcials see (Figure 3). Diﬀerences in how locals and oﬃcials see eruptive activity at Fuego may be through a diﬀerence in recognition of activity as an eruption or not (Section 4.1.1), or by comparison of an eruption to a previous, larger eruption (Section 4.1.2). Even if locals and authorities recognize the same eruptive activity, it is uncertain how inﬂuential this experience is in locals’ decision to evacuate. In 2019, some locals stated that in a future eruptive crisis they would self-evacuate when they considered volcanic risk “high enough”. This ini- tially suggests that self-evacuation is a viable policy. However, when asked to describe a speciﬁc situation that would require them to evacuate, locals could not do so (with the exception of arena/ceniza). Locals’ lack of a deﬁnite threshold for volcanic risk tolerance has previously been observed both at Fuego [León-Ramírez Carné 2012] and at analogous volcanoes like Mt Mer- api [Mei et al. 2013]. At Mt Merapi, locals who were less familiar with oﬃcial disaster risk reduction strate- gies were more reluctant to comply with evacuation or- ders, and several from under-educated areas were trag- ically killed in 2010 [Mei et al. 2013]. At Fuego, a fur- ther complication regarding direct experience of activ- ity and its inﬂuence on self-evacuation occurs with the diﬀerence between communities on the east and west ﬂanks that has until now not been acknowledged. In multiple studies of evacuation from natural hazards, the only two signiﬁcant predictors of preparedness to act were (i) severity of previous direct experience of hazards, and (ii) trust and communication with outside stakeholders [Wachinger et al. 2013]. These predictors are also the two most explicit diﬀerences between east- ern and western communities at Fuego (through, re- spectively, experiences of 20th-century eruptions and Other than the three main assumptions implicit in the self-evacuation policy, other factors such as com- munication may hinder the success of self-evacuation in future eruptive crisis. 5.2.2 Responsibility for decision-making and self- evacuation policy At Fuego the ability to self-evacuate is not supported by evidence: for example, evacuation is greatly facilitated by access to vehicles, but most people do not have ac- cess to transport and would have to evacuate on foot. This is slow and dangerous and would be greatly com- plicated by factors such as an eruption at night or hav- ing to carry young children or elderly relatives. The self-evacuation policy at Fuego also assumes availability of resources that expedite evacuation. Of- ﬁcial 4 gives the example of escaping a house on ﬁre to promote self-reliance – but one should wait for the ar- rival of ﬁreﬁghters, conﬁdent that their knowledge of the hazard and resources to quell the ﬂames are greater than those of the person at risk, who holds no oﬃcial responsibility for dealing with the ﬁre. Similarly, while the eﬀorts of CONRED to improve local knowledge of volcanic hazards and encourage resilience through The third assumption of the self-evacuation policy (as examined through Figure 5 and the quotations of Section 4.2.2) relates to direct experience of volcanic activity. If local residents experience a change in vol- canic activity that represents increased volcanic risk, they will then decide to evacuate. This approach is comparable to the hazard-perception approach stud- ied in Section 2, where individuals perceive changes in Presses universitaires de rasbourg Page 221 Page 221 Fireside tales: Volcán de Fuego Naismith et al., 2020 Figure 5: Infographic issued by CONRED promoting self-evacuation [left]; English translation [right]. Figure 5: Infographic issued by CONRED promoting self-evacuation [left]; English translation [right]. links with INSIVUMEH/CONRED). Therefore, a future eruption of Fuego may have very diﬀerent outcomes between communities that are only a few kilometres apart, just as in 2018 we witnessed the diﬀerent fates of San Miguel Los Lotes and La Reunión. By assum- ing that local residents will recognize eruptive activ- ity at Fuego in the same way as oﬃcials do, the self- evacuation policy may have critical implications: if a community’s knowledge and preparation are overesti- mated, they may be expected to organize an evacuation that they cannot achieve. Older residents referenced fall of arena as indication to evacuate (Section 4.1.2). However, the fall of arena would suggest the eruption had already reached a critical stage associated with haz- ards that may inhibit evacuation (e.g. through descent of pyroclastic ﬂows and/or lahars preventing escape by road). 6 Conclusions Furthermore, experiences of local resi- dents were not homogenous but diverged signiﬁcantly between residents of communities on Fuego’s west- ern ﬂanks and residents of communities further east. This study has revealed a previously unreported diﬀer- ence between residents Fuego’s west and east ﬂanks in terms of (1) direct experience of the eruptions of 1966 and 1974, and (2) trust and communication with IN- SIVUMEH/CONRED. These ﬁndings suggest that vol- canic risk may be even more localized than previously considered, and that experiences of previous eruptions can inﬂuence response to activity decades after the ini- tial event. The diﬀerence between Fuego’s west and east ﬂanks highlights the importance of responsibility and choice in response to eruptive activity. Local people are highly restricted by their responsibilities to their land and livelihoods, they may hold a strong attachment to their home, and the choices that they make are inﬂu- enced by this. Previous experiences of livelihood dev- astation in the 20th century appears to have inﬂuenced, and continues to inﬂuence, the choices of local people on the west ﬂanks of Fuego in a way it has not further east. The diﬀerent relationships between INSIVUMEH and CONRED with communities on diﬀerent sides of the volcano provides them with diﬀerent information and diﬀerent sources of trust that in turn shape an in- dividual’s risk at Fuego. As in analogous environments such as Tungurahua, volcanic risk at Fuego is not static but variable with time and with location. An individual living in Ceilán and volunteering in its COLRED lives with a completely diﬀerent risk from that of his elderly isolated aunt whom he visits in Panimaché Uno. O f h l l id i Interviewer: And what do you think of . . . have there been changes since the tragedy in June? If another eruption happens now, what are the things that have changed most? Interviewer: And what do you think of . . . have there been changes since the tragedy in June? If another eruption happens now, what are the things that have changed most? Oﬃcial 1: Well, it’s that the people know that they are living in a volcanic area. That they can- not conﬁde in it as before, “Ah, it’s having an- other eruption”. Another eruption comes – “Ah, it’s only making a noise!”. Now, I think that they . . . they are the best volcanologists now. 6 Conclusions 6 Trust between locals and authorities is heterogeneous around the volcano, suggesting that in the event of a future eruptive crisis, both knowledge of hazards and a willingness to evacuate may be inconsistent between communities, as seen at Mt Merapi [Donovan et al. 2012; Mei et al. 2013]. Furthermore, the three assump- tions implicit in self-evacuation policy—knowledge, resources, and experience—are not validated by our ﬁndings from interviews with local residents. Self- evacuation dictates that in the event of crisis, infor- mation and advice will issue from INSIVUMEH and CONRED, but local authorities must make the deci- sion to evacuate themselves. However, a lack of shared vocabulary between locals and authorities (e.g. in the term “pyroclastic ﬂow”) suggests that the information INSIVUMEH and/or CONRED share during height- ened activity may not be understood in the manner in- tended. Furthermore, in a future eruption conditions may prevail that prevent local people from resources that give them further information of volcanic activity, similar to what occurred on 3rd June 2018 (e.g. low visi- bility due to cloud, poor phone signal). The lack of such information may discourage locals from leaving in the spirit of prevention encouraged by self-evacuation. In addition, people in rural communities face signiﬁcant logistical barriers to evacuating that those in La Re- unión do not have (e.g. diﬃculty in reuniting families (men generally work in the ﬁeld, women at home); lack of transporting vehicles). During the last days of ﬁeld- work in 2019, an oﬃcial from INSIVUMEH expressed some ﬁnal thoughts on the hopes and challenges that lie ahead: In recent decades, volcanological research has increas- ingly recognized the importance of including local knowledge to obtain a holistic understanding of vol- canic risk. Meanwhile, research on evacuation from natural hazards shows that trust in authorities and di- rect previous experience are factors that strongly inﬂu- ence the decision to evacuate. Our research contributes to the debate of direct experience of eruptive activity and evacuation through an in-depth case study con- ducted over two years at the active Volcán de Fuego in Guatemala. Our study conﬁrmed ﬁndings in previ- ous literature that eruptive activity is experienced dif- ferently by diﬀerent people. At Volcán de Fuego, local residents’ experiences of previous eruptive activity dif- fer signiﬁcantly from authorities’ experiences over the same period. 6 Conclusions They can distinguish now between pyroclastic ﬂows, lava ﬂows, everything. Now they know perfectly. 5.2.2 Responsibility for decision-making and self- evacuation policy For example, local residents in 2019 stated they would rely on advice from IN- SIVUMEH or CONRED, consistent with results from Escobar-Wolf [2013]. As seen on many previous occa- sions at Fuego, this advice is either slow to arrive or does not come. From these ﬁndings we conclude that there are many deterrents to a community successfully organizing its self-evacuation from Fuego. The self-evacuation policy has merit in avoiding bringing more people into an area of high risk. In many environments threatened by persistent volcanic activity, repeated temporary evacuations can be suc- cessful, given good trust between locals and author- ities and security of domestic resources that encour- ages co-operative or proactive evacuation by locals [An- dreastuti et al. 2019; Armijos et al. 2017; Few et al. 2017]. However, the results of this study show that in 2019 these conditions were not in place at Fuego. Presses universitaires de rasbourg Page 222 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica Acknowledgements We would like to give thanks to all people at Volcán de Fuego who were willing to participate in this investiga- tion. Funding for this project was provided by NERC (grant number NE/L002434/1). AKN and IMW would like to acknowledge the assistance of NERC Urgency Grant NE/S011498/1. Armijos, M. T., J. Phillips, E. Wilkinson, J. Barclay, A. Hicks, P. Palacios, P. Mothes, and J. Stone (2017). “Adapting to changes in volcanic behaviour: Formal and informal interactions for enhanced risk manage- ment at Tungurahua Volcano, Ecuador”. Global Envi- ronmental Change 45, pp. 217–226. doi: 10.1016/j. gloenvcha.2017.06.002. Associated Press (2018). “Guatemala ups number of missing to 332 in volcano eruption”. Associated Press News. [online]. References Albino, F., J. Biggs, R. Escobar-Wolf, A. Naismith, M. Watson, J. Phillips, and G. C. Marroquin (2020). “Us- ing TanDEM-X to measure pyroclastic ﬂow source lo- cation, thickness and volume: Application to the 3rd June 2018 eruption of Fuego volcano, Guatemala”. Journal of Volcanology and Geothermal Research 406, p. 107063. doi: 10.1016/j.jvolgeores.2020.107063. Fuego remains a highly active volcano, and the pop- ulations close to its summit continue to increase. If risk mitigation policy fails to recognize the diﬀerent ways of experiencing eruptive activity, or ignores the social and economic pressures that may disincentivize locals from making the decision to evacuate, it is uncertain whether the rich knowledge of hazards and qualitative assessments of risk that local residents include in their experiences of Fuego’s activity may translate into them taking suﬃcient protective measures to preserve life in the case of a future explosive eruption. Andreastuti, S., E. Paripurno, H. Gunawan, A. Bu- dianto, D. Syahbana, and J. Pallister (2019). “Char- acter of community response to volcanic crises at Sinabung and Kelud volcanoes”. Journal of Volcanol- ogy and Geothermal Research 382, pp. 298–310. doi: 10.1016/j.jvolgeores.2017.01.022. Armijos, M. T. and R. Few (2015). “Living with volcanic risk: Vulnerability, knowledge and adaptation in the slopes of Tungurahua, Ecuador”. DEV Report and Pol- icy Papers Series. School of International Development, University of East Anglia, Norwich, UK. Interviewer: The lived experience. Oﬃcial 1: Yes. That’s it, they know now what can happen and where they mustn’t be. Because it was all . . . well, I’m aware that also CONRED people have been working around the volcano for some time. But the people weren’t interested, they didn’t go to the meetings, no. But now I think that . . . it’s the opportunity for CONRED to work with the communities. And with us too . . . rais- ing consciousness in people while they can. Our assessment of how local residents experience eruptive activity and trust in authorities allow us to consider (1) how these two factors inﬂuence their de- cision to evacuate from eruptive crisis; and (2) how CONRED’s current policy of self-evacuation should be viewed in light of the ﬁndings. Diﬀerences in how people experience eruptions may impact the success of future evacuation eﬀorts because diﬀerent people dis- agree in the threshold of volcanic risk that can be toler- Presses universitaires de rasbourg Page 223 Fireside tales: Volcán de Fuego Naismith et al., 2020 ated before a decision to evacuate is made. At other vol- canoes, acceptance of both local and oﬃcial experiences has created eﬀective adaptive volcanic risk mitigation strategies. We argue that at Fuego, local residents’ di- rect experiences of previous activity are an under-used resource; acknowledging these experiences by includ- ing them in training and policy may empower locals and encourage their collaboration with INSIVUMEH and CONRED in strengthening existing risk mitigation strategies. ated before a decision to evacuate is made. At other vol- canoes, acceptance of both local and oﬃcial experiences has created eﬀective adaptive volcanic risk mitigation strategies. We argue that at Fuego, local residents’ di- rect experiences of previous activity are an under-used resource; acknowledging these experiences by includ- ing them in training and policy may empower locals and encourage their collaboration with INSIVUMEH and CONRED in strengthening existing risk mitigation strategies. Copyright notice The current policy of self-evacuation encourages community empowerment through a transfer of re- sponsibility for deciding to evacuate to local commu- nities. It also has the additional beneﬁt of avoiding bringing more people into a high-risk zone. Never- theless, our study showed that self-evacuation contains implicit assumptions about local residents’ knowledge, resources, and experiences. These assumptions were not conﬁrmed by interviews with local residents. Lo- cal residents are knowledgeable of Fuego’s activity but lack knowledge of speciﬁc hazards such as pyroclastic ﬂows to make the decision to evacuate without diﬃ- culty. Other factors as lack of resources and security fears further complicate the decision to evacuate. © The Author(s) 2020. This article is distributed un- der the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, pro- vided you give appropriate credit to the original au- thor(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Data availability Questionnaire design (in English) is included as Sup- plementary Material alongside the online version of this article. Research design and questionnaire design (in Spanish) available on request from the correspond- ing author. Presses universitaires de rasbourg Author contributions PhD thesis. Michigan Technological University, US. Blaikie, P., T. Cannon, I. Davis, and B. Wisner (2014). At risk: natural hazards, people’s vulnerability and dis- asters. Routledge. Ferres, D. and R. Escobar-Wolf (2018). Informe Téc- nico: Volcán de Fuego. Acción Contra El Hambre y Coop- eración Española. Tech. rep. Acción Contra el Hambre (Guatemala). Bonis, S. and O. Salazar (1973). “The 1971 and 1973 eruptions of Volcán Fuego, Guatemala, and some socio-economic considerations for the volcanologist”. Bulletin Volcanologique 37.3, pp. 394–400. doi: 10. 1007/bf02597636. Few, R., M. T. Armijos, and J. Barclay (2017). “Living with Volcan Tungurahua: The dynamics of vulnera- bility during prolonged volcanic activity”. Geoforum 80, pp. 72–81. doi: 10.1016/j.geoforum.2017.01. 006. Cashman, K. V. and S. J. Cronin (2008). “Welcom- ing a monster to the world: Myths, oral tradition, and modern societal response to volcanic disasters”. Journal of Volcanology and Geothermal Research 176.3, pp. 407–418. doi: 10.1016/j.jvolgeores.2008.01. 040. Flyvbjerg, B. (2006). “Five Misunderstandings About Case-Study Research”. Qualitative Inquiry 12.2, pp. 219–245. doi: 10.1177/1077800405284363. Gaillard, J.-C. (2008). “Alternative paradigms of vol- canic risk perception: The case of Mt. Pinatubo in the Philippines”. Journal of Volcanology and Geother- mal Research 172.3-4, pp. 315–328. doi: 10.1016/j. jvolgeores.2007.12.036. Christie, R., O. Cooke, and J. Gottsmann (2015). “Fear- ing the knock on the door: critical security studies insights into limited cooperation with disaster man- agement regimes”. Journal of Applied Volcanology 4.1. doi: 10.1186/s13617-015-0037-7. Gaillard, J.-C. and J. Mercer (2013). “From knowledge to action”. Progress in Human Geography 37.1, pp. 93– 114. doi: 10.1177/0309132512446717. Cresswell, J. W. (2014). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches. Graves, K. L. (2007). “Risk Perception of Natural Hazards in the Volcanic Regions of Ecuador and Guatemala”. MA thesis. Michigan Technological University, US. Cronin, S. J. and K. V. Cashman (2016). “Volcanic oral traditions in hazard assessment and mitigation”. Liv- ing Under the Shadow. Ed. by J. Grattan and R. Tor- rence. Routledge, pp. 185–212. Hayes, J. (2018). “Stories and Standards: The Impact of Professional Social Practices on Safety Decision Mak- ing”. Beyond Safety Training: Embedding Safety in Pro- fessional Skills. Ed. by C. Bieder, C. Gilbert, and B. J. anf Hervé Laroche. Springer, Cham, pp. 73–83. Cronin, S. J., M. G. Petterson, P. W. Taylor, and R. Biliki (2004). “Maximising Multi-Stakeholder Participa- tion in Government and Community Volcanic Haz- ard Management Programs; A Case Study from Savo, Solomon Islands”. Natural Hazards 33.1, pp. 105– 136. doi: 10.1023/b:nhaz.0000035021.09838.27. Author contributions AN designed the study, led ﬁeldwork, analysed tran- scripts, translated interview data, and led drafting of the manuscript. TA guided the study design through- out ﬁeldwork and contributed to the development of the manuscript. WC and EB assisted in data collec- tion. MW assisted with ﬁeldwork and contributed to writing. All authors have made a substantial and in- tellectual contribution to the work and approved it for publication. Atkinson, R. and J. Flint (2004). “Snowball sampling”. The SAGE encyclopedia of social science research meth- ods. Ed. by M. S. Lewis-Beck, A. Bryman, and T. Fut- ing Liao. Thousand Oaks, CA: Sage Publications. doi: 10.4135/9781412950589.n931. Bankoﬀ, G. (2007). “Living with risk; coping with dis- asters”. Education about Asia 12.2, pp. 26–29. Presses universitaires de rasbourg Page 224 Page 224 3(2): 205 – 226. doi: 10.30909/vol.03.02.205226 Volcanica Barclay, J., R. Few, M. T. Armijos, J. C. Phillips, D. M. Pyle, A. Hicks, S. K. Brown, and R. E. A. Robertson (2019). “Livelihoods, Wellbeing and the Risk to Life During Volcanic Eruptions”. Frontiers in Earth Sci- ence 7. doi: 10.3389/feart.2019.00205. mal Research 172.3-4, pp. 288–299. doi: 10.1016/j. jvolgeores.2007.12.014. Donovan, A., J. R. Eiser, and R. S. J. Sparks (2014). “Sci- entists’ views about lay perceptions of volcanic haz- ard and risk”. Journal of Applied Volcanology 3.1. doi: 10.1186/s13617-014-0015-5. Barclay, J., K. Haynes, T. Mitchell, C. Solana, R. Teeuw, A. Darnell, H. S. Crosweller, P. Cole, D. Pyle, C. Lowe, C. Fearnley, and I. Kelman (2008). “Framing vol- canic risk communication within disaster risk reduc- tion: ﬁnding ways for the social and physical sciences to work together”. Geological Society, London, Special Publications 305.1, pp. 163–177. doi: 10.1144/sp305. 14. Donovan, K., A. Suryanto, and P. Utami (2012). “Map- ping cultural vulnerability in volcanic regions: The practical application of social volcanology at Mt Merapi, Indonesia”. Environmental Hazards 11.4, pp. 303–323. doi: 10.1080/17477891.2012.689252. Dove, M. R. (2008). “Perception of volcanic eruption as agent of change on Merapi volcano, Central Java”. Journal of Volcanology and Geothermal Research 172.3- 4, pp. 329–337. doi: 10.1016/j.jvolgeores.2007. 12.037. Bird, D. K. (2009). “The use of questionnaires for ac- quiring information on public perception of natu- ral hazards and risk mitigation – a review of current knowledge and practice”. Natural Hazards and Earth System Sciences 9.4, pp. 1307–1325. doi: 10.5194/ nhess-9-1307-2009. Escobar-Wolf, R. P. (2013). “Volcanic processes and hu- man exposure as elements to build a risk model for Volcan de Fuego, Guatemala”. Author contributions Haynes, K., J. Barclay, and N. Pidgeon (2008). “Whose reality counts? Factors aﬀecting the perception of volcanic risk”. Journal of Volcanology and Geother- mal Research 172.3-4, pp. 259–272. doi: 10 . 1016 / j.jvolgeores.2007.12.012. Dibben, C. J. (1999). “Looking beyond eruptions for an explanation of volcanic disasters: vulnerability in volcanic environments”. PhD thesis. University of Bedfordshire, UK. – (2008). “Leaving the city for the suburbs—The dom- inance of ‘ordinary’ decision making over volcanic risk perception in the production of volcanic risk on Mt Etna, Sicily”. Journal of Volcanology and Geother- – (2008). “Leaving the city for the suburbs—The dom- inance of ‘ordinary’ decision making over volcanic risk perception in the production of volcanic risk on Mt Etna, Sicily”. Journal of Volcanology and Geother- Hewitt, K. (1983). “The Idea of Calamity in a Tech- nocratic Age”. Interpretations of Calamity. Ed. by K. Hewitt. Allen and Unwin Inc. Presses universitaires de rasbourg Page 225 Page 225 Fireside tales: Volcán de Fuego Naismith et al., 2020 Noticias ONU (2018). “La erupción del volcán de Fuego afecta a 1,7 millones de personas en Guatemala”. Noticias ONU. [online, in Spanish]. Hicks, A., M. T. Armijos, J. Barclay, J. Stone, R. Robert- son, and G. P. Cortés (2017). “Risk communication ﬁlms: Process, product and potential for improving preparedness and behaviour change”. International Journal of Disaster Risk Reduction 23, pp. 138–151. doi: 10.1016/j.ijdrr.2017.04.015. Palinkas, L. A., S. M. Horwitz, C. A. Green, J. P. Wis- dom, N. Duan, and K. Hoagwood (2015). “Purposeful Sampling for Qualitative Data Collection and Analy- sis in Mixed Method Implementation Research”. Ad- ministration and Policy in Mental Health and Mental Health Services Research 42.5, pp. 533–544. doi: 10. 1007/s10488-013-0528-y. Hicks, A., J. Barclay, P. Simmons, and S. Loughlin (2014). “An interdisciplinary approach to volcanic risk reduction under conditions of uncertainty: a case study of Tristan da Cunha”. Natural Hazards and Earth System Sciences 14.7, pp. 1871–1887. doi: 10. 5194/nhess-14-1871-2014. Pistrang, N. and C. Barker (2012). “Varieties of qual- itative research: A pragmatic approach to selecting methods.” APA handbook of research methods in psy- chology, Vol 2: Research designs: Quantitative, qualita- tive, neuropsychological, and biological. American Psy- chological Association, pp. 5–18. doi: 10 . 1037 / 13620-001. Hoepﬂ, M. C. (1997). “Choosing Qualitative Research: A Primer for Technology Education Researchers”. Journal of Technology Education 9.1. doi: 10.21061/ jte.v9i1.a.4. Jóhannesdóttir, G. and G. Gísladóttir (2010). Author contributions “People living under threat of volcanic hazard in southern Iceland: vulnerability and risk perception”. Natural Hazards and Earth System Sciences 10.2, pp. 407–420. doi: 10.5194/nhess-10-407-2010. Ritchie, J. (2003). “The applications of qualitative methods to social research”. Qualitative research prac- tice: A guide for social science students and researchers. Ed. by J. Ritchie and J. Lewis. SAGE Publications Ltd, pp. 24–46. Johnston, D. M., M. S. B. C.-D. Lai, B. F. Houghton, and D. Paton (1999). “Volcanic hazard perceptions: com- parative shifts in knowledge and risk”. Disaster Pre- vention and Management: An International Journal 8.2, pp. 118–126. doi: 10.1108/09653569910266166. Rose, W. I., S. Self, P. J. Murrow, C. Bonadonna, A. J. Durant, and G. G. J. Ernst (2008). “Nature and sig- niﬁcance of small volume fall deposits at composite volcanoes: Insights from the October 14, 1974 Fuego eruption, Guatemala”. Bulletin of Volcanology 70.9, pp. 1043–1067. doi: 10.1007/s00445-007-0187-5. Lane, L. R., G. A. Tobin, and L. M. Whiteford (2003). “Volcanic hazard or economic destitution: hard choices in Banños, Ecuador”. Environmental Hazards 5.1, pp. 23–34. doi: 10.1016/j.hazards.2004.01.001. RT (2018). “VIDEO: Momento exacto de la erupción del Volcán de Fuego en Guatemala”. RT. [online, in Spanish]. León-Ramírez Carné, A. X. L. (2012). “Evaluación de la vulnerabilidad asociada a la amenaza del Volcán de Fuego en la aldea Panimaché”. MA thesis. Universi- dad de San Carlos de Guatemala (USAC), Guatemala. Stone, J., J. Barclay, P. Simmons, P. D. Cole, S. C. Lough- lin, P. Ramón, and P. Mothes (2014). “Risk reduction through community-based monitoring: the vigıas of Tungurahua, Ecuador”. Journal of Applied Volcanology 3.1. doi: 10.1186/s13617-014-0011-9. Lyons, J. J., G. P. Waite, W. I. Rose, and G. Chigna (2010). “Patterns in open vent, strombolian behavior at Fuego volcano, Guatemala, 2005–2007”. Bulletin of Volcanology 72.1, pp. 1–15. doi: 10.1007/s00445- 009-0305-7. Tobar, J. E. (2018). “Las alertas ignoradas por Conred y la cifra dudosa de desaparecidos”. Nómada. [online, in Spanish]. Mei, E. T. W., F. Lavigne, A. Picquout, E. de Bélizal, D. Brunstein, D. Grancher, J. Sartohadi, N. Cholik, and C. Vidal (2013). “Lessons learned from the 2010 evacuations at Merapi volcano”. Journal of Volcanol- ogy and Geothermal Research 261, pp. 348–365. doi: 10.1016/j.jvolgeores.2013.03.010. p Van Loon, A. F., I. Lester-Moseley, M. Rohse, P. Jones, and R. Day (2020). “Creative practice as a poten- tial tool to build drought and ﬂood resilience in the Global South”. doi: 10.5194/gc-2020-11. van Manen, S. M. (2014). Author contributions “Hazard and risk percep- tion at Turrialba volcano (Costa Rica);implications for disaster risk management”. Applied Geography 50, pp. 63–73. doi: 10.1016/j.apgeog.2014.02.004. Naismith, A. K. (2020). “Todo se oscureció: uniting re- mote sensing observations and human experiences to understand recent eruptive activity of Volcán de Fuego, Guatemala”. PhD thesis. University of Bris- tol, UK. Wachinger, G., O. Renn, C. Begg, and C. Kuhlicke (2013). “The Risk Perception Paradox-Implications for Governance and Communication of Natural Haz- ards”. Risk Analysis 33.6, pp. 1049–1065. doi: 10. 1111/j.1539-6924.2012.01942.x. Naismith, A. K., I. M. Watson, R. Escobar-Wolf, G. Chigna, H. Thomas, D. Coppola, and C. Chun (2019). “Eruption frequency patterns through time for the current (1999–2018) activity cycle at Volcán de Fuego derived from remote sensing data: Evidence for an accelerating cycle of explosive paroxysms and poten- tial implications of eruptive activity”. Journal of Vol- canology and Geothermal Research 371, pp. 206–219. doi: 10.1016/j.jvolgeores.2019.01.001. White, G. (1942). “Human adjustment to ﬂoods: A geographical approach to the ﬂood problem in the United States”. PhD thesis. Department of Geogra- phy, University of Chicago, USA. Presses universitaires de rasbourg Page 226 Page 226
https://openalex.org/W4366963756	https://www.e3s-conferences.org/articles/e3sconf/pdf/2023/19/e3sconf_unsat2023_07003.pdf	English	null	An analytical method to evaluate the capacity of laterally loaded piles in unsaturated soils	E3S web of conferences	2,023	cc-by	4,755	1 Introduction The soil above the ground water table (GWT) is often unsaturated with pores partly filled by air and partly filled by water. The difference between pore air and water pressures, 𝑠= 𝑢𝑎 − 𝑢𝑤 is named matric suction or, simply, suction (as in the following part of this paper). In most applications, the pore air pressure is atmospheric (i.e. equal to zero) while the pore water pressure is tensile (i.e. negative), meaning that the suction is equal to the pore water pressure changed of sign. Pile foundations are frequently chosen in geotechnical design to support a combination of axial forces, lateral forces and concentrated moments. Consideration of lateral forces is crucial, for example, in the design of tall buildings, bridge abutments, earth-retaining structures and wind turbines. Experimental, analytical and numerical studies show that the response of flexible piles is largely influenced by the pile-soil stiffness ratio whereas, for short and rigid piles, the behaviour is governed by both stiffness and slenderness ratio [1]. The non-linear soil behaviour also strongly influences design parameters such as head displacement, maximum bending moment and critical length. The increase in soil strength generated by capillarity above the GWT is modelled through the introduction of an extra cohesion, which is named “apparent cohesion” [8]. The apparent cohesion is often calculated by the expression 𝑆𝑟𝑠tan 𝜑 where 𝑆𝑟 is the degree of saturation and 𝜑 is the friction angle of the soil. The above expression of apparent cohesion is obtained from the Mohr-Coulomb failure criterion written in terms of Bishop stress [9] with = 𝑆𝑟 as: The kinematics of pile foundations under lateral loading may affect a significant volume of soil several diameters below ground level. In many applications, most of the interested soil lies above the ground water table (GWT) and is therefore partly saturated. Although traditional methods for the design of laterally loaded piles assume that the soil is either dry or fully saturated, the effect of partial saturation is lately attracting fresh research interest [2-6]. Recently, Lalicata et al. [7] extended the well-known Broms method for dry or saturated soils to unsaturated soils by taking into account the combined effects of the position of the GWT and the cohesion of the unsaturated soil above it. Leonardo Maria Lalicata1 , Agostino Walter Bruno1 and Domenico Gallipoli1 nt of Civil, Chemical and Environmental Engineering, University of Genoa, 16145 Genoa, Italy Abstract. The growing pressures of climate change, increased usage and unprecedented geo-hazards impose a modification in the way civil engineering structures are designed and constructed. This is particularly true for geotechnical works, which are very sensitive to changes of environmental conditions. For instance, the response of a pile under lateral loading is strongly influenced by the stiffness and strength of the first few metres of soil below the surface, which are often partly saturated. To consider this effect, the present paper describes an analytical method, which extends the well-known Broms approach to predict the lateral capacity of piles in unsaturated soils. More specifically, the proposed method considers the combined effects of the position of the ground water table and the extra strength of the partially saturated soil above it. Compared to Broms approach, the solution introduces four additional non-dimensional parameters that relate the soil-water retention behaviour to the geometry of the pile. The method provides a direct evaluation of the lateral pile capacity in partly saturated soils, which can be used as a basis for more accurate design. © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 https://doi.org/10.1051/e3sconf/202338207003 DP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ses/by/4 0/) An analytical method to evaluate the capacity of laterally loaded piles in unsaturated soils Leonardo Maria Lalicata1 , Agostino Walter Bruno1 and Domenico Gallipoli1 1 Introduction (1) 𝑓= 𝑐+ (+ 𝑆𝑟𝑠)tan 𝜑= 𝑐+  tan 𝜑+ 𝑆𝑟𝑠tan 𝜑 where 𝑐 is the effective cohesion and  tan 𝜑 is the frictional strength which depends on total stresses. In eq. (1) the apparent cohesion 𝑆𝑟𝑠tan 𝜑 can be expressed in terms of the suction s via the Soil Water Retention Curve (SWRC), which is a material law linking the degree of saturation 𝑆𝑟 of the soil to the suction s. In this paper, the main features of the method by Lalicata et al. [7] are briefly recalled and the influence of the soil-water retention parameters on the computed pile capacity are highlighted before validating results against field data. Among the many SWRCs proposed in the literature, the following Van Genuchten [10] equation has been chosen in this study: E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 E3S Web of Conferences 382, 07003 (2023) https://doi.org/10.1051/e3sconf/202338207003 Broms solution was derived for either purely frictional (Fig. 1a) or purely cohesive (Fig. 1b) soils. Conversely, a cohesive-frictional soil is here considered due to the apparent cohesion generated by the suction above the GWT. Note also that, for simplicity, the effective cohesion 𝑐 of eq. (2) is taken equal to zero. 𝑆𝑟= 𝑆𝑟,𝑟𝑒𝑠+ (1 −𝑆𝑟,𝑟𝑒𝑠) (1 + ( 𝑠 𝑠𝑒) 𝑛 ) −𝑚 (2) 𝑆𝑟= 𝑆𝑟,𝑟𝑒𝑠+ (1 −𝑆𝑟,𝑟𝑒𝑠) (1 + ( 𝑠 𝑠𝑒) 𝑛 ) −𝑚 (2) where 𝑆𝑟,𝑟𝑒𝑠 is the residual degree of saturation and 𝑠𝑒 (kPa) approximates the air entry suction. Note that Lalicata et al. [7] assumed 𝑆𝑟,𝑟𝑒𝑠 equal to zero for sake of simplicity. The present work instead considers SWRCs of three real soils that may exhibit a non-null residual degree of saturation (see Section 5). The air entry suction also coincides with the air expulsion suction due to the uniqueness of the relationship between degree of saturation and suction. The non-dimensional parameters 𝑛 and 𝑚 control the shape of the SWRC curve where, to simplify calibration, 𝑚 is related to 𝑛 by the relationship 𝑚= 1 − 1 𝑛. Fig. 1. Soil resistance profiles for a) purely cohesive soil according to Broms [11]; b) purely frictional soil according to Broms [12]. l h i il 3 Extension of Broms method to unsaturated soils Broms [11, 12] provides the lateral pile capacity and the corresponding maximum bending moment as a function of: i) the pile geometry, i.e. the length 𝐿 and load eccentricity 𝑒, ii) the pile yielding moment of the pile 𝑀𝑦 iii) the soil resistance, i.e. the friction angle 𝜑 or undrained shear strength 𝑠𝑢 and iv) the type of failure, i.e. short, intermediate, or long pile failure mechanism. In the short pile mechanism, only the ultimate soil resistance is attained, and the pile reacts with its whole length. Fig. 1. Soil resistance profiles for a) purely cohesive soil according to Broms [11]; b) purely frictional soil according to Broms [12]. 3.1 Including apparent cohesion into Broms method p g Conversely, in the case of intermediate or long pile mechanisms, both the soil and pile ultimate resistance are mobilised. The pile resistance is attained in one or two cross-sections depending on both the head restraint and the specific failure mechanism. This corresponds to the formation of one or two plastic hinges at the ground surface and/or at some depth below it. In the presence of a buried plastic hinge, the pile reacts only with the section comprised between the ground surface and the plastic hinge. Both pile and soil are assumed rigid-perfectly plastic so that, once the limit soil resistance 𝑝𝑙𝑖𝑚 and pile yielding moment 𝑀𝑦 are defined, the solution can be found by imposing the translational and rotational equilibrium. The pore water pressure profile, and hence the suction profile, above the GWT depends on the hydraulic boundary conditions at the ground surface (Fig. 2). Once the soil-water retention properties of eq. (2) and the suction profile are known, the apparent cohesion profile 𝑆𝑟𝑠tan 𝜑 is readily calculated as shown in Fig. 2c and 2d. To solve analytically the translational and rotational equilibrium, the apparent cohesion profile is here idealised by a simpler design profile, e.g. a constant or triangular design profile, having the same integral area of the actual one (Fig. 2c and 2d). equilibrium. Fig. 2. Evaluation of apparent cohesion above the GWT (modified from [7]). Fig. 2. Evaluation of apparent cohesion above the GWT (modified from [7]). 2 2 E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 https://doi.org/10.1051/e3sconf/202338207003 The equivalence of integral areas ensures that the soil resisting force is correctly calculated by the design profile, though the corresponding lever arm is likely to differ from the actual one. dimensional groups listed in Table 1. Compared to Broms, the solution includes additional non-dimensional groups linked to the GWT depth and the SWRC parameters. More specifically, the apparent cohesion is taken into account through the non-dimensional group 𝐶, which is defined in terms of 𝑐𝑎𝑝𝑝 as shown in Table 1. It can be further shown that the value of 𝑐𝑎𝑝𝑝 in eq. (4) depends on the non-dimensional GWT depth 𝑧𝑤𝑑 ⁄ , air entry suction 𝑠𝑒𝛾𝑑 ⁄ , unit weight of water 𝛾𝑤𝛾 ⁄ , slope of the SWRC 𝑛, residual degree of saturation 𝑆𝑟,𝑟𝑒𝑠 and friction angle 𝜑 (see Lalicata et al. [7] for further details). 3.1 Including apparent cohesion into Broms method Interestingly, the non-dimensional group 𝑠𝑒𝛾𝑑 ⁄ indicates that, for a given soil and GWT depth, the effect of the air entry suction on the pile horizontal capacity decreases with increasing values of the pile diameter. Both constant and triangular design profiles of Fig. 2c and 2d therefore respect the translational equilibrium but they do not necessarily respect the rotational equilibrium. The triangular design profile (Fig. 2d) has the advantage of correctly calculating a zero apparent cohesion in correspondence of the GWT, where suction reduces to zero. Conversely, the constant design profile (Fig. 2c) has the advantage of being more conservative as the lever arm of the resulting force, with respect to the ground surface, is bigger and therefore produces smaller values of lateral pile capacity than the triangular one. Thus, in the interest of safety, only the constant design profile is considered hereafter. The constant design apparent cohesion 𝑐𝑎𝑝𝑝 is calculated as the average of the 𝑆𝑟𝑠tan 𝜑 profile over the reacting unsaturated depth 𝐿𝑢 as: Table 1. Non-dimensional groups. Geometry Slenderness 𝐿 𝑑 Eccentricity 𝑒 𝑑 Load and moment Limit load 𝐻𝑙𝑖𝑚 𝐾𝑝 𝛾 𝑑3 Limit moment 𝑀𝑙𝑖𝑚 𝐾𝑝 𝛾 𝑑4 Yielding moment 𝑀𝑦 𝐾𝑝 𝛾 𝑑4 GWT Submerged soil unit weight 𝛾′ 𝛾 Water unit weight 𝛾𝑤  Water table depth 𝑧𝑤 𝑑 SWRC Cohesive term C= 9√𝐾𝑝 𝐾𝑝 𝑐𝑎𝑝𝑝 𝛾 𝑑 Air entry value of suction 𝑠𝑒 𝛾 𝑑 Slope of SWRC 𝑛 Residual degree of saturation 𝑆𝑟,𝑟𝑒𝑠 Table 1. Non-dimensional groups. as: 𝑐𝑎𝑝𝑝= ∫ (𝑆𝑟𝑠tan )𝑑𝑧 𝐿𝑢 0 𝐿𝑢 (3) (3) The reacting unsaturated depth 𝐿𝑢 coincides with the GWT depth if the failure mechanism extends below the unsaturated layer while it is smaller than the GWT depth if the failure mechanism is strictly contained within the unsaturated layer. This latter circumstance happens when the pile is shorter than the depth of the GWT or if plastic hinges form above the GWT. 3.1 Including apparent cohesion into Broms method Under the hypothesis of drained soil response, the soil reaction 𝑝𝑙𝑖𝑚 is calculated above and below the GWT as: 𝑝𝑙𝑖𝑚= (3𝐾𝑝 𝛾 𝑧+ 9√𝐾𝑝 𝑐𝑎𝑝𝑝) 𝑑 𝑓𝑜𝑟 𝑧< 𝑧𝑤 (4a) 𝑝𝑙𝑖𝑚= (3𝐾𝑝 𝛾 𝑧𝑤 + 3𝐾𝑝 𝛾′(𝑧−𝑧𝑤) )𝑑 𝑓𝑜𝑟 𝑧> 𝑧𝑤 (4b) (4a) (4b) where 𝑧𝑤 is the GWT depth, 𝑑 is the pile diameter, 𝛾 is the soil bulk unit weight, 𝛾′ = 𝛾−𝛾𝑤 is the soil submerged unit weight (where 𝛾𝑤 is the water unit weight) and 3𝐾𝑝 = 3 tan2 (45° +  2) is the passive earth pressure coefficient where the factor of 3 accounts for three- dimensional effects. Following Cecconi et al. [13], the three-dimensional passive earth pressure coefficient for the cohesive term is taken equal to 9√𝐾𝑝. Inspection of eq. (4) indicates that: i) the apparent cohesion 𝑐𝑎𝑝𝑝 only contributes to the strength of the top unsaturated layer and ii) the frictional term depends on the bulk unit weight 𝛾 in the top unsaturated layer and on the submerged unit weight 𝛾′ = 𝛾−𝛾𝑤 in the bottom saturated layer. In the sake of simplicity, the bulk unit weight 𝛾 is here assumed to be identical in both unsaturated and saturated layers. The solution of the translational and rotational equilibrium is presented by Lalicata et al. [7] in a dimensionless form that provides the horizontal capacity 𝐻𝑙𝑖𝑚 and the corresponding maximum bending moment 𝑀𝑙𝑖𝑚 in terms of the non- 4 Validation against field data The proposed method has been validated against field tests on free head short rigid piles embedded in three distinct unsaturated soils, i.e. a weathered granite [14], a sand with clay and silt [15] and a fine sand [16], . In all cases, the water table was deeper than the toe of the pile (𝑧𝑤> 𝐿) so that the unsaturated reacting depth coincides with the pile length (i.e. 𝐿𝑢= 𝐿). The authors of the above testing campaigns also reported the average degree of saturation along the pile length as obtained from site investigations. Due to the small dimensions of the piles 3 E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 https://doi.org/10.1051/e3sconf/202338207003 Fig. 3. Comparison between measured and predicted values of lateral pile capacity. 5 Influence of partial saturation on the lateral pile capacity The importance of the soil-water retention properties is Table 2. Material and geometrical parameters of laterally loaded pile tests. Reference 𝑳/𝒅 (-) 𝒆/𝒅 (-) 𝜸 (kN/m3) 𝝋 (°) 𝒄𝒂𝒑𝒑 (kPa) Choi et al. [14] 𝑑 = 0.4m 3.0 5.0 18 30 10.9 6.0 5.0 5.46 6.0 0.4 Truong [15] 𝑑 = 0.127m 10.9 2.0 17.5 35 33.2 11.1 2.0 10.6 1.7 11.7 2.0 Wang et al. [16] 𝑑 = 0.273m 3.7 1.2 16.4 38 2.34 3.7 1.2 5.5 1.3 Wang et al. [16] 𝑑 = 0.457m 2.2 0.7 2.2 0.8 3.3 0.7 and the deep GWT, the degree of saturation can be assumed constant along the pile length and equal to the average value. The corresponding suction was then computed according to the SWRCs declared by the above authors. Finally, the apparent cohesion 𝑆𝑟𝑠tan 𝜑 was obtained from the corresponding values of degree of saturation and suction, thus assuming a constant design profile as shown in Fig. 2c. Table 2. Material and geometrical parameters of laterally loaded pile tests. Reference 𝑳/𝒅 (-) 𝒆/𝒅 (-) 𝜸 (kN/m3) 𝝋 (°) 𝒄𝒂𝒑𝒑 (kPa) Choi et al. [14] 𝑑 = 0.4m 3.0 5.0 18 30 10.9 6.0 5.0 5.46 6.0 0.4 Truong [15] 𝑑 = 0.127m 10.9 2.0 17.5 35 33.2 11.1 2.0 10.6 1.7 11.7 2.0 Wang et al. [16] 𝑑 = 0.273m 3.7 1.2 16.4 38 2.34 3.7 1.2 5.5 1.3 Wang et al. [16] 𝑑 = 0.457m 2.2 0.7 2.2 0.8 3.3 0.7 Table 2. Material and geometrical parameters of laterally loaded pile tests. The material and geometrical parameters of the testing campaigns are summarised in Table 2. 4 Validation against field data Calibration details are reported in both the original publications [14-16] and in Lalicata et al. [7]. The lateral capacity of free head short piles is computed by using eq. (5a) and imposing the rotational equilibrium around the pile toe as: 𝐻𝑙𝑖𝑚∙(𝑒+ 𝐿) = 3𝐾𝑝𝛾𝑑𝐿∙𝐿 2 ∙𝐿 3 + 9√𝐾𝑝𝑐𝑎𝑝𝑝𝑑𝐿∙𝐿 2 (5a) (5a) or in non-dimensional form: or in non-dimensional form: Fig. 3. Comparison between measured and predicted values of lateral pile capacity. 𝐻𝑙𝑖𝑚 𝐾𝑝𝛾𝑑= (1 2 (𝐿 𝑑) 3 + 1 2 𝐶(𝐿 𝑑) 2 ) ∙ 1 (𝑒 𝑑+ 𝐿 𝑑) (5b) (5b) where the non-dimensional group 𝐶 is defined in Table 1. Compared to Broms solution for purely frictional soils, eq. (5b) incorporates the additional term 1 2 𝐶( 𝐿 𝑑) 2 that accounts for the extra cohesive strength of the unsaturated soil layer. Fig. 3 compares the measured and predicted values of lateral capacity showing that the proposed method captures well the experimental data, particularly those by Truong [15]. In this case, the good estimate of the lateral pile capacity relies on the accurate measurement of the soil-water retention behaviour. Conversely, a higher discrepancy (>20%) between measured and predicted values is observed for two of the three piles tested by Choi et al. [14]. This difference may be due to the lack of information about the variation of suction with depth, which was not provided by the authors. Inspection of Fig. 3 confirms that accounting for the extra soil strength due to partial saturation leads to a more accurate estimate of the lateral pile capacity compared with Broms solution, which underestimates all experimental data. Fig. 3. Comparison between measured and predicted values of lateral pile capacity. 5 Influence of partial saturation on the lateral pile capacity Material se (kPa) n (-) Sr,res (-) se/ d B-grade kaolin clay [17] 58.0 1.37 0 8.06 Jossigny Silt [18] 8.0 1.27 0 1.11 Rudlingen sand [19] 1.2 1.88 0.41 0.17 Table 3. Parameter values of Van Genuchten SWRC. Fig. 4. Restrained head pile: failure mechanism and soil resistance profiles for the case of short pile. Fig. 5. Interpolation of experimental data by Van Genuchten SWRC. The non-dimensional lateral capacity of the pile is computed from the horizontal equilibrium and is given by eqs. (6a) and (6b) when 𝐿> 𝑧𝑤 and 𝐿≤𝑧𝑤, respectively: 𝐻𝑙𝑖𝑚 𝐾𝑝𝛾𝑑3 = [3 2 (𝑧𝑤 𝑑) 2 + 3 𝑧𝑤 𝑑(𝐿 𝑑−𝑧𝑤 𝑑) + 3 2 𝛾′ 𝛾(𝐿 𝑑−𝑧𝑤 𝑑) 2 ] + 𝐶𝑧𝑤 𝑑 (6a) 𝐻𝑙𝑖𝑚 𝐾𝑝𝛾𝑑3 = 3 2 (𝐿 𝑑) 2 + 𝐶𝐿 𝑑 (6b) (6a) (6b) Eqs. (6a) and (6b) indicate that the horizontal capacity depends on the geometrical parameters 𝐿/𝑑, 𝑒/𝑑 and 𝑧𝑤/𝑑, on the normalised bulk unit weight of the soil 𝛾′ 𝛾 ⁄ and on the SWRC properties via the parameter 𝐶. In detail, the first term on the right and side of eqs. (6a) and (6b) describes the contribution of the frictional strength while the second term relates to the cohesive strength of the unsaturated layer above the GWT. Eqs. (6a) and (6b) indicate that the horizontal capacity depends on the geometrical parameters 𝐿/𝑑, 𝑒/𝑑 and 𝑧𝑤/𝑑, on the normalised bulk unit weight of the soil 𝛾′ 𝛾 ⁄ and on the SWRC properties via the parameter 𝐶. In detail, the first term on the right and side of eqs. (6a) and (6b) describes the contribution of the frictional strength while the second term relates to the cohesive strength of the unsaturated layer above the GWT. Fig. 5. Interpolation of experimental data by Van Genuchten SWRC. Fig. 6 shows the lateral pile capacity computed by the present work and Broms solution for different values of 𝑧𝑤/𝑑 from full soil saturation (𝑧𝑤/𝑑 = 0) to a GWT depth well below the pile toe (𝑧𝑤/𝑑 = 15). The solution for 𝐶 = 0 corresponds to the assumption of a dry soil above the GWT. Inspection of Fig. 6 indicates that, for shallow GWT depths (i.e. small values of 𝑧𝑤/𝑑), the present work calculates smaller values of the lateral pile capacity compared to Broms dry soil solution. 5 Influence of partial saturation on the lateral pile capacity The importance of the soil-water retention properties is here explored for the common case of restrained head pile. For the sake of brevity, the analysis is restricted to the case of short pile, i.e. the case where the pile yielding moment is always larger than the maximum bending moment. The interested reader can find the general solution in Lalicata et al. [7]. Equation (5b) highlights the effect of the non- dimensional cohesive group 𝐶 on the lateral capacity of piles. It is worth noting that even a small amount of apparent cohesion may significantly increase the lateral pile capacity. With reference to the tests of Choi et al. [14] and Wang et al. [16], the apparent cohesion is relatively low (it ranges from 2.34 to 10.9), but its effect on the predicted lateral capacity of the pile is very evident in Fig. 3. In the tests from Choi et al. [14], the non-dimensional cohesive group 𝐶 is responsible for 40% to 72% of the pile capacity depending on the value of 𝐿/𝑑 while in the test from Wang et al. [16] it accounts for 35% to 87% of the pile capacity. As shown in Fig. 4, the failure mechanism of a short restrained head pile is a rigid horizontal translation that involves the whole pile length and the maximum bending moment is therefore located at the pile head. E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 https://doi.org/10.1051/e3sconf/202338207003 Fig. 4. Restrained head pile: failure mechanism and soil resistance profiles for the case of short pile. The calibration of the Van Genuchten SWRC against experimental data is illustrated in Fig. 5 while the corresponding parameters are listed in Table 3. The calibration of the Van Genuchten SWRC against experimental data is illustrated in Fig. 5 while the corresponding parameters are listed in Table 3. Fig. 5. Interpolation of experimental data by Van Genuchten SWRC. Table 3. Parameter values of Van Genuchten SWRC. Material se (kPa) n (-) Sr,res (-) se/ d B-grade kaolin clay [17] 58.0 1.37 0 8.06 Jossigny Silt [18] 8.0 1.27 0 1.11 Rudlingen sand [19] 1.2 1.88 0.41 0.17 Table 3. Parameter values of Van Genuchten SWRC. References References 1. L.M. Lalicata, G. M. Rotisciani, A. Desideri, F. Casini. Geosc. 12(1), 1 (2021) https://doi.org/10.3390/geosciences12010001 2. L.M. Lalicata, A. Desideri, F. Casini, L. Thorel. Can. Geotech. J. 56(11), 1545–1556 (2019) https://doi.org/10.1139/cgj-2018-0322 3. X. Cheng, S.K. Vanapalli. Comp. Geotech. 140, 104480 (2021) https://doi.org/10.1016/j.compgeo.2021.104480 4. M. Ghazavi, E. Mahmoodi, H. El Naggar. A. Geo. 17 (2022) http://dx.doi.org/10.1007/s11440-022-01647- w 5. L.M. Lalicata, G. M. Rotisciani, A. Desideri, F. Casini, L. Thorel. In: Geotechnical Research for Land Protection and Development. CNRIG 2019, Lecture Notes in Civil Engineering, Springer, Cham, 40, 713-722 (2019) https://doi.org/10.1007/s11440- Fig. 6. Predicted non-dimensional lateral pile capacity as a function of GWT depth. 5 Influence of partial saturation on the lateral pile capacity It is worth noting that, since the lateral capacity is evaluated by imposing the horizontal equilibrium, eqs. (6a) and (6b) are rigorously true because their unsaturated cohesive terms (i.e. 𝐶 𝑧𝑤 𝑑 and 𝐶 L 𝑑, respectively) coincide with those calculated via the integration of the apparent cohesion profile 𝑆𝑟𝑠tan . However, as the GWT deepens and the resisting contribution of 𝑐𝑎𝑝𝑝 grows larger, the trend reverses and the present work provides larger values of the lateral pile capacity compared to Broms dry soil solution. Over this range, low values of 𝑠𝑒𝛾𝑑 ⁄ and large values of 𝑛 reduce the increase of lateral pile capacity predicted by the present method compared to Broms dry soil solution. For the Rudlingen sand, however, this effect is partly counteracted by the high residual degree of saturation (Table 3). Even for the relatively low value of 𝑠𝑒𝛾𝑑 ⁄ = 0.17 of Rudlingen sand, the non-dimensional lateral capacity predicted by the present method is larger than Broms dry soil solution by 4.7% at 𝑧𝑤/𝑑 = 10 and by 23% at 𝑧𝑤/𝑑 = 15. To better illustrate the importance of accounting for the apparent cohesion of the unsaturated layer, eqs. (7a) and (7b) are applied to the prediction of the lateral capacity of a 0.4 m diameter pile having an embedded length of 6m, i.e. a slenderness ratio 𝐿/𝑑= 15. The retention parameters of three real soils are considered, namely a fine silt (B-grade kaolin clay [17]), a silty sand (Jossigny Silt [18]) and a sand with clay (Rudlingen sand [19]). To simplify comparison, the bulk unit weight of the soil and the friction angle have been assumed identical in all cases and equal to 𝛾 = 18 kN/m3 and 𝜑 = 28°, respectively (i.e. 𝐾𝑝 = 2.77). The unit weight of water is 𝛾𝑤 = 10kN/m3 while the depth of the the GWT ranges between 0 (fully saturated conditions) and 12 m, so that 𝑧𝑤/𝑑 spans from 0 to 30. The pore pressure profile is assumed hydrostatic above and below the GWT, i.e. 𝑢𝑤= 𝑤(𝑧−𝑧𝑤). Finally, the 𝑧𝑤/𝑑 threshold marking the transition from smaller to larger predictions of lateral pile capacity compared to Broms dry soil solution depends on the SWRC properties (i.e. on the values of 𝑠𝑒𝛾𝑑 ⁄ , 𝑛 and 𝑆𝑟,𝑟𝑒𝑠). 5 E3S Web of Conferences 382, 07003 (2023) UNSAT 2023 E3S Web of Conferences 382, 07003 (2023) https://doi.org/10.1051/e3sconf/202338207003 Fig. 6. 6 Conclusions 6. X. Zou, Z. Yang, W. Wu. S. Dyn. Earthquake Eng. 165, 107672 (2023) http://dx.doi.org/10.1016/j.soildyn.2022.107672 Broms method for the calculation of the horizontal pile capacity has been extended to include the presence of a groundwater table (GWT) and the cohesive strength of the unsaturated layer above it. Above the GWT, the soil resistance is given by two components: a frictional term, which depends on the total stress, and a cohesive term (apparent cohesion), which depends on the soil-water retention properties. To allow a closed-form solution, the apparent cohesion has here been assumed constant and equal to the average value of the actual apparent cohesion profile above the GWT. 65, 0767 ( 0 3) http://dx.doi.org/10.1016/j.soildyn.2022.107672 7. L.M. Lalicata, A.W. Bruno, D. Gallipoli. Comp. Geotech. 155 (2023) https://doi.org/10.1016/j.compgeo.2022.105189 8. S.K. Vanapalli, D.G. Fredlund, D.E. Pufahl, A.W. Clifton. Can. Geotech. J. 33, 3, 379–392. (1996) https://doi.org/10.1139/t96-060. 9. A. W. Bishop, G. E. Blight. Géotechnique 13, 3, 177–197 (1963) https://doi.org/10.1680/geot.1963.13.3.177 10. M.Th. van Genuchten. S. Sc. Soc. Am. J. 44, 5, 892– 898 (1980). https://doi.org/10.2136/sssaj1980.036159950044000 50002x The method has been successfully validated against the values of the lateral capacity of piles as measured in works published in literature. Results confirm that neglecting the partially saturated state of the soil above the GWT leads to a significant underestimation of the lateral pile capacity. 11. B. B. Broms. J. Soil Mech. Found. Div. 90, 3, 123- 156 (1964) 12. B. B. Broms. J. S. Mech. Found. Div. 90, 2, 27-63 (1964) In the present solution, additional non-dimensional groups have been introduced to represent the influence of unsaturated conditions, i.e. the normalised GWT depth 𝑧𝑤/𝑑, the slope of the soil-water retention curve 𝑛, the residual degree of saturation 𝑆𝑟,𝑟𝑒𝑠 and the normalised air entry value of suction 𝑠𝑒/𝑑 (linking the air entry value of suction to the vertical stress at one diameter of depth). 13. M. Cecconi, V. Pane, A. Vecchietti, D. Bellavita. Soils Found. 59, 4, 840–856 (2019) https://doi.org/10.1016/j.sandf.2019.01.007 14. H.Y. Choi, S.R. Lee, H.I. Park, D.H. Kim. (2013). J. Geotech. and Geoenv. Eng. 139(9), 1477–1489. https://doi.org/10.1061/(ASCE)GT.1943- The effect of apparent cohesion increases with growing values of 𝑠𝑒/𝑑 and 𝑆𝑟,𝑟𝑒𝑠 while decreases with growing values of 𝑛. For reasonably low values of 𝑛, even the apparent cohesion of an unsaturated sandy soil can provide a significant increase in horizontal capacity (20 – 40%) for 𝑧𝑤/𝑑 > 15 compared to Broms dry soil solution. 5 Influence of partial saturation on the lateral pile capacity Predicted non-dimensional lateral pile capacity as a function of GWT depth. Fig. 6. Predicted non-dimensional lateral pile capacity as a function of GWT depth Fig. 6. Predicted non-dimensional lateral pile capacity as a function of GWT depth. References 1. L.M. Lalicata, G. M Casini. Geosc. 12(1 https://doi.org/10.33 2. L.M. Lalicata, A. D Geotech. J. 56(11), https://doi.org/10.1 3. X. Cheng, S.K. Van 104480 (2021) https://doi.org/10.10 4. M. Ghazavi, E. Mah (2022) http://dx.doi w 5. L.M. Lalicata, G. M Casini, L. Thorel. In Land Protection an Lecture Notes in Ci 40 713-722 (2019) 5606.0000831 15. P. Truong. Experimental investigation on the behaviour of laterally loaded piles in soft clay, sand and residual soils. PhD thesis. (2017) 16. H. Wang, B. M. Lehane, M.F. Bransby, L.Z. Wang, Y. Hong. A. Geo. 1-12. (2022) The present method provides an expeditious, and yet effective, estimation of the influence of partial saturation on the capacity of laterally loaded piles. It is hoped that this method could serve as a basis to further disseminate the application of unsaturated soils mechanics principles among engineering practitioners. 17. L.M. Lalicata, G. M. Rotisciani, A. Desideri, F. Casini, L. Thorel. In: Eu. Conf. Unsat. Soils. E3S Web of Conferences, 2020, 195, 02012 10.1051/e3sconf/202019502012 18. G.M. Rotisciani, G. Sciarra, F. Casini, A. Desideri. Int. J. Num. An. Meth. Geomech. 39, 11, 1212-1234 (2015) http://dx.doi.org/10.1002/nag.2359 19 P Sitarenios F Casini A Askarinejad S 18. G.M. Rotisciani, G. Sciarra, F. Casini, A. Desideri. Int. J. Num. An. Meth. Geomech. 39, 11, 1212-1234 (2015) http://dx.doi.org/10.1002/nag.2359 19. P. Sitarenios, F. Casini, A. Askarinejad, S. Springman. Géotechnique 71, 2, 96–109 (2019) https://doi.org/10.1680/jgeot.18.P.188 6
https://openalex.org/W2257228561	https://europepmc.org/articles/pmc6243352?pdf=render	English	null	Brachial pseudoaneurysm of the neonate with partial response to thrombin injections and late spontaneous thrombosis and regression during expectant management	BJR case reports	2,016	cc-by	2,408	Cite this article as: P D Z k R C Cite this article as: Parra D, Zuker R, Connolly B. Brachial pseudoaneurysm of the neonate with partial response to thrombin injections and late spontaneous thrombosis and regression during expectant management. BJR Case Rep 2016; 2: 20150383. 1DIMITRI PARRA, MD, 2RONALD ZUKER, MD and 1BAIRBRE CONNOLLY, MB, FRCPC Department of Diagnostic Imaging, Division of Image Guided Therapy, The Hospital for Sick Children, University of Toronto Toronto, Canada Department of Plastic Surgery, The Hospital for Sick Children, University of Toronto, Toronto, Canada Department of Plastic Surgery, The Hospital for Sick Children, Address correspondence to: Dr Dimitri Parra E-mail: dimitri.parra@sickkids.ca Address correspondence to: Dr Dimitri Parra E-mail: dimitri.parra@sickkids.ca E-mail: dimitri.parra@sickkids.ca ABSTRACT We illustrate the case of a brachial pseudoaneurysm in a 32-week preterm baby male who presented with a large pulsatile mass in the right antecubital fossa, with no clear aetiologic factor. The management of this type of lesion has been controversial and based mainly on case reports. In this case, after obtaining partial response with two thrombin injections, it spontaneously regressed during expectant management. This combination of therapeutic options may be an alternative for the management of complex lesions and, to the best of our knowledge, has not been previously reported. INTRODUCTION another institution. As the baby was very small, the initial consideration was to manage the lesion without surgery until he was a few months old. However, there was consider- able clinical and family unease about the progression or cat- astrophic rupture of the PSA. Given the safety profile of ultrasound compression and thrombin injection in the liter- ature,1,2 and after considerable discussion, we decided to intervene initially with ultrasound-guided compression, and if unsuccessful, with percutaneous thrombin injection. Endovascular management was not considered owing to the Pseudoaneurysms (PSAs) develop from a disruption of the arterial wall, which can be secondary to different aetiolo- gies, including inflammation, trauma and iatrogenic causes.1 They can have variable presentations, and cur- rently it is accepted that image-guided occlusion methods have replaced surgery as the first choice of treatment.2 Brachial artery PSA is a rare condition in neonates.3 It has been associated with trauma to the artery during veni- puncture.3,4 Its management is controversial,3–5 with reports of use of surgery and thrombin injection as therapeutic options. Figure 1. Clinical image of the patient showing a large pulsatile mass in the right antecubital fossa (arrows). Brachial pseudoaneurysm of the neonate with partial response to thrombin injections and late spontaneous thrombosis and regression during expectant management 1DIMITRI PARRA, MD, 2RONALD ZUKER, MD and 1BAIRBRE CONNOLLY, MB, FRCPC 1DIMITRI PARRA, MD, 2RONALD ZUKER, MD and 1BAIRBRE CONNOLLY, MB, FRCPC Received: 28 September 2015 Revised: 29 January 2016 Accepted: 04 February 2016 Cite this article as: Parra D, Zuker R, Connolly B. Brachial pseudoaneurysm of the neonate with partial response to thrombin injections and late spontaneous thrombosis and regression during expectant management. BJR Case Rep 2016; 2: 20150383. BJR\|case reports http://dx.doi.org/10.1259/bjrcr.2015038 Received: 28 September 2015 Revised: 29 January 2016 Accepted: 04 February 2016 Cite this article as: Parra D, Zuker R, Connolly B. Brachial pseudoaneurysm of the neonate with partial response to thrombin injections and late spontaneous thrombosis and regression during expectant management. BJR Case Rep 2016; 2: 20150383. BJR\|case reports http://dx.doi.org/10.1259/bjrcr.2015038 CASE REPORT A 32-week preterm twin baby male was referred to us at 7 weeks of age (3 kg) with a large, readily palpable pulsatile mass in the upper portion of the right antecubital fossa, which was increasing since day 14 of life (Figure 1). He had been normal at birth, but had had mild respiratory distress for 1 week and prior surgery for hypertrophic pyloric steno- sis. Ultrasound examination of the lesion showed a large (3.2 2.3 1.8 cm) PSA secondary to a small defect in the posterior aspect of the distal brachial artery (visible, wide and short neck) (Figure 2). CT imaging of the area was per- formed for treatment planning (Figure 3). There was no clear history of trauma, apart from routine venipunctures at © 2016 The Authors. Published by the British Institute of Radiology. This is an open access article under the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited. © 2016 The Authors. Published by the British Institute of Radiology. This is an open access article under the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited BJR\|case reports BJR\|case reports Parra et al Figure 2. (a) Longitudinal view of the pulsatile mass on colour Doppler ultrasound. This study demonstrated typical features of a large PSA. (b) Transverse view on grey scale ultrasound imaging illustrating that the PSA was secondary to a defect in the posterior wall of the brachial artery (arrows). (c) Colour Doppler study of the same area as in (b) showing that the neck of the PSA was large and wide (arrows), with significant flow (F) into the lesion. PSA, pseudoaneurysm. small size of the parent artery (less than 2 mm) and its critical role in supplying blood to the forearm. thrombosis and absence of flow was observed 3 months after the second injection (Figure 8). The baby’s forearm and hand func- tion remained normal. small size of the parent artery (less than 2 mm) and its critical role in supplying blood to the forearm. thrombosis and absence of flow was observed 3 months after the second injection (Figure 8). The baby’s forearm and hand func- tion remained normal. CASE REPORT Informed consent was obtained and the patient was anaesthe- tized. Two attempts at ultrasound-guided compression using a 12 Mhz linear probe with a large foot piece placed over the neck of the PSA for 10 min at each attempt were unsuccessful. Under sterile conditions and using sonographic guidance, 5 units of thrombin were injected with a 25-gauge needle into the sac of the PSA, (Figure 4) obtaining complete thrombosis (Figure 5); however, there was immediate compromise of the distal perfu- sion of the arm. Initial concern was for a thromboembolic event in the arteries of the forearm, although no clot migration was seen on ultrasound during the injection. Later, it became evident that it was owing to the hard mass effect due to the sudden clot- ting within the sac over the adjacent brachial artery. Intravenous heparin and topical nitroglycerin were used, resulting in total recovery of the extremity; however, the PSA reopened in 24 h (Figure 6). A second injection of 45 units of thrombin was given 7 days later, obtaining good initial results, but partial reopening was observed the following day (Figure 7), mainly in the area adjacent to the neck of the PSA. As there was no functional loss or enlargement of the lesion, conservative expectant manage- ment was then adopted. Follow-up with ultrasound imaging showed a persistent decrease in the size of the lesion. Complete Figure 3. CT angiogram confirmed the sonographic findings. (a) On this longitudinal MIP reformat the PSA is seen displacing the brachial artery (*). (b) Three-dimensional reformat of the lesion performed for surgical planning showing its relationship with the brachial artery () and the humerus (H). MIP, maxi- mum intensity projection; PSA, pseudoaneurysm. Figure 4. (a) Thrombin was injected into the lesion under real- time sonographic guidance and sterile conditions with a 25-gauge needle. (b) Transverse sonographic view showing the needle (arrow) positioned in the sac of the PSA prior to the injection. PSA, pseudoaneurysm. DISCUSSION PSAs are common vascular abnormalities secondary to a disrup- tion in arterial wall continuity.1 They can be secondary to differ- ent aetiologies, including iatrogenic causes.1 They can present as a pulsatile mass, and management by interventional radiology has replaced surgery as the first line of treatment.2 Brachial artery PSA is a rare condition in neonates.3 It has been associated with trauma to the artery during venipuctures.3,4 Its management is controversial,3–5 and based on case reports, surgery is the most commonly used treatment option.3–4 The surgical approach requires resection of the lesion with vascular grafting, end-to-end anastomosis or repair of the donor artery,4 and is asso- ciated with a very good success rate, which has improved with the continuous advancements in microsurgical techniques. Ultrasound-guided thrombin injection for the treatment of superficial PSA had been widely used with a success rate that ranges from 91% to 100%.6 Pezzullo and Wallach5 successfully treated a brachial PSA in a 3-week-old patient with 300 units of thrombin. In our case, surgery was not initially considered owing to the age and weight of the patient. Technically, it did not appear possible to use an endovascular approach owing to the size of Figure 3. CT angiogram confirmed the sonographic findings. (a) On this longitudinal MIP reformat the PSA is seen displacing the brachial artery (*). (b) Three-dimensional reformat of the lesion performed for surgical planning showing its relationship with the brachial artery () and the humerus (H). MIP, maxi- mum intensity projection; PSA, pseudoaneurysm. Figure 4. (a) Thrombin was injected into the lesion under real- time sonographic guidance and sterile conditions with a 25-gauge needle. (b) Transverse sonographic view showing the needle (arrow) positioned in the sac of the PSA prior to the injection. PSA, pseudoaneurysm. BJR Case Rep;2:20150383 2 of 4 birpublications.org/bjrcr Case report: Brachial pseudoaneurysm in a neonate BJR\|case reports Figure 5. Tranverse sonographic view of the pseudoaneurysm after injecting thrombin (5 units). It shows complete and abrupt cessation of flow in the colour Doppler image within the lesion owing to clot (C) formation. Figure 7. Colour Doppler study in a longitudinal plane after the second thrombin injection. It shows that the PSA reop- ened again with normal perfusion of the arm, normal flow through the brachial artery (**) and without loss of function. PSA, pseudoaneurysm. the parent artery. Therefore, our option was initially to attempt ultrasound-guided thrombin injection. While performing the first injection, we noted quick thrombosis with significant hardening of the lesion. We believe that this hardening increased the pressure in the compartment of the arm, with secondary compression of the brachial artery and decreased flow in the distal arm. This necessitated heparinization and use of topical nitroglycerin. With the second injection, we only obtained a partial response, probably secondary to the wide and short neck and the associated high intralesional blood flow. Expectant management was undertaken, waiting for the patient to gain weight for a surgical repair. However, the lesion resolved without any further interventions. It is matter of spec- ulation whether or not the occlusion may have been assisted by the partial intralesional thrombosis obtained with the throm- bin injections. Expectant management has been mentioned as a therapeutic option; however, it is rarely recommended.2 In conclusion, this case illustrates the use of a combination of ultrasound-guided thrombin injection and expectant manage- ment in managing a technically challenging PSA from interven- tional radiology (neck characteristics and size of the parent artery) and surgical (weight of the patient, size and location of the lesion) perspectives. In conclusion, this case illustrates the use of a combination of ultrasound-guided thrombin injection and expectant manage- ment in managing a technically challenging PSA from interven- tional radiology (neck characteristics and size of the parent artery) and surgical (weight of the patient, size and location of the lesion) perspectives. Figure 6. Colour Doppler study performed 24 h after the initial thrombin injection showing partial reopening of the lesion: col- our Doppler flow (F) around the clot (C). Figure 6. Colour Doppler study performed 24 h after the initial thrombin injection showing partial reopening of the lesion: col- our Doppler flow (F) around the clot (C). LEARNING POINTS 1. Brachial PSA in neonates is an uncommon condition, which has been associated with trauma to the artery during venipuncture. CONSENT 2. Its management can be challenging, especially if the PSA neck is wide. Informed consent was obtained from the parents of the patient for publication of this case report and unidentified pictures. This was recorded in the patient’s chart as per our hospital policy. 3. As treatment may risk the arterial supply to the arm, a multidisciplinary approach, on a case-to-case basis, is recommended. BJR\|case reports Parra et al 1. Brachial PSA in neonates is an uncommon condition, which has been associated with trauma to the artery during venipuncture. Figure 8. (a) Colour Doppler study on a transverse plane per- formed after 3 months of expectant management. It shows complete thrombosis (T) and absence of flow within the pseudoaneurysm. (b) Same finding as that in (a) on a longitu- dinal plane. In this image, good colour Doppler signal can be seen in the brachial artery (). 3 of 4 of 4 birpublications.org/bjrcr BJR Case Rep;2:20150383 REFERENCES 1. Saad NE, Saad WE, Davies MG, Waldman DL, Fultz PJ, Rubens DJ. Pseudoaneurysms and the role of minimally invasive techniques in their management. Radiographics 2005; 25 Suppl 1: S173–S189. doi: http://dx.doi.org/10.1148/rg. 25si055503 2. Keeling AN, McGrath FP, Lee MJ. Interventional radiology in the diagnosis, management, and follow-up of pseudoaneurysms. Cardiovasc Intervent Radiol 2009; 32: 2–18. doi: http://dx.doi.org/ 10.1007/s00270-008-9440-3 3. Landau D, Schreiber R, Szendro G, Golcman L. Brachial artery pseudoaneurysm in a premature infant. Arch Dis Child Fetal Neonatal Ed 2003; 88: F152–3. doi: http://dx.doi.org/10. 1136/fn.88.2.F152 4. Dzepina I, Unusic J, Mijatovic D, Bulic K. Pseudoaneurysms of the brachial artery following venipuncture in infants. Pediatr Surg Int 2004; 20: 594–7. doi: http://dx.doi. org/10.1007/s00383-004-1238-z 5. Pezzullo JA, Wallach MT. Successful percutaneous thrombin injection of a brachial artery pseudoaneurysm in a neonate. AJR Am J Roentgenol 2002; 178: 244–5. doi: http://dx. doi.org/10.2214/ajr.178.1.1780244 6. Middleton WD, Dasyam A, Teefey SA. Diagnosis and treatment of iatrogenic femoral artery pseudoaneurysms. Ultrasound Q 2005; 21: 3–17. 1. Saad NE, Saad WE, Davies MG, Waldman DL, Fultz PJ, Rubens DJ. Pseudoaneurysms and the role of minimally invasive techniques in their management. Radiographics 2005; 25 Suppl 1: S173–S189. doi: http://dx.doi.org/10.1148/rg. 25si055503 2. Keeling AN, McGrath FP, Lee MJ. Interventional radiology in the diagnosis, management, and follow-up of pseudoaneurysms. Cardiovasc Intervent Surg Int 2004; 20: 594–7. doi: http://dx.doi. org/10.1007/s00383-004-1238-z 5. Pezzullo JA, Wallach MT. Successful percutaneous thrombin injection of a brachial artery pseudoaneurysm in a neonate. AJR Am J Roentgenol 2002; 178: 244–5. doi: http://dx. doi.org/10.2214/ajr.178.1.1780244 6. Middleton WD, Dasyam A, Teefey SA. Diagnosis and treatment of iatrogenic femoral artery pseudoaneurysms. Ultrasound Q 2005; 21: 3–17. 4 of 4 birpublications.org/bjrcr 4 of 4 BJR Case Rep;2:20150383
https://openalex.org/W3058844134	https://link.springer.com/content/pdf/10.1007/s11071-020-05886-y.pdf	English	null	On the nonlinear effects of the mean wind force on the galloping onset in shallow cables	Nonlinear dynamics	2,020	cc-by	13,572	On the nonlinear effects of the mean wind force on the galloping onset in shallow cables Daniele Zulli · Giuseppe Piccardo · Angelo Luongo Received: 26 April 2020 / Accepted: 5 August 2020 / Published online: 18 August 2020 © The Author(s) 2020 Abstract The critical aeroelastic behavior of hori- zontal, suspended, and shallow cables is analyzed via a continuous model accounting for both external and internal damping. Quasi-steady aerodynamic forces are considered, including their stationary contribution (mean wind force). This latter induces a rotation of the cable (steady swing) around the line connecting the suspension points, together with a deformation of the initial equilibrium proﬁle under self-weight. First, by using perturbation methods, the nontrivial equilib- rium conﬁguration is analytically determined as a non- linear function of the wind velocity. Then, the wind critical values at which bifurcations take place and the corresponding modal shapes are determined by solv- ing a boundary value problem in the complex ﬁeld. Numerical investigations are carried out to validate the perturbation solution. A preliminary nonlinear gallop- ing analysis is also performed to verify the galloping onset in terms of non-trivial equilibrium path and crit- ical modes. The nonlinear terms related to the funda- mental path, from which bifurcations take place, play a keyrolerevealingnewinsights.Theyareabletoheavily inﬂuence the system bifurcation, making unstable con- ﬁgurations which instead would be aerodynamically stable without considering the effect of the mean wind force. Keywords Shallow cable · Galloping onset · Swing · Bifurcation from a nontrivial path · Perturbation solutions Nonlinear Dyn (2021) 103:3127–3148 https://doi.org/10.1007/s11071-020-05886-y ORIGINAL PAPER ORIGINAL PAPER D. Zulli · G. Piccardo (B) · A. Luongo M&MoCS - University of L’Aquila, Loc. Monteluco di Roio, 67100 L’Aquila, AQ, Italy e-mail: daniele.zulli@univaq.it G. Piccardo e-mail: giuseppe.piccardo@unige.it A. Luongo e-mail: angelo.luongo@univaq.it G. Piccardo DICCA - University of Genoa, Via Montallegro 1, 16145 Gen- ova, GE, Italy 1 Introduction In the case of inclined cables, additional properties that cannot be obtained by using horizontal cable modeling can be observed (e.g., [8]). ( g ) The wind-induced vibration of cables is a clas- sic topic in structural applications [9]. Various aeroe- lastic phenomena may arise including vortex-induced vibrations (e.g., [10,11]), rain-wind-induced vibrations (e.g., [12,13]), dry galloping (e.g., [14,15]), ice gallop- ing (e.g., [16,17]), and wake galloping (e.g., [18,19]). When operating in cold regions, an important role is played by ice galloping. It occurs when there is an ice accretion on the cable surface that breaks the symme- try of the cross section. As a consequence, the lift force may induce aerodynamic instabilities even for moder- atewind.Thisphenomenonisﬁrstdescribedin[20]and is largely analyzed in the literature, in the framework of the quasi-steady theory [21–23]. Without wanting to be exhaustive, a few examples of iced cable galloping are given in [16,24–26]. Effects of possible internal reso- nances, in discrete or continuous nonlinear models, are investigated by the authors in [17,27], whereas a stiff modeling, able to take into account in a consistent way both bending and torsional stiffness, are introduced in [28–30]. Concerning ﬂexible cables, the cable rotation, and resulting variation of their wind attitude are how- ever very important for changes in aerodynamic prop- erties of the cross section (e.g., [27,31]). Finite-element solutions of the problem can be found, for example, in [32]. The role of the internal and external damping con- tributions is evaluated in [33], where in-plane gallop- ing analysis of ﬂexible cables is performed, pursuing a direct approach on the nonlinear pde’s of motion. 1 Introduction Cables are widely used structures in civil and indus- trial applications. They assume a fundamental role in the carrying capacity of suspended and stayed bridges, as well as in the realization of cable-cars, electrical transmission lines and many other structures. A gen- eral characterization of statics and dynamics of cables can be found in the classic Irvine’s book [1], where sev- eral applications are analyzed as well. Recent advances in the statics of cables under forces exclusively verti- cal but not necessary uniform are proposed in [2], then extended to the case of general 3D forces in [3], where a perturbation asymptotic expression for the elastic prob- lemsolutionis proposed. At thesametime, thedynamic characterization of suspension cable systems is very important (e.g., [4]). Concerning nonlinear vibrations of suspended cables, the quadratic terms play a fun- damental role in presence of internal resonances (e.g., D. Zulli · G. Piccardo (B) · A. Luongo M&MoCS - University of L’Aquila, Loc. Monteluco di Roio, 67100 L’Aquila, AQ, Italy e-mail: daniele.zulli@univaq.it G. Piccardo e-mail: giuseppe.piccardo@unige.it A. Luongo e-mail: angelo.luongo@univaq.it G. Piccardo DICCA - University of Genoa, Via Montallegro 1, 16145 Gen- ova, GE, Italy 3 3 3128 D. Zulli et al. In this paper, considering a perfectly ﬂexible (i.e., with evanescent ﬂexural stiffness) and torsionally rigid cable arranged horizontally, the inﬂuence of mean wind actions on ice galloping stability is carefully addressed in the context of a continuous approach. For this pur- pose, the structural model of cable is taken from [33], and here extended to possible occurrence of out-of- plane motions. Due to the acting forces, a conﬁguration change of the static equilibrium of the cable occurs and bifurcations take place starting from a nonlinear, non- trivial fundamental path. After modeling the aeroelastic ﬂexible cable (Sect. 2), the aim is, therefore, twofold: (1) to ﬁnd the nontrivial path as a function of the mean wind velocity (Sect. 3), and, along this path, (2) to search for the critical wind velocities (Sect. 4). Both of these objectives will be pursued analytically with the use of appropriate perturbation techniques. Compar- isons with a numerical solution are however performed (Sect. 5) in both linear and nonlinear ﬁeld in order to verify the accuracy of the results obtained. Some ﬁnal considerations are reported in the ending Sect. 6. [5,6]),differentlyfromtautstringsforwhichonlycubic terms are present (e.g., [7]). On the nonlinear effects of the mean wind force... 3129 Fig. 1 Horizontal cable under self-weight and binormal wind ﬂow Fig. 1 Horizontal cable under self-weight and binormal wind ﬂow Fig. 1 Horizontal cable under self-weight and binormal wind ﬂow Fig. 1 Horizontal cable under self-weight and binormal wind ﬂow ciated with the transversal ones. Moreover, bending and torsional stiffness is ignored, consistently with the model of ﬂexible cables, and differently than what it is done in case of stiff cables [28–30]. When adapted to the problem at hand, the equations of motion read: while here the procedure is extended to possible out- of-plane motions as well). In particular, both external and internal damping contributions are considered. The latter is essential to avoid, in a continuum approach, inﬁnitely many coincident critical wind velocities, one for each mode and, therefore, to split the coalescence of bifurcation points. (T0 + E Ae) v′′ + E A¯ke −m ¨v + f d n + f a n = 0 (T0 + E Ae) w′′ −m ¨w + f d b + f a b = 0 e + ¯k l l 0 vds −1 2l l 0 v′2 + w′2 ds = 0 vA = 0, vB = 0 wA = 0, wB = 0 Speciﬁcally,theexternaldampingisassumedaspro- portional to velocities: f de n = −ce ˙v, f de b = −ce ˙w, with ce a damping coefﬁcient. This leads to an exter- nal damping operator proportional to the mass opera- tor. Conversely, the internal damping, which accounts for various dissipative phenomena occurring in the material, here is assumed as proportional to the stiff- ness operator, to make the analysis as simple as pos- sible (see [33] for further comments on the internal damping contribution). Therefore, since from Eq. On the nonlinear effects of the mean wind force... (1), the geometric and elastic linear force components are: (1) Here, T0 is the prestress, assumed uniform along the abscissa s ∈[0,l]; l is the length of the cable, taken nearly equal to the chord (i.e., the distance between the supports); E A is the axial stiffness; m is the mass per unit length (including possible ice coating); ¯k := mg T0 is the prestress curvature, also assumed uniform; e(t) is the dynamic unit extension, uniform on s, from which the dynamic tension is evaluated as ˜T = E Ae; f d n , f d b are damping forces and f a n , f a b aerodynamic forces, all per unit length, acting in the normal and binormal direction, respectively; the prime stands for s-derivative and the dot for time derivative. Here, T0 is the prestress, assumed uniform along the abscissa s ∈[0,l]; l is the length of the cable, taken nearly equal to the chord (i.e., the distance between the supports); E A is the axial stiffness; m is the mass per unit length (including possible ice coating); ¯k := mg T0 is the prestress curvature, also assumed uniform; e(t) is the dynamic unit extension, uniform on s, from which the dynamic tension is evaluated as ˜T = E Ae; f d n , f d b are damping forces and f a n , f a b aerodynamic forces, all per unit length, acting in the normal and binormal direction, respectively; the prime stands for s-derivative and the dot for time derivative. f el n f el b := T0 ∂2 ∂s2 v w + E A T0 ¯ke 0 (2) (2) then the damping force components become: then the damping force components become: f di n f di b := ζ ∂2 ∂s2 ˙v ˙w + E A T0 ¯k ˙e 0 (3) (3) Models for dampingandaerodynamicforces aredis- cussed below. with ζ an internal damping coefﬁcient. with ζ an internal damping coefﬁcient. By superimposing the internal and external mechanisms, the Rayleigh model of damping is obtained: 2.1 Equations of motion A horizontal ﬂexible cable, having small sag-to-span ratio and under the effect of gravity and uniform wind, is analyzed here. As long as the sole self- weight is considered, i.e., no wind blows, the cable hangs in the ¯ax, ¯ay -plane (vertical), in the con- ﬁguration referred to as ¯C (see Fig. 1, solid thin line). A local triad is deﬁned in such a conﬁgura- tion, constituted by the tangent, normal and binor- mal unit vectors, and denoted by (¯at, ¯an, ¯ab). When a uniform wind of velocity U = U ¯az is consid- ered as well, it induces a time-dependent displace- ment of the generic point of the center line of the cable from the conﬁguration ¯C, which is indicated as u = u¯at + v¯an + w¯ab (solid thick line in Fig. 1); the speciﬁc effect of the wind is described in detail in Sect. 2.3. Focusing the attention on galloping critical condi- tions, the 1:1 internal resonance effects are pointed out in [16,34,35] as concerns two degree-of-freedom mod- els.Averyrecentpaper[36]presentsaperturbationgal- loping stability criterion for a three-degree-of-freedom coupled motion of an iced conductor. A study of mul- timodal galloping on an iced transmission line is car- ried out in [37] using a reduced model including the ﬁrst four in-plane modes and the ﬁrst torsional mode, which is obtained by using Galerkin spatial discretiza- tion. Critical conditions and nonlinear responses are determined in a purely numerical way; the actual role of the rotation caused by the static aerodynamic forces is not completely unveiled since it seems related only to the torsional mode considered. The equations of motion for the cable, which are well-known in the literature [1,5,27,38], are written here in terms of normal and binormal displacement only (v, w), i.e., with the tangent displacement (u) con- densed, under the usual hypothesis that the celerity of the longitudinal waves is much greater than that asso- 123 123 2.2 Internal and external damping forces The forces are assumed uniform along the centerline of the cable, due to its small sag-to-span ratio (¯an ≃¯ay) and to the uniform shape of the ice coating. Ur = U −˙u = −˙v¯an + (U −˙w)¯az (7) (7) In Eq. (7) the contribution of the spin ˙θ is ignored, as typically done for compact shapes, it being small as compared to ˙u. As a consequence, it turns out that: In Eq. (7) the contribution of the spin ˙θ is ignored, as In Eq. (7) the contribution of the spin ˙θ is ignored, as typically done for compact shapes, it being small as compared to ˙u. As a consequence, it turns out that: typically done for compact shapes, it being small as compared to ˙u. As a consequence, it turns out that: Ur = ˙v2 + (U −˙w)2 aur = −˙v Ur ¯an + U −˙w Ur ¯az (8) As previously said, the gravity force ¯b = −mg¯ay (g is the gravity acceleration), which includes the own and the ice weights, lets the cable lie in the vertical plane (conﬁguration ¯C, light gray plane in Fig. 2a). The initial orientation of the iced cross section is described by the direction ¯as, which is inclined by the angle ϕ0 to the binormal axis ¯ab. In other words, ϕ0 represents the initial orientation of the cross section when the cable hangs on the vertical plane, in the absence of wind (Fig. 2b). (8) In Eq. (6), Cd(γ + ϕ0) and Cl(γ + ϕ0) are the drag and lift coefﬁcients, neglecting their intrinsic uncer- tainty [39]; note that, if the angle of attack γ = 0, i.e., if the cross section is in its ﬁxed initial orientation, the face exposed to the wind would be described by the angle ϕ0. With reference to Fig. 2b, the expression for γ is evaluated from: The wind, which blows normally to the cable plane, produces aerodynamic forces which can be decom- posed in steady plus unsteady parts. The steady part, i.e., the one not depending on the velocity of the body, can be combined to the gravity giving rise to a uni- form and constant force ﬁeld ˘b, that results parallel to an inclined plane and modiﬁes the equilibrium conﬁg- uration. 2.2 Internal and external damping forces f d n = −ce ˙v + ζ ˙v′′ + E A T0 ¯k ˙e f d b = −ce ˙w + ζ ˙w′′ (4) The damping scheme follows what it is proposed in [33], in the spirit of the Kelvin-Voigt rheological model (note that, in [33], only in-plane motions are involved (4) 123 3 D. Zulli et al. 3130 The equations of motion (1) accordingly read: referred to as ˘C (dark gray plane in Fig. 2a). The rotated plane will be called swung plane in the following. As speciﬁed in Sect. 2.1, and contrary to what it is done in [30], the vertical planar conﬁguration ¯C is taken here as reference, and displacements are measured from it. T0 1 + ζ T0 ∂ ∂t v′′ + E A¯k 1 + ζ T0 ∂ ∂t e + E Aev′′ −m ¨v −ce ˙v + f a n = 0 T0 1 + ζ T0 ∂ ∂t w′′ + E Aew′′ −m ¨w −ce ˙w + f a b = 0 e+ ¯k l l 0 vds −1 2l l 0 v′2 + w′2 ds = 0 vA = 0, vB = 0 wA = 0, wB = 0 Under the aforementioned hypotheses, the aerody- namic force acting on the cross section, which is the sum of a drag (fd) and a lift (fl) component, is (see Fig. 2b): (5) fa = 1 2ρU2 r bCd(γ + ϕ0)aur fd + 1 2ρU2 r bCl(γ + ϕ0)¯at × aur fl (6) (6) where ρ is the air density, b a characteristic radius of the cross section, × is the cross product, Ur = Uraur the relative velocity between the ﬂow and the cable, deﬁned as where ρ is the air density, b a characteristic radius of the cross section, × is the cross product, Ur = Uraur the relative velocity between the ﬂow and the cable, deﬁned as 2.3 Aerodynamic forces 2.3 Aerodynamic forces 2.3 Aerodynamic forces Aerodynamic forces are formulated in the framework of the quasi-steady theory (e.g., [21,22]). They are inﬂuenced by the occurrence, at any point, of an ice coating which breaks the axis-symmetry of the circu- lar cross section of the cable. 2.2 Internal and external damping forces More speciﬁcally, a static swing ˘θ(U) of the cable is induced, i.e., the cable swings to this new plane, which still passes through the line connecting the sup- ports but is inclined of the angle ˘θ(U) with respect to the vertical one. This new equilibrium conﬁguration is tan( ˘θ + γ ) = −Ur · ¯an Ur · ¯az (9) (9) where the dot indicates the scalar product, so that, from Eqs. (7) it turns out: γ = −˘θ + arctan ˙v U −˙w (10) (10) From Eq. (10) (and from Fig. 2b), it is evident the con- tribution of the swing angle ˘θ which, summed to the 123 On the nonlinear effects of the mean wind force... 3131 (a) (b) Fig. 2 a Swing of the cable plane under gravity and steady part of the binormal wind ﬂow (light gray: vertical plane, conﬁgura- tion ¯C; dark gray: swung plane, conﬁguration ˘C); b Cross section of the cable with drag and lift components of the aerodynamic forces ((˘an, ˘ab) are normal and binormal unit vectors in the con- ﬁguration ˘C, and ¯at is orthogonal to the plane of the picture, coming out of it) (a) Fig. 2 a Swing of the cable plane under gravity and steady part of the binormal wind ﬂow (light gray: vertical plane, conﬁgura- tion ¯C; dark gray: swung plane, conﬁguration ˘C); b Cross section of the cable with drag and lift components of the aerodynamic forces ((a ﬁguratio coming o (a) (b) (a) (b) (b) forces ((˘an, ˘ab) are normal and binormal unit vectors in the con- ﬁguration ˘C, and ¯at is orthogonal to the plane of the picture, coming out of it) (b) (a) Fig. 2 a Swing of the cable plane under gravity and steady part of the binormal wind ﬂow (light gray: vertical plane, conﬁgura- tion ¯C; dark gray: swung plane, conﬁguration ˘C); b Cross section of the cable with drag and lift components of the aerodynamic forces ((˘an, ˘ab) are normal and binormal unit vectors in the con- ﬁguration ˘C, and ¯at is orthogonal to the plane of the picture, coming out of it) initial orientation of the cross section ϕ0, changes the face of the cross section to the wind. 2.2 Internal and external damping forces where the terms uniquely depending on ˘θ are collected in ˘fα, while those which depend on the velocity com- ponents ˙v, ˙w as well, are collected in ˜fα, α = n, b. The former are the steady components, depending on theswingangle(and,parametrically,onthewindveloc- ity); the latter are the non-steady contributions, depend- ing on structural velocities (and, parametrically, on the swingangleandwindvelocity).Inparticular,thesteady forces are: When the previous results are expanded for small velocity ratios ˙v/U, ˙w/U up to the ﬁrst order (i.e., considering a purely linear contribution from aeroelas- tic actions), it is found: Ur = U −˙w (11) aur = −˙v U ¯an + ¯az (12) γ = −˘θ + ˙v U (13) Ur = U −˙w (11) aur = −˙v U ¯an + ¯az (12) γ = −˘θ + ˙v U (13) (11) (12) ˘fn ˘fb : = f a n0 f a b0 U 2 − ka 21 ka 31 U 2 ˘θ − n211 n311 U 2 ˘θ2 − n2111 n3111 U 2 ˘θ3 (15) (13) (15) Consistently with the mechanical model of ﬂexi- ble cable, twist and torsional moment are not deﬁned. From Eqs. 2.2 Internal and external damping forces (6) and (11)–(13), after series expansion with respect to the variables ˘θ, ˙v/U, ˙w/U, the normal and binormal components of the vector of aerodynamic forces, referred to as f a n and f a b , respectively, become: and the non-steady ones are: and the non-steady ones are: and the non-steady ones are: ˜fn ˜f a b := − U ca 22 ca 23 ca 32 ca 33 + ˘θU n212 n213 n312 n313 + ˘θ2U n2112 n2113 n3112 n3113 ˙v ˙w (16) f a n f a b := ˘fn( ˘θ;U) ˘fb( ˘θ;U) + ˜fn(˙v, ˙w; ˘θ,U) ˜fb(˙v, ˙w; ˘θ,U) (14) ˜fn ˜f a b := − U ca 22 ca 23 ca 32 ca 33 + ˘θU n212 n213 n312 n313 + ˘θ2U n2112 n2113 n3112 n3113 ˙v ˙w (16) (16) f a n f a b := ˘fn( ˘θ;U) ˘fb( ˘θ;U) + ˜fn(˙v, ˙w; ˘θ,U) ˜fb(˙v, ˙w; ˘θ,U) (14) fb c32 c33 n312 n313 + ˘θ2U n2112 n2113 n3112 n3113 ˙v ˙w (16) f a n f a b := ˘fn( ˘θ;U) ˘fb( ˘θ;U) + ˜fn(˙v, ˙w; ˘θ,U) ˜fb(˙v, ˙w; ˘θ,U) f a n f a b := ˘fn( ˘θ;U) ˘fb( ˘θ;U) + ˜fn(˙v, ˙w; ˘θ,U) ˜fb(˙v, ˙w; ˘θ,U) (14) b (14) 123 123 3132 D. Zulli et al. Fig. 3 Equilibrium of the steady forces in the (¯ay, ¯az)-plane and determination of the swing angle ˘θ Fig. 3 Equilibrium of the steady forces in the (¯ay, ¯az)-plane and determination of the swing angle ˘θ Fig. 3 Equilibrium of the steady forces in the (¯ay, ¯az)-plane and determination of the swing angle ˘θ The coefﬁcients in Eqs. (15) and (16) are given in Appendix A. The expression (16) of the aerodynamic forces, as a function of the swing angle, is asymptot- ically correct till ˘θ is small of the same order of the (nondimensionalized) velocities ˙v/U, ˙w/U. Series expansions (15) and (16) are expected to be sufﬁciently accurate as long as ˘θ is small. However, when ˘θ is large, a different approach must be followed, namely: (a) the exact (not expanded) expression of the steady forces are evaluated, with Eqs. 2.2 Internal and external damping forces (7)-(10), from Eq.(6),afterneglectingallthestructuralvelocityterms; (b) the non-steady forces are determined by a series expansion carried out on the structural velocities only, and not on the swing angle. In this case, the steady forces, which substitute Eqs. (15), become: In order to evaluate the equilibrium path, that is the steady solution ˘v(s), ˘w(s), ˘e as a function of the wind velocityU, Eqs. (19) are solved. Due to the dependence of the steady aerodynamic forces on ˘θ, a further equa- tion which states that the cable lies on the swung plane at equilibrium, i.e., relating ˘v(s) and ˘w(s) to ˘θ, should be considered. ˘fn ˘fb := −1 2ρU 2bCl(−˘θ + ϕ0) 1 2ρU 2bCd(−˘θ + ϕ0) (17) (17) However, as more convenient alternative, the swing angle can be ﬁrst evaluated through a physical consider- ation on forces, and then, once the steady aerodynamic forces are known, Eqs. (19) are taken on, in order to obtain the displacement and strain. In particular, at the equilibrium, the total steady force, uniformly acting on the cable, is ˘b = ( ˘fn −mg)¯ay + ˘fb ¯az (see Fig. 3), and this induces the cable to assume the planar parabolic conﬁguration ˘C in the swung plane, whose trigonomet- ric tangent is the ratio between the two force compo- nents, namely: They have the important drawback of not being explicit in ˘θ and U, thus calling for a purely numeric approach. The unsteady forces, which substitute Eqs. (16), are: ˜fn ˜f a b := 1 2ρUb∗ −(Cd(−˘θ + ϕ0) + C′ l(−˘θ + ϕ0)) 2Cl(−˘θ + ϕ0) −(Cl(−˘θ + ϕ0) −C′ d(−˘θ + ϕ0)) −2Cd(−˘θ + ϕ0) ˙v ˙w ˜fn ˜f a b := 1 2ρUb∗ (18) ˘θ = arctan ˘fb ˘fn −mg (20) (20) 2.4 The equilibrium equations Fig. 3 Equilibrium of the steady forces in the (¯ay, ¯az)-plane and determination of the swing angle ˘θ 2.5 The linear incremental equations Two main approaches can be possibly followed to get to the ﬁnal goal, which is the evaluation of the bifur- cation conditions of Eqs. (22). The ﬁrst is valid for large ˘θ and it is purely numerical using the following steps: 1) pick of a wind velocity U; 2) evaluation of the swing angle solving Eq. (20), by means of the Newton- Raphsonmethod,usingtheexactandimplicitdeﬁnition of the forces (17); 3) evaluation of the equilibrium solu- tion solving Eqs. (19) by means of the ﬁnite difference method; 4) solution of the inﬁnite dimensional eigen- value problem (22), with deﬁnition (18) for the forces and still using ﬁnite difference method. This approach requires a parametric sweep in terms of the wind veloc- ity U, i.e., the procedure is entirely repeated for differ- ent values of U until critical conditions are found. The whole procedure is implemented in MATLAB c⃝[40], and requires a very short computational time (a few seconds) for each wind velocity. The general dynamic response to Eqs. (1) can be expressed in incremental form from equilibrium, by splitting the solution as: v(s, t) :=˘v(s) + ˜v(s, t) w(s, t) := ˘w(s) + ˜w(s, t) e (t) :=˘e + ˜e (t) (21) (21) where ˜v(s, t), ˜w(s, t), ˜e (t) are incremental variables. By substituting Eqs. (21) in Eqs. (5), accounting for Eqs. (19) and linearizing in the increments, the lin- ear incremental equations of motion are obtained (tilde omitted on the incremental variables): where ˜v(s, t), ˜w(s, t), ˜e (t) are incremental variables. By substituting Eqs. (21) in Eqs. (5), accounting for Eqs. (19) and linearizing in the increments, the lin- ear incremental equations of motion are obtained (tilde omitted on the incremental variables): (T0 + ˘T ) 1 + ζ T0 ∂ ∂t v′′ w′′ + E A ¯k + ˘v′′ ˘w′′ 1 + ζ T0 ∂ ∂t e −ce ˙v ˙w −m ¨v ¨w + ˜f a n ˜f a b = 0 0 e = −1 l l 0 ˘v′′v + ˘w′′w + ¯kv ds vA = 0, vB = 0 wA = 0, wB = 0 (22) The second approach is analytical and makes use of perturbation methods; it is valid for moderately small ˘θ and entails: 1) perturbation expansion and chain- solving of Eq. 3 The equilibrium path l 0 ˘v′v′ + ˘w′w′ ds = − l 0 ˘v′′v + ˘w′′w ds + ˘v′v + ˘w′w l 0 (23) 2.5 The linear incremental equations (20) to obtain the swing angle as a function of the wind velocity, using the approximated expression of forces in Eqs. (15); 2) perturbation solv- ing of Eqs. (19) to obtain the equilibrium solution; 3) analytical seeking of the critical solution of the linear problem (22), with deﬁnition (16) for the aerodynamic forces. (22) e = −1 l l 0 ˘v′′v + ˘w′′w + ¯kv ds In the following Sects. 3 and 4, the second, ana- lytical procedure is described in detail, and ﬁnally, in Sect. 5, outcomes are compared to those given by the ﬁrst, numeric, procedure. where the steady tensile force of the cable is ˘T = E A˘e. Note that, in writing Eq. (22-b), an integration by parts is performed, i.e., 2.4 The equilibrium equations Equation (20), which is transcendent in ˘θ, once solved in terms of the parameter U, allows one to evaluate the steady aerodynamic forces. Since it must be ˘fn −mg < 0 as the cable remains in tension, it is ˘θ < 0 in the velocity ﬁeld of interest. The nontrivial equilibrium conﬁguration in the swung plane is described by the steady response, indicated as ˘v(s), ˘w(s), ˘e. This is the solution of the nonlinear equi- librium equations, which are obtained from Eqs. (1) when inertia, damping and non-steady components of the aerodynamic forces are neglected, namely: Note that, in the swung plane, the intensity of the normal load per unit length changes from mg to ( ˘fn −mg)2 + ˘f 2 b . The difference of load entails elastic strain in the cable, which modiﬁes its sag (the proﬁle still remaining parabolic) and its stress. Conse- quently, the dynamic characteristics of the swung cable modify, with respect to those originally owned in the vertical plane. T0 ˘v′′ ˘w′′ + E A˘e ¯k + ˘v′′ ˘w′′ + ˘fn ˘fb = 0 0 ˘e = − ¯k l l 0 ˘vds + 1 2l l 0 ˘v′2 + ˘w′2 ds ˘vA = 0, ˘vB = 0 ˘wA = 0, ˘wB = 0 (19) (19) As a further comment on Eq. (20), it turns out that the value of the swing angle is independent of the initial curvature of the cable. 123 3133 On the nonlinear effects of the mean wind force... 2.5 The linear incremental equations 2.5 The linear incremental equations 3.2 Evaluation of the static displacements (19-a,b), the following nonlinear algebraic system in αn, αb, ˘e is obtained: αn αb + E A T0 ˘e αn −1 αb = ⎛ ⎝ ˘fn(U) mg ˘fb(U) mg ⎞ ⎠ ˘e = 1 12 ¯kl 2 −αn + 1 2 α2 n + α2 b (30) ( g) (26) (26) (26) from which ˘θ(U) is reconstituted as: αb T0 αb ⎝ fb(U) mg ⎠ ˘e = 1 12 ¯kl 2 −αn + 1 2 α2 n + α2 b (30) from which ˘θ(U) is reconstituted as: ⎝ mg ⎠ ˘e = 1 12 ¯kl 2 −αn + 1 2 α2 n + α2 b ( from which ˘θ(U) is reconstituted as: ˘e 1 ¯kl 2 from which ˘θ(U) is reconstituted as: ˘θ = − f a b0 mg U2 − f a b0(ka 31 + f a n0) (mg)2 U4 + f a b0[( f a b0)2 −3( f a b0 + ka 31)2 −3 f a b0(ka 21 −n311)] 3(mg)3 U6 (27) 12 2 After eliminating ˘e, the system reduces to: Λ2 1 2 2 ( 1) ˘θ = − f a b0 mg U2 − f a b0(ka 31 + f a n0) (mg)2 U4 + f a b0[( f a b0)2 −3( f a b0 + ka 31)2 −3 f a b0(ka 21 −n311)] 3(mg)3 U6 After eliminating ˘e, the system reduces to: (27) (27) αn + Λ2 12 −αn + 1 2 α2 n + α2 b (αn −1) = ˘fn(U) mg αb + Λ2 12 −αn + 1 2 α2 n + α2 b αb = ˘fb(U) mg (31) Withthis result, thesteadyforces canbeexpressedin termsofthewindvelocityonly.BysubstitutingEq.(27) in Eq. (15), one ﬁnally gets: (31) ˘fn(U) ˘fb(U) = f a n0 f a b0 U 2 − ka 21 ka 31 ( ˘θ1(U) + ˘θ2(U) + ˘θ3(U))U 2 − n211 n311 U 2( ˘θ2 1 (U) + 2 ˘θ1(U) ˘θ2(U))2 − n2111 n3111 U 3 ˘θ3 1 (U) + h.o.t. where Λ2 := E A T0 ¯kl 2 is the Irvine parameter [1]. where Λ2 := E A T0 ¯kl 2 is the Irvine parameter [1]. 3.1 Evaluation of the swing angle The equilibrium equation (20), together with the deﬁ- nitions (15) of the steady forces, implicitly determines ˘θ as function of U. It reads: (23) (mg −˘fn( ˘θ;U)) tan ˘θ + ˘fb( ˘θ;U) = 0 (24) (24) where the boundary terms go to zero. Equations (22) govern the small oscillations of the cable around the nontrivial equilibrium conﬁguration, with the incre- mental displacement components referred to the orig- inal basis located in the vertical plane and the aerody- namic forces deﬁned in Eq. (16). Since the forces are expressed in polynomial form (Eqs. (15)), a perturbation method is conveniently applied. By rescaling the forces at order ϵ (i.e., by let- ting ˘fα →ϵ ˘fα, so that ˘θ →0 when ϵ →0), expand- ing tan ˘θ = ˘θ + ˘θ3 3 +. . . and using the series expansion 12 3134 D. Zulli et al. ˘θ = ϵ ˘θ1 +ϵ2 ˘θ2 +ϵ3 ˘θ3, the substitution in Eq. (24) and vanishing of the coefﬁcients of different powers of ϵ produces the following perturbation equations: 3.2 Evaluation of the static displacements To ﬁnd the steady response of the cable, accounting for elasticity, the equilibrium equations (19) must be solved. The following trial solution is used: ϵ1 : mg ˘θ1 = −f a b0U 2 ϵ2 : mg ˘θ2 = ka 31 + f a n0 U 2 ˘θ1 ϵ3 : mg ˘θ3 = (ka 31 + f a n0)U 2 ˘θ2 + (n311 −ka 21)U 2 ˘θ2 1 −1 3mg ˘θ3 1 (25) ˘v ˘w = −y(s) αn αb (29) (29) (25) where y(s) = ¯k 2(s −l)s is the parabolic initial proﬁle, with ¯k = mg T0 , and αn, αb are (small) nondimensional amplitudes, to be determined. Since yA = yB = 0, the boundary conditions in Eqs. (19) are satisﬁed. Chain solution leads to: Chain solution leads to: Chain solution leads to: Chain solution leads to: ˘θ1 = − f a b0 mg U2 ˘θ2 = − f a b0(ka 31 + f a n0) (mg)2 U4 ˘θ3 = f a b0[( f a b0)2 −3( f a b0 + ka 31)2 −3 f a b0(ka 21 −n311)] 3(mg)3 U6 By substituting Eq. (29) into Eqs. 3.2 Evaluation of the static displacements (30-c); the associated increment of tension ˘T = E A˘e is: ˘T = T0 Λ2 12 −αn + 1 2 α2 n + α2 b (35) (35) ϵ1 : 1 + Λ2 12 αn1 = ˘fn(U) mg αb1 = ˘fb(U) mg ϵ2 : 1 + Λ2 12 αn2 = Λ2 24 3α2 n1 + α2 b1 αb2 = Λ2 12 αn1αb1 ϵ3 : 1 + Λ2 12 αn3 = Λ2 12 αb1αb2 + 1 2(αn1α2 b1 + α3 n1 −6αn1αn2) αb3 = Λ2 12 αn1αb2 + αn2αb1 −1 2 α2 n1 + α2 b1 αb1 (33) ϵ1 : 1 + Λ2 12 αn1 = ˘fn(U) mg αb1 = ˘fb(U) mg It is worth noticing from Eqs. (33)-(35) that, when Λ2 →∞(inextensible cable) the linear part of the normal displacement ampliture αn1 disappears, as it happens in an inextensible pendulum performing small swing; furthermore the (nonlinear) elastic strain tends to zero, in a perturbation sense, when Λ2 →∞(i.e., it contains terms of order ϵ4 if the perturbation procedure is stopped at order ϵ3), entailing no proﬁle changes, while ˘T T0 → ˘fn mg, i.e., the tension ratio equates the load ratio; therefore a static effect only occurs. On the other hand, when ¯k →0 (taut string), Λ2 →0 too, so that a deﬂection takes place (αn ̸= 0), while ˘T →0, i.e., kinematic effect only occurs. ϵ2 : 1 + Λ2 12 αn2 = Λ2 24 3α2 n1 + α2 b1 αb2 = Λ2 12 αn1αb1 : 1 + Λ2 12 αn3 = Λ2 12 αb1αb2 + 1 2(αn1α2 b1 + α3 n1 −6αn1αn2) αb3 = Λ2 12 αn1αb2 + αn2αb1 −1 2 α2 n1 + α2 b1 αb1 ϵ3 : 3.2 Evaluation of the static displacements T0 It is worth noticing that, as a check for the validity of the trial solution (29), since both ˘v(s), ˘w(s) are pro- portional to y(s), this solution keeps the cable planar. Indeed, denoting by x(s) := (x(s)+ ˘u(s))¯ax +(y(s)+ ˘v(s))¯ay + ˘w(s)¯az the position of a generic point in the swung conﬁguration, the vector n := ¯ax × x(s) = y(s)[(1−αn)¯az +αb ¯ay] keeps its direction constant for any s, coincident with the normal to the swung plane. The rotation of the plane is the angle formed by n and ¯az, i.e.: (28) (28) where the positions (26) hold and h.o.t. stands for higher order terms. 123 On the nonlinear effects of the mean wind force... 3135 αb = ˘fb(U) 1 mg + ˘fn(U)Λ2 (Λ2 + 12)(mg)2 − Λ2 2 Λ2 + 12 3 (mg)3 ∗ ˘fb(U)2 Λ2 + 12 2 −2 ˘fn(U)2 Λ4 + 24Λ2 −72 (34) ˘θ = −arctan αb 1 −αn (32) (32) Moreover, by combining Eqs. (31), it follows that αb 1−αn = ˘fb mg−˘fn , so that Eqs. (32) and (20) are con- sistent. Therefore, ﬁnding αb, αn and using Eq. (32) gives an alternative method (to Eq. (24)) to evaluate ˘θ. (34) Substitution of this results in Eq. (29) provides the static deﬂection of the cable, including rigid and elastic effects. The solution of Eqs. (31) is now sought with a per- turbation method. By rescaling the forces at order-ϵ and expanding the unknowns as αh = ϵαh1 + ϵ2αh2 + ϵ3αh3, substitution in Eqs. (31) and vanishing the coef- ﬁcients of equal powers of ϵ gives the following per- turbation equations: Concerning the increment of strain ˘e, this can be evaluated by Eq. 4 The critical wind velocity To compute the critical wind velocity and the corre- sponding critical mode, the incremental equations of motion (22) must be considered. By using Eq. (16) for the expression of the aerodynamic forces and Eqs. (29) for the equilibrium solution, they become: (33) After chain-solving the linear systems (33), and recon- stituting, the following expressions are obtained: T0 + ˘T 1 + ζ T0 ∂ ∂t v′′ w′′ + E A¯k 1 −αn −αb 1 + ζ T0 ∂ ∂t e − U ca 22 ca 23 ca 32 ca 33 + U ˘θ n212 n213 n312 n313 + U ˘θ2 n2112 n2113 n3112 n3113 ˙v ˙w −m ¨v ¨w −ce ˙v ˙w = 0 0 αn = 1 2 Λ2 + 12 5 (mg)3 2 ˘fb(U)2 ˘fn(U)Λ2 Λ4 +24Λ2 −72 Λ2 + 12 2 + ˘fb(U)2mgΛ2 Λ2 + 12 4 +6912 ˘fn(U)3Λ2 2Λ2 −3 +432 ˘fn(U)2mgΛ2 Λ2 + 12 2 +24 ˘fn(U) Λ2 + 12 4 (mg)2 123 123 D. Zulli et al. 3136 e = ¯k l l 0 [(1 −αn)v −αbw] ds vA = 0, vB = 0 wA = 0, wB = 0 (36) ﬁeld equation assumes the fo ˆu = (C1 exp (β1(s −l)) + C + (C3 exp (β2(s −l)) + e = ¯k l l 0 [(1 −αn)v −αbw] ds vA = 0, vB = 0 wA = 0, wB = 0 (36) ﬁeld equation assumes the form: ˆu = (C1 exp (β1(s −l)) + C2 exp (−β1s)) φ1 + (C3 exp (β2(s −l)) + C4 exp(−β2s)) φ2 e = ¯k l l 0 [(1 −αn)v −αbw] ds vA = 0, vB = 0 wA = 0, wB = 0 (36) ) ] d ﬁeld equation assumes the form: ﬁeld equation assumes the form: ﬁeld equation assumes the form: ˆu = (C1 exp (β1(s −l)) + C2 exp (−β1s)) φ1 + (C3 exp (β2(s −l)) + C4 exp(−β2s)) φ2 + ˆe ˆu⋆ (41) (36) (41) It is worth noticing that, in addition to the effects dis- cussed in the previous section, the change in the equi- librium conﬁguration (swing) induces a change also in the aerodynamic damping matrix, thus accounting for the modiﬁed exposure to wind of the cross section. 4 The critical wind velocity Fur- thermore, it is important to point out that the analytical solving of the problem (36) is possible only as conse- quence of the evaluation of the non-trivial equilibrium path using the proposed perturbation method, as carried out in the previous Section. where β1, β2, φ1, φ2 are the eigenvalues and eigenvec- tors of the following 2 × 2 algebraic problem: B (λ,U) + β2 A (λ) φ = 0 (42) (42) Moreover ˆu⋆:= −B−1 (λ,U) b (λ) is a particular solution to the non-homogeneous problem (consider- ing ˆe as a known term), and C1, . . . , C4 are arbitrary constants. By substituting the components of ˆu in the equation for ˆe (38), this latter is drawn in terms of the arbitrary constants. Finally, from the boundary condi- tions (39), a 4 × 4 homogeneous algebraic problem follows. Explicit expressions of the solution are cum- bersome, but help of an algebraic manipulator makes the problem easy to be solved. In the critical condition, it is λ = iωc, U = Uc; by splitting the characteristic equations in the real and imaginary parts, two transcen- dent real equations for ωc,Uc, which are the critical frequency and wind velocity, respectively, are derived and solved by the Newton-Raphson method. To solve the problem (36), variable separation is used, namely the substitution (v(s, t), w(s, t), e(t)) = (ˆv(s), ˆw(s), ˆe) exp(λt), which produces: A (λ,U) ˆu′′ + B (λ,U) ˆu + b (λ) ˆe = 0 (37) ˆe = ¯k l l 0 (1 −αn)ˆv −αb ˆw ds (38) ˆuA = 0, ˆuB = 0 (39) (37) (38) (39) where: where: 5 Numerical results A (λ,U) := T0 + ˘T 1 + λ ζ T0 I 2 5.1 Description of the case studies Moreover, for both the cases S and L, the analyses are carried out for two different values of initial angle of the ice coating, which correspond to conﬁgurations not prone to galloping in the Den Hartog meaning [20], i.e., where Cd(ϕ) + C′ l(ϕ) > 0: they are (1) ϕ0 = π/8 rad = 22.5◦and (2) ϕ0 = 11π/48 rad = 41.25◦(red dashed lines in Fig. 4b). Summarizing, a total number of four cases are analyzed, namely S.1, S.2, L.1, and L.2, where 1 or 2 indicates the initial angle of the ice coating. In all the cases, the outcomes of the perturbation procedures are compared to those of the numeric approach, as performed through a ﬁnite difference code implemented in MATLAB c⃝[40]. four analyzed cases. As reasonable, the binormal com- ponent is always larger than the normal one. More- over, among the four cases, the one in Fig. 6b, i.e., S.2 appears to have larger amplitudes in both the compo- nents. The perturbation outcomes are in good agree- ment with those relevant to the numeric procedure, even if some discrepancies are obtained for the αn component in case S.2 for wind velocities larger than U = 8 m/s, due to the loss of accuracy in the evaluation of the swing angle. It is worth noticing how, in cases S.1 and S.2 (Fig. 6), the binormal displacements are much smaller than in cases L.1 and L.2 (Fig. 7), as a consequence of the different sag-to-span ratios. Anyway, no signiﬁcant differences occur keeping ﬁxed the sag-to-span ratio and changing the initial angle of the ice coating, i.e., between case S.1 and S.2, or L.1 and L.2. It is interesting to evaluate the total tensile force of the cable at equilibrium in the swung plane, namely T0+ ˘T , as a function of the wind velocity. This is shown in Fig. 8a for cases S.1 and L.1, and in Fig. 8b for cases S.2 and L.2. It can be observed that, in most of the range of the considered velocities (U ≤8 m/s), the agreement between analytical and numeric results is very good, as well as modiﬁcations in the tensile force are of a few percent. On the other hand, bad agreement is found for larger velocities, particularly for cases S.2 and L.2 (Fig. 8b). 5.1 Description of the case studies Geometrical and mechanical parameters of the cable assumed as case study are given here. The unstretched length isl = 267 m, the mass per unit length is m = 4.4 kg/m, and the axial stiffness is E A = 4.0×108 N. Two different values of initial sags d are considered, the ﬁrst onegivingrisetoaverytautconﬁgurationandindicated as case S (smaller sag), the second one to a moderately taut conﬁguration, referred to as case L (larger sag). The values for case S are: d = 3 m, which corresponds to an initial curvature ¯κ = 3.37 × 10−4 m−1, initial tensile force T0 = 128.213 kN and the Irvine parameter /(2π) = 0.79, i.e., at the left of the ﬁrst crossover point; the values for case L are: d = 26 m, i.e., initial curvature ¯κ = 2.92 × 10−3 m−1, initial tensile force T0 = 14.734 kN and the Irvine parameter /(2π) = 20.39, i.e., far to the right of the ﬁrst cross-over point. In ( , ) −λ ⎡ ⎢⎢⎣ ce + U(ca 22 + ˘θn212 U(ca 23 + ˘θn213 + ˘θ2n2112) + ˘θ2n2113) U(ca 32 + ˘θn312 ce + U(ca 33 + ˘θn313 + ˘θ2n3112) + ˘θ2n3113) ⎤ ⎥⎥⎦ b (λ) := E A¯k 1 + λ ζ T0 1 −αn −αb , ˆu := ˆv ˆw (40) (40) The eigenvalue problem (37)-(39) must be solved in the complex set; indeed, the galloping (critical) modes, which are the eigenfunctions associated to the eigen- values with zero real part, are generally complex, due to the non-proportional nature of the damping operator (through its aerodynamic counterpart). Solution to the On the nonlinear effects of the mean wind force... 3137 the two cases, different values of external and internal damping coefﬁcients are assumed, respectively, so as to have equal modal structural damping ratio for the ﬁrst two natural modes (without wind), in the amount of 0.5%: they are ce = 0.056 kg×s/m and ζ = 239.4 kg×m×s for case S, ce = 0.020 kg×s/m and ζ = 73.0 kg×m×s for case L. The aerodynamic parameters are: air mass per unit volume ρ = 1.25 kg/m3, radius b = 0.102 m; the drag Cd(ϕ) and lift Cl(ϕ) coefﬁcients refer to the NDT cross section in [16,41], and they are shown in Fig. 4a. 5.2 Equilibrium conﬁguration As a ﬁrst result, the swing angle ˘θ evaluated from the analytical solution, Eq. (27), is shown in Fig. 5, as a function of the wind velocity U. Note that ˘θ is inde- pendent of the sag-to-span ratio, therefore Fig. 5a holds for both cases S.1 and L.1, while Fig. 5b for both cases S.2 and L.2. It turns out that the perturbation solu- tion (in blue solid line) is in good agreement with the numeric one (in red dashed line) up to wind veloci- ties which do not overtake about U = 8 m/s in Fig. 5a and U = 10 m/s in Fig. 5b, corresponding to angles of magnitude 0.1 rad and 0.2 rad, respectively. These angles represent the upper limit where the cubic series expansion is valid; moreover, different limit values are obtained for the two initial angles as a consequence of the different magnitudes of the derivatives of the aerodynamic coefﬁcients, which lead to a better cor- respondence to the assumed ordering for the case (2). As a further comment, the swing angle is negative as expected for the convention assumed. 5.1 Description of the case studies Actually, the perturbation solution gives a totally wrong outcome in case L.2, and this is due to the evaluation of ˘T as a term proportional to the difference of αn and 1 2(α2 n + α2 b) (see Eq. (35)), which both assume similar values, very close to 0.025 at U = 10 m/s (αb about 0.25), in violation of the perturbation order. In these cases, further perturbation ordersshouldbeconsidered,soastogiveconsistencyto the analytical solution even for larger wind velocities. 5.3 Critical conditions The critical modes and relevant critical conditions in terms of wind velocity and modal frequency are now analyzed, as solutions of the eigenvalue problem (22). In particular, in Table 1, the critical conditions are reported for both analytical and numeric solutions, showing good agreement, speciﬁcally in the modal fre- quency where the percentage error is practically zero. The amplitudes of the parabolic shape at equilib- rium, αn and αb, as given by Eq. (34), are shown in Figs. 6 and 7, as function of the wind velocity, for the For case S.1, the corresponding modal shape is shown in Fig. 9. It is anti-symmetric and real, involv- 12 3 3 3138 D. Zulli et al. (a) (b) Fig. 4 Aerodynamic coefﬁcients for the NDT section, as taken from [41]: a Drag and lift coefﬁcients (black dots are experimental data); b values of Cd + C′ l and initial angles of the ice-coating ϕ0 (red dashed lines) (b) (a) (b) (a) Fig. 4 Aerodynamic coefﬁcients for the NDT section, as taken from [41]: a Drag and lift coefﬁcients (black dots are experimental data); b values of Cd + C′ l and initial angles of the ice-coating ϕ0 (red dashed lines) (a) (b) Fig. 5 Evolution of the swing angle with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 (a) (b) Fig. 6 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case S.1; b case S.2 (a) (b) Fig. 5 Evolution of the swing angle with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 (b) (a) (b) (a) Fig. 5 Evolution of the swing angle with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 (b) (a) (b) Fig. 6 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case S.1; b case S.2 (a) (b) (a) Fig. 6 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case S.1; b case S.2 123 On the nonlinear effects of the mean wind force... 3139 (a) (b) Fig. 7 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case L.1; b case L.2 (a) (b) Fig. 5.3 Critical conditions 8 Evolution of the tensile force at the equilibrium conﬁguration with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 (a) (b) Fig. 7 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case L.1; b case L.2 (a) (b) (a) Fig. 7 Evolution of the amplitudes of the equilibrium conﬁguration with the wind velocity: a case L.1; b case L.2 (a) (a) (b) Fig. 8 Evolution of the tensile force at the equilibrium conﬁguration with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 (b) (a) (b) ensile force at the equilibrium conﬁguration with the wind velocity: a cases S.1 and L.1; b cases S.2 and L.2 Fig. 8 Evolution of the tensile force at the equilibrium conﬁguration with the wind velocity: a cases S.1 a Table 1 Critical modal conditions: velocity in m/s, frequency in rad/s case Uc pert. Uc num. err.% ωc pert. ωc num. err.% S.1 4.42 4.53 2.4 3.98 3.98 0.0 S.2 6.08 6.27 3.0 3.43 3.43 0.0 L.1 3.40 3.25 4.6 1.35 1.35 0.0 L.2 6.09 6.33 3.8 1.91 1.91 0.0 Table 1 Critical modal conditions: velocity in m/s, frequency in rad/s ing both normal (ˆv) and binormal ( ˆw) components with the same order of amplitude, in good agreement with numeric outcomes. Moreover, in Fig. 10 the evolution of all the eigenvalues which become unstable are shown in the complex plane, as the parameter U is increased and assumes the values in Table 2; concurrently, the corresponding modal shapes are shown at the right side of the same Figure. From this Figure, which is real- ized with data coming from the numeric procedure, it turns out that the eigenvalues which become unstable are four; furthermore, a regain of stability is achieved as the wind velocity is increased, due to the progres- sive change of orientation of the cross section with the increasing of the swing angle and, as a consequence, to the modiﬁcation of the intensity of the aerodynamic forces. It is worth noticing that the order in which the eigenvalues regain stability is different than that of their loss of stability. Moreover, the regain of stability is not 12 3 D. Zulli et al. 3140 (a) (b) Fig. 9 Critical mode for case S.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Fig. 5.3 Critical conditions 10 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.1 necessarilyassociatedtoacomebackoftheevolutionof the system on the trivial dynamic solution The system particular for the normal component (Fig. 11a). The eigenvalues which lose stability are shown in the com (a) (b) Fig. 9 Critical mode for case S.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (a) (a) (b) Fig. 9 Critical mode for case S.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Fig. 10 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.1 necessarilyassociatedtoacomebackoftheevolutionof the system on the trivial dynamic solution. The system might evolve on other possible, stable and coexisting attractors, deriving from secondary bifurcations here not investigated. For case S.2, the critical mode is shown in Fig. 11: it turns out to be symmetric and complex, with only one semi-wave along the span, coherently with a conﬁgura- tion slightly at the left of the ﬁrst crossover point. In this case the agreement with numeric outcomes is good in particular for the normal component (Fig. 11a). The eigenvalues which lose stability are shown in the com- plex plane in Fig. 12, at the wind velocities reported in Table 3; relevant eigenfunctions are shown as well at the right part of the ﬁgure. In this case, the eigenvalues which lose stability are six, and the order in which they regain stability is the opposite of that in which they lose it. For case L.1, in Fig. 13, the critical mode is again anti symmetric while it turns out to be symmetric for Fig. 9 Critical mode for case S.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Fig. 10 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.1 Fig. 10 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.1 necessarilyassociatedtoacomebackoftheevolutionof th t th t i i l d i l ti Th t particular for the normal component (Fig. 11a). The i l hi h l t bilit h i th necessarilyassociatedtoacomebackoftheevolutionof the system on the trivial dynamic solution. The system might evolve on other possible, stable and coexisting attractors, deriving from secondary bifurcations here not investigated. particular for the normal component (Fig. 11a). 5.3 Critical conditions The eigenvalues which lose stability are shown in the com- plex plane in Fig. 12, at the wind velocities reported in Table 3; relevant eigenfunctions are shown as well at the right part of the ﬁgure. In this case, the eigenvalues which lose stability are six, and the order in which they regain stability is the opposite of that in which they lose it. For case S.2, the critical mode is shown in Fig. 11: it turns out to be symmetric and complex, with only one semi-wave along the span, coherently with a conﬁgura- tion slightly at the left of the ﬁrst crossover point. In this case, the agreement with numeric outcomes is good, in For case L.1, in Fig. 13, the critical mode is again anti-symmetric, while it turns out to be symmetric for 123 123 123 As a summary, for both cases S.1 and L.1, i.e., for the ﬁrst value of ice-coating initial angle, and inde- pendently of the sag-to-span ratio, it turns out that the critical mode is anti-symmetric and real; on the other hand, for the second value of ice-coating initial angle, the critical mode is symmetric and complex. Therefore, it appears that the modal shape of the cable at bifurca- tion is strictly related to its initial attitude to wind and, then, to the value of the aerodynamic forces, which have the ability to signiﬁcantly change the structure of the Hamiltonian dynamical system and, consequently, the nature of the eigenfunctions, triggering symmetric or anti-symmetric modal shapes. Consistently, critical modes might be strongly different from natural ones, in absence of wind. Concerning the case study S.1, Fig. 16 shows the time histories for steady oscillations at a mean wind velocity equal to 5 m/s, slightly higher than the ﬁrst critical speed. Cable oscillations occur around the equi- librium conﬁguration (marked by a red dashed line in the ﬁgure), with values fully consistent with those of Fig. 6a and Eq. (29), and are characterized by a single frequency corresponding to the ﬁrst critical one (see Table 1). From the direct integration of the nonlinear motion equations, the cable dynamics during a period of oscillation, subtracted to the equilibrium conﬁgu- ration determined by the mean wind, is presented in Fig. 17: the shapes are entirely in accordance with the ﬁrst (normalized) critical modes of the system (Fig. 9), pointing out a not negligible coupling between the two planes, already highlighted by the critical modes. It should be noted that the vibration shapes remain almost constant during the oscillation conﬁrming the real nature of the corresponding eigenfunctions (the imaginary part of which is almost zero, see Fig. 9). In this respect, it is interesting to look at the value of the steady angle under which the ice-coating faces the wind at bifurcation, i.e., −˘θ + ϕ0, checking the corre- sponding value of the Den Hartog’s coefﬁcient Cd +C′ l, valid for 1-d.o.f. vertical galloping (see Fig. 15): it is evident how the effect of ˘θ (and, then, of the mean wind force) is to reduce the Den Hartog coefﬁcient, so as to take the cross section closer to the 1-d.o.f. instability region, even if it can still be non-negative (e.g., for cases S.2 and L.2). 123 On the nonlinear effects of the mean wind force... 3141 (a) (b) Fig. 11 Critical mode for case S.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Fig. 12 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.2 (a) (b) Fig. 11 Critical mode for case S.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (b) (a) (a) (b) ase S.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (a) Fig. 11 Critical mode for case S.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Fig. 12 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.2 Fig. 12 Evolution of the eigenvalues of Eqs. (36) and corresponding eigenfunctions for increasing wind velocity, case S.2 3 D. Zulli et al. 3142 Table 2 Numerical values of the velocities for the positions in Fig. 10, case S.1 position 1 2 3 4 5 6 7 8 9 10 11 U (m/s) 3.20 4.53 5.89 6.02 6.61 7.90 9.25 9.58 9.62 10.20 10.4 equations of motion (5) is carried out and a numeri- cal integration of the resulting equations in time is per- formed (details on the numerical procedure are given in Appendix 2). In this way, it is also possible to fully validate, in a completely independent way, the results obtained with respect to the equilibrium conﬁgura- tions (Sect. 5.2) and to the ﬁrst critical modes (Sect. 5.3). Results are limited to the case S for the sake of brevity; anyway, outcomes show that aerodynamic forces greatly inﬂuence the critical response despite the feature of the Hamiltonian system. case L.2, in Fig. 14, having a three semi-waves shape, which is coherent with the conﬁguration at the right of the cross-over. In the latter case, as for case S.2, still the normal component has better agreement with the numeric outcomes than the binormal component, which presents some quantitative differences, even if the qualitative agreement is always very satisfactory. 123 Speciﬁcally, it is evident that, due to the generally concurrent presence of both the modal com- ponents ˆv, ˆw, the mechanism of instabilization is more complicated than that which only relies on the Den Har- tog’s coefﬁcient (see, e.g., [35] for 2-d.o.f. systems). Moving on to the case study S.2, Fig. 18 shows the steady oscillations at a mean wind velocity equal to 6.4 m/s, immediately after the ﬁrst bifurcation point, and justpriortothesecondone.Alsointhiscasetheaverage value of the oscillations is completely consistent with the values obtained for the equilibrium conﬁguration (Fig. 6b and Eq. (29)), with a clear prevalence of the horizontal component. The motion is monomodal peri- odic, with the presence of a single vibration frequency (i.e., the ﬁrst critical frequency in Table 2). The corre- sponding cable deformations during a vibration period 5.4 Nonlinear galloping In order to have a ﬁrst insight into the cable nonlin- ear behavior close to the ﬁrst bifurcation point, a spa- tial ﬁnite-difference approximation of the nonlinear 123 123 123 On the nonlinear effects of the mean wind force... 3143 On the nonlinear effects of the mean wind force... 3143 Table 3 Numerical values of the velocities for the positions in Fig. 12, case S.2 position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 U (m/s) 6.00 6.27 6.51 6.60 6.85 7.11 7.46 8.00 8.55 8.77 8.90 9.02 9.04 9.24 9.40 (a) (b) Fig. 13 Critical mode for case L.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (a) (b) Fig. 14 Critical mode for case L.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw are shown in Fig 19 still subtracted to the equilibrium the horizontal critical mode was highlighted and just Table 3 Numerical values of the velocities for the positions in Fig. 12, case S.2 position 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 U (m/s) 6.00 6.27 6.51 6.60 6.85 7.11 7.46 8.00 8.55 8.77 8.90 9.02 9.04 9.24 9.40 (a) (b) Fig. 13 Critical mode for case L.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw Table 3 Numerical values of the velocities for the positions in Fig. 12, case S.2 (b) (a) (a) Fig. 13 Critical mode for case L.1: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (b) (a) (b) Fig. 14 Critical mode for case L.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw (a) (a) (b) Fig. 14 Critical mode for case L.2: a real and imaginary parts of ˆv; b real and imaginary parts of ˆw the horizontal critical mode was highlighted, and just an order of magnitude lower than the vertical mode (which instead was almost real, as conﬁrmed by this nonlinear galloping analysis). are shown in Fig. 19, still subtracted to the equilibrium conﬁguration associated to the mean wind, allowing one to highlight the dynamics of the motion. It occurs almost completely in the vertical plane with a modest coupling in the horizontal direction. 5.4 Nonlinear galloping Moreover, the ver- tical oscillations maintain almost the same shape over time, while the horizontal oscillations present remark- able variations in shape during a vibration period. This result is in excellent agreement with what was obtained in the critical condition analysis (Fig. 11 - normal- ized critical mode), in which the complex nature of 6 Conclusions and ﬁnal remarks This paper deals with the research of critical conditions of ice galloping in suspended cables. The key point of the work is the nonlinear contribution of the non-trivial equilibrium path that inﬂuences the terms governing 123 3144 D. Zulli et al. Fig. 15 Values of Cd + C′ l, initial angles of the ice-coating ϕ0 (red dashed lines) and values of the steady angle −˘θ + ϕ0 at bifurcation (marked by points; the black one is underneath the green one) the proposed approach can be easily used in the pres- ence of non-planarity of the static conﬁguration of the cable. A second strength of the work is that the search for critical conditions is proposed analytically through the use of perturbation techniques. In particular, the non- linear, non-trivial fundamental path from which the bifurcation takes place is ﬁrst determined in two steps: (a) determination of the swing angle that identiﬁes the plane on which the cable lies due to the effect of the mean wind force (swung plane), and (b) evaluation of the corresponding static displacements of the cable, together with the associated increment of tension. An eigenvalue problem is then deduced starting from the continuous formulation of the problem, which leads to transcendental characteristic equations in analytical form. Fig. 15 Values of Cd + C′ l, initial angles of the ice-coating ϕ0 (red dashed lines) and values of the steady angle −˘θ + ϕ0 at bifurcation (marked by points; the black one is underneath the green one) Concerningnumericalresults,somepointsareworth noting: the vibration modes obtained are naturally com- plex due to the non-classical nature of aerodynamic damping; the agreement of the analytical solution of the problem with the numerical one is excellent on critical instability conditions (i.e., when the eigenval- ues become unstable), while a greater error could be present for higher mean wind speeds due to the ana- lytical calculation of the swing angle (truncated to the third order); thanks to the continuous formulation, the inﬂuence of steady swing is naturally present on all modes of the system (unlike what happens in the lit- erature for discrete multimodal galloping, e.g., [37]); therefore, all critical eigenvalues return to the stabil- ity domain for sufﬁciently high mean wind velocities; galloping onset conditions. 6 Conclusions and ﬁnal remarks It reveals new insights on the bifurcation phenomenon, being able to make unsta- ble cable conﬁgurations that instead would be aerody- namically stable without the effects of the mean wind force. Concerning the modeling aspects, the main novelty of the paper is the use of a continuous model, unlike what happens in current literature. In this way it is not necessary to select in advance a certain number of modal shapes (as usually happens in discrete Galerkin models, e.g., [37]). The proposed method is able to nat- urally follow the evolution of critical conditions as the wind speed increases. The problem is for now restricted to the case of horizontal cables to have an exact plane description of the non-trivial fundamental path, even if (a) (b) Fig. 16 Time evolution at the quarter-span node at U = 5.0 m/s for case S.1, where the red line indicates the equilibrium position: a v( l 4, t); b w( l 4, t). The time range is suitably chosen to consider steady-state oscillations (b) (a) (b) (a) (b) (a) Fig. 16 Time evolution at the quarter-span node at U = 5.0 m/s for case S.1, where the red line indicates the equilibrium position: a v( l 4, t); b w( l 4, t). The time range is suitably chosen to consider steady-state oscillations 123 12 On the nonlinear effects of the mean wind force... 3145 (a) (b) Fig. 17 Evolution of the solution during one time period at U = 5.0 m/s for case S.1: a v cleaned of the equilibrium value ˘v; b w cleaned of the equilibrium value ˘w (b) (a) (a) (b) Fig. 17 Evolution of the solution during one time period at U = 5.0 m/s for case S.1: a v cleaned of the equilibrium value ˘v; b w cleaned of the equilibrium value ˘w (b) (a) (b) Fig. 18 Time evolution at the mid-span node at U = 6.4 m/s for case S.2, where the red line indicates the equilibrium position: a v( l 2, t); b w( l 2, t). The time range is suitably chosen to consider steady-state oscillations (a) (b) (a) (b) (a) Fig. 18 Time evolution at the mid-span node at U = 6.4 m/s for case S.2, where the red line indicates the equilibrium position: a v( l 2, t); b w( l 2, t). 6 Conclusions and ﬁnal remarks The time range is suitably chosen to consider steady-state oscillations (a) (b) Fig. 19 Evolution of the solution during one time period at U = 6.4 m/s for case S.2: a v cleaned of the equilibrium value ˘v; b w cleaned of the equilibrium value ˘w (b) (a) (b) (a) Fig. 19 Evolution of the solution during one time period at U = 6.4 m/s for case S.2: a v cleaned of the equilibrium value ˘v; b w cleaned of the equilibrium value ˘w 123 12 3 D. Zulli et al. 3146 the numerical integration of the nonlinear equations of motion (in both mechanical and aerodynamic terms) has allowed a further validation of the goodness of the results obtained from the critical conditions, at least limited to the ﬁrst bifurcation point. Indeed, through a completely independent numerical procedure, full cor- respondences on both equilibrium path values and crit- ical modes of the system are determined, even in cases where the oscillation modes assume signiﬁcant com- plex representations. In fact, the steady swing strongly inﬂuences the aerodynamic forces, since it changes the exposition to wind of the body. Therefore, its consistent inclusion in the model is fundamental, and the nonlin- ear terms arising from it turn out to strongly affect the onset of galloping. analysis of the limit-cycle arising at successive Hopf bifurcations. However, this is a problem which appears far from being trivial, since a cluster of several modes (from four to six in the examined examples) could in principle interact in the postcritical ﬁeld. Moreover, the regain of stability of the fundamental nontrivial path does not exclude the existence of supercritical compet- itive attractors, produced by the previous bifurcations. Funding Open access funding provided by Universitá degli Studi di Genova within the CRUI-CARE Agreement. Conﬂict of interest The authors declare that they have no con- ﬂict of interest. Conﬂict of interest The authors declare that they have no con- ﬂict of interest. A special comment deserves the fact that, in the pre- sented examples, the unstable mode can be symmet- ric or anti-symmetric despite the lower vibration fre- quency of the cable always corresponds to a symmet- ric out-of-plane mode for any conﬁguration (smaller or larger sag) considered. This can be attributable to the presence of the aerodynamic damping operator, which in turn depends on the non-trivial equilibrium path through the aerodynamic coefﬁcients, capable of signiﬁcantly modifying the critical behavior of the sys- tem. This result, which may signiﬁcantly affect design choices for suspended cables in cold regions, is worthy of future investigations, also using particular perturba- tion techniques (e.g., [42]). In this context it is also worth noting that, as a consequence of the proposed approach, the motion must be described in terms of incremental variables, which are superimposed to the static conﬁguration in order to ﬁnd the critical condi- tions and carry out possible nonlinear bifurcation anal- yses. The incremental variables are here expressed in terms of the original extrinsic basis, in the vertical plane where the cable lies in absence of wind. Therefore, the locutions ’in-plane’ and ’out-of-plane’, relating to cable dynamic displacements, lose their meaning since they should be referred to the vertical, not to the cur- rentlyrotatedplanewhenthewindisblowing.Then,the cable natural modes will appear coupled in the two dis- placement components. But this approach, apparently more complicated than the intrinsic one (e.g., [29]), has the great advantage not to require any change of basis, nor projection of forces, while keeping the orig- inal meaning of the variables. Open Access This article is licensed under a Creative Com- mons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com- mons licence, and indicate if changes were made. The images or otherthirdpartymaterialinthisarticleareincludedinthearticle’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Cre- ative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/ by/4.0/. 123 References n2111 := −1 12ρbC′′′ l (ϕ0), n3111 := 1 12ρbC′′′ d (ϕ0) n2112 := 1 4ρb(C′′ d(ϕ0) + C′′′ l (ϕ0)), n3112 := −1 4ρb(C′′′ d (ϕ0) −C′′ l (ϕ0)) n2113 := −1 2ρbC′′ l (ϕ0), n3113 := 1 2ρbC′′ d(ϕ0) (45) 1. Irvine, H.: Cable Structures. MiT Press, Cambridge (1981) 2. Luongo, A., Zulli, D.: Statics of shallow inclined elastic cables under general vertical loads: a perturbation approach. Mathematics 6(2), 63–72 (2018a) 3. Luongo,A.,Zulli,D.:Staticperturbationanalysisofinclined shallow elastic cables under general 3D-loads. Curv. Lay- ered Struct. 5, 250–259 (2018b) (45) 4. Fei, H., Danhui, D., Yiqing, Z., Huan, L.: Experimental and theoretical study on cable-supporting system. Mech. Syst. Signal Process. 140, 106638 (2020) 5. Rega, G.: Nonlinear vibrations of suspended cables-Part I: modeling and analysis. Appl. Mech. Rev. 57(6), 443–478 (2004) 6. Lee, C., Perkins, N.: Nonlinear oscillations of suspended cables containing a two-to-one internal resonance. Nonlin- ear Dyn. 3, 465–490 (1992) 7. Ferretti, M., Piccardo, G., Luongo, A.: Semi-analytical approaches for the nonlinear dynamics of a taut string sub- ject to a moving load. Nonlinear Dyn. 98(4), 2463–2474 (2019) Appendix 1: Aerodynamic coefﬁcients The explicit expression of the coefﬁcients deﬁning the components of the aerodynamic forces in Eq. (15) are: f a n0 := −1 2ρbCl(ϕ0), ka 21 := −1 2ρbC′ l(ϕ0), ca 22 := 1 2ρb Cd(ϕ0) + C′ l(ϕ0) , ca 32 := 1 2ρb Cl(ϕ0) −C′ d(ϕ0) ca := ρbCl(ϕ0) f a n0 := −1 2ρbCl(ϕ0), f a b0 := 1 2ρbCd(ϕ0) ka 21 := −1 2ρbC′ l(ϕ0), ka 31 := 1 2ρbC′ d(ϕ0) ca 22 := 1 2ρb Cd(ϕ0) + C′ l(ϕ0) , ca 32 := 1 2ρb Cl(ϕ0) −C′ d(ϕ0) ca 23 := −ρbCl(ϕ0), ca 33 := ρbCd(ϕ0) (43) ca 23 := −ρbCl(ϕ0), ca 33 := ρbCd(ϕ0) (43) ca 33 := ρbCd(ϕ0) (43) ca 33 := ρbCd(ϕ0) (43) (43) where the prime stands for derivative with respect to the argument. Furthermore, the expressions of the coefﬁ- The ﬁrst results of nonlinear galloping oscillations, limited to the ﬁrst bifurcation, has to be extended to the 123 123 3147 On the nonlinear effects of the mean wind force... cients of the quadratic polynomial na 2 are: by using standard techniques. The numerical results presented in Sect. 5.4 are obtained with a spatial dis- cretization of 41 nodes, which we found to be a suitable compromise between computational effort and preci- sion of the solution. It is noteworthy that the nonlinear expression of aerodynamic forces is not expanded, as usual, in MacLaurin series of structural velocities but it is rigorously calculated using Eqs. (6)-(8), where the aerodynamic coefﬁcients Cd and Cl are calculated in their instantaneous angle values, which are equal to (γ + ϕ0), being the instantaneous angle of attack γ expressed by Eq. (10). n211 := 1 4ρbC′′ l (ϕ0), n311 := −1 4ρbC′′ d(ϕ0) n212 := −1 2ρb(C′ d(ϕ0) + C′′ l (ϕ0)), n312 := 1 2ρb(C′′ d(ϕ0) −C′ l(ϕ0)) n213 := ρbC′ l(ϕ0), n313 := −ρbC′ d(ϕ0) (44) (44) and those of the cubic polynomial na 3 are Appendix 2: Numerical integration of the nonlinear equations of motion Jones, K.: Coupled vertical and horizontal galloping. J. Eng. Mech. 118(1), 92–107 (1992) 32. Foti, F., Martinelli, L.: Finite element modeling of cable galloping vibrations. Part II: application to an iced cable in 1:2 multiple internal resonance. J. Vib. Control 24(7), 1322– 1340 (2018) 17. Luongo, A., Piccardo, G.: Non-linear galloping of sagged cables in 1:2 internal resonance. J. Sound Vib. 214(5), 915– 940 (1998) 18. Cigada, A., Diana, G., Falco, M., Fossati, F., Manenti, A.: Vortex shedding and wake-induced vibrations in single and bundle cables. J. Wind Eng. Ind. Aerodyn. 72, 253–263 (1997) 33. Ferretti, M., Zulli, D., Luongo, A.: A continuum approach to the nonlinear in-plane galloping of shallow ﬂexible cables. Adv. Math. Phys. 2019, 6865730 (2019) 34. Luongo, A., Piccardo, G.: Linear instability mechanisms for coupled translational galloping. J. Sound Vib. 288(4–5), 1027–1047 (2005) 19. Tokoro, S., Komatsu, H., Nakasu, M., Mizuguchi, K., Kasuga, A.: Study on wake-galloping employing full aeroe- lastic twin cable model. J. Wind Eng. Ind. Aerodyn. 88, 247–261 (2000) 35. Nikitas, N., Macdonald, J.: Misconceptions and generaliza- tions of the Den Hartog galloping criterion. J. Eng. Mech. ASCE 140(4), 04013005 (2014) 20. Den Hartog, J.: Mechanical Vibrations. Dover Publications, New York (1985) 36. Lou, W., Wu, D., Xu, H., Yu, J.: Galloping stability criterion for 3-DOF coupled motion of an ice-accreted conductor. J. Struct. Eng. ASCE 146(5), 0402007 (2020) 21. Blevins, R.: Flow-induced Vibration, 2nd edn. Krieger Pub- lishing Company, Florida (2001) Struct. Eng. ASCE 146(5), 0402007 (2020) 22. Païdouissis, M., Price, S., de Langre, E.: Fluid-structure Interactions-Cross-Flow Induced Instabilities. Cambridge University Press, Cambridge (2011) 37. Liu, X., Huo, B.: Nonlinear vibration and multimodal inter- action analysis of transmission line with thin ice accretions. Int. J. Appl. Mech. 7(1), 1550007 (2015) y g 23. Piccardo, G., Pagnini, L., Tubino, F.: Some research per- spectives in galloping phenomena: critical conditions and post-critical behavior. Continuum Mech. Thermodyn. 27, 261–285 (2015) 38. Nayfeh,A.,Pai,P.:LinearandNonlinearStructuralMechan- ics. Wiley, New York (2004) 39. Pagnini, L., Freda, A., Piccardo, G.: Uncertainties in the evaluation of one degree-of-freedom galloping onset. Eur. J. Environ. Civil Eng. 21(7–8), 1043–1063 (2017) ( ) 24. Yu, P., Desai, Y.M., Shah, A.H., Popplewell, N.: Three- degree-of-freedom model for galloping. Part I: formulation. J. Eng. Mech. 119(12), 2404–2425 (1993a) 40. MATLAB, version R2019b. The MathWorks Inc., Natick, Massachusetts (2019) 25. Yu, P., Desai, Y.M., Popplewell, N., Shah, A.H.: Three- degree-of-freedom model for galloping. Part II: solutions. J. Eng. Appendix 2: Numerical integration of the nonlinear equations of motion First, once ﬁxed the mean wind velocity U, the cor- responding swing angle ˘θ is evaluated using Eq. (20). Then, a spatial ﬁnite-difference method is established for Eqs. (5), which are both mechanically and aerody- namically nonlinear, by dividing the material domain [0,l] of the cable in n intervals of the same length, with the corresponding deﬁnition of (n + 1) nodes; central ﬁnite differences are deﬁned for both ﬁrst and second s−derivatives of the central nodes whereas for- ward and backward ﬁnite differences are used for bor- der nodes. In this way a differential-algebraic systems of equations is deduced, whose algebraic part can be solved separately in a symbolic way by obtaining a dis- crete expression for the dynamic unit extension e(t). Then, a numerical integration of the resulting system of ordinary differential equations in time is performed 8. Triantafyllou, M., Grinfogel, L.: Natural frequencies and modes of inclined cables. J. Struct. Eng. ASCE 112(1), 139– 148 (1986) 9. Jafari, M., Hou, F., Abdelkeﬁ, A.: Wind-induced vibration of structural cables. Nonlinear Dyn. (2020). https://doi.org/ 10.1007/s11071-020-05541-6 10. Violette, R., de Langre, E., Szydlowski, J.: A linear stability approach to vortex-induced vibrations and waves. J. Fluids Struct. 26(3), 442–466 (2010) 11. Gao, D., Chen, W.L., Zhang, R.T., Huang, Y.W., Li, H.: Multi-modal vortex- and rain-wind- induced vibrations of an inclined ﬂexible cable. Mech. Syst. Signal Process. 118, 245–258 (2019) 12. Matsumoto, M., Yagi, T., Goto, M., Sakai, S.: Rain-wind induced vibration of inclined cables at limited high reduced wind velocity region. J. Wind Eng. Ind. Aerodyn. 91, 1–12 (2003) 13. Luongo, A., Zulli, D.: Dynamic instability of inclined cables under combined wind ﬂow and support motion. Nonlinear Dyn. 67(1), 71–87 (2012) 12 3 3 3148 D. Zulli et al. 14. Macdonald, J., Larose, G.: Two-degree-of-freedom inclined cable galloping-part 1: general formulation and solution for perfectly tuned system. J. Wind Eng. Ind. Aerodyn. 96, 291– 307 (2008) 29. Luongo, A., Zulli, D., Piccardo, G.: On the effect of twist angle on nonlinear galloping of suspended cables. Comput. Struct. 87, 1003–1014 (2009) 30. Luongo, A., Zulli, D., Piccardo, G.: Analytical and numeri- cal approaches to nonlinear galloping of internally resonant suspended cables. J. Sound Vib. 315(3), 375–393 (2008) ( ) 15. Piccardo,G.,Zulli,D.,Luongo,A.:Drygallopingininclined cables: linear stability analysis. Procedia Eng. 199, 3164– 3169 (2017) 31. McComber, P., Paradis, A.: A cable galloping model for thin ice accretions. Atmos. Res. 46(1–2), 13–25 (1998) 16. Appendix 2: Numerical integration of the nonlinear equations of motion Mech. 119(12), 2426–2448 (1993b) 41. Novak, M., Davenport, A., Tanaka, H.: Vibration of towers due to galloping of iced cables. J. Eng. Mech. Div. ASCE 104(2), 457–473 (1978) 26. He, M., Macdonald, J.: Aeroelastic stability of a 3DOF sys- tem based on quasi-steady theory with reference to inertial coupling. J. Wind Eng. Ind. Aerodyn. 171, 319–329 (2017) 42. Luongo, A., Zulli, D.: A paradigmatic system to study the transition from zero/Hopf to double-zero/Hopf bifurcation. Nonlinear Dyn. 70(1), 111–124 (2012) 27. Luongo, A., Piccardo, G.: A continuous approach to the aeroelastic stability of suspended cables in 1:2 internal res- onance. J. Vib. Control 14(1–2), 135–157 (2008) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. 28. Luongo, A., Zulli, D., Piccardo, G.: A linear curved-beam model for the analysis of galloping in suspended cables. J. Mech. Mater. Struct. 2(4), 675–694 (2007) 123
https://openalex.org/W3011824779	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0229956&type=printable	English	null	Trajectories toward maximum power and inequality in resource distribution networks	PloS one	2,020	cc-by	10,305	PLOS ONE RESEARCH ARTICLE Natalie DavisID1,2, Andrew Jarvis1, M. J. Aitkenhead2, J. Gareth Polhill2 1 Lancaster Environment Centre, Lancaster University, Lancaster, United Kingdom, 2 The James Hutton Institute, Aberdeen, United Kingdom 1 Lancaster Environment Centre, Lancaster University, Lancaster, United Kingdom, 2 The James Hutton Institute, Aberdeen, United Kingdom ndavis.research@gmail.com a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Editor: Rachata Muneepeerakul, University of Florida, UNITED STATES Editor: Rachata Muneepeerakul, University of Florida, UNITED STATES Received: October 25, 2019 Accepted: February 17, 2020 Published: March 10, 2020 Received: October 25, 2019 Accepted: February 17, 2020 Published: March 10, 2020 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0229956 Trajectories toward maximum power and inequality in resource distribution networks Natalie DavisID1,2*, Andrew Jarvis1, M. J. Aitkenhead2, J. Gareth Polhill2 OPEN ACCESS Citation: Davis N, Jarvis A, Aitkenhead MJ, Polhill JG (2020) Trajectories toward maximum power and inequality in resource distribution networks. PLoS ONE 15(3): e0229956. https://doi.org/ 10.1371/journal.pone.0229956 Citation: Davis N, Jarvis A, Aitkenhead MJ, Polhill JG (2020) Trajectories toward maximum power and inequality in resource distribution networks. PLoS ONE 15(3): e0229956. https://doi.org/ 10.1371/journal.pone.0229956 PLOS ONE PLOS ONE Abstract Resource distribution networks are the infrastructure facilitating the flow of resources in both biotic and abiotic systems. Both theoretical and empirical arguments have proposed that physical systems self-organise to maximise power production, but how this trajectory is related to network development, especially regarding the heterogeneity of resource distribu- tion in explicitly spatial networks, is less understood. Quantifying the heterogeneity of resource distribution is necessary for understanding how phenomena such as economic inequality or energetic niches emerge across socio-ecological and environmental systems. Although qualitative discussions have been put forward on this topic, to date there has not been a quantitative analysis of the relationship between network development, maximum power, and inequality. This paper introduces a theoretical framework and applies it to simu- late the power consumption and inequality in generalised, spatially explicit resource distribu- tion networks. The networks illustrate how increasing resource flows amplify inequality in power consumption at network end points, due to the spatial heterogeneity of the distribution architecture. As increasing resource flows and the development of hierarchical branching can both be strategies for increasing power consumption, this raises important questions about the different outcomes of heterogeneous distribution in natural versus human-engi- neered networks, and how to prioritise equity of distribution in the latter. Introduction Inequality in human society is typically conceived as an outcome of combined social, politi- cal, psychological, and economic influences. Although many theories about the origins of inequality include discussion of resources, such as their economic defensibility, most theories still invoke cultural or technological arguments as well [4]. Additionally, even arguments based on instincts and social behaviour rarely connect these to resource distribution explicitly [5], despite the essential role of resource movement in giving rise to any cultural, technological, and social forces. This gives the appearance of resource distribution and emergent inequality in social systems as having fundamentally different causes than hierarchies in environmental and biological systems, or energetic niches in ecosystems. Moreover, while energy consump- tion is not typically the named objective of economic management, the drive toward ever- increasing economic growth still requires energetic resources to build and maintain the infra- structure that generates returns [6], paralleling the energy used for growth and maintenance within natural systems. As both natural and human-engineered systems rely on resource dis- tribution networks to relocate energetic resources, it seems logical to consider heterogeneity within the networks and resources themselves as potentially foundational causes of inequality [7]. However, a formal, quantitative linkage between RADE network architecture, inequality in resource distribution, and the rate of increase of that inequality during network develop- ment, has not yet been elucidated. RADE networks are theorised to develop in a way that maximises the availability of resources to points of end use, such that these end consumers capture the maximum free energy for their own purposes in doing ‘useful work’ [2], such as increases in growth, develop- ment, or storage [8]. This is formalised in the Maximum Power Principle (MPP), which states that, given adequate degrees of freedom, a system will self-organise so as to maximise its power output, or capture and use of free energy per unit time [9]. An explanation for why such behav- iour would emerge is that increasing the availability of useful energy currently within a system allows the system to capture more free energy in the future, such that MPP is simply the expression of a growth-orientated positive feedback, which inevitably evolves to some bound- ary or constraint. Often these constraints can be considered thermodynamic limits on effi- ciency [10]. Introduction Copyright: © 2020 Davis et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Both biotic and abiotic systems require energy for maintenance and growth, necessitating the relocation of energetic resources from points of supply to points of consumption and end use. This need for energy drives the development of resource acquisition, distribution, and end-use (RADE) networks [1] in all earth systems. RADE networks are by definition spatial structures, constructed with both physical materials, such as asphalt, wire, or connective tissue, and infor- mational cues, such as scent trails or memories. Additionally, all RADE networks can be con- ceptualised as a collection of resources, where the energy flow is generated and supplied; end- use consumers, where the energy flow is required; and the links between them. The construc- tion, maintenance, and use of these networks inevitably requires a considerable proportion of the resources available to consumers. As it is evolutionarily advantageous to maximise the net Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Funding: ND acknowledges funding from a joint Lancaster University/James Hutton Institute PhD PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 1 / 19 PLOS ONE Maximum power and inequality in resource distribution networks resources available for further growth and development [2,3], there is significant adaptive pres- sure to drive RADE network development toward increasing efficiency. Additionally, these networks often share common forms such as hierarchical branching, and serve end consumers operating in highly heterogeneous states. Rarely, if ever, are these two observations explicitly associated, but given the role of RADE networks in determining the states of the consumers they support, correlation between network topology and variance in supply to these points of end use should be expected. Establishing this connection is crucial, not only in natural systems as a means of accounting for variability, but especially in social systems where inequality is of such profound importance. studentship. JGP acknowledges funding from the Scottish Government Rural Affairs, Food and Environment Strategic Research Programme 2016–2021 (https://www2.gov.scot/Topics/ Research/About/EBAR), Work Packages 2.4 "Rural Industries" and 3.3 "Food Security." The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Introduction Hereon, this maximisation of energy consumption and power production will be referred to as ‘maximum power,’ to include the transfer or capture of free energy, and its con- sumption in performing useful work. MPP is closely related or equivalent in many systems to the Maximum Entropy Production Principle (MEPP) and related thermodynamic extremisa- tion principles (see e.g. [11,12]). While criticisms of both MPP and MEPP [13–15] have been put forward, these have mostly been resolved through clarification and restrictions to the theo- ries [16,17]. As such, these extremisation principles provide a framework and directionality for evolution and systems progression, and can be used to help understand broader trajectories for systems development, and network development within that [12]. Specifically, systems often maximise power via changing state with respect to available energy inputs and constraints; changing network architecture to take advantage of untapped resources or minimise energy consumption in transporting resources; or both. Some theorise PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 2 / 19 PLOS ONE Maximum power and inequality in resource distribution networks that the development of self-similar hierarchical branched networks, seen in a diverse array of naturally-occurring and human-engineered systems, including vascular networks in plants and animals, power grids, and river basins, is an example of the latter strategy [18,19]. Resource flows transmit energy using a mass carrier, such as food or electrons; and during transmission these carriers experience frictional dissipation when moving over distances. This creates the evolutionary pressure to minimise transmission distance to maximise the energy transferred, hence the development of optimal space-filling structures such as hierarchical branching. Despite the theoretical universal drive toward increasing levels of energy consump- tion, there has been limited study on the relationship between this increasing trajectory, the architectures favourable to it, and the impact that has on the inequality of energy distribution in ecological and socio-ecological systems, as introduced above. Since frictional dissipation derives from distance, spatially explicit modelling of RADE net- works is crucial to understanding their development and dynamics, and the impacts these have on inequality. The dynamics of energy-mass flows over distances are described by a group of phenomenological linear flow laws, including Ohm’s law for electrical current, Darcy’s law for fluid flow, Fick’s law for diffusion, and Fourier’s law for heat transport. Introduction These flow laws state how force and flux are closely related to one another [20], making them useful for modelling a diverse range of energy-mass flow systems. It is hypothesised that, when viewed from the appropriate perspective, physical systems such as ecosystems and socio-eco- logical systems should all follow these force-flux relationships [21]. Odum in particular made extensive use of electrical analogue modelling, which calculated the flows through a system using Ohm’s law, by identifying the analogous concepts to voltage, current, and resistance or conductance in a system (see e.g. [22,23]). While his focus was on interactional models such as food webs, less work has been done applying this type of modelling to spatially-explicit net- works, where the friction or resistance term, or equivalently the latter’s inverse, conductance, is related to the physical distance the flows must cover (although see specific case studies in [24,25]). Drawing analogies between resource flows in complex coupled socio-ecological sys- tems and electrical circuits can be criticized because the formulas underlying analysis of elec- trical systems are linear, while those of the former are nonlinear. However, Wang et al. (2012) argue that many systems show linear behaviour at macroscales or microscales, and these can be modelled individually and recombined. Such linear models thus remain useful analogies for exploring generalized realistic systems [26], and may still result in the emergence of complex properties. Exploration of the effects of nonlinear formulas on these observations is the poten- tial subject of future work. Given the theoretical argument and empirical evidence for systems to evolve toward a state of maximum power, this paper will explore the potential relationships between the trajectory towards maximum power, RADE network structure, and inequality. It will thereby generate fur- ther insight into the characteristics of complex spatially explicit RADE networks as they develop toward and operate at maximum power. Specifically, systems will be modelled with representa- tive electrical circuits to elucidate the dynamics and characteristics of generalised RADE net- works evolving toward maximum power transfer, explore characteristics of those networks and the evolutionary levers employed in their development, and discuss how these relate to the exis- tence and development of inequality between end consumers in those networks. Inequality as a function of network architecture and resource flows Modelling framework using an electrical analogue In mass-flow networks, the flow through the network is generally conceptualised as a function of the driving potential gradient, and the characteristics of the material through which it flows. 3 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks As introduced above, this relationship can be represented in a given system using an analogue of one of the phenomenological flow laws, such as Ohm’s law, I ¼ DV R : ð1Þ I ¼ DV R : ð1Þ In the framework here, ΔV is the potential gradient driving the flow between two points in the network, I is the resource flow, and R is the resistance of the associated link, a measure of the friction encountered by the flow, given by the ratio of link length to link strength or capac- ity, R ¼ L S. The power output P delivered to a given end-point consumer ci in the network, or final power, is defined as PCi ¼ ICiVCi: ð2Þ ð2Þ Alternatively, ΔV can be conceptualised as the energy consumed in transport, whether active or passive, as the power consumed in transport between two points, PL, is given by com- bining Eqs 1 and 2 as PL ¼ IDV ¼ I2R: ð3Þ ð3Þ The relationship between this power consumption in transport and the spatially-related resistance term clarifies the evolutionary pressure for a system to minimise resistance, such as through the development of increasingly efficient structures that are hypothesised to minimise frictional losses [18]. Specifically, minimising the frictional losses maximises the rate of energy transfer, or power, at the spatially disparate points of final dissipation or consumption. Along with reorganisation of network architecture to minimise resistance, systems can evolve toward higher final power by adapting network state with respect to the quantity and potential of available resources. For example, the increased availability of resources in summer months allows mammals to operate at a higher metabolism and in a greater geographic range, whereas hibernation is an adaptation to decreased resource availability in the same range dur- ing winter months [27]. In the framework here, adaptation of network state can be represented by changing I, or by changing the potentials that comprise VC. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 Inequality as a function of network architecture and resource flows Modelling framework using an electrical analogue In the former case, increasing I causes PCi to increase (Eq 2), until the increased frictional losses from higher resource flow (Eq 3) causes a large enough increase in ΔV, such that PCi decreases. In this way, the trade-off between I and ΔV is mediated by R, again providing evolutionary pressure for a system to develop lower resistance, as it increases resource flows. Due to this trade-off between I and ΔV, maximum final power occurs in this framework when the potential at the consumer is half the potential at the resource (see S1 Text for deriva- tion). This is consistent with the Maximum Power Transfer theorem for electrical circuits [28], empirical findings of maximum power in natural systems such as streamflow [29], muscle con- traction [30], sediment transport [11], and the Maximum Power Principle as extended to gen- eralised interacting components [9]. In electrical circuits, simplification algorithms such as The´venin’s theorem [28] allow for complex circuits to be represented by simpler equivalents. Similarly, in the framework presented here, the relationship between consumer and resource potential can be extended over the entire network using the network mean values for power consumption, resource flow, resistance, and potentials. Specifically, the network-wide maxi- mum final power state is then VC ¼ VR 2 ; ð4Þ VC ¼ VR 2 ; ð4Þ 4 / 19 4 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks where VC and VR are the network mean values for consumer and resource potential, respectively. In order to extend this framework to explore the heterogeneity among consumers within the network, and how this is affected by increasing consumption and changing network orga- nisation, the relationship between consumer potential, resource flow, and resistance can also be expressed in terms of the respective standard deviations. Although it is more common to use the Gini coefficient or other relative measures to quantify inequality in economic and simi- lar analyses, these can obfuscate increases in absolute inequality when the relationship between variables stays constant [31,32]. For example, if each number in a distribution is increased by 50%, the standard deviation of the distribution increases by 50% and the range by 67%, but the Gini coefficient remains the same as the relative relationships are unchanged. Inequality as a function of network architecture and resource flows Modelling framework using an electrical analogue Moreover, the Gini coefficient and similar metrics are unitless measures, whereas the standard deviation has the same units as the mean. Any relationships elucidated involving standard deviation will therefore be more consistent with those identified above using means. The distributions of consumer potentials and final power consumption in a network are the result of the spatial distribution of consumer and resource nodes and links, and the magnitude of resource flow. In networks where there is a single direct connection from each consumer to a resource point, equal resource flow to all consumers, and no interconnections between consum- ers, the standard deviation of the consumer potentials is sVC ¼ sRI, derived from Eq 1, and the standard deviation of consumer final power is sPC ¼ sRI2. Therefore, in networks with equal resistances along all links, such as an idealised radial burst network, the standard deviations of consumer potentials and final power consumption would be zero. In contrast, increasing resource flows along links with unequal resistance would cause an increase in the standard deviations of potential and final power consumption, due to unequal decreases in consumer potentials. In more interconnected networks, however, the standard deviations of consumer potential and final power consumption are complex properties, as changes in potential at one node would propagate to interconnected nodes throughout the entire network. As such, determin- ing the baseline structural heterogeneity of the network helps isolate the effects of spatial distri- bution and connectivity from those of resource flow in increasing the distributions of consumer potential and final power consumption. Here, the ‘effective resistance’ REi is the resource flow-normalised drops in potential from a resource to a given consumer i, REi ¼ VR VCi Ii : ð5Þ ð5Þ As opposed to the traditional measure of resistance, which is calculated for a given link, the effective resistance is calculated along the whole path between a given consumer and resource, even if the two nodes are connected indirectly via multiple links. The effective resistance there- fore considers the interaction effects along the links, as well as the real resistances of the link or links between a consumer and resource: its standard deviation relates the heterogeneity in physical distances around the network that the flows cross, network connectivity, and the quantity of flow, to the disparities in consumer potential or power. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 Inequality as a function of network architecture and resource flows Modelling framework using an electrical analogue In the special case of direct connections between consumers and a resource, the effective resistance simplifies to the link resistance. In all networks, therefore, the standard deviation of effective resistance is the constant of proportionality between the standard deviation of con- sumer potential or power, and resource flow, such that sVC ¼ sREI, and sPC ¼ sREI2. As with the traditional measure of resistance, effective resistance and its standard deviation are station- ary for any quantity of resource flow through a given link in the network architecture. It is clearly influenced by the connectivity and symmetry of the network, as asymmetry in path 5 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks length, Euclidean distance, or number of intermediary or downstream nodes all increase the inequality in consumer potential and final power consumption. Notably, since effective resis- tance includes the effects of both physical structure and connectivity, it could potentially be a useful mapping between spatial and relational dimensions of networks, which have typically been analysed separately. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 https://doi.org/10.1371/journal.pone.0229956.g001 Simulations to illustrate framework To illustrate these described dynamics of resource flow in networks, generalised RADE net- works were modelled using the relationships presented above. Initially, the networks were com- prised of only two types of nodes distributed in space: resource supply nodes and consumer nodes. Consumer nodes could be connected to one another, such that the consumers who were more directly connected to resource nodes passed resource flow along the network to more dis- tant consumers. However, this was limited to the excess resource flow remaining after the initial consumers had met their requirement: consumer nodes could not act as resources to generate additional flow. The resistance was held constant across all links, and was modelled as the ratio of link length to strength, as described previously. The networks were evolved toward maximum power by increasing the resource flows through them and determining the distribution of power consumption across the network using a matrix inversion. The full details are provided in Section 5. This approach, modified from load flow analysis in electrical grids [33], ensured that the resource flows calculated for each node were consistent with the constraints of the first and second laws of thermodynamics, as resource flow was conserved, and power losses around the network were proportional to the size of the network. A sample of the networks simulated is shown in Fig 1, and complete results are in S1 Table. Fig 1. A sample of the networks used to simulate evolution toward maximum power. The green squares are resource nodes, and the blue circles are consumer nodes. The grey lines are links between them. Maximum power was calculated by varying the resource flow through the network and calculating the total final power across all consumer nodes. Fig 1. A sample of the networks used to simulate evolution toward maximum power. The green squares are resource nodes, and the blue circles are consumer nodes. The grey lines are links between them. Maximum power was calculated by varying the resource flow through the network and calculating the total final power across all consumer nodes. https://doi.org/10.1371/journal.pone.0229956.g001 6 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks The outcomes of a representative sample of the simulations are shown in Fig 2. Simulations to illustrate framework As consis- tent with Eq 4 above, in all simulations maximum power occurs when the mean consumer operates at 50% of the potential of the mean resource (Fig 2A). Moreover, the relationship between resource flow per consumer squared, I2 c , and the standard deviation of consumer power, sPC; is linear (Fig 2B), with slope sRE, as calculated by least-squares regression and plot- ted against the estimate using Eq 5 (Fig 2C). This heterogeneity of distances and connections between the consumers and resource causes a distribution of consumer potentials, reflected in sPC: The relationship between consumer potential and power heterogeneity for the different networks, and the resource flow per consumer, is also shown in Fig 3, where increasing IC over the course of the simulation, and hence decreasing VC=VR, causes sPC to increase. The networks that show more heterogeneity in the Euclidean distance, path distance, or both, and less connectivity between consumers, have higher inequality as measured by sRE (Figs 1 and 2C). This suggests that connectivity among consumers can also play a role in limit- ing the inequality in frictional losses and the resultant consumer potential heterogeneity. This mechanism is perhaps similar to the translocation of nutrients through fungi, where symbiotic connections between the mycelium and plant root systems allow for the redistribution of het- erogeneously-located nutrients, providing more remote portions of the mycelial network greater access to resources [34]. While the resource nodes in these simulations operated at a constant potential, similarly to time-averaged behaviour of renewable resources, or a system observed over a short timeframe, these results suggest that inequality would increase even more quickly in systems with dimin- ishing resources. This would be because the less optimally located and connected consumers would experience larger decreases in power, due to the decreasing resource potential amplify- ing the effects of their higher effective resistance. This is a current line of investigation for an extension of this work. Branching as a strategy to increase the maximum power of a system Although changes in state variables, such as potential, allow any given network architecture to achieve its maximum power, this maximum can be increased further through the evolution of the network architecture itself, as discussed. In the framework presented here, this would be illustrated by network reorganisation or otherwise reducing R, such that higher resource flows do not cause as much frictional loss (Eq 3). This does not necessarily require decreasing sRE however, as theoretically the distribution of effective resistances could remain the same for a different configuration of actual resistances. One means by which systems evolve toward increased consumption through network change is through self-organisation into hierarchical branching structures, which are prevalent in both naturally-occurring and human-engineered systems [18,19]. In these networks, multiple downstream consumers may draw resource flow from the same resource, although this causes increased frictional losses by increasing the I term in Eq 3. This is offset in many systems by the development of higher-capacity links along shared pathways, such as preferential flow paths [11,35]. This is equivalent to varying the link strength in the equation for R (see Methods). Branching simulations The slope of (c) is exactly 1. Units are generalised units of power, potential, and resource flow. A copy of (a) with raw data is included in S1 Fig. slope of (b) and the standard deviation of effective resistance (σRE). These illustrate the main equations derived in the presentation of the modelling framework. Here, each coloured point range represents a different network topology over which the simulations were run. The slope of (c) is exactly 1. Units are generalised units of power, potential, and resource flow. A copy of (a) with raw data is included in S1 Fig. https://doi.org/10.1371/journal.pone.0229956.g002 slope of (b) and the standard deviation of effective resistance (σRE). These illustrate the main equations derived in the presentation of the modelling framework. Here, each coloured point range represents a different network topology over which the simulations were run. The slope of (c) is exactly 1. Units are generalised units of power, potential, and resource flow. A copy of (a) with raw data is included in S1 Fig. https://doi.org/10.1371/journal.pone.0229956.g002 https://doi.org/10.1371/journal.pone.0229956.g002 branching pattern around a single resource (‘fully branched’ network, Fig 4A). In the second, a branching network was artificially evolved from a nearly radial burst pattern, by adding in consecutive levels of non-demand junctions or ‘branch points,’ and re-calculating the con- sumption (‘evolved branching’ networks, Fig 4B). In the ‘evolved branching’ networks, at each iteration of the evolution, the average link length became shorter, and the network became more similar to a fractal branching structure. This was done to observe how power consump- tion was affected by changing the architecture to reflect known optimal distribution patterns, without increasing the number of consumers in the network. In the ‘fully branched’ network, both the total quantity of power consumption, and inequal- ity of consumer potential and power, were considerably higher than in the other architectures illustrated in Fig 5 (see S1 Table). In contrast, the ‘evolved branching’ simulations showed lower total power consumption and no inequality present in the final stage of network evolu- tion, as the consumers were all placed equal path distances from the resource, despite being at slightly differing Euclidean distances. This demonstrates that the self-similar branching archi- tecture itself does not lead to inequality, but rather the hierarchical or otherwise heterogeneous distribution of consumers. Branching simulations To illustrate the dynamics of branching networks more clearly, another set of simulations was performed, featuring idealised self-similar hierarchical branching networks. In these simula- tions, two networks were constructed. In the first, the network had consumers arranged in a 7 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks Fig 2. For the six example networks, (a) the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (VC=VR), (b) the relationship between the standard deviation of consumer final power (σPC) and resource flow squared (I2), and (c) the relationship between the OS ONE Maximum power and inequality in resource distribution networks OS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 8 / 19 Fig 2. For the six example networks, (a) the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (VC=VR), (b) the relationship between the standard deviation of consumer final power (σPC) and resource flow squared (I2), and (c) the relationship between the Fig 2. For the six example networks, (a) the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (V =V ) (b) the relationship between the standard deviation of consumer final power (σPC) and resource flow squared (I2) and (c) the relationship between the Fig 2. For the six example networks, (a) the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (VC=VR), (b) the relationship between the standard deviation of consumer final power (σPC) and resource flow squared (I2), and (c) the relationship between the Fig 2. For the six example networks, (a) the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (VC=VR), (b) the relationship between the standard deviation of consumer final power (σPC) and resource flow squared (I2), and (c) the relationship between the PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 8 / 19 PLOS ONE Maximum power and inequality in resource distribution networks slope of (b) and the standard deviation of effective resistance (σRE). These illustrate the main equations derived in the presentation of the modelling framework. Here, each coloured point range represents a different network topology over which the simulations were run. Branching simulations https://doi.org/10.1371/journal.pone.0229956.g004 Fig 4. (a) A ‘fully branched’ network, with consumers at each junction, and (b) ‘evolved branching’ networks illustrating the addition of branch points and links over the course of the simulations. In each network, the green square is the resource, and the blue circles are consumers. In the ‘evolved branching’ networks, the branch points, represented by triangles, and links of the same colour denote when they were added during the evolution of branching: black links are the original network with no branch points, purple links and branch points are the first level of branching, and gold links and branch points are the second level, which also includes some branch points from previous levels. The network shown here is simplified for illustration purposes: the simulated ‘evolved branching’ networks contained seven levels of branch points at the final stage of development, and 512 consumers. https://doi.org/10.1371/journal.pone.0229956.g004 https://doi.org/10.1371/journal.pone.0229956.g004 network maximum final power, which appears to show power-law properties (Fig 6). While the focus of the work here is on spatial networks, hierarchies can also emerge in relational ‘scale-free’ networks. These are often represented as hub-and-spoke topologies, with power Fig 5. The total final power consumption (P) against the ratio of mean consumer potential to mean resource potential (VC=VR) for each level of the ‘evolved branching’ networks. With additional levels of branching, the network became more similar to a fractal branching structure: average link length shortened, and resource flow was concentrated onto fewer, more shared links. Here, each coloured point series represents the trajectory of final power consumption as the network became more branched: Level 0 had no branch points, and Level 7 was a fully self-similar fractal. Relative total final power is the sum of final power consumption at all consumer nodes, normalised by the maximum power achieved by the network, which preserves relative differences. A copy of the figure with raw data is included in S2 Fig. https://doi.org/10.1371/journal.pone.0229956.g005 Fig 5. The total final power consumption (P) against the ratio of mean consumer potential to mean resource potential (VC=VR) for each level of the ‘evolved branching’ networks. With additional levels of branching, the network became more similar to a fractal branching structure: average link length shortened, and resource flow was concentrated onto fewer, more shared links. Branching simulations In the fully branched network, the underlying hierarchical spatial distribution of consumer nodes and links led to a highly skewed distribution of consumer potentials and final power at Fig 3. Density plots of normalised consumer final power (PC) for the six example networks, shown over decreasing ratios of mean consumer to mean resource potential, VC=VR. Each plot shows the density for the normalised consumer final power at VC=VR = 0.75, 0.5, and 0.25, from left to right, as the ratio decreases due to increased resource flow during the simulation. The data were normalised by subtracting the mean consumer power at each ratio level, and dividing by the standard deviation of consumer power at VC=VR = 0.75, such that the width of the first subplot for each network is one standard deviation. https://doi.org/10.1371/journal.pone.0229956.g003 Fig 3. Density plots of normalised consumer final power (PC) for the six example networks, shown over decreasing ratios of mean consumer to mean resource potential, VC=VR. Each plot shows the density for the normalised consumer final power at VC=VR = 0.75, 0.5, and 0.25, from left to right, as the ratio decreases due to increased resource flow during the simulation. The data were normalised by subtracting the mean consumer power at each ratio level, and dividing by the standard deviation of consumer power at VC=VR = 0.75, such that the width of the first subplot for each network is one standard deviation. https://doi.org/10.1371/journal.pone.0229956.g003 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 9 / 19 PLOS ONE Maximum power and inequality in resource distribution networks Fig 4. (a) A ‘fully branched’ network, with consumers at each junction, and (b) ‘evolved branching’ networks illustrating the addition of branch points and links over the course of the simulations. In each network, the green square is the resource, and the blue circles are consumers. In the ‘evolved branching’ networks, the branch points, represented by triangles, and links of the same colour denote when they were added during the evolution of branching: black links are the original network with no branch points, purple links and branch points are the first level of branching, and gold links and branch points are the second level, which also includes some branch points from previous levels. The network shown here is simplified for illustration purposes: the simulated ‘evolved branching’ networks contained seven levels of branch points at the final stage of development, and 512 consumers. https://doi.org/10.1371/journal.pone.0229956.g005 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 Branching simulations Here, each coloured point series represents the trajectory of final power consumption as the network became more branched: Level 0 had no branch points, and Level 7 was a fully self-similar fractal. Relative total final power is the sum of final power consumption at all consumer nodes, normalised by the maximum power achieved by the network, which preserves relative differences. A copy of the figure with raw data is included in S2 Fig. 10 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks Fig 6. The frequency distribution of consumer potentials (VC) at maximum network final power for a hierarchically branched network, plotted on log-log axes. The highly heterogeneous consumer potentials are due to the hierarchical network structure shown in Fig 4A. https://doi.org/10.1371/journal.pone.0229956.g006 Fig 6. The frequency distribution of consumer potentials (VC) at maximum network final power for a hierarchically branched network, plotted on log-log axes. The highly heterogeneous consumer potentials are due to the hierarchical network structure shown in Fig 4A. Fig 6. The frequency distribution of consumer potentials (VC) at maximum network final power for a hierarchically branched network, plotted on log-log axes. The highly heterogeneous consumer potentials the hierarchical network structure shown in Fig 4A. https://doi.org/10.1371/journal.pone.0229956.g006 https://doi.org/10.1371/journal.pone.0229956.g006 law distributions of node degrees. Power law or similarly heavy-tailed distributions in physical systems are typically described as resulting from interactions between interdependent compo- nents [36], but the simulations here demonstrate how this distribution can also occur as a result of the spatial organisation of interacting components. It is therefore possible that similar processes give rise to scale-free characteristics both spatially, as in self-similar hierarchical branching, and relationally, as in a power-law distribution of node degrees. Notably, although self-similar hierarchical branching networks such as the ‘fully branched’ network can achieve a higher maximum power at the network level, most individual consum- ers would have higher power if they had direct links to the resource, such as in the radial burst networks. Therefore, branching is still only energetically advantageous to the overall system, and those positioned close to the resource within the network architecture. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 be enforced despite being sub-optimal for more distant consumers, and potentially the net- work as a whole (Fig 7). be enforced despite being sub-optimal for more distant consumers, and potentially the net- work as a whole (Fig 7). Branching simulations Each series represents the relative final power consumption of consumers at that level in the network, where Level 0 is the consumers closest to the resource, and Level 7 are the consumers furthest from the resource in the network. As the resource flow increases across the network, the more distant consumers experience disproportionally greater frictional losses and therefore power losses, while consumers closer to the resource continue to increase in power. Values have been normalised by the maximum consumer final power and maximum consumer resource flow, which preserves relative differences. A copy of the figure with raw data is included in S3 Fig. Branching simulations In addition, these optimally located and connected consumers experience increased final power even after the total network final power begins to decrease (Fig 7), due to the larger frictional losses experi- enced by the more distant consumers along the bottom level of the network, who have higher effective resistance (Fig 4A). This suggests that hierarchical organisation is only beneficial to the system if the consumers located further from the resource benefit from the overall system operating at a higher maximum power: the more peripheral elements need to gain some of the system-level returns. One example of hierarchical branching as a system-optimal configuration in this way is in the circulatory system of some organisms, where more distant organs and limbs may benefit from the hierarchical organisation of the whole system, even if their individ- ual blood pressure and oxygen levels are lower. Alternatively, if the consumers in more ener- getically privileged locations exerted enough dominance over the system, the hierarchy could PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 11 / 19 Maximum power and inequality in resource distribution networks PLOS ONE Fig 7. The relative final power of consumers (PC) at each level of the ‘fully branched’ network, as related to the relative resource flow to each consumer (IC). The vertical black line denotes the relative resource flow associated with network-wide maximum final power. Each series represents the relative final power consumption of consumers at that level in the network, where Level 0 is the consumers closest to the resource, and Level 7 are the consumers furthest from the resource in the network. As the resource flow increases across the network, the more distant consumers experience disproportionally greater frictional losses and therefore power losses, while consumers closer to the resource continue to increase in power. Values have been normalised by the maximum consumer final power and maximum consumer resource flow, which preserves relative differences. A copy of the figure with raw data is included in S3 Fig. https://doi.org/10.1371/journal.pone.0229956.g007 Fig 7. The relative final power of consumers (PC) at each level of the ‘fully branched’ network, as related to the relative resource flow to each consumer (IC). The vertical black line denotes the relative resource flow associated with network-wide maximum final power. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 Conclusion This work has explored the characteristics of complex networks evolving toward maximum power production, and the relationship between the development and dynamics of these net- works and the inequality of resource distribution through them. The derived equations and illustrative simulations related the potential, resource flow, power, and resistance across a net- work of resources and consumers, and illustrated how those relationships changed as the net- work evolved toward maximum power, through adaptation in network state, architecture, or both. Specifically, it was shown that if the network structure consists of unequal link resis- tances, resulting from heterogeneity in path distance or connectivity in the network, the inequality of resource distribution will increase as the quantity of resource flow across the net- work increases. The potential for this architecturally-driven inequality is seen most promi- nently in hierarchical structures, such as the branching architectures common across in biological, environmental, and human-engineered systems (see e.g. [19,37–39]). Additionally, this hierarchical branching was shown to only increase the energy transferred through the network at maximum power at the scale of the entire network, and for the con- sumers located and connected closely to the resources. In contrast, more distant consumers in PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 12 / 19 PLOS ONE Maximum power and inequality in resource distribution networks these architectures experienced rapid decreases in energy consumption as the resource flow through the network increased, due to higher frictional losses of energy in transport. While prescription is not a focus of the current work, it has illustrated how RADE networks, and spe- cifically hierarchical branching architectures, can be fundamentally linked to the deep inequal- ity experienced by those served by these networks. Explicitly structuring these networks in an attempt to equalise distribution could take the form of co-locating resources and end-users to the greatest extent possible, such as locating solar panels or other forms of renewable energy on homes and businesses [40], or increasing the integration of locally-sourced products into a community’s food system [41]. Additional efforts, such as intentionally improving RADE net- work infrastructure to currently underserved populations of end users [42], could also be a sig- nificant step in the right direction. The question remains, however, as to whether even the best efforts at improving equality of distribution can offset the argued thermodynamic trajectory for systems to develop increasing patterns of consumption and dissipation (see e.g. Conclusion [12]), which appears to be most effectively facilitated by inherently unequal distribution networks. https://doi.org/10.1371/journal.pone.0229956.t001 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 Required simulation inputs 13 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks The topologies simulated here included planes, spheres, and sphere surfaces. Planes and spheres can be classed as two- and three-dimensional spaces, respectively, while sphere sur- faces are of a more ambiguous dimension (see e.g. [1]). The exploration of these three relevant topologies, commonly used to represent idealised spaces in physical systems, allowed identifi- cation of any effect on power consumption or resource distribution due to dimensionality. In these networks, the size of the topology, measured in generalised units as the radius of the sphere or sphere surface, or one side of the square plane, was determined by the number of nodes of each type, Size ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 10nC 100nR p ; ð6Þ ð6Þ Size ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 10nC 100nR p ; fifififififififififififififififififififififififififififi where nC is the number of consumers, and nR is the number of resources. This was chosen as it allowed for meaningfully large distances between nodes in networks with multiple consumer and resource nodes. The branched networks had set lengths for each link, such that topology size was not a factor. The relationship between spatial size and power distribution and consumption was not directly explored, such as by spreading the same network architecture across a larger area, but the linearity of the equation for resistance with unit-strength links suggests that inequality in consumer potential would increase linearly, and power consumption would decrease linearly, with increases in topological size. Similarly, the resource potentials were chosen to provide a clear visualisation of the maximum power ‘curve’ (Fig 2A), but a range was not explored, as increasing or decreasing the resource potential(s) would simply linearly increase or decrease the consumer potentials (see Eq 1). In all simulations with random and radial burst topologies, link strength was set to 1. In the branching simulations, it was set to be proportional to the resource flow, squared, to offset the increased frictional losses from higher resource flow along shared links. Specifically, by re- arranging Eq 1, the potential gradient along a link can be calculated as a product of resource flow I and link resistance R. Required simulation inputs Recall that power loss along a link PL is a product of this potential gradient and resource flow along it (Eq 3), which when combined with Eq 1 gives PL ¼ I2L S : ð7Þ ð7Þ Since losses are proportional to the resource flow squared, it rapidly dominates the energy losses. Therefore, as branching networks combine resource flows onto shared branches, they experience higher flow-driven losses on those shared links, despite having lower total network resistance, due to the shared links shortening the total path length around the network. It fol- lows that, for the branching to be energetically advantageous, the link strength must be a func- tion of resource flow, S = f{I}. If PL/I2, then S/I2, resulting in the power loss becoming a function exclusively of link length (Eq 7). This allows the advantages of shorter total link length in a branching network to be realised. Required simulation inputs The simulation code required a Comma-Separated Values (CSV) file to specify parameterisa- tion, including the number of nodes of each type, the size and shape of the spatial topology where they were distributed, whether links were all unit strength or potentially heterogeneous, and the file paths of the CSV files storing the locations of the nodes, or specifying random con- sumer placement. A complete list of the parameters required, and a description of each, is listed in Table 1. Table 1. Modified load flow methodology input parameters and description. Parameter name Description topology The name of the shape in or on which the nodes are distributed. Values: SPHERE (nodes located within a sphere of a given radius), SPHERE_SURFACE (nodes located on the surface of a sphere a given radius),PLANE (nodes located on the surface of a plane). pNoConnection The probability of two nodes not connecting, in a network with random links. noConnection The placeholder value in the connections matrix for non-connected nodes. resourcesFile The file path of the CSV file storing the coordinate locations and potentials of the resources. planeMaxCoords The maximum coordinates of the plane, stored as a pair of values separated with a semi-colon (e.g. 100;100). sphereR The radius of the sphere, or sphere surface. nBranchPoints The number of branch points. nConsumers The number of consumers. useStrength Whether or not to use link strength in calculating the resistance between nodes. Values: TRUE/FALSE. strengthExponent The exponent to which the link strength, if used, should be raised. manualNetwork Whether to read in a pre-specified connections matrix or generate the links randomly. Values: TRUE/FALSE. randomConsumers Whether to distribute the consumers randomly in the topology or use specified locations. Values: TRUE/FALSE. consumersFile The file path of the CSV file storing the coordinate locations of the consumers (if not random). matrixFile The file path of the CSV file storing the connections matrix, if a pre-specified one is used. branchPointsFile The file path of the CSV file storing the coordinate locations of the branch points (if used). outputCSV The file path to the CSV file where the output of the code run is stored. Includes the resource flow specification per consumer, the power and potential at each resource and consumer, and the total link length of the network. Table 1. Modified load flow methodology input parameters and description. Simulation code operation An overview of the simulation code is shown in Fig 8. After the program read in the specified parameters above, it created a customised data structure to store the node locations, resource potentials, and connections matrix. If the consumer locations were random, the program placed each consumer in space by drawing each coordinate from a uniform distribution bounded by the maximum topology coordinates supplied. If the links were random, the pro- gram put a link between each node, except for resources, with the probability of 1 –pNoCon- nection parameter described above (Table 1). PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 14 / 19 PLOS ONE Maximum power and inequality in resource distribution networks Fig 8. Code flow diagram for flow calculator program. The main controller of the program reads in the paramete and creates the network, and eventually terminates the program when complete, while the main calculations of the program are based on an iterative matrix inversion process in the flow calculator class. https://doi.org/10.1371/journal.pone.0229956.g008 Fig 8. Code flow diagram for flow calculator program. The main controller of the program reads in the parameters and creates the network, and eventually terminates the program when complete, while the main calculations of the program are based on an iterative matrix inversion process in the flow calculator class. https://doi.org/10.1371/journal.pone.0229956.g008 To calculate the resource flows and power consumption of the network, the program con- structed a Jacobian matrix representing the conductance of the network, or the inverse of the resistance, using the connections matrix. This was inverted to solve for the mismatch between specified and received resource flow at each consumer node, based on the consumer poten- tials. These were determined by the load flow equations using a matrix form of Eq 1. The potentials at the consumers, and branch points if used, were then adjusted to counter the mismatch. The matrix inversion and mismatch calculations were repeated until the PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 15 / 19 PLOS ONE Maximum power and inequality in resource distribution networks mismatches were within the specified error threshold of 0.001. Simulation outputs The program output was a single CSV file, with the potential at each resource and consumer, and the power production and consumption of each resource and consumer, respectively, at each resource flow specification tested. It also included the total link length of the network, which does not change over the duration of the simulations. Simulation code operation After convergence, the total power consumption of the network PNetwork was calculated as the sum of the power consump- tion at each consumer PCi, which was the product of potential VCi and resource flow ICi (Eq 2): PNetwork ¼ PnConsumers i¼1 PC: ð8Þ ð8Þ Initially, the total power consumption was calculated for 1 unit of resource flow arriving at each consumer. With each iteration, the specification was incremented by 0.1 unit, and the resource flows were re-calculated. This was repeated until either 1000 units of resource flow was arriving at each consumer, or the power consumption of the network was negative, due to the inverse relationship between consumer potential and resource flow (Eq 1). In the evolving branching simulations, an additional level of branch points was added, and the links between nodes re-arranged, after the iterations had completed for a given network, until there were 7 levels of branch points between the resource and consumers (Fig 4B). Code availability statement and languages used A complete copy of the code, along with usage instructions, a sample parameter file, and sam- ple resource, consumer, branch point, and matrix CSVs, is available upon request. The code is written in Java Version 8. All figures and analyses were generated using R, including the base R package version 3.6.1 [43], ggplot2 [44], and the rgl package [45]. Project administration: Natalie Davis. Supervision: Andrew Jarvis, M. J. Aitkenhead, J. Gareth Polhill. Visualization: Natalie Davis. Writing – original draft: Natalie Davis. Writing – original draft: Natalie Davis. Writing – original draft: Natalie Davis. Writing – review & editing: Natalie Davis, Andrew Jarvis, M. J. Aitkenhead, J. Gareth Polhill. S1 Text. Maximum power derivation. (DOCX) Non-normalised version of Fig 7, showing final power of consumers (PC) at each level of the ‘fully branched’ network, as related to the resource flow to each consumer (IC). The units are generalised units of power and resource flow, rather than units only applicable to a specific type or types of resource distribution network. (TIF) Author Contributions Conceptualization: Natalie Davis, Andrew Jarvis, J. Gareth Polhill. Data curation: Natalie Davis. Formal analysis: Natalie Davis. Investigation: Natalie Davis. Methodology: Natalie Davis. Project administration: Natalie Davis. Software: Natalie Davis. Supervision: Andrew Jarvis, M. J. Aitkenhead, J. Gareth Polhill. Visualization: Natalie Davis. Writing – original draft: Natalie Davis. Writing – review & editing: Natalie Davis, Andrew Jarvis, M. J. Aitkenhead, J. Gareth P Conceptualization: Natalie Davis, Andrew Jarvis, J. Gareth Polhill. Data curation: Natalie Davis. Formal analysis: Natalie Davis. Investigation: Natalie Davis. Methodology: Natalie Davis. Project administration: Natalie Davis Conceptualization: Natalie Davis, Andrew Jarvis, J. Gareth Polhill. Methodology: Natalie Davis. S1 Text. Maximum power derivation. (DOCX) S1 Table. Parameterisation and power consumption details of networks simulated: a) branched networks, and b) random and radial burst networks. The slope of the linear relation- ship between sPC and I2 shown in Fig 1 was calculated for each plot using least-squares regres- sion, and compared to the value of sRE calculated using Eq 5 and the consumer and resource potentials for each network, shown in the table here. The least-squares regression estimate of sRE is shown in brackets below the original estimate using consumer and resource potentials, for the networks plotted. S1 Fig. Non-normalised version of Fig 2A, showing the relationship between total final power (P) and the ratio of mean consumer potential to mean resource potential (VC=VR) for six example networks. Each coloured point range represents a different network topology over which the simulations were run. The units are generalised units of power, rather than units only applicable to a specific type or types of resource distribution network. (TIF) S2 Fig. Non-normalised version of Fig 5, showing total final power consumption (P) against the ratio of mean consumer potential to mean resource potential (VC=VR), for the PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 16 / 19 PLOS ONE Maximum power and inequality in resource distribution networks ‘evolved branching’ networks. The units are generalised units of power, rather than units only applicable to a specific type or types of resource distribution network. (TIF) ‘evolved branching’ networks. The units are generalised units of power, rather than units only applicable to a specific type or types of resource distribution network. (TIF) S3 Fig. Non-normalised version of Fig 7, showing final power of consumers (PC) at each level of the ‘fully branched’ network, as related to the resource flow to each consumer (IC). The units are generalised units of power and resource flow, rather than units only applicable to a specific type or types of resource distribution network. (TIF) S3 Fig. Non-normalised version of Fig 7, showing final power of consumers (PC) at each level of the ‘fully branched’ network, as related to the resource flow to each consumer (IC). The units are generalised units of power and resource flow, rather than units only applicable to a specific type or types of resource distribution network. (TIF) S3 Fig. References A General Model for the Origin of Allometric Scaling Laws in Biology. Science (80-) [Internet]. 1997; 276(April):122–126. Available from: http://www.sciencemag.org/cgi/ content/abstract/276/5309/122 19. Banavar JR, Maritan A, Rinaldo A. Size and form in efficient transportation networks. Nature [Internet]. 1999 May; 399(6732):130–2. Available from: http://www.nature.com/articles/20144 https://doi.org/10. 1038/20144 PMID: 10335841 20. Kirkham MB. Principles of Soil and Plant Water Relations [Internet]. 2nd ed. Cambridge, Massachu- setts: Elsevier; 2014. 598 p. Available from: https://linkinghub.elsevier.com/retrieve/pii/C20130128711 21. Odum HT. Environment, Power and Society. New York: Wiley; 1971. 331 p. 22. Odum HT. Biological circuits and the marine systems of Texas. In: Olson TA, Burgess FJ, editors. Pollu- tion and Marine Ecology. New York: Wiley Interscience; 1967. p. 99–157. 23. Odum HT. Explanations of ecological relationships with energy systems concepts. Ecol Modell [Inter- net]. 2002; 158(3):201–11. Available from: http://www.sciencedirect.com/science/article/pii/ S0304380002002326 24. Collier CG. Modelling a river catchment using an electrical circuit analogue. Hydrol Earth Syst Sci. 2010; 2(1):9–18. 25. Wang L, Zou C, O’Donnell F, Good S, Franz T, Miller GR, et al. Characterizing ecohydrological and bio- geochemical connectivity across multiple scales: a new conceptual framework. Ecohydrology [Internet]. 2012 Mar; 5(February):221–33. Available from: http://www3.interscience.wiley.com/journal/ 122653919/abstract 26. Levins R. The strategy of model building in population biology. Am Sci. 1966; 54(4):421–31. 27. Humphries MM, Thomas DW, Kramer DL. The Role of Energy Availability in Mammalian Hibernation: A Cost-Benefit Approach. Physiol Biochem Zool [Internet]. 2003 Mar; 76(2):165–79. Available from: https://doi.org/10.1086/367950 PMID: 12794670 28. Paul CR. Fundamentals of Circuit Analysis. Hoboken, New Jersey: John Wiley and Sons; 2001. 29. Bejan A. Maximum power from fluid flow. Int J Heat Mass Transf. 1996; 39(6):1175–81 30. Santilla´n M, Angulo-Brown F. A thermodynamic approach to the compromise between power and effi- ciency in muscle contraction. J Theor Biol. 1997; 189(4):391–8. https://doi.org/10.1006/jtbi.1997.0526 PMID: 9446748 31. Sutcliffe B. A Converging or Diverging World? [Internet]. Vol. 63, DESA Working Paper. 2005. Available from: www.un.org/esa/papers 32. Niño-Zarazu´a M, Roope L, Tarp F. Global Inequality: Relatively Lower, Absolutely Higher. Rev Income Wealth. 2017; 63(4):661–84. 33. von Meier A. Power Flow Analysis. In: Electric Power Systems: A Conceptual Introduction. 2006. p. 195–228. 34. Boswell GP, Jacobs H, Davidson FA, Gadd GM, Ritz K. Functional consequences of nutrient transloca- tion in mycelial fungi. J Theor Biol [Internet]. 2002; 217(4):459–77. Available from: http://www.ncbi.nlm. nih.gov/pubmed/12234753 https://doi.org/10.1006/jtbi.2002.3048 PMID: 12234753 35. Zehe E, Blume T, Kleidon A, Ehret U, Scherer U, Westhoff M. References 1. Jarvis AJ, Jarvis SJ, Hewitt CN. Resource acquisition, distribution and end-use efficiencies and the growth of industrial society. Earth Syst Dyn. 2015; 6(2):689–702. 2. Lotka AJ. Contribution to the Energetics of Evolution. Proc Natl Acad Sci [Internet]. 1922; 8(6):147–51. Available from: https://doi.org/10.1073/pnas.8.6.147 PMID: 16576642 3. Boltzmann L V. Popula¨re Schriften. Leipzig: J.A. Barth; 1905. 454 p. 4. Mattison SM, Smith EA, Shenk MK, Cochrane EE. The evolution of inequality. Evol Anthropol Issues, News, Rev [Internet]. 2016 Jul; 25(4):184–99. Available from: http://doi.wiley.com/10.1002/evan.21491 5. Charlton BG. The Inequity of Inequality. J Health Psychol [Internet]. 1997 Jul; 2(3):413–25. Available from: https://doi.org/10.1177/135910539700200309 PMID: 22013030 6. Georgescu-Roegen N. The Entropy Law and the Economic Process in Retrospect. East Econ J. 1986; 12(1):3–25. 7. Bejan A, Errera MRR. Wealth inequality: The physics basis. J Appl Phys [Internet]. 2017 Mar 28; 121 (12):124903. Available from: http://aip.scitation.org/doi/10.1063/1.4977962 8. Ulanowicz RE. Some steps toward a central theory of ecosystem dynamics. Comput Biol Chem [Inter- net]. 2011 Dec; 27:523–30. Available from: http://link.springer.com/10.1007/978-0-8176-4904-3 9. Odum HT, Pinkerton RC. Time’s speed regulator: The optimum efficiency for maximum power output in physical and biological systems. Am Sci. 1955; 43(2):331–43. 10. Kleidon A. Thermodynamic Foundations of the Earth System. First. Cambridge, UK: Cambridge Uni- versity Press; 2016. 396 p. 11. Kleidon A, Zehe E, Ehret U, Scherer U. Thermodynamics, maximum power, and the dynamics of prefer- ential river flow structures at the continental scale. Hydrol Earth Syst Sci. 2013; 17(1):225–51. 12. Kleidon A, Malhi Y, Cox PM. Maximum entropy production in environmental and ecological systems. Philos Trans R Soc B Biol Sci. 2010; 365(1545):1297–302. 17 / 19 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 March 10, 2020 PLOS ONE Maximum power and inequality in resource distribution networks 13. Mansson BA, McGlade JM. Ecology, thermodynamics and H.T. Odum’s conjectures. Oecologia. 1993; 93(4):582–96. https://doi.org/10.1007/BF00328969 PMID: 28313829 14. Ross J, Corlan AD, Mu¨ller SC. Proposed Principles of Maximum Local Entropy Production. J Phys Chem B [Internet]. 2012 Jul 12; 116(27):7858–65. Available from: https://doi.org/10.1021/jp302088y PMID: 22746154 15. Polettini M. Fact-checking Ziegler’s maximum entropy production principle beyond the linear regime and towards Steady States. Entropy. 2013; 15(7):2570–84. 16. Odum HT. Maximum power and efficiency: A rebuttal. Ecol Modell [Internet]. 1983 Oct; 20(1):71–82. Available from: https://linkinghub.elsevier.com/retrieve/pii/0304380083900327 17. Martyushev LM, Seleznev VD. The restrictions of the maximum entropy production principle. Phys A Stat Mech its Appl [Internet]. 2014; 410:17–21. Available from: http://link.springer.com/10.1007/978-3- 642-40154-1 18. West GB, Brown JH, Enquist BJ. PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 References Connected flow paths as first order con- trol on critical zone water flows: coincidence or self-organized optimality? Hydrol Earth Syst Sci Discuss. 2012; 9(9):10595–655. 36. Parunak HVD, Brueckner S, Savit R. Universality in multi-agent systems. In: Proceedings of the Third International Joint Conference on Autonomous Agents and Multi-Agent Systems. New York; 2004. 37. West GB, Brown JH, Enquist BJ. The Fourth Dimension of Life: Fractal Geometry and Allometric Scal- ing of Organisms. Science (80-) [Internet]. 1999; 284(5420):1677–9. Available from: http://www. sciencemag.org/cgi/doi/10.1126/science.284.5420.1677 PLOS ONE \| https://doi.org/10.1371/journal.pone.0229956 March 10, 2020 18 / 19 PLOS ONE Maximum power and inequality in resource distribution networks 38. Hines P, Blumsack S, Sanchez EC, Barrows C. The topological and electrical structure of power grids. Proc Annu Hawaii Int Conf Syst Sci. 2010; 39. Tero A, Takagi S, Saigusa T, Ito K, Bebber DP, Fricker MD, et al. Rules for biologically inspired adaptive network design. Science (80-). 2010; 327(5964):439–42. 40. Alstone P, Gershenson D, Kammen DM. Decentralized energy systems for clean electricity access. Nat Clim Chang [Internet]. 2015; 5(4):305–14. Available from: http://dx.doi.org/10.1038/nclimate2512 41. Martinez S. Local Food Systems: Concepts, Impacts, and Issues. Darby, Pennsylvania: DIANE Pub- lishing; 2010. 42. Brelsford C, Martin T, Hand J, Bettencourt LMA. Toward cities without slums: Topology and the spatial evolution of neighborhoods. Sci Adv [Internet]. 2018; 4(8):1–9. Available from: http://advances. sciencemag.org/lookup/doi/10.1126/sciadv.aar4644 43. R Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna, Austria; 2018. Available from: https://www.r-project.org/ 44. Wickham H. ggplot2: Elegant Graphics for Data Analysis [Internet]. New York: Springer-Verlag; 2009. Available from: http://ggplot2.org 45. Adler D, Murdoch D. rgl: 3D Visualization using OpenGL. 2019. 45. Adler D, Murdoch D. rgl: 3D Visualization using OpenGL. 2019. 19 / 19
https://openalex.org/W4306661063	https://www.researchsquare.com/article/rs-2097082/latest.pdf	English	null	Clinical nurse competence and its effect on patient safety culture: A Systematic Review	Research Square (Research Square)	2,022	cc-by	6,962	Background Unsafe health practices are highly regarded to cause disability and death. It is estimated that the chance that unsafe practice can cause harm for the patient is 1 in 300 chance (World Health Organizaion, 2019b). Nearly 400,000 deaths occur annually in the United States due to several reversible adverse events, such as medication error, infection transmission, and fall events (Armando Nahum, 2021). In addition, poor quality care can cause death and a global health burden (Kruk et al., 2018). Patient safety is a health care discipline that evolved as a result of the increasing sophistication of health care systems and the increasing in adverse outcomes in health-care facilities. Its goal is to avoid and decrease risks, mistakes, and harm to patients while providing health care. Therefore, reliable, safe, equitable, effective, and highly standardized patient-centered care has become the ultimate goal of all health care institutions worldwide (World Health Organizaion, 2019a). Besides that, patient safety culture focuses on organizational culture issues related to patient safety, , patient safety culture is concerned with internalizing safety beliefs, values, and attitudes, translating them into health care practices and commitment to maintaining an error-free health setting and emphasizing reporting culture (Khoshakhlagh et al., 2019). Competent nurses are key contributors to maintaining safe and effective health care services by integrating knowledge, skills, and attitudes that enable them to adapt to dynamic health environments (Fukada, 2018). Patient safety competencies are a core competency in the continuum of professional development activities that protect patients from unnecessary risks and hazards (Ayoung Huh & Juh Hyun Shin, 2021). A high level of competence promotes the achievement and compliance with the patient safety goal. Several studies found that patient safety culture and nurse safety competency are affected by many factors. For example, workplace regulations and climate, nursing fatigue, satisfaction, stress, demographics, type of health institution type, teamwork and learning opportunity, specialty, degree of bedside involvement, and job description are all factors that affect safety culture (Al Ma'mari et al., 2020; Halabi et al., 2021; Khoshakhlagh et al., 2019; Wang et al., 2019; Yan et al., 2021). Any improvement strategies to modulate these factors are unnegotiable. However, a better understanding of nurse competence and patient safety culture is essential to improve safe practice and professional development and minimize adverse events (Wami et al., 2016). Study Design We conducted a systematic review of the published literature from April 1, 2022, and April 11, 2022, by the electronic search from January 1, 2018, through May 1, 2022. We conducted a systematic review of the published literature from April 1, 2022, and April 11, 2022, by the elect May 1, 2022. Clinical nurse competence and its effect on patient safety culture: A Systematic Review Rasha Abu Zaitoun (  rasha.zaitoun@najah.edu ) An-Najah National University Hospital Nizar B. Said An-Najah National University Lila de Tantillo Jacksonville University Version of Record: A version of this preprint was published at BMC Nursing on May 19th, 2023. See the published version at https://doi.org/10.1186/s12912- 023-01305-w. Version of Record: A version of this preprint was published at BMC Nursing on May 19th, 2023. See the published version at https://doi.org/10.1186/s12912- 023 01305 Page 1/11 Page 1/11 Abstract Background: Unsafe health practices are one of the leading causes of disability and even death. Competent nurses are crucial to ensure safe and high-quality health care services. The patient safety culture is concerned with internalizing safety beliefs, values, and attitudes, translating them into health care practices, and committing to maintain an error-free health environment. A high level of competence ensures the achievement and compliance with the safety culture goal. Objective This systematic review aims to identify the relationship between the level of nursing competence and the safety culture score and perception among nurses at their workplace. Methods: Four international online databases were searched to find relevant studies published between 2018 and 2022. Peer-reviewed articles using quantitative methods, targeting nursing staff, written in English were included. After reviewing 117 identified articles, 16 full-text studies were included. The PRISMA 2020 checklist for systematic reviews was used. Results: Evaluation of the articles indicates safety culture, competency, and perception are assessed using a wide variety of instruments. Safety culture was generally perceived as positive. No unique and standard tool has been developed to investigate the effect of safety competency on the perception of the safety culture in a standardized way. Conclusions: Existing research provides evidence of a positive correlation between the nursing competence and patient safety score. Future research is recommended to investigate ways to measure the effect of nursing competency level on safety culture in health care institutions. Relevance for Clinical Practice Clinicians and leaders in the clinical setting may consider applying the findings of this study to address nursing competence as a component of addressing the patient safety culture in their practice setting. Background Therefore, this study aims to address the gap in knowledge between nursing competencies and the perception of patient safety among nurses in their workplace. This knowledge may be used to providing in-depth explanation to the following research questions: 1. Is there a valid and reliable international tool that helps nurses self-report their core competencies, especially in reliable international tool that helps nurses self-report their core competencies, especially in the safety domains? 2. How does the competency level affect the patient safety culture among nurses? 2. How does the competency level affect the patient safety culture among nurses? The design of the reviewed studies Of the 16 qualifying articles, 15 were cross-sectional descriptive studies, and only one applied a quasi-experimental, pretest-posttest design (Letourneau & McCurry, 2019). Three out of the sixteen articles were conducted in South Korea (Han & Roh, 2020; Han et al., 2020; A. Huh & J. H. Shin, 2021). Three studies originated from Saudi Arabia (Abdualrahman Saeed Alshehry, 2022; Halabi et al., 2021; Mahsoon & Dolansky, 2021). Two studies from Iran (Aghil Habibi Soola et al., 2022; Edris Kakemam et al., 2021). In addition, a study was conducted in each Australia (Connell et al., 2021), Palestine (Najjar et al., 2018), Brazil (Lousada et al., 2020), Jordan (Abdullah Khamaiseh et al., 2020), Spain (Peñataro-Pintado et al., 2022), England (Letourneau & McCurry, 2019), and China (Yan et al., 2021). Finally, one study had an unidentified setting or country of origin. Search Strategy Four international online databases, MEDLINE (via PubMed), CINHAL (via EBSCOhost), Scopus (via Elsevier), and Embase were searched for published studies that describing the relationship between nurse competencies and patient safety culture. Search terms were developed based on experience and keywords from similar research. The search was structured using Boolean operators (AND, OR) and consisted of MeSH terms and free terms related to nursing, patient safety, competency, and safety culture. The term 'OR' was used between keywords or comparable MeSH phrases; meanwhile, the Boolean operator ‘AND’ was used to Page 2/11 Page 2/11 connect phrases or keywords with different meanings to refine the search (see Appendix A). The search approach used for PubMed-MEDLINE is described in detail Appendix A. For the other databases, the strategy was essentially the same but adapted as appropriate to the characteristics of each. The review process was presented using the PRISMA statement 2020 (Page et al., 2021) (Fig. 1). connect phrases or keywords with different meanings to refine the search (see Appendix A). The search approach used for PubMed-MEDLINE is described in detail Appendix A. For the other databases, the strategy was essentially the same but adapted as appropriate to the characteristics of each. The review process was presented using the PRISMA statement 2020 (Page et al., 2021) (Fig. 1). The sampling technique Given the sampling technique, eight articles used the convenience sampling technique (Abdualrahman Saeed Alshehry, 2022; Connell et al., 2021; Aghil Habibi Soola et al., 2022; Halabi et al., 2021; A. Huh & J. H. Shin, 2021; Mahsoon & Dolansky, 2021; Son et al., 2019; Yan et al., 2021). The highest sample size was 2,836 participants in the study by Najjar et al. (2018), while the least was 56 participants in (Lousada et al., 2020) study. Measurement Tools For Safety Culture And Nursing Competency The construct concepts of patient safety culture, safety climate, and patient safety competency were operationally measured using various tools or instruments throughout the sixteen studies. Nursing safety competency was measured using the Patient Safety Competency Self-Evaluation Tool, the Health Professional Education in Patient Safety Survey (H-PEPSS) (Han et al., 2020), and the Nurses’ Attitudes and Skills Safety scale, the latest version of NASUS (Mahsoon & Dolansky, 2021). Furthermore, Abdualrahman Saeed Alshehry (2022) used the Health Professional Education in Patient Safety Survey ‘ (H- PEPSS) to assess safety competency. Halabi et al. (2021) used the short version of Nurse Professional Competence (NPC).The self-reported Patient Safety Competency Nurse Evaluation Scale (PSCNES) was used by Yan et al. (2021). Three studies applied the Hospital Survey on Patient Safety Culture (HSOPSC) developed by the Agency for Healthcare Research and Quality to measure safety culture (Han et al., 2020; E. Kakemam et al., 2021; Najjar et al., 2018). The safety climate was measured using the Safety Attitudes Questionnaire (SAQ) and the Safety Climate Survey (SCS); (Connell et al., 2021; A. Khamaiseh et al., 2020; Lousada et al., 2020) Results After applying the inclusion and exclusion criteria and removing duplicate studies, the remaining retrieved articles were examined thoroughly for further refinement of the results and review of the inclusion criteria. The net result of this thorough approach was sixteen articles qualifying for systematic qualitative analysis. Risk of Bias Assessment The methodological quality of the studies was evaluated by two reviewers (LZ and RZ). Disputes about the risk of bias were resolved through discussion and consensus between the two reviewers, with any remaining disagreements addressed to a third author (ES). Inclusion and Exclusion Criteria Each retrieved study was evaluated by title and abstract and included according to the following criteria: studies using the quantitative approach; targeting nursing staff; published in English from January 1, 2018, through May 1, 2022, and full free full text available. The following studies were excluded: studies in the form of letters, editorials, essays, case studies, comments or narrative, systematic reviews, and conference abstracts; articles published in other languages than English; studies with nursing students, and studies using qualitative approach; and studies of pre-hospital and ambulatory care. Measurement of safety nursing competency Eight of the studies, or 50% of the literature, in this review showed that nurses were more competent in communicating effectively than working in teams with colleagues, and their overall safety culture score was positive. Likewise, Zabin et al. (2022) found that both organizational learning and continuous improvement, as well as cooperation within units, received the highest composite frequency of patient safety perception. In the current findings, the reviewed literature mentioned four self-reported measurement tools. Despite that, there was no consensus on the best tool for measuring safety competencies, and the lack of a key self-reported tool for measuring safety culture and linking it to nursing safety competency limited the ability to directly assess the effect of nursing competency on safety culture (Arzahan et al., 2022). As a result, more research is needed to enrich the literature, improve the understanding of the effect of safety competency on safety culture scores among nurses, and help in providing more clear operational definition of safety culture and nursing safety competency (Kalteh et al., 2021). Although we used broad keywords to search different online databases, the retrieved articles did not specifically discuss the relationship between the dimensions of the safety culture, perception, and competency in nursing safety. Additionally, the lack of a standardized tool to measure the concepts of safety culture and nursing safety competency made it difficult to find a robust number of targeted studies and limited our ability to find specific operational definitions to the concepts of safety culture and nursing safety competency that was used consistently across the literature. Moreover, using self-reported surveys and relying on convenient sampling eased data collection and provide more objective data for many of the studies in this review. However, there were drawbacks that limited the generalizability and might not cover all aspects of the studies’ content. Therefore, future use of mixed designs with the use of qualitative methods is highly recommended to deepen the study issue and explore the unique relationship between the dimensions of the safety culture and nursing competency that can play a pivotal role in improving safety practices. This review study recognizes the importance of conducting additional searches and reviews and broadening the scope of keywords used to search online databases to focus on the core of this study. Relationship between Nursing Competencies and Safety Culture None of the reviewed articles explicitly explored the relationship between nursing competencies and the safety culture dimensions and how they affect each other. Furthermore, studies identified no specific comprehensive tool specifically to assess the relationship between nursing safety competency and the dimension of culture or safety climate safety. Rather, the included articles examined patient safety culture, health care safety climate, and other nursing competencies that influenced or affect safety climate independently. The Patient Safety Culture Patient safety culture was mentioned in the title of six studies (Han et al., 2020; E. Kakemam et al., 2021; A. Khamaiseh et al., 2020; Lousada et al., 2020; Mahsoon & Dolansky, 2021; Najjar et al., 2018). The was considerable variation of how this concept was perceived among nurses for example, (Lousada et al., 2020) study found that professionals working in home care services perceived higher scores related to safety culture compared with those working in primary care services. The accredited primary center in Jordan had an average positive response rate in some safety cultures ranging from 58.54–75.63% (A. Khamaiseh et al., 2020), A total of 32 Iranian teaching hospitals out of 150 reported poor patient safety culture (E. Kakemam et al., 2021). On the other hand, and with regard to the safety climate, the study by Connell et al. (2021) revealed that novice-competent nurses in Australian emergency departments rated the safety climate higher than expert nurses in all domains except stress recognition. Patient Safety Competency Patient Safety Competency Page 3/11 Page 3/11 Six studies examined patient safety competency and several connected safety competencies with the domain of safety culture. For example, Yan et al. (2021) assessed the safety competence of Chinese associate degree nurses and found it moderate. Aghil Habibi Soola et al. (2022) found a positive correlation among Iranian emergency between safety competence nurses with teamwork, psychological safety, leadership, communication, mutual support, situation monitoring, and team structure. Han and Roh (2020) found that the night shift among emergency nurses in Korean hospitals negatively affected safety competence and was significantly and positively correlated with the experience of the emergency department and reported adverse events by others. Six studies examined patient safety competency and several connected safety competencies with the domain of safety culture. For example, Yan et al. (2021) assessed the safety competence of Chinese associate degree nurses and found it moderate. Aghil Habibi Soola et al. (2022) found a positive correlation among Iranian emergency between safety competence nurses with teamwork, psychological safety, leadership, communication, mutual support, situation monitoring, and team structure. Han and Roh (2020) found that the night shift among emergency nurses in Korean hospitals negatively affected safety competence and was significantly and positively correlated with the experience of the emergency department and reported adverse events by others. In a medical city in Saudi Arabia, A. S. The Patient Safety Culture Alshehry (2022) studied the correlation between the conflict between nurses and the patient, and safety competencies. The study revealed that nurses were highly competent in effective communication, but they had the lowest competency in ‘working in teams with other health professionals.’ Nurses perceived ‘mistrust of motivations’ and “contradictory communication” got the poorer self-reported safety competency. Measurement of patient safety culture In this review, two self-reported questionnaires were found to have been applied to understand the dimensions and scores in health institutions. The Safety Attitude Questionnaire (SAQ) (A. Khamaiseh et al., 2020; Lousada et al., 2020) and the Hospital Survey on Patient Safety Culture (HSOPSC) (Han et al., 2020; E. Kakemam et al., 2021; Mahsoon & Dolansky, 2021; Najjar et al., 2018) were two assessment tools that employed Likert scales. However, in their study, Alsalem et al. (2018) revealed that five instruments are used to assess the patient safety culture and climate in health institutions. Furthermore, these tools vary in their psychometric properties and scope. The most common study design among the studies was a questionnaire-based descriptive quantitative approach. Using different study designs, such as qualitative research and a variety of data collection methods, could help improve understanding of the safety culture and health care provider perceptions and would be required to address the existing relationship between safety culture and nursing competency (Alqattan et al., 2019) f f Discussion The purpose of this study was to provide a systematic review of the literature investigating the relationship between nursing competencies and perception of patient safety among nurses in their workplace. In this systematic review, and after a thorough analysis of the entire manuscript of the retrieved articles, we selected and discussed sixteen articles based on their conformity with the inclusion criteria. Measurement of safety nursing competency In addition, it is recommended to investigate the effect of nursing competency on perceptions of dimensions of safety culture and to make valid comparisons between demographics and cultures. More research would provide a better understanding and may have a greater clinical impact and aid in improving and delivering highly effective, safe, and efficient care. Furthermore, the findings would support synchronicity between academic clinical programs and nursing staff safety practices. For example, nursing students receive competency-based training that allows them to live in a safe environment and directly implement the dimension of safety culture (Debourgh, 2012). Page 4/11 Page 4/11 Conclusions This systematic review draws several conclusions from the sixteen reviewed articles. First, the study showed no specific tool to measure the safety culture and nursing safety competency dimensions. Additionally, no study explicitly discussed the effect of nursing safety competency on safety culture scores among nursing staff. However, most of the studies employed a questionnaire-based descriptive approach. Conducting more research with different study designs such as the experimental, qualitative, and longitudinal approaches may enhance the understanding and assist in constructing a valid and more reliable tool to measure the effect of nursing safety competency on safety culture. Second, rigorous research needed to establish a well-designed competency-based training program to improve safety scores among more diverse demographics and cultures is needed. Finally, the findings can motivate administrators to promote safety culture in different health care facilities, as well as increase professional awareness of the factors that impact safety culture and, consequently, patient safety. Limitations This systematic review study has several limitations. First, we restricted the databases to four primary resources considered suitable for gathering eligible articles for the study purpose. The second is that the delimiters of this review included only articles in English, so some related articles may not have been included. Another limitation that should be highlighted is that we reviewed articles published between 2018 and 2022 to include the most recent data, but this also restricted the number of retrieved studies. Another potential limitation was that the results of the reviewed studies could not be generalized. The articles were only published in peer-reviewed journals to ensure the high quality of evidence and the reported findings, which omitted many worthwhile studies such as grey and unpublished studies. Additionally, the study designs of the majority of the retrieved papers were descriptive, which restricts the generalizability of their findings and prevents access to many relevant studies. Therefore, it is recommended to conduct further systematic or integrative reviews that might include qualitative and descriptive studies as well as expand the inclusion criteria for other types of literature. References Retrieved March 31, from https://blogs.cdc.gov/safehealthcare/patient-safety-action-plan/ 5. Armando Nahum. (2021, November 8, 2021). Patient safety representatives unite to implement global patient safety action plan in the united states. Centre for Disease Control and Prevention. Retrieved March 31, from https://blogs.cdc.gov/safehealthcare/patient-safety-action-plan/ 5. Armando Nahum. (2021, November 8, 2021). Patient safety representatives unite to implement global patient safety action plan in the united states. Centre for Disease Control and Prevention. Retrieved March 31, from https://blogs.cdc.gov/safehealthcare/patient-safety-action-plan/ 6. Arzahan, I. S. N., Ismail, Z., & Yasin, S. M. (2022). Safety culture, safety climate, and safety performance in healthcare facilities: a systematic review. Safety science, 147, 105624 . 6. Arzahan, I. S. N., Ismail, Z., & Yasin, S. M. (2022). Safety culture, safety climate, and safety performance in healthcare facilities: a systematic review. Safety science, 147, 105624 . 7. Connell, C. J., Cooper, S., & Endacott, R. (2021). Measuring the safety climate in an Australian emergency department. Int Emerg Nurs, 58, 101048. https://doi.org/10.1016/j.ienj.2021.101048 7. Connell, C. J., Cooper, S., & Endacott, R. (2021). Measuring the safety climate in an Australian emergency depar https://doi.org/10.1016/j.ienj.2021.101048 7. Connell, C. J., Cooper, S., & Endacott, R. (2021). Measuring the safety climate in an Australian emergency department. Int Emerg Nurs, 58, 101048. https://doi.org/10.1016/j.ienj.2021.101048 8. Debourgh, G. A. (2012). Synergy for patient safety and quality: academic and service partnerships to promote e practice. J Prof Nurs, 28(1), 48-61. https://doi.org/10.1016/j.profnurs.2011.06.003 9. Fukada, M. (2018 .)Nursing competency: Definition, structure and development. Yonago acta medica, 61(1) 2018 .)Nursing competency: Definition, structure and development. Yonago acta medica, 61(1), 001-007 . 10. Habibi Soola, A., Ajri-Khameslou, M., Mirzaei, A., & Bahari, Z. (2022). Predictors of patient safety competency among emergency nurses in Iran: a cross- sectional correlational study. BMC Health Services Research 22(1), 1-10.547. https://doi.org/10.1186/s12913-022-07962-y 10. Habibi Soola, A., Ajri-Khameslou, M., Mirzaei, A., & Bahari, Z. (2022). Predictors of patient safety competency among emergency nurses in Iran: a cross- sectional correlational study. BMC Health Services Research 22(1), 1-10.547. https://doi.org/10.1186/s12913-022-07962-y 0. Habibi Soola, A., Ajri-Khameslou, M., Mirzaei, A., & Bahari, Z. (2022). Predictors of patient safety competency am sectional correlational study. BMC Health Services Research 22(1), 1-10.547. https://doi.org/10.1186/s12913-0 11. Halabi, J. O., Nilsson, J., & Lepp, M. (2021). Professional competence among registered nurses working in hospitals in saudi arabia and their experien f li f i d i f J T l N 32(4) 425 433 h //d i /10 1177/1043659621992845 11. Halabi, J. References 1. Al Ma'mari, Q., Sharour, L. A., & Al Omari, O. (2020). Fatigue, burnout, work environment, workload and perceived patient safety culture among critical care nurses. Br J Nurs, 29(1), 28-34. https://doi.org/10.12968/bjon.2020.29.1.28 1. Al Ma'mari, Q., Sharour, L. A., & Al Omari, O. (2020). Fatigue, burnout, work environment, workload and perceived patient safety culture among critical care nurses. Br J Nurs, 29(1), 28-34. https://doi.org/10.12968/bjon.2020.29.1.28 2. Alqattan, H., Morrison, Z., & Cleland, J. A. (2019). A narrative synthesis of qualitative studies conducted to assess patient safety culture in hospital settings. Sultan Qaboos Univ Med J, 19(2), e91-e98. https://doi.org/10.18295/squmj.2019.19.02.002 2. Alqattan, H., Morrison, Z., & Cleland, J. A. (2019). A narrative synthesis of qualitative studies conducted to assess patient safety culture in hospital settings. Sultan Qaboos Univ Med J, 19(2), e91-e98. https://doi.org/10.18295/squmj.2019.19.02.002 2. Alqattan, H., Morrison, Z., & Cleland, J. A. (2019). A narrative synthesis of qualitative studies conducted to assess patient safety culture in hospital settings. Sultan Qaboos Univ Med J, 19(2), e91-e98. https://doi.org/10.18295/squmj.2019.19.02.002 3. Alsalem, G., Bowie, P & ,.Morrison, J. (2018). Assessing safety climate in acute hospital settings: a systematic review of the adequacy of the psychometric properties of survey measurement tools. BMC Health Serv Res, 18(1), 353. https://doi.org/10.1186/s12913-018-3167-x 3. Alsalem, G., Bowie, P & ,.Morrison, J. (2018). Assessing safety climate in acute hospital settings: a systematic review of the adequacy of the psychometric properties of survey measurement tools. BMC Health Serv Res, 18(1), 353. https://doi.org/10.1186/s12913-018-3167-x 3. Alsalem, G., Bowie, P & ,.Morrison, J. (2018). Assessing safety climate in acute hospital settings: a systematic review of the adequacy of the psychometric properties of survey measurement tools. BMC Health Serv Res, 18(1), 353. https://doi.org/10.1186/s12913-018-3167-x 4. Alshehry, A. S. (2022). Nurse–patient/relatives conflict and patient safety competence among nurses. INQUIRY: The Journal of Health Care Organization, Provision, and Financing, 59, 00469580221093186 . 4. Alshehry, A. S. (2022). Nurse–patient/relatives conflict and patient safety competence among nurses. INQUIRY Provision, and Financing, 59, 00469580221093186 . 4. Alshehry, A. S. (2022). Nurse–patient/relatives conflict and patient safety competence among nurses. INQUIRY: The Journal of Health Care Organization, Provision, and Financing, 59, 00469580221093186 . 4. Alshehry, A. S. (2022). Nurse–patient/relatives conflict and patient safety competence among nurses. INQUIRY: The Journal of Health Care Organization, Provision, and Financing, 59, 00469580221093186 . 5. Armando Nahum. (2021, November 8, 2021). Patient safety representatives unite to implement global patient safety action plan in the united states. Centre for Disease Control and Prevention. References O., Nilsson, J., & Lepp, M. (2021). Professional competence among registered nurses working in hospitals in saudi arabia and their experiences of quality of nursing care and patient safety. J Transcult Nurs, 32(4), 425-433. https://doi.org/10.1177/1043659621992845 11. Halabi, J. O., Nilsson, J., & Lepp, M. (2021). Professional competence among registered nurses working in hospitals in saudi arabia and their experience of quality of nursing care and patient safety. J Transcult Nurs, 32(4), 425-433. https://doi.org/10.1177/1043659621992845 12. Han, J. H., & Roh, Y. S. (2020). Teamwork, psychological safety, and patient safety competency among emergency nurses. Int Emerg Nurs, 51, 100892. https://doi.org/10.1016/j.ienj.2020.100892 12. Han, J. H., & Roh, Y. S. (2020). Teamwork, psychological safety, and patient safety competency among emergency nurses. Int Emerg Nurs, 51, 100892. https://doi.org/10.1016/j.ienj.2020.100892 3. Han, Y., Kim, J. S., & Seo, Y. (2020). Cross-sectional study on patient safety culture, patient safety competency, a 32-40. https://doi.org/10.1177/0193945919838990 13. Han, Y., Kim, J. S., & Seo, Y. (2020). Cross-sectional study on patient safety culture, patient safety competency, and adverse events. West J Nurs Res, 42(1), 32-40. https://doi.org/10.1177/0193945919838990 13. Han, Y., Kim, J. S., & Seo, Y. (2020). Cross-sectional study on patient safety culture, patient safety competency, and adverse events. West J Nurs Res, 42(1), 32-40. https://doi.org/10.1177/0193945919838990 14. Huh, A., & Shin, J. H. (2021). Person-Centered Care Practice, Patient Safety Competence, and Patient Safety Nursing Activities of Nurses Working in Geriatric Hospitals Int J Environ Res Public Health 18(10) https://doi org/10 3390/ijerph18105169 n-Centered Care Practice, Patient Safety Competence, and Patient Safety Nursing Activities of Nurses Working in 14. Huh, A., & Shin, J. H. (2021). Person-Centered Care Practice, Patient Safety Competence, and Patient Safety Geriatric Hospitals. Int J Environ Res Public Health, 18(10). https://doi.org/10.3390/ijerph18105169 14. Huh, A., & Shin, J. H. (2021). Person-Centered Care Practice, Patient Safety Competence, and Patient Safety Nursing Activities of Nurses Working in Geriatric Hospitals. Int J Environ Res Public Health, 18(10). https://doi.org/10.3390/ijerph18105169 15. Kakemam, E., Gharaee, H., Rajabi, M. R., Nadernejad, M., Khakdel, Z., Raeissi, P., & Kalhor, R. (2021). Nurses’ perception of patient safety culture and its relationship with adverse events: a national questionnaire survey in Iran. BMC nursing, 20(1), 60. https://doi.org/10.1186/s12912-021-00571-w 15. Kakemam, E., Gharaee, H., Rajabi, M. R., Nadernejad, M., Khakdel, Z., Raeissi, P., & Kalhor, R. (2021). Nurses’ perception of patient safety culture and its relationship with adverse events: a national questionnaire survey in Iran. BMC nursing, 20(1), 60. https://doi.org/10.1186/s12912-021-00571-w 6. References Retrieved March 28, from https://www.who.int/news-room/fact-sheets/detail/patient-safety 29. World Health Organizaion. (2019a). Patient Safety. WHO. Retrieved March 28, from https://www.who.int/news-room/fact-sheets/detail/patient-safety 9. World Health Organizaion. (2019a). Patient Safety. WHO. Retrieved March 28, from https://www.who.int/news- 29. World Health Organizaion. (2019a). Patient Safety. WHO. Retrieved March 28, from https://www.who.int/news-room/fact-sheets/detail/patient-safety 30. World Health Organizaion. (2019b). Patient safety Retrieved March 31, from https://www.who.int/news-room/facts-in-pictures/detail/patient-safety 30. World Health Organizaion. (2019b). Patient safety Retrieved March 31, from https://www.who.int/news-room/facts-in-pictures/detail/patient-safety 31. Yan, L., Yao, L., Li, Y., & Chen, H. (2021). Assessment and analysis of patient safety competency of Chinese nurses with associate degrees: A cross- sectional study. Nurs Open, 8(1), 395-403. https://doi.org/10.1002/nop2.640 30. World Health Organizaion. (2019b). Patient safety Retrieved March 31, from https://www.who.int/news-roo 31. Yan, L., Yao, L., Li, Y., & Chen, H. (2021). Assessment and analysis of patient safety competency of Chinese nurses with associate degrees: A cross- sectional study. Nurs Open, 8(1), 395-403. https://doi.org/10.1002/nop2.640 31. Yan, L., Yao, L., Li, Y., & Chen, H. (2021). Assessment and analysis of patient safety competency of Chinese nurses with associate degrees: A cross- sectional study. Nurs Open, 8(1), 395-403. https://doi.org/10.1002/nop2.640 32. Zabin, L. M., Zaitoun, R. S. A., & Abdullah, A. A. (2022). Patient safety culture in Palestine: university hospital nurses' perspectives. BMC Nurs, 21(1), 204. https://doi.org/10.1186/s12912-022-00987-y 32. Zabin, L. M., Zaitoun, R. S. A., & Abdullah, A. A. (2022). Patient safety culture in Palestine: university hospital nurses' perspectives. BMC Nurs, 21(1), 204 https://doi.org/10.1186/s12912-022-00987-y References Kalteh, H .O., Mortazavi, S. B., Mohammadi, E., & Salesi, M. (2021). The relationship between safety culture and systematic review. Int J Occup Saf Ergon, 27(1), 206-216. https://doi.org/10.1080/10803548.2018.1556976 16. Kalteh, H .O., Mortazavi, S. B., Mohammadi, E., & Salesi, M. (2021). The relationship between safety culture and safety climate and safety performance: a systematic review. Int J Occup Saf Ergon, 27(1), 206-216. https://doi.org/10.1080/10803548.2018.1556976 16. Kalteh, H .O., Mortazavi, S. B., Mohammadi, E., & Salesi, M. (2021). The relationship between safety culture and safety climate and safety performance: a systematic review. Int J Occup Saf Ergon, 27(1), 206-216. https://doi.org/10.1080/10803548.2018.1556976 Page 5/11 Page 5/11 17. Khamaiseh, A., Al-Twalbeh, D., & Al-Ajlouni, K. (2020). Patient safety culture in Jordanian primary health-care centres as perceived by nurses: a cross- sectional study. East Mediterr Health J, 26(10), 1242-1250. https://doi.org/10.26719/emhj.20.044 18. Khoshakhlagh, A. H., Khatooni, E., Akbarzadeh, I., Yazdanirad, S., & Sheidaei, A. (2019). Analysis of affecting factors on patient safety culture in public and private hospitals in Iran. BMC Health Serv Res, 19(1), 1009. https://doi.org/10.1186/s12913-019-4863-x 19. Kruk, M. E., Gage, A. D., Joseph, N. T., Danaei, G., García-Saisó, S., & Salomon, J. A. (2018). Mortality due to low-quality health systems in the universal health coverage era: a systematic analysis of amenable deaths in 137 countries. The Lancet, 392(10160), 2203-2212 . 20. Letourneau, R .M., & McCurry, M. K. (2019). The effect of transition to practice programs on the self-assessment of newly licensed registered nurses' confidence in quality and safety competency attainment. Nurs Educ Perspect, 40(3), 151-156. https://doi.org/10.1097/01.Nep.0000000000000438 21. Lousada, L. M., da Silva Dutra, F. C., da Silva, B. V., de Oliveira, N. L. L., Bastos, I. B., de Vasconcelos, P. F., & de Carvalho, R. (2020). Patient safety culture in primary and home care services. BMC Fam Pract, 21(1), 188. https://doi.org/10.1186/s12875-020-01263-1 22. Mahsoon, A. N., & Dolansky, M. (2021). Safety culture and systems thinking for predicting safety competence and safety performance among registered nurses in Saudi Arabia: a cross-sectional study. J Res Nurs, 26(1-2 .19-32 ,)https://doi.org/10.1177/1744987120976171 23. Najjar, S., Baillien, E., Vanhaecht, K., Hamdan, M., Euwema, M., Vleugels, A., Sermeus, W., Schrooten, W., Hellings, J., & Vlayen, A. (2018). Similarities and differences in the associations between patient safety culture dimensions and self-reported outcomes in two different cultural settings: a national cross- sectional study in Palestinian and Belgian hospitals. BMJ Open, 8(7), e021504. https://doi.org/10.1136/bmjopen-2018-021504 24. Page, M. J., McKenzie ,J. E., Bossuyt, P. References M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., & Brennan, S. E. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. International Journal of Surgery, 88 .105906 , 24. Page, M. J., McKenzie ,J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., & Brennan, S. E. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. International Journal of Surgery, 88 .105906 , 25. Peñataro-Pintado, E., Rodríguez-Higueras, E., Llauradó-Serra, M., Gómez-Delgado, N., Llorens-Ortega, R., & Díaz-Agea, J. L. (2022). Development and validation of a questionnaire of the perioperative nursing competencies in patient safety. Int J Environ Res Public Health, 19(5). https://doi.org/10.3390/ijerph19052584 25. Peñataro-Pintado, E., Rodríguez-Higueras, E., Llauradó-Serra, M., Gómez-Delgado, N., Llorens-Ortega, R., & Díaz-Agea, J. L. (2022). Development and validation of a questionnaire of the perioperative nursing competencies in patient safety. Int J Environ Res Public Health, 19(5). https://doi.org/10.3390/ijerph19052584 26. Son, Y. J., Lee, E. K., & Ko, Y. (2019). association of working hours and patient safety competencies with adverse nurse outcomes: a cross-sectional study. Int J Environ Res Public Health, 16(21). https://doi.org/10.3390/ijerph16214083 26. Son, Y. J., Lee, E. K., & Ko, Y. (2019). association of working hours and patient safety competencies with adverse nurse outcomes: a cross-sectional study. Int J Environ Res Public Health, 16(21). https://doi.org/10.3390/ijerph16214083 27. Wami, S. D., Demssie, A. F., Wassie, M. M., & Ahmed, A. N. (2016). Patient safety culture and associated factors: A quantitative and qualitative study of healthcare workers' view in Jimma zone Hospitals, Southwest Ethiopia. BMC Health Serv Res, 16, 495. https://doi.org/10.1186/s12913-016-1757-z 27. Wami, S. D., Demssie, A. F., Wassie, M. M., & Ahmed, A. N. (2016). Patient safety culture and associated factors: A quantitative and qualitative study of healthcare workers' view in Jimma zone Hospitals, Southwest Ethiopia. BMC Health Serv Res, 16, 495. https://doi.org/10.1186/s12913-016-1757-z 28. Wang, K. Y., Chou, C. C., & Lai, J. C. (2019). A structural model of total quality management, work values, job satisfaction and patient-safety-culture attitude among nurses. J Nurs Manag, 27(2), 225-232. https://doi.org/10.1111/jonm.12669 28. Wang, K. Y., Chou, C. C., & Lai, J. C. (2019). A structural model of total quality management, work values, job satisfaction and patient-safety-culture attitude among nurses. J Nurs Manag, 27(2), 225-232. https://doi.org/10.1111/jonm.12669 29. World Health Organizaion. (2019a). Patient Safety. WHO. Table Table 1. Summary of articles. Page 6/11 Page 6/11 Author /Citation Research title Research question or aim Sample Sample number Setting Country Design Questionnaire collection (Habibi Soola et al., 2022) Predictors of patient safety competency among emergency nurses in Iran: a cross-sectional correlational study. Assess predictors of patient safety competency among emergency nurses. The convenience sampling method was used to recruit emergency department nurses 254 Five emergency departments of Ardabil Teaching Hospitals, northwest of Iran Iran Cross- sectional correlational research Patient Safet Nursing Educ Questionnaire (Huh & Shin, 2021) Person-Centered Care Practice, Patient Safety Competence, and Patient Safety Nursing Activities of Nurses Working in Geriatric Hospitals Investigate the relationships between person- centered care practice, patient safety competence, and patient safety during nursing activities in geriatric hospitals. Convenience sampling was used to recruit geriatric nurses 186 in 12 geriatric hospitals 12 geriatric hospitals Canada a descriptive survey design Patient Safet Competence Assessment T Nurses. (Khamaiseh et al., 2020) Patient safety culture in Jordanian primary health- care centers as perceived by nurses: a cross- sectional study Determine nurses’ attitudes towards patient safety culture in primary health centers in Jordan. All nurses who were willing to participate in the study at the time of data collection and had at least 1 year of experience in governmental health settings; having worked in the center concerned for a minimum of 6 months; having Jordanian nationality; 644 91 accredited primary health care centers in Jordan d in 12 governorates Jordanian A cross- sectional descriptive study SAQ (Alshehry, 2022) Nurse– Patient/Relatives Conflict and Patient Safety Competence Among Nurses To assess self-reported nurse and patients/relatives’ conflicts among nurses and investigate its association with the nurses’ confidence in patient safety competencies. Table Registered nurse in Saudi Arabia; employed in the university hospital for at least 6 months; and provided direct care to patients; 320 King Saud University Medical City (KSUMC) in Riyadh City, Tertiary Saudi Arabia descriptive and cross-sectional “Healt Profes Educa Patien Safety Survey PEPSS (Connell et al., 2021) Measuring the safety climate in an Australian emergency department Measure perceptions of the safety climate in an Australian The convenience sampling method was used to recruit the survey participants who 127 a mixed metropolitan emergency department (adult and Australia Not mentioned The Safety C Survey (SCS) Page 7/11 Author /Citation Research title Research question or aim Sample Sample number Setting Country Design Questionnaire collection (Kakemam et al., 2021) Nurses’ perception of patient safety culture and its relationship with adverse events: a national questionnaire survey in Iran Assess the association between nurses’ perception of patient safety culture and their perceived proportion of adverse events. Nurses employed in 32 teaching hospitals in Iran 2295 32 teaching hospitals in Iran Iran A nationwide cross- sectional study Persian versi the Hospital S of Patient Sa Culture (HSO (Halabi et al., 2021) Professional Competence Among Registered Nurses Working in Hospitals in Saudi Arabia and Their Experiences of Quality of Nursing Care and Patient Safety Assess nurses' self- reported professional competence and illuminate experiences of the quality of nursing care and patient safety. Nurses working in different unit 469 Two public hospitals and Regions Kingdom of Saudi Arabia cross- sectional design The Nurse Professional Competence (Han & Roh, 2020) Teamwork, psychological safety, and patient safety competency among emergency nurses This study aimed to identify the levels of teamwork, psychological safety, and patient safety competency, and determine the associations among these variables in emergency nurses. Nurses with more than 12 months of ED experience in. New graduate nurses with less than one year of experience were excluded from the study 200 15 tertiary or secondary hospitals located in Seoul and the metropolitan areas of South Korea Korea A descriptive, correlational study used cross- sectional survey methodology. Patient safety competency' evaluation to (Letourneau & McCurry, 2019) The Effect of Transition to Practice Programs on the Self-Assessment of Newly Licensed Registered Nurses’ Confidence in Quality and Safety Competency Attainment To evaluate the effectiveness of transition to practice programs (TPPs) in self- assessment of quality and safety competency development in newly licensed registered nurses (NLRN). Table nurses who graduated from an accredited prelicensure nursing program within the last 12 months; were licensed as RNs; and currently practiced in a hospital setting, with no more than six months of experience Sixty- four Participants were recruited from three acute care institutions. England A quasi- experimental. Pretest- posttest design The Nursing Q and Safety Se Inventory (NQ (Lousada et al., 2020) Patient safety culture in primary and home care services To evaluate safety culture in primary and home care settings and to verify relationships between the SAQ domains and variables related to gender, type of service (primary or home care) and time of Professionals from nine districts covered by one home care program and six primary healthcare centers 56 One home care service and six primary care centers located in the metropolitan region of Fortaleza, Brazil. Brazil A cross- sectional study The Safety At Questionnaire Page 8/11 Author /Citation Research title Research question or aim Sample Sample number Setting Country Design Questionnaire collection (Mahsoon & Dolansky, 2021) Safety culture and systems thinking for predicting safety competence and safety performance among registered nurses in Saudi Arabia: a cross-sectional study. Explore the relationships among systems thinking, educational level, safety culture, safety competence, and safety performance among registered nurses working in medical and surgical units in Saudi Arabia. Convenience sampling was used to recruit registered nurses from inpatient medical and surgical units in a hospital. 84 Surgical units in a hospital. Saudi Arabia correlational cross- sectional design Safety comp was measure the nurse’s at and skills saf scale the late version of NA Safety cultur measured by hospital surve patient safety culture (HSOP (Najjar et al., 2018) Similarities and differences in the associations between patient safety culture' dimensions and self-reported outcomes in two different cultural settings: a national cross- sectional study in Palestinian and Belgian hospitals. Investigate the relationships between patient safety culture (PSC) dimensions and PSC self- reported outcomes of the PSC in different cultures and to gain insight into cultural differences regarding PSC. Health care professionals. 2836 90 Belgian hospitals and 13 Palestinian hospitals. Belgium and Palestine. Observational, cross- sectional study The HSOPSC (Peñataro- Pintado et al., 2022) Development and Validation of a Questionnaire of the Perioperative Nursing Competencies in Patient Safety. Table to design, validate and analyze the psychometric properties of the Perioperative Nursing Safety Competencies Perioperative nurses from 55 hospitals 415 55 hospitals Spain An instrumental, quantitative and descriptive study divided into two phases Questionnaire Perioperative Nursing Safe Competencie (Son et al., 2019) Association of Working Hours and Patient Safety Competencies with Adverse Nurse Outcomes: A Cross-Sectional Study Identify the differences in PS competencies and adverse nurse outcomes in different working hour's groups; and to explore the association of working hours and PS competencies with adverse nurse outcomes as an indicator of organizational performance in South Korean nurses. Convenience sampling. Participants were registered nurses of all levels. 364 Acute care hospital setting. Nurses working in three tertiary care hospitals affiliated with Soonchunhyang University in South Korea South Korean A cross- sectional descriptive design ‘Health Profes Education in Survey” Page 9/11 Author /Citation Research title Research question or aim Sample Sample number Setting Country Design Questionnaire collection (Han et al., 2020) Cross Sectional Study on Patient Safety Culture, Patient Safety Competency, and Adverse Events The study has the following objectives: (a) identify nurses’ perceptions of patient safety culture, patient safety competency, and adverse events; (b) examine associations between nurses’ perceptions of patient safety culture and adverse events; and (c) clarify associations between nurses’ perceptions of patient safety competency and adverse events. Convenient sampling was used. Sample was nurses who had been employed for more than 1 year 212 Two university hospitals located in two provinces of South Korea. Korea A cross- sectional research design The Hospital on Patient Sa Culture (HSO Patient safety competency w measured us Health Profes Education in Safety Survey PEPSS) (Yan et al., 2021) Assessment and analysis of patient safety competency of Chinese nurses with associate degrees: A cross‐ sectional study Analyze patient safety competency (PSC) of Chinese nurses with associate degrees (ADNs) and explore factors. A convenience sample 451 18 hospitals China Cross sectional Patient Safet Competency Evaluation Sc (PSCNES). Note. ED: Emergency department. PS: Patient safety. Note. ED: Emergency department. PS: Patient safety. Figures Figures Page 10/11 Page 10/11 Figure 1 PRISMA flow diagram of the systematic review Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. AppendixAsearchstartegy.pdf PRISMA flow diagram of the systematic review AppendixAsearchstartegy.pdf Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. AppendixAsearchstartegy.pdf Page 11/11 Page 11/11
https://openalex.org/W2988184041	https://www.biorxiv.org/content/biorxiv/early/2019/02/16/551523.full.pdf	English	null	Honey bees are important pollinators of South African blueberries despite their inability to sonicate	South African journal of botany	2,021	cc-by	12,963	. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 1 Honey bees, Apis mellifera, are important pollinators of the highbush blueberry variety Ventura 2 despite the inability to sonicate 3 Honey bees are important pollinators of blueberries despite their inability to sonicate 4 Keanu Martin1, Bruce Anderson1, Corneile Minnaar1, Marinus de Jager2 5 1Department of Botany and Zoology, Stellenbosch University, Matieland 7602, Cape Town, Western 6 Cape, South Africa 7 2Department of Conservation Ecology and Entomology, Stellenbosch University, Matieland 7602, 8 Cape Town, Western Cape, South Africa 9 Corresponding author: keanum@sun.ac.za 3 Honey bees are important pollinators of blueberries despite their inability to sonicate 4 Keanu Martin1*, Bruce Anderson1, Corneile Minnaar1, Marinus de Jager2 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 10 Abstract ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 31 Introduction 32 Over 30% of the world’s agricultural crops depend on animal-mediated pollination, an essential 33 ecosystem service which is valued at approximately €153 billion [1]. The dependence of crops on 34 pollinators varies, with fruit formation in certain crops being extremely dependent on pollinators [2]. 35 For example, atemoya, Brazil nut, cantaloupe, cocoa, kiwi, macadamia nut, passion fruit, pawpaw, 36 rowanbarry, sapodilla, squashes and pumpkins, vanilla and watermelon, show a ca. 90% reduction in 37 produce when pollinators are absent [2]. 32 Over 30% of the world’s agricultural crops depend on animal-mediated pollination, an essential 33 ecosystem service which is valued at approximately €153 billion [1]. The dependence of crops on 34 pollinators varies, with fruit formation in certain crops being extremely dependent on pollinators [2]. 35 For example, atemoya, Brazil nut, cantaloupe, cocoa, kiwi, macadamia nut, passion fruit, pawpaw, 36 rowanbarry, sapodilla, squashes and pumpkins, vanilla and watermelon, show a ca. 90% reduction in 37 produce when pollinators are absent [2]. 38 Blueberry production is also highly dependent on pollinators for the production of high- 39 quality fruit [3–5]. This is a partly a result of floral architecture, where the pollen of blueberry 40 flowers is concealed within poricidal anthers, making pollen transfer both within and between 41 flowers unlikely without pollinators. For example, Campbell et al. [6] found that blueberry fruit were 42 ca. 47% heavier when plants had access to pollinators (including honey bees, Bombus spp., H. 43 laboriosa and Xylocopa spp.), than when pollinators were excluded. Effective pollinators of 44 blueberries are typically large bees such as bumble bees [7–10], blueberry bees [6,11] and mining 45 bees [7,8], which are able to extract pollen from anthers by vibrating their bodies at high frequency. 46 This causes pollen to dehisce from the pores inside the blueberry anthers [11]. 38 Blueberry production is also highly dependent on pollinators for the production of high- 39 quality fruit [3–5]. This is a partly a result of floral architecture, where the pollen of blueberry 40 flowers is concealed within poricidal anthers, making pollen transfer both within and between 41 flowers unlikely without pollinators. For example, Campbell et al. [6] found that blueberry fruit were 42 ca. 47% heavier when plants had access to pollinators (including honey bees, Bombus spp., H. 43 laboriosa and Xylocopa spp.), than when pollinators were excluded. 10 Abstract 11 Animal-mediated pollination is an essential ecosystem service which over a third of the world’s 12 agricultural crops depend on. Blueberry fruit production is highly dependent on pollinators and in 13 their native range they are pollinated mostly by bumble bees (Bombus spp.). Demand for blueberries 14 has increased in recent years due to their perceived health benefits. Consequently, blueberry 15 cultivation has expanded well beyond their native range, including several regions where bumble 16 bees are not present. In many areas, honey bees may be the only commercially available pollinators 17 of blueberries because many countries ban the importation of bumble bees. This study aimed to 18 determine the benefits of honey bee pollination on blueberry fruit quality and quantity for the 19 variety Ventura by comparing yields of honey-bee-pollinated flowers to flowers where pollinators 20 had been excluded. Honey bees significantly increased berry mass and diameter. Our results suggest 21 that the presence of honey bee pollinators potentially increases revenue by approximately $864 22 501/ha in areas without bumble bees. We conclude that Ventura is reliably pollinated by honey 23 bees, and that honey bee pollination may be a useful substitute for bumble bees in areas where 24 bumble bees are absent. We also determined the extent to which blueberry yields could still be 25 improved by comparing fruit quality and quantity under honey bee pollination to fruit quality and 26 quantity achieved through ideal hand pollination. We found that blueberry yields may be still be 27 significantly increased relative to ideal hand pollination and we discuss potential ways to improve 28 the efficiency of honeybee pollination in the future. Additional research is required to study how 29 beneficial honey bees are to fruit yield on varieties as the benefits of honey bees are likely to vary 30 across different varieties. 1 1 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. 31 Introduction Effective pollinators of 44 blueberries are typically large bees such as bumble bees [7–10], blueberry bees [6,11] and mining 45 bees [7,8], which are able to extract pollen from anthers by vibrating their bodies at high frequency. 46 This causes pollen to dehisce from the pores inside the blueberry anthers [11]. 47 This “buzz pollination” strategy is employed by approximately 15,000–20,000 plant species 48 [12,13]. Buzz pollination may have evolved to exclude less efficient or wasteful pollinators, such as 49 honey bees (Apis mellifera), which are unable to obtain pollen from these specialized anthers 50 through buzzing [7,10,12]. For example, Javorek et al. [7] found honey bees deposited approximately 51 three times less pollen on blueberry stigmas during a single visit than bumble bees. However, the 52 large numbers of foragers in honey bee colonies may enable effective pollination of blueberry 53 flowers, as they may achieve increased flower visitation rates relative to bumble bees. This may 54 explain why many commercial blueberry farmers completely, or partially, depend on honey bees as 47 This “buzz pollination” strategy is employed by approximately 15,000–20,000 plant species 48 [12,13]. Buzz pollination may have evolved to exclude less efficient or wasteful pollinators, such as 49 honey bees (Apis mellifera), which are unable to obtain pollen from these specialized anthers 50 through buzzing [7,10,12]. For example, Javorek et al. [7] found honey bees deposited approximately 51 three times less pollen on blueberry stigmas during a single visit than bumble bees. However, the 52 large numbers of foragers in honey bee colonies may enable effective pollination of blueberry 53 flowers, as they may achieve increased flower visitation rates relative to bumble bees. This may 54 explain why many commercial blueberry farmers completely, or partially, depend on honey bees as 2 2 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 55 pollinators. 31 Introduction Even within the native range of bumble bees, North American blueberry farmers 56 frequently add honey bee hives to supplement bumble bee pollination [4,14–16]. 57 The pervasive use of honey bees in the blueberry industry may also be a result of the 58 prohibition on importation and use of non-native, commercially-produced bumble bees in parts of 59 USA and many other regions where bumble bees are not native, such as southern Africa and 60 Australasia [17,18]. These strict laws, prohibiting the movement of bumble bees, are in place 61 because their introduction can have catastrophic effects on native fauna and flora [19,20]. For 62 example, Bombus ruderatus and Bombus terrestris were introduced to Chile for agricultural 63 pollination; these species have subsequently invaded southern South America, including Argentina, 64 which has since banned commercial importation of bumble bees [21–24]. In Argentina, the highly 65 invasive Bombus terrestris has caused a reduction in geographic range of the largest bumble bee in 66 the world, and the sole native Patagonian bumble bee, Bombus dahlbomii [22]. Other potential 67 impacts include pathogen transmission to native bumble bees, nectar robbing and flower damage 68 [21,22,24]. Despite the potentially harmful effects of introducing bumble bees for agricultural 69 pollination, the widely-held contention that honey bees are inferior pollinators of blueberries, drives 70 the industry to place pressure on governments to allow bumble bee importation. 71 To alleviate the temptation to introduce bumble bees into new areas it is pertinent to 55 pollinators. Even within the native range of bumble bees, North American blueberry farmers 56 frequently add honey bee hives to supplement bumble bee pollination [4,14–16]. 71 To alleviate the temptation to introduce bumble bees into new areas, it is pertinent to 72 quantify the actual benefit of honey bees as commercial pollinators of blueberries, so that policies 73 regarding the importation of bumble bees are based on substantive evidence and not impressions. 74 Further, it may be possible to optimize the efficiency of honey bee pollination, so that the 75 advantages of introducing bumble bees to new ranges are reduced. It is therefore important to 76 quantify how well different blueberry varieties perform under honey bee pollination, while also 77 estimating the potential for improvement by comparing blueberry yields under honeybee pollination 78 to yields achieved under optimal hand pollination. 3 3 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 79 We aim to study the effects of honey bees on the production of blueberries in the variety 80 Ventura, which is extensively planted in South America and South Africa [25]. More specifically, we 81 compare the benefits (in terms of fruit quality, yield and revenue) of having honey bees as the only 82 pollinators with blueberry yields achieved in the absence of pollinators. Despite their inability to 83 buzz-pollinate, honey bees still transfer pollen between flowers and are capable of increasing fruit 84 production in a variety of blueberry crops [4,5,26,27]. Consequently, we hypothesize that managed 85 honey bees significantly increase fruit quality (i.e., mass and diameter) and decrease fruit 86 development time, compared to flowers without access to pollinators. Next, we determine whether 87 there is room to improve blueberry yields beyond that which is achieved when honey bee pollinators 88 are allowed access to flowers. 55 pollinators. Even within the native range of bumble bees, North American blueberry farmers 56 frequently add honey bee hives to supplement bumble bee pollination [4,14–16]. Although honey bees can transfer blueberry pollen between flowers 89 [27], we expect that the inefficiency of honey bee pollination on blueberry flowers should result in 90 significant potential for improvement, and consequently fruit quality and yield resulting from honey 91 bee pollination should be lower than by hand pollination, which maximizes the transfer of pollen. 79 We aim to study the effects of honey bees on the production of blueberries in the variety 80 Ventura, which is extensively planted in South America and South Africa [25]. More specifically, we 81 compare the benefits (in terms of fruit quality, yield and revenue) of having honey bees as the only 82 pollinators with blueberry yields achieved in the absence of pollinators. Despite their inability to 83 buzz-pollinate, honey bees still transfer pollen between flowers and are capable of increasing fruit 84 production in a variety of blueberry crops [4,5,26,27]. Consequently, we hypothesize that managed 85 honey bees significantly increase fruit quality (i.e., mass and diameter) and decrease fruit 86 development time, compared to flowers without access to pollinators. Next, we determine whether 87 there is room to improve blueberry yields beyond that which is achieved when honey bee pollinators 88 are allowed access to flowers. Although honey bees can transfer blueberry pollen between flowers 89 [27], we expect that the inefficiency of honey bee pollination on blueberry flowers should result in 90 significant potential for improvement, and consequently fruit quality and yield resulting from honey 91 bee pollination should be lower than by hand pollination, which maximizes the transfer of pollen. 79 4 4 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 92 Materials and Methods 105 106 Benefit of honey bees 107 To determine whether honey bees are commercially beneficial pollinators for berry production in 108 Ventura, we compared blueberry production in plants where all pollinators were excluded 109 (pollinator exclusion) with blueberry production in plants where honey bees had access to flowers 110 (honey bee pollination) (See Fig. 1). 111 Pollinator exclusion (bagged) 112 Pollinators were excluded from visiting some blueberry flowers by placing a fine mesh bag over 20 113 individual virgin flowers (one flower per plant). Consequently, seed production in this treatment was 114 the result of autonomous pollination and/or parthenocarpy (the production of fruit in the absence of 115 fertilization), and not pollinator visitation. This provides an estimate of the yields expected in the 107 To determine whether honey bees are commercially beneficial pollinators for berry production in 108 Ventura, we compared blueberry production in plants where all pollinators were excluded 109 (pollinator exclusion) with blueberry production in plants where honey bees had access to flowers 110 (honey bee pollination) (See Fig. 1). 92 Materials and Methods 93 This study was conducted in a one-hectare block of Ventura plants (7500) stocked with 15 honey bee 94 hives on Backsberg blueberry farm (Western Cape, South Africa, 33°48'30.7"S 18°54'09.8"E). Our 95 experiment consisted of three treatments: pollinator exclusion, open honey bee pollination, and 96 optimized pollination (by hand). By comparing fruit quality among these three treatments, we 97 determined whether the addition of honey bees was beneficial to blueberry production as well as 98 the extent to which pollination by honey bees could potentially be improved upon (see Fig. 1 and 99 treatment descriptions below). The three treatments were replicated across 20 plants, with each 100 treatment applied once to each individual plant. 93 This study was conducted in a one-hectare block of Ventura plants (7500) stocked with 15 honey bee 94 hives on Backsberg blueberry farm (Western Cape, South Africa, 33°48'30.7"S 18°54'09.8"E). Our 95 experiment consisted of three treatments: pollinator exclusion, open honey bee pollination, and 96 optimized pollination (by hand). By comparing fruit quality among these three treatments, we 97 determined whether the addition of honey bees was beneficial to blueberry production as well as 98 the extent to which pollination by honey bees could potentially be improved upon (see Fig. 1 and 99 treatment descriptions below). The three treatments were replicated across 20 plants, with each 100 treatment applied once to each individual plant. 93 This study was conducted in a one-hectare block of Ventura plants (7500) stocked with 15 honey bee 94 hives on Backsberg blueberry farm (Western Cape, South Africa, 33°48'30.7"S 18°54'09.8"E). Our 95 experiment consisted of three treatments: pollinator exclusion, open honey bee pollination, and 96 optimized pollination (by hand). By comparing fruit quality among these three treatments, we 97 determined whether the addition of honey bees was beneficial to blueberry production as well as 98 the extent to which pollination by honey bees could potentially be improved upon (see Fig. 1 and 99 treatment descriptions below). The three treatments were replicated across 20 plants, with each 100 treatment applied once to each individual plant. 101 102 Fig 1: Hypothetical figure showing how the three treatments (pollinator exclusion, honey bee pollination and 103 optimized pollination) may be useful metrics in determining the benefits of using honey bees to pollinate 104 crops, and how yields could potentially be increased under more optimal pollination environments. 111 Pollinator exclusion (bagged) This treatment provides an estimate of the yield resulting from honey 125 bee pollination and is expected to be similar to yields currently obtained by farmers of the Ventura 126 variety across South Africa (realized yield). 127 128 Pollination deficit 129 To determine the extent to which honey bee pollination can potentially be improved, we compared 130 the honey bee pollination treatment (above) to a hand-pollinated treatment (optimized pollination). 131 We assumed that hand pollination would result in the best fruit production possible by maximizing 132 the deposition of pollen. 133 Optimized pollination (hand pollination) 134 Prior to hand-pollination, Ventura flowers were emasculated. We did this by removing part of the 135 corolla with a scalpel, before removing all stamens with a pair of fine forceps, thus ensuring that no 136 self-pollination could take place. To ensure that the experimental flowers were virgin, buds that 137 were about to open were bagged three days before hand pollination. The day before pollination, 138 pollen was collected from pollen donor flowers (approximately five flowers per pollen application) 139 [3]. Pollen was extracted from donor flowers by removing the corolla with a scalpel and agitating the 111 Pollinator exclusion (bagged) 112 Pollinators were excluded from visiting some blueberry flowers by placing a fine mesh bag over 20 113 individual virgin flowers (one flower per plant). Consequently, seed production in this treatment was 114 the result of autonomous pollination and/or parthenocarpy (the production of fruit in the absence of 115 fertilization), and not pollinator visitation. This provides an estimate of the yields expected in the 112 Pollinators were excluded from visiting some blueberry flowers by placing a fine mesh bag over 20 113 individual virgin flowers (one flower per plant). Consequently, seed production in this treatment was 114 the result of autonomous pollination and/or parthenocarpy (the production of fruit in the absence of 115 fertilization), and not pollinator visitation. This provides an estimate of the yields expected in the 5 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 116 absence of pollinators, and we expect that this should result in the poorest yield of our treatments, 117 which we have termed “minimum yield”. Once flowers wilted, the bags were removed, and fruit 118 maturation was allowed to continue normally. 116 absence of pollinators, and we expect that this should result in the poorest yield of our treatments, 117 which we have termed “minimum yield”. Once flowers wilted, the bags were removed, and fruit 118 maturation was allowed to continue normally. 119 Honey bee pollination (control) 120 Some flowers were left unbagged, allowing access to commercial honey bees placed on the farm. In 121 South Africa, honey bees are generally the only pollinators observed on blueberries. The hive 122 densities used for this treatment were 15 hives/ha, which corresponds to the densities of hives 123 actually used by commercial blueberry farms [15,27,28]. Each flower was labelled to distinguish its 124 fruit from other treatments. 148 Measurements of fruit quality 149 After pollination treatments were applied, we checked fruit development once a week to determine 150 whether fruit were mature and ready for fruit quality measurement. Fruits were considered mature 151 when the entire fruit turned a uniform dark blue. Mature fruit were harvested by hand and 152 subsequently weighed, and the diameter of each fruit was measured on the day it was harvested. By 153 checking fruit every week, we could also determine the developmental period for each fruit (the 154 number of weeks from pollination to harvesting) for each treatment. Apart from fruit mass and 155 diameter, the developmental period is an important determinant of fruit quality, as early fruit are 156 more valuable than late fruit; earlier fruit can be sold at higher prices when market demand is not 157 saturated [3,29]. The percentage fruit set per treatment was also calculated. 149 After pollination treatments were applied, we checked fruit development once a week to determine 150 whether fruit were mature and ready for fruit quality measurement. Fruits were considered mature 151 when the entire fruit turned a uniform dark blue. Mature fruit were harvested by hand and 152 subsequently weighed, and the diameter of each fruit was measured on the day it was harvested. By 153 checking fruit every week, we could also determine the developmental period for each fruit (the 154 number of weeks from pollination to harvesting) for each treatment. Apart from fruit mass and 155 diameter, the developmental period is an important determinant of fruit quality, as early fruit are 156 more valuable than late fruit; earlier fruit can be sold at higher prices when market demand is not 157 saturated [3,29]. The percentage fruit set per treatment was also calculated. 119 Honey bee pollination (control) This pollen mix was applied to recipient stigmas by dipping the stigma 143 into the Petri dish containing pollen, and visually confirming that the stigma was saturated with 144 pollen. Such careful hand pollination is likely to result in the maximum yield possible. Following hand 145 pollination, a fine mesh bag was placed over hand-pollinated flowers to prevent honey bees from 146 depositing additional pollen of unknown origin onto the stigma. This bag was again removed after 147 the flower wilted. 140 anthers with forceps, causing the poricidal anthers to release pollen into a Petri dish. Pollen from 141 donor flowers on different individual plants was mixed together so that recipient flowers received 142 pollen from multiple donors. This pollen mix was applied to recipient stigmas by dipping the stigma 143 into the Petri dish containing pollen, and visually confirming that the stigma was saturated with 144 pollen. Such careful hand pollination is likely to result in the maximum yield possible. Following hand 145 pollination, a fine mesh bag was placed over hand-pollinated flowers to prevent honey bees from 146 depositing additional pollen of unknown origin onto the stigma. This bag was again removed after 147 the flower wilted. 119 Honey bee pollination (control) 120 Some flowers were left unbagged, allowing access to commercial honey bees placed on the farm. In 121 South Africa, honey bees are generally the only pollinators observed on blueberries. The hive 122 densities used for this treatment were 15 hives/ha, which corresponds to the densities of hives 123 actually used by commercial blueberry farms [15,27,28]. Each flower was labelled to distinguish its 124 fruit from other treatments. This treatment provides an estimate of the yield resulting from honey 125 bee pollination and is expected to be similar to yields currently obtained by farmers of the Ventura 126 variety across South Africa (realized yield). 120 Some flowers were left unbagged, allowing access to commercial honey bees placed on the farm. In 121 South Africa, honey bees are generally the only pollinators observed on blueberries. The hive 122 densities used for this treatment were 15 hives/ha, which corresponds to the densities of hives 123 actually used by commercial blueberry farms [15,27,28]. Each flower was labelled to distinguish its 124 fruit from other treatments. This treatment provides an estimate of the yield resulting from honey 125 bee pollination and is expected to be similar to yields currently obtained by farmers of the Ventura 126 variety across South Africa (realized yield). 129 To determine the extent to which honey bee pollination can potentially be improved, we compared 130 the honey bee pollination treatment (above) to a hand-pollinated treatment (optimized pollination). 131 We assumed that hand pollination would result in the best fruit production possible by maximizing 132 the deposition of pollen. 6 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 140 anthers with forceps, causing the poricidal anthers to release pollen into a Petri dish. Pollen from 141 donor flowers on different individual plants was mixed together so that recipient flowers received 142 pollen from multiple donors. 158 Estimating the economic impact of honey bees 159 In addition to fruit quality, we also determined how differences in pollination treatments could 160 translate into differences in revenue gained. Firstly, to estimate the number of flowers, and thus 161 potential fruits per plant, we used a constant flower number of 11 016 per plant [15]. This number 162 serves only as an estimate of the total flowers for the highbush variety, Ventura, since we were 163 unable to perform flower counts on our experimental plants. To calculate fruit yield per plant, we 164 took the product of 11 016, the proportion of fruit set, and the predicted mass of fruits (taken from 159 In addition to fruit quality, we also determined how differences in pollination treatments could 160 translate into differences in revenue gained. Firstly, to estimate the number of flowers, and thus 161 potential fruits per plant, we used a constant flower number of 11 016 per plant [15]. This number 162 serves only as an estimate of the total flowers for the highbush variety, Ventura, since we were 163 unable to perform flower counts on our experimental plants. To calculate fruit yield per plant, we 164 took the product of 11 016, the proportion of fruit set, and the predicted mass of fruits (taken from 7 7 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 165 our linear mixed-effects model, see below) produced by individual plants for each treatment. Fruit 166 yield allows us to determine whether differences between treatments found in fruit quality actually 167 translates to yield, as it takes into account flowers that did not set fruit set as well. To calculate the 168 per-hectare economic value of fruit for each treatment, we multiplied yield per plant by the number 169 of plants per hectare and by the US dollars obtained per kilogram of fruit ($7.48) [30] (Eq. 1). 158 Estimating the economic impact of honey bees 170 Revenue = X X (Equation 1) 𝑌𝑖𝑒𝑙𝑑 𝑝𝑒𝑟 𝑝𝑙𝑎𝑛𝑡 𝑃𝑙𝑎𝑛𝑡𝑠 ℎ𝑎 𝑈𝑆$ 𝑘𝑔 171 172 Statistical analyses 173 To test for differences in fruit mass, diameter, and developmental period between treatments, we 174 used a linear mixed-effects model with treatment nested in plant ID as the repeated random factor. 175 Linear contrasts (Tukey) were used to test for treatment differences. To test how well our model 176 explained the variance in our data, we used Nakagawa R2 values [31], which provides both 177 conditional variance (R2c) and marginal variance (R2m) estimates than can be equated to traditional 178 R2 values. Conditional R2 values show the variance explained by the entire model (fixed effects and 179 random effects), whereas marginal R2 values show the variance explained by the random effects 180 alone. The fixed effect was treatment and the random effect was treatment nested in plant ID. To 181 test the overall effect of treatment on fruit set, we performed a log-likelihood ratio test between 182 two mixed-effects logistic regression models, one with and one without treatment as a fixed effect. 183 To test for differences in fruit yield (see calculation above) between treatments, we used a linear 184 model. All statistical analyses were conducted in R (version 3.3.2) [32] using the packages nlme [33], 185 multcomp [34], ggplot2 [35], sjPlot [36], car [37], lme4 [38], grid [32], gridExtra [39], lattice [40], 186 MuMIn [41], plyr [42] and plotrix [43]. 165 our linear mixed-effects model, see below) produced by individual plants for each treatment. Fruit 166 yield allows us to determine whether differences between treatments found in fruit quality actually 167 translates to yield, as it takes into account flowers that did not set fruit set as well. To calculate the 168 per-hectare economic value of fruit for each treatment, we multiplied yield per plant by the number 169 of plants per hectare and by the US dollars obtained per kilogram of fruit ($7.48) [30] (Eq. 1). 173 To test for differences in fruit mass, diameter, and developmental period between treatments, we 174 used a linear mixed-effects model with treatment nested in plant ID as the repeated random factor. 175 Linear contrasts (Tukey) were used to test for treatment differences. 158 Estimating the economic impact of honey bees To test how well our model 176 explained the variance in our data, we used Nakagawa R2 values [31], which provides both 177 conditional variance (R2c) and marginal variance (R2m) estimates than can be equated to traditional 178 R2 values. Conditional R2 values show the variance explained by the entire model (fixed effects and 179 random effects), whereas marginal R2 values show the variance explained by the random effects 180 alone. The fixed effect was treatment and the random effect was treatment nested in plant ID. To 181 test the overall effect of treatment on fruit set, we performed a log-likelihood ratio test between 182 two mixed-effects logistic regression models, one with and one without treatment as a fixed effect. 183 To test for differences in fruit yield (see calculation above) between treatments, we used a linear 184 model. All statistical analyses were conducted in R (version 3.3.2) [32] using the packages nlme [33], 185 multcomp [34], ggplot2 [35], sjPlot [36], car [37], lme4 [38], grid [32], gridExtra [39], lattice [40], 186 MuMIn [41], plyr [42] and plotrix [43]. 8 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 188 Benefit of bees 189 The linear mixed-effects model explained the majority of the conditional variance for developmental 190 period (R2c=0.999). The model also accounted for some of the marginal variance in developmental 191 period (R2m=0.283), demonstrating that individual plants had different responses depending on 192 treatment. The presence of honey bee pollinators did not significantly decrease the ripening period 193 of blueberry fruits in comparison to treatments where honey bees were excluded (Table 1, Fig 2A). 194 195 Fig 2: A) Mean number of weeks needed for the fruit to ripen for each treatment. B) The mean mass of fruits for 196 each treatment. C) The mean diameter of fruits for every treatment. D) Percentage fruit set for each treatment. 197 E) The mean yield per plant for each treatment. Letters indicate significance (p<0.05) of linear contrasts (Tukey 198 HSD). Error bars indicate standard error. BB = benefit of bees, this is the percentage difference between flowers 199 with no access to pollinators compared to flowers which had access to honey bees. PD = pollination deficit, this 200 is the percentage difference between flowers with access to honey bees compared to flowers were hand 201 pollinated. 202 203 Table 1: Outcome of three pollination treatments on blueberry fruit ripening, quality and yield. Benefit of bees (BB) Pollination deficit (PD) Pollinator exclusion – optimized pollination Ripening period z=-1.505, p=0.2885 z=-2.734, p=0.0173 z=-4.165, p<0.001 Mass z=4.346, p<0.001 z=2.449, p=0.0381 z=6.771, p<0.001 Diameter z=4.617, p<0.001 z=2.338, p=0.0508 z=6.938, p<0.001 Yield t=4.870, p<0.001 t=3.164, p=0.0078 t=7.992, p<0.001 204 Significance indicated by bold type 205 The linear mixed-effects model explained the majority of conditional variance for both mass 206 (R2c=0.999) and diameter (R2c=0.999). The model explained a large proportion of the marginal 195 Fig 2: A) Mean number of weeks needed for the fruit to ripen for each treatment. B) The mean mass of fruits for 196 each treatment. C) The mean diameter of fruits for every treatment. D) Percentage fruit set for each treatment. 197 E) The mean yield per plant for each treatment. Letters indicate significance (p<0.05) of linear contrasts (Tukey 198 HSD). Error bars indicate standard error. BB = benefit of bees, this is the percentage difference between flowers 199 with no access to pollinators compared to flowers which had access to honey bees. 188 Benefit of bees There was no difference in the fruit set of 214 flowers with access to honey bees relative to flowers which had no access to honey bees or hand 215 pollinated flowers (Chi-square=1.36, df=4, p=0.507) (Fig 2D). 216 The beneficial effects of honey bees on fruit mass and fruit set also translated into 217 differences in total yield as calculated using Isaac’s [15] average flowers produced per plant, with the 218 linear model explaining 54% of the variance in yield (R2=0.5355, F (2,44) =27.52, p<0.001). Here, yield 219 increased substantially (152%) from 10.11kg ± 2.2kg (mean ± SE predicted plant yield) per plant 220 when pollinators were excluded to 25.52kg ± 3kg (mean ± SE predicted plant yield) when honey bees 221 were allowed to forage on blueberries (Table 1, Fig 2E). Using a value of $7.48/kg [30] of blueberries 222 and a density of 7500 plants per hectare, adding honey bee hives in areas lacking natural blueberry 223 pollinators can potentially increase blueberry revenue by $864 501/ha (152%) compared to if 224 pollinators were excluded. 225 226 Pollination deficit 227 Hand pollination significantly shortened the ripening period of blueberry fruit by approximately two 228 weeks or 15% (Table 1), from 15 ± 0.65 weeks (mean ± SE weeks to ripen) when pollinated by honey 229 bees to 13 ± 0.64 (mean ± SE weeks to ripen) when pollinated by hand (Fig 2A). Hand-pollinated 208 responses in fruit mass and diameter per treatment. The presence of honey bees substantially 209 increased fruit mass (mean ± SE fruit mass = 2.89g ± 0.23g, Table 1), with a 72% increase in mass per 210 fruit compared to treatments without access to pollinators (mean ± SE fruit mass = 1.68g ± 0.25g, Fig 211 2B). Similarly, honey bees also caused a mean increase of 4mm (31%) per fruit (mean ± SE fruit 212 diameter = 18.69mm ± 0.65mm) compared to treatments where honey bees were excluded (mean ± 213 SE fruit diameter = 14.32mm ± 0.69mm, Fig 2C, Table 1). There was no difference in the fruit set of 214 flowers with access to honey bees relative to flowers which had no access to honey bees or hand 215 pollinated flowers (Chi-square=1.36, df=4, p=0.507) (Fig 2D). 214 flowers with access to honey bees relative to flowers which had no access to honey bees or hand 215 pollinated flowers (Chi-square=1.36, df=4, p=0.507) (Fig 2D). 188 Benefit of bees PD = pollination deficit, this 200 is the percentage difference between flowers with access to honey bees compared to flowers were hand 201 pollinated. 203 Table 1: Outcome of three pollination treatments on blueberry fruit ripening, quality and yield. Benefit of bees (BB) Pollination deficit (PD) Pollinator exclusion – optimized pollination Ripening period z=-1.505, p=0.2885 z=-2.734, p=0.0173 z=-4.165, p<0.001 Mass z=4.346, p<0.001 z=2.449, p=0.0381 z=6.771, p<0.001 Diameter z=4.617, p<0.001 z=2.338, p=0.0508 z=6.938, p<0.001 Yield t=4.870, p<0.001 t=3.164, p=0.0078 t=7.992, p<0.001 204 Significance indicated by bold type able 1: Outcome of three pollination treatments on blueberry fruit ripening, quality and yield. 205 The linear mixed-effects model explained the majority of conditional variance for both mass 206 (R2c=0.999) and diameter (R2c=0.999). The model explained a large proportion of the marginal 207 variance for mass (R2m=0.503) and diameter (R2m=0.517), suggesting variation in individual plant 205 The linear mixed-effects model explained the majority of conditional variance for both mass 206 (R2c=0.999) and diameter (R2c=0.999). The model explained a large proportion of the marginal 207 variance for mass (R2m=0.503) and diameter (R2m=0.517), suggesting variation in individual plant 9 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 208 responses in fruit mass and diameter per treatment. The presence of honey bees substantially 209 increased fruit mass (mean ± SE fruit mass = 2.89g ± 0.23g, Table 1), with a 72% increase in mass per 210 fruit compared to treatments without access to pollinators (mean ± SE fruit mass = 1.68g ± 0.25g, Fig 211 2B). Similarly, honey bees also caused a mean increase of 4mm (31%) per fruit (mean ± SE fruit 212 diameter = 18.69mm ± 0.65mm) compared to treatments where honey bees were excluded (mean ± 213 SE fruit diameter = 14.32mm ± 0.69mm, Fig 2C, Table 1). 188 Benefit of bees 216 The beneficial effects of honey bees on fruit mass and fruit set also translated into 217 differences in total yield as calculated using Isaac’s [15] average flowers produced per plant, with the 218 linear model explaining 54% of the variance in yield (R2=0.5355, F (2,44) =27.52, p<0.001). Here, yield 219 increased substantially (152%) from 10.11kg ± 2.2kg (mean ± SE predicted plant yield) per plant 220 when pollinators were excluded to 25.52kg ± 3kg (mean ± SE predicted plant yield) when honey bees 221 were allowed to forage on blueberries (Table 1, Fig 2E). Using a value of $7.48/kg [30] of blueberries 222 and a density of 7500 plants per hectare, adding honey bee hives in areas lacking natural blueberry 223 pollinators can potentially increase blueberry revenue by $864 501/ha (152%) compared to if 224 pollinators were excluded. 225 226 Pollination deficit 227 Hand pollination significantly shortened the ripening period of blueberry fruit by approximately two 228 weeks or 15% (Table 1), from 15 ± 0.65 weeks (mean ± SE weeks to ripen) when pollinated by honey 229 bees to 13 ± 0.64 (mean ± SE weeks to ripen) when pollinated by hand (Fig 2A). Hand-pollinated 230 fruits were significantly heavier (Table 1) at 3.69g ± 0.23g (mean ± SE fruit mass), than flowers 231 pollinated by honey bees at 2.89g ± 0.23g (mean ± SE fruit mass, Fig 2B), a ca.27% increase. Hand 216 The beneficial effects of honey bees on fruit mass and fruit set also translated into 217 differences in total yield as calculated using Isaac’s [15] average flowers produced per plant, with the 218 linear model explaining 54% of the variance in yield (R2=0.5355, F (2,44) =27.52, p<0.001). Here, yield 219 increased substantially (152%) from 10.11kg ± 2.2kg (mean ± SE predicted plant yield) per plant 220 when pollinators were excluded to 25.52kg ± 3kg (mean ± SE predicted plant yield) when honey bees 221 were allowed to forage on blueberries (Table 1, Fig 2E). Using a value of $7.48/kg [30] of blueberries 222 and a density of 7500 plants per hectare, adding honey bee hives in areas lacking natural blueberry 223 pollinators can potentially increase blueberry revenue by $864 501/ha (152%) compared to if 224 pollinators were excluded. 188 Benefit of bees 216 The beneficial effects of honey bees on fruit mass and fruit set also translated into 217 differences in total yield as calculated using Isaac’s [15] average flowers produced per plant, with the 218 linear model explaining 54% of the variance in yield (R2=0.5355, F (2,44) =27.52, p<0.001). Here, yield 219 increased substantially (152%) from 10.11kg ± 2.2kg (mean ± SE predicted plant yield) per plant 220 when pollinators were excluded to 25.52kg ± 3kg (mean ± SE predicted plant yield) when honey bees 221 were allowed to forage on blueberries (Table 1, Fig 2E). Using a value of $7.48/kg [30] of blueberries 222 and a density of 7500 plants per hectare, adding honey bee hives in areas lacking natural blueberry 223 pollinators can potentially increase blueberry revenue by $864 501/ha (152%) compared to if 224 pollinators were excluded. 227 Hand pollination significantly shortened the ripening period of blueberry fruit by approximately two 228 weeks or 15% (Table 1), from 15 ± 0.65 weeks (mean ± SE weeks to ripen) when pollinated by honey 229 bees to 13 ± 0.64 (mean ± SE weeks to ripen) when pollinated by hand (Fig 2A). Hand-pollinated 230 fruits were significantly heavier (Table 1) at 3.69g ± 0.23g (mean ± SE fruit mass), than flowers 231 pollinated by honey bees at 2.89g ± 0.23g (mean ± SE fruit mass, Fig 2B), a ca.27% increase. Hand 10 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is mad The copyright holder for this preprint (whi this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. 188 Benefit of bees ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 232 pollination did not significantly increase the size of fruits compared to fruits resulting from honey 233 bee pollination (Table 1, Fig 2C). 234 However, when both mass and fruit set were incorporated into a model to calculate total 235 yield, hand pollinations resulted in significantly greater yields than honey bee pollination (Table 1). 236 Calculated per plant, optimizing pollination (i.e. hand pollination) can potentially increase yields 237 from 25.52kg ± 3kg (mean ± SE predicted plant yield, after honey bee pollination) to 34.47kg ± 2.9kg 238 (mean ± SE predicted plant yield, Fig. 2E), approximately 35%. This could lead to additional revenue 239 amounting to $502 095/ha (35%). 11 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 240 Discussion 241 This study revealed that the pollination environment has the potential to strongly affect the quality 242 of fruit produced by the highbush blueberry variety Ventura. In particular, yields are greatly 243 increased by the addition of honey bees in areas where bumble bee pollinators do not occur 244 naturally, and importation is illegal. Honey bees were the only pollinators at this site and therefore 245 the effects shown are a direct result of access to honey bees, rather than other unaccounted wild 246 pollinators. This provides valuable information for the pollination of commercial blueberries, 247 particularly with respect to the underutilized role played by honey bees, and suggests some 248 important directions for research on blueberry pollination. 241 This study revealed that the pollination environment has the potential to strongly affect the quality 242 of fruit produced by the highbush blueberry variety Ventura. In particular, yields are greatly 243 increased by the addition of honey bees in areas where bumble bee pollinators do not occur 244 naturally, and importation is illegal. Honey bees were the only pollinators at this site and therefore 245 the effects shown are a direct result of access to honey bees, rather than other unaccounted wild 246 pollinators. This provides valuable information for the pollination of commercial blueberries, 247 particularly with respect to the underutilized role played by honey bees, and suggests some 248 important directions for research on blueberry pollination. 249 We show for the first time that Ventura can produce fruit without pollinators. However, 250 these fruits are of lower quality than the fruits of flowers exposed to honey bee pollinators. This 251 ability is not unique to Ventura, as other highbush blueberry varieties can also produce fruit in the 252 absence of pollinators. However, these fruits are also of noticeably poorer quality than fruits 253 produced by flowers with access to pollinators [5,6,15]. The presence of honey bees significantly 254 increased blueberry yields by improving fruit quality through greater fruit diameter and mass (Fig. 2). 255 This shows that despite honey bees’ inability to buzz-pollinate, they do extract blueberry pollen from 256 anthers and transfer it to stigmas. Thus, in areas lacking native blueberry pollinators, the addition of 257 honey bees may increase blueberry yields by more than 150% (Fig 2E). This translates to an 258 economic value of approximately $864 501/ha. 269 Hand pollinations are not a realistic maximum yield 270 The magnitude of the pollination deficit is contingent on what honey bee-pollination yields are being 271 compared with (in this case hand-pollinations). It is possible that blueberry yields resulting from 272 honey bee pollination are already close to the maximum that can be achieved through animal 273 pollination and that no amount of tinkering is likely to reduce this perceived deficit. For example, 274 pollination by bumble bees may not result in smaller pollination deficits than pollination by honey 275 bees. Unfortunately, there is a lack of data comparing blueberry fruit yields after pollination by 276 different bee species to fruit yields after hand pollinations. Consequently, it is unclear how 277 pollination deficits under different pollinator environments are likely to vary and if honey bee hives 278 are any less effective than bumble bee colonies as commercial pollinators of blueberries. This 279 represents an important first step in determining whether the pollination deficit can be reduced. 270 The magnitude of the pollination deficit is contingent on what honey bee-pollination yields are being 271 compared with (in this case hand-pollinations). It is possible that blueberry yields resulting from 272 honey bee pollination are already close to the maximum that can be achieved through animal 273 pollination and that no amount of tinkering is likely to reduce this perceived deficit. For example, 274 pollination by bumble bees may not result in smaller pollination deficits than pollination by honey 275 bees. Unfortunately, there is a lack of data comparing blueberry fruit yields after pollination by 276 different bee species to fruit yields after hand pollinations. Consequently, it is unclear how 277 pollination deficits under different pollinator environments are likely to vary and if honey bee hives 278 are any less effective than bumble bee colonies as commercial pollinators of blueberries. This 279 represents an important first step in determining whether the pollination deficit can be reduced. 240 Discussion CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 265 research on blueberry pollination. There could be several reasons for the sub-optimal yields 266 produced by honey bee pollination and future research needs to concentrate on these to optimize 267 the pollination environment. Below we discuss four potential reasons for sub-optimal yields, each of 268 which should be targeted in future studies in an attempt to improve blueberry yields. 265 research on blueberry pollination. There could be several reasons for the sub-optimal yields 266 produced by honey bee pollination and future research needs to concentrate on these to optimize 267 the pollination environment. Below we discuss four potential reasons for sub-optimal yields, each of 268 which should be targeted in future studies in an attempt to improve blueberry yields. 240 Discussion Consequently, honey bees may be extremely 259 beneficial, potentially eliminating the need to import bumble bees in countries which do not have 260 native blueberry pollinators. 249 We show for the first time that Ventura can produce fruit without pollinators. However, 250 these fruits are of lower quality than the fruits of flowers exposed to honey bee pollinators. This 251 ability is not unique to Ventura, as other highbush blueberry varieties can also produce fruit in the 252 absence of pollinators. However, these fruits are also of noticeably poorer quality than fruits 253 produced by flowers with access to pollinators [5,6,15]. The presence of honey bees significantly 254 increased blueberry yields by improving fruit quality through greater fruit diameter and mass (Fig. 2). 255 This shows that despite honey bees’ inability to buzz-pollinate, they do extract blueberry pollen from 256 anthers and transfer it to stigmas. Thus, in areas lacking native blueberry pollinators, the addition of 257 honey bees may increase blueberry yields by more than 150% (Fig 2E). This translates to an 258 economic value of approximately $864 501/ha. Consequently, honey bees may be extremely 259 beneficial, potentially eliminating the need to import bumble bees in countries which do not have 260 native blueberry pollinators. 261 Despite this benefit, there is still a pollination deficit of approximately 27%, which suggests 262 that there may be room to optimize pollination. However, it is unclear exactly why honeybee 263 pollination results in sub-optimal fruit yields and whether native bumble bee pollinators result in 264 greater yields than high densities of honey bees. These represent important directions for future 261 Despite this benefit, there is still a pollination deficit of approximately 27%, which suggests 262 that there may be room to optimize pollination. However, it is unclear exactly why honeybee 263 pollination results in sub-optimal fruit yields and whether native bumble bee pollinators result in 264 greater yields than high densities of honey bees. These represent important directions for future 12 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . 280 Blueberry attractiveness 281 If honey bee pollination really is less effective than other modes of pollination, it may be the result 282 of low visitation rates to blueberry flowers. Blueberry flowers may be less attractive to honeybees 283 than wild or other agricultural flowers that surround blueberry farms. Low attractiveness relative to 284 other flowers may occur because blueberries are adapted to larger bees and both their nectar and 285 pollen rewards may be more difficult for honeybees to access [44]. Ventura has a long floral tube 286 length (11.39mm ± 0.4mm) which may make it difficult for honey bees to access nectar at the 287 bottom of the flower as bees would need to insert nearly half of their bodies into the flower to reach 288 the nectar. This increases honey bees energy expenditure and may cause honey bees to search for 281 If honey bee pollination really is less effective than other modes of pollination, it may be the result 282 of low visitation rates to blueberry flowers. Blueberry flowers may be less attractive to honeybees 283 than wild or other agricultural flowers that surround blueberry farms. Low attractiveness relative to 284 other flowers may occur because blueberries are adapted to larger bees and both their nectar and 285 pollen rewards may be more difficult for honeybees to access [44]. Ventura has a long floral tube 286 length (11.39mm ± 0.4mm) which may make it difficult for honey bees to access nectar at the 287 bottom of the flower as bees would need to insert nearly half of their bodies into the flower to reach 288 the nectar. This increases honey bees energy expenditure and may cause honey bees to search for 13 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is mad The copyright holder for this preprint (wh this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . 280 Blueberry attractiveness CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 289 more favourable flowers. A possible way to overcome this is through the development of varieties 290 with shorter or wider corollas. 291 292 Honeybees may not pick up pollen 293 High visitation by honey bees may still result in low fruit set if they forage for nectar but remove very 294 little pollen from the poricidal anthers. This could occur because honey bees are not capable of buzz 295 pollinating and hence extract few pollen grains per visit [7,10,12]; honey bees do appear to deposit 296 very few pollen grains per visit to blueberry flowers [7]. Other than trying to develop varieties with 297 pollen which is more easily released, there is very little one can do to improve this potential 298 problem. 299 300 Further research needs to explore the efficiency of honey bees as commercial pollinators of different 301 blueberry varieties, especially comparative studies with bumble bees. More detailed investigations 302 of behavioural interactions of honey bees with blueberry flowers may highlight traits which 303 determine honey bee preferences for different varieties. Studies on the mechanical fit of honey bees 304 to the flowers of different blueberry varieties may also illuminate which varieties are best suited to 305 honey bee pollination. 292 Honeybees may not pick up pollen 293 High visitation by honey bees may still result in low fruit set if they forage for nectar but remove very 294 little pollen from the poricidal anthers. This could occur because honey bees are not capable of buzz 295 pollinating and hence extract few pollen grains per visit [7,10,12]; honey bees do appear to deposit 296 very few pollen grains per visit to blueberry flowers [7]. Other than trying to develop varieties with 297 pollen which is more easily released, there is very little one can do to improve this potential 298 problem. 14 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 306 Acknowledgements 307 We would like to thank BerryWorld South Africa for providing access to the Backsberg farm. We 308 would also like to thank our funders, the National Research Foundation (South Africa) under grant 309 number 112277 and South African Berry Producers Association (KM), the Claude Leon Foundation 310 (MDJ), the Eva Crane Trust (ECTA_20170609 to CM and ECTA_20170905 to MDJ) and the National 311 Research Foundation (South Africa) (105987 to BA and 111979 to CM). 15 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 312 References 313 1. Gallai N, Salles JM, Settele J, Vaissière BE. Economic valuation of the vulnerability of world 314 agriculture confronted with pollinator decline. Ecol Econ. 2009;68: 810–821. 315 2. Klein A-M, Vaissiere BE, Cane JH, Steffan-Dewenter I, Cunningham SA, Kremen C, et al. 316 Importance of pollinators in changing landscapes for world crops. Proc R Soc B Biol Sci. 317 2007;274: 303–313. 318 3. Dogterom MH, Winston ML, Amy M. Effect of pollen load size and source (self, outcross) on 319 seed and fruit production in highbush blueberry cv. “Bluecrop” (Vaccinium corymbosum; 320 Ericaceae). Am J Bot. 2000;87: 1584–1591. 321 4. Aras P, de Oliveira D, Savoie L. Effect of a honey bee (Hymenoptera: Apidae) gradient on the 322 pollination and yield of lowbush blueberry. J Econ Entomol. 1996;89: 1080–1083. 323 5. MacKenzie K. Pollination requirements of three highbush blueberry cultivars. Journal of 324 American Society of Horticultural Science. 1997. pp. 891–896. 325 6. Campbell JW, Kimmel CB, Bammer M, Stanley-Stahr C, Daniels JC, Ellis JD. Managed and wild 326 bee flower visitors and their potential contribution to pollination services of low-chill 327 highbush blueberry (Vaccinium corymbosum L.; Ericales: Ericaceae). J Econ Entomol. 328 2018;111: 2011–2016. 329 7. Javorek S, MacKenzie K, Vander Kloet S. Comparative pollination effectiveness among bees 330 (Hymenoptera: Apoidea) on lowbush blueberry (Ericaceae: Vaccinium angustifolium). Ann 331 Entomol Soc Am. 2002;95: 345–351. 332 8. Mackenzie KE, Eickwort GC. Diversity and abundance of bees (Hymenoptera: Apoidea) 333 foraging on highbush blueberry (Vaccinium corymbosum L .) in central New York. J Kansas 334 Entomol Soc. 1996;69: 185–194. 313 1. Gallai N, Salles JM, Settele J, Vaissière BE. Economic valuation of the vulnerability of world 314 agriculture confronted with pollinator decline. Ecol Econ. 2009;68: 810–821. 318 3. Dogterom MH, Winston ML, Amy M. Effect of pollen load size and source (self, outcross) on 319 seed and fruit production in highbush blueberry cv. “Bluecrop” (Vaccinium corymbosum; 320 Ericaceae). Am J Bot. 2000;87: 1584–1591. 321 4. Aras P, de Oliveira D, Savoie L. Effect of a honey bee (Hymenoptera: Apidae) gradient on the 322 pollination and yield of lowbush blueberry. J Econ Entomol. 1996;89: 1080–1083. 321 4. Aras P, de Oliveira D, Savoie L. Effect of a honey bee (Hymenoptera: Apidae) gradient on the 322 pollination and yield of lowbush blueberry. J Econ Entomol. 1996;89: 1080–1083. 323 5. MacKenzie K. Pollination requirements of three highbush blueberry cultivars. . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 312 References Journal of 324 American Society of Horticultural Science. 1997. pp. 891–896. 325 6. Campbell JW, Kimmel CB, Bammer M, Stanley-Stahr C, Daniels JC, Ellis JD. Managed and wild 326 bee flower visitors and their potential contribution to pollination services of low-chill 327 highbush blueberry (Vaccinium corymbosum L.; Ericales: Ericaceae). J Econ Entomol. 328 2018;111: 2011–2016. 332 8. Mackenzie KE, Eickwort GC. Diversity and abundance of bees (Hymenoptera: Apoidea) 333 foraging on highbush blueberry (Vaccinium corymbosum L .) in central New York. J Kansas 334 Entomol Soc. 1996;69: 185–194. 16 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 335 9. Campbell JW, O’Brien J, Irvin JH, Kimmel CB, Daniels JC, Ellis JD. Managed bumble bees 336 (Bombus impatiens) (Hymenoptera: Apidae) caged with blueberry bushes at high density did 337 not increase fruit set or fruit weight compared to open pollination. Environ Entomol. 2017;46: 338 237–242. 339 10. Stubbs CS, Drummond FA. Bombus impatiens (Hymenoptera: Apidae): an alternative to Apis 340 mellifera (Hymenoptera: Apidae) for lowbush blueberry pollination. J Econ Entomol. 2001;94: 341 609–616. 342 11. Cane JH, Payne JA. Regional, annual, and seasonal variation in pollinator guilds: intrinsic traits 343 of bees (Hymenoptera: Apoidea) underlie their patterns of abundance at Vaccinium ashei 344 (Ericaceae). Ann Entomol Soc Am. 1993;56: 577–588. 345 12. De Luca PA, Vallejo-Marín M. What’s the “buzz” about? The ecology and evolutionary 346 significance of buzz-pollination. Curr Opin Plant Biol. 2013;16: 429–435. 347 13. Buchmann SL, Jones CE, Little RJ. Buzz pollination in angiosperms. Handbook of experimental 348 pollination biology. New York, NY, USA: Van Nostrand Reinhold Company; 1983. pp. 73–113. 349 14. Fulton M, Jesson LK, Bobiwash K, Schoen DJ. Mitigation of pollen limitation in the lowbush 350 blueberry agroecosystem: effect of augmenting natural pollinators. Ecosphere. 2015;6: 1–19. 351 15. Isaacs R, Kirk AK. 312 References CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 358 18. Winter K, Adams L, Thorp R, Inouye DW, Day L, Ascher J, et al. Importation of non-native 359 bumble bees into North America: potential consequences of using Bombus terrestris and 360 other non-native bumble bees for greenhouse crop pollination in Canada, Mexico, and the 361 United States [Internet]. A White Paper of the North American Pollinator Protecting 362 Campaign. 2006. 363 19. Inoue MN, Yokoyama J. Competition for flower resources and nest sites between Bombus 364 terrestris (L .) and Japanese native bumblebees. Appl Entomol Zool. 2010;45: 29–35. 365 20. Goka K. Introduction to the special feature for ecological risk assessment of introduced 366 bumblebees: status of the European bumblebee, Bombus terrestris, in Japan as a beneficial 367 pollinator and an invasive alien species. Appl Entomol Zool. 2010;45: 1–6. 368 21. Aizen MA, Smith-Ramírez C, Morales CL, Vieli L, Sáez A, Barahona-Segovia RM, et al. 369 Coordinated species importation policies are needed to reduce serious invasions globally: the 370 case of alien bumblebees in South America. J Appl Ecol. 2018; 1–7. 371 22. Morales CL, Arbetman MP, Cameron SA, Aizen MA. Rapid ecological replacement of a native 372 bumble bee by invasive species. Front Ecol Environ. 2013;11: 529–534. 373 23. Geslin B, Morales CL. New records reveal rapid geographic expansion of Bombus terrestris 374 Linnaeus, 1758 (Hymenoptera: Apidae), an invasive species in Argentina. Check List. 2015;11: 375 1620. 376 24. Schmid-Hempel R, Eckhardt M, Goulson D, Heinzmann D, Lange C, Plischuk S, et al. The 377 invasion of southern South America by imported bumblebees and associated parasites. J 378 Anim Ecol. 2014;83: 823–837. 379 25. Palau M, Bargain V, Valenzuela M del M, Leebutterworth D, Bammatoua N, Mishina A, et al. 380 All about blueberries in Latin America. Eurofresh Distribution. 2016: 38. 312 References Pollination services provided to small and large highbush blueberry fields by 352 wild and managed bees. J Appl Ecol. 2010;47: 841–849. 353 16. Lang G a, Danka RG. Honey-bee mediated cross- versus self-pollination of ‘Sharpblue’ 354 blueberry increases fruit size and hastens ripening. J Am Soc Hortic Sci. 1991;116: 770–773. 355 17. Velthuis HHW, van Doorn A. A century of advances in bumblebee domestication and the 356 economic and environmental aspects of its commercialization for pollination. Apidologie. 357 2006;37: 421–451. 335 9. Campbell JW, O’Brien J, Irvin JH, Kimmel CB, Daniels JC, Ellis JD. Managed bumble bees 336 (Bombus impatiens) (Hymenoptera: Apidae) caged with blueberry bushes at high density did 337 not increase fruit set or fruit weight compared to open pollination. Environ Entomol. 2017;46: 338 237–242. 342 11. Cane JH, Payne JA. Regional, annual, and seasonal variation in pollinator guilds: intrinsic traits 343 of bees (Hymenoptera: Apoidea) underlie their patterns of abundance at Vaccinium ashei 344 (Ericaceae). Ann Entomol Soc Am. 1993;56: 577–588. 342 11. Cane JH, Payne JA. Regional, annual, and seasonal variation in pollinator guilds: intrinsic traits 343 of bees (Hymenoptera: Apoidea) underlie their patterns of abundance at Vaccinium ashei 344 (Ericaceae). Ann Entomol Soc Am. 1993;56: 577–588. 345 12. De Luca PA, Vallejo-Marín M. What’s the “buzz” about? The ecology and evolutionary 346 significance of buzz-pollination. Curr Opin Plant Biol. 2013;16: 429–435. 345 12. De Luca PA, Vallejo-Marín M. What’s the “buzz” about? The ecology and evolutionary 346 significance of buzz-pollination. Curr Opin Plant Biol. 2013;16: 429–435. 347 13. Buchmann SL, Jones CE, Little RJ. Buzz pollination in angiosperms. Handbook of experimental 348 pollination biology. New York, NY, USA: Van Nostrand Reinhold Company; 1983. pp. 73–113. 349 14. Fulton M, Jesson LK, Bobiwash K, Schoen DJ. Mitigation of pollen limitation in the lowbush 350 blueberry agroecosystem: effect of augmenting natural pollinators. Ecosphere. 2015;6: 1–19. 351 15. Isaacs R, Kirk AK. Pollination services provided to small and large highbush blueberry fields by 352 wild and managed bees. J Appl Ecol. 2010;47: 841–849. 353 16. Lang G a, Danka RG. Honey-bee mediated cross- versus self-pollination of ‘Sharpblue’ 354 blueberry increases fruit size and hastens ripening. J Am Soc Hortic Sci. 1991;116: 770–773. 355 17. Velthuis HHW, van Doorn A. A century of advances in bumblebee domestication and the 356 economic and environmental aspects of its commercialization for pollination. Apidologie. 357 2006;37: 421–451. 17 17 . 312 References CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 382 visiting rabbiteye blueberry. J Econ Entomol. 2000;93: 1726–1731. 383 27. Hoffman GD, Lande C, Rao S. A novel pollen transfer mechanism by honey bee foragers on 384 highbush blueberry (Ericales: Ericaceae). Environ Entomol. 2018;47: 1465–1470. 385 28. Gibbs J, Elle E, Bobiwash K, Haapalainen T, Isaacs R. Contrasting pollinators and pollination in 386 native and non-native regions of highbush blueberry production. PLoS One. 2016;11: 1–24. 387 29. Sobekova K, Thomsen MR, Ahrendsen BL. Market trends and consumer demand for fresh 388 berries. Agroinform Publ House. 2013;2: 11–14. 389 30. Nicholson CC, Ricketts TH. Wild pollinators improve production, uniformity, and timing of 390 blueberry crops. Agric Ecosyst Environ. Elsevier; 2019;272: 29–37. 391 31. Nakagawa S, Schielzeth H. A general and simple method for obtaining R2 from generalized 392 linear mixed-effects models. Methods Ecol Evol. 2013;4: 133–142. 393 32. R Core Team. R: a language and environment for statistical computing [Internet]. Vienna; 394 2017. Available: https://www.r-project.org/ 395 33. Pinheiro J, Bates D, DebRoy S, Sarkar D, Team RC. Nlme: linear and nonlinear mixed effects 396 models [Internet]. 2017. Available: https://cran.r-project.org/package=nlme 397 34. Hothorn T, Bretz F, Westfall P. Simultaneous inference in general parametric models. 398 Biometrical J. 2008;50: 346–363. 399 35. Wickham H. Ggplot2: elegant graphics for data analysis. New York: Springer-Verlag; 2009. 400 36. Lüdecke D. SjPlot: data visualization for statistics in social science [Internet]. 2017. Available: 401 https://cran.r-project.org/package=sjPlot 402 37. Fox J, Weisberg S. An {R} companion to applied regression. Second Edi. California: Sage; 2011 403 38. Bates D, Maechler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J 382 visiting rabbiteye blueberry. J Econ Entomol. 2000;93: 1726–1731. 382 383 27. Hoffman GD, Lande C, Rao S. A novel pollen transfer mechanism by honey bee foragers on 384 highbush blueberry (Ericales: Ericaceae). Environ Entomol. 2018;47: 1465–1470. 385 28. Gibbs J, Elle E, Bobiwash K, Haapalainen T, Isaacs R. Contrasting pollinators and pollination in 386 native and non-native regions of highbush blueberry production. PLoS One. 2016;11: 1–24. 387 29. Sobekova K, Thomsen MR, Ahrendsen BL. Market trends and consumer demand for fresh 388 berries. Agroinform Publ House. 312 References 381 26 Sampson B Cane J Pollination efficiencies of three bee (Hymenoptera: Apoidea) species 358 18. Winter K, Adams L, Thorp R, Inouye DW, Day L, Ascher J, et al. Importation of non-native 359 bumble bees into North America: potential consequences of using Bombus terrestris and 360 other non-native bumble bees for greenhouse crop pollination in Canada, Mexico, and the 361 United States [Internet]. A White Paper of the North American Pollinator Protecting 362 Campaign. 2006. 363 19. Inoue MN, Yokoyama J. Competition for flower resources and nest sites between Bombus 364 terrestris (L .) and Japanese native bumblebees. Appl Entomol Zool. 2010;45: 29–35. 365 20. Goka K. Introduction to the special feature for ecological risk assessment of introduced 366 bumblebees: status of the European bumblebee, Bombus terrestris, in Japan as a beneficial 367 pollinator and an invasive alien species. Appl Entomol Zool. 2010;45: 1–6. 368 21. Aizen MA, Smith-Ramírez C, Morales CL, Vieli L, Sáez A, Barahona-Segovia RM, et al. 369 Coordinated species importation policies are needed to reduce serious invasions globally: the 370 case of alien bumblebees in South America. J Appl Ecol. 2018; 1–7. 371 22. Morales CL, Arbetman MP, Cameron SA, Aizen MA. Rapid ecological replacement of a native 372 bumble bee by invasive species. Front Ecol Environ. 2013;11: 529–534. 373 23. Geslin B, Morales CL. New records reveal rapid geographic expansion of Bombus terrestris 374 Linnaeus, 1758 (Hymenoptera: Apidae), an invasive species in Argentina. Check List. 2015;11: 375 1620. 376 24. Schmid-Hempel R, Eckhardt M, Goulson D, Heinzmann D, Lange C, Plischuk S, et al. The 377 invasion of southern South America by imported bumblebees and associated parasites. J 378 Anim Ecol. 2014;83: 823–837. 379 25. Palau M, Bargain V, Valenzuela M del M, Leebutterworth D, Bammatoua N, Mishina A, et al. 380 All about blueberries in Latin America. Eurofresh Distribution. 2016: 38. 18 381 26. Sampson B, Cane J. Pollination efficiencies of three bee (Hymenoptera: Apoidea) species 18 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . 312 References 2013;2: 11–14. 389 30. Nicholson CC, Ricketts TH. Wild pollinators improve production, uniformity, and timing of 390 blueberry crops. Agric Ecosyst Environ. Elsevier; 2019;272: 29–37. 391 31. Nakagawa S, Schielzeth H. A general and simple method for obtaining R2 from generalized 392 linear mixed-effects models. Methods Ecol Evol. 2013;4: 133–142. 391 31. Nakagawa S, Schielzeth H. A general and simple method for obtaining R2 from generalized 392 linear mixed-effects models. Methods Ecol Evol. 2013;4: 133–142. 393 32. R Core Team. R: a language and environment for statistical computing [Internet]. Vienna; 394 2017. Available: https://www.r-project.org/ 393 32. R Core Team. R: a language and environment for statistical computing [Internet]. Vienna; 394 2017. Available: https://www.r-project.org/ 395 33. Pinheiro J, Bates D, DebRoy S, Sarkar D, Team RC. Nlme: linear and nonlinear mixed effects 396 models [Internet]. 2017. Available: https://cran.r-project.org/package=nlme 19 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint 404 Stat Softw. 2015;67: 1–48. 405 39. Auguie B. GridExtra: miscellaneous functions for “grid” graphics [Internet]. 2017. Available: 406 https://cran.r-project.org/package=gridExtra 407 40. Deepayan S. Lattice: multivariate data visualization with R. New York: Springer; 2008. 408 41. Barton K. MuMIn: multi-model inference [Internet]. 2017. Available: https://cran.r- 409 project.org/package=MuMIn 410 42. Wickham H. The split-apply-combine strategy for data analysis. J Stat Softw. 2011;40: 1–29. 411 43. Lemon J. Plotrix: a package in the red light district of R. R-News. 2006;6: 8–12. 412 44. Courcelles DMM, Button L, Elle E. Bee visit rates vary with floral morphology among highbush 413 blueberry cultivars (Vaccinium corymbosum L.). J Appl Entomol. 2013;137: 693–701. 414 404 410 42. Wickham H. The split-apply-combine strategy for data analysis. J Stat Softw. 2011;40: 1–29. 412 44. Courcelles DMM, Button L, Elle E. Bee visit rates vary with floral morphology among highbush 413 blueberry cultivars (Vaccinium corymbosum L.). J Appl Entomol. 2013;137: 693–701. 414 20 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted February 16, 2019. ; https://doi.org/10.1101/551523 doi: bioRxiv preprint uity. It is made preprint (which
https://openalex.org/W2061296641	https://www.scielo.br/j/motriz/a/dGRhJJXWLbQznM3DdKRKjDH/?lang=pt&format=pdf	Portuguese	null	Necessidades de formação para o exercício profissional na perspectiva do treinador de Futebol em função da sua experiência e nível de formação	Motriz	2,010	cc-by	7,535	doi: http://dx.doi.org/10.5016/1980-6574.2010v16n4p931 doi: http://dx.doi.org/10.5016/1980-6574.2010v16n4p931 Motriz, Rio Claro, v.16 n.4 p.931-941, out./dez. 2010 Artigo Original 1 Mestre pela Faculdade de Desporto, Universidade do Porto, Portugal 2 Professora Associada com Agregação da Faculdade de Desporto, Universidade do Porto, Portugal 3 Professor Associado com Agregação da Faculdade de Motricidade Humana Universidade Técnica de Lisboa, Portugal 4 Licenciado pela Faculdade de Desporto, Universidade do Porto, Portugal 1 Mestre pela Faculdade de Desporto, Universidade do Porto, Portugal 2 Professora Associada com Agregação da Faculdade de Desporto, Universidade do Porto, Portugal 3 Professor Associado com Agregação da Faculdade de Motricidade Humana Universidade Técnica de Lisboa, Portugal 4 Licenciado pela Faculdade de Desporto, Universidade do Porto, Portugal Resumo: O objetivo do presente estudo consistiu no reconhecimento das necessidades de formação das competências profissionais de treinadores de Futebol em função dos anos de experiência profissional e da formação federativa. A amostra foi constituída por 81 treinadores portugueses. A recolha dos dados foi realizada através da aplicação de um questionário após validação de construção e de conteúdo. Como procedimentos estatísticos, recorremos ao teste paramétrico One-Way ANOVA para a comparação entre dois grupos e para mais de dois grupos foram estimados os Post-Hocs com recurso ao teste de Tukey. Os resultados mostraram que os treinadores com formação federativa mais elevada, níveis II e III, reconhecem menores necessidades de formação que os treinadores sem formação nas áreas do Treino, Competição e Papel de Formador. Por sua vez, os treinadores mais experientes reconheceram menores necessidades de formação nos domínios do Treino, da Gestão Desportiva e do Papel de Formador em relação aos menos experientes. Palavras-chave: Treinadores. Competências profissionais. Necessidades de formação. Formação federativa. Experiência profissional Training needs for professional practice in the perspective of the football coach according to its experience and level of education Training needs for professional practice in the perspective of the football coach according to its experience and level of education Abstract: The aim of the present study consisted in the recognition of Football coaches’ perceptions of training needs related to the professional competences according to their experience and level of certification, as coach. A sample of 81 Portuguese coaches answered a questionnaire applied after its construction and content validation. The statistical procedure used was One-Way ANOVA to make comparison between two groups and the Tukey' test for post-hoc comparisons. The results showed that the coaches with the highest level of certification, levels II and III, recognized lower training needs than coaches without certification in the areas of Training, Competition and Coach Education. Moreover, the more experienced coaches recognized, also, lower training needs than less experienced coaches in the areas of Training, Sport Management and Coach Education. Key Words: Coaches. Professional Competences. Training needs. Certification level. Coaching actualmente, remetam para entidades diferenciadas. O termo “competência”, mais amplo, está associado a capacidade, aptidão, resolução e, como tal, distingue-se do conhecimento memorizado de termos, factos e procedimentos básicos (ABRAHAM; COLLINS, 1998; KIRSCHNER et al., 1997; WESTERA, 2001). A definição de “competência” pressupõe a aquisição de um conjunto de conhecimentos e de processos que conduzam à compreensão, interpretação e resolução de problemas actualmente, remetam para entidades diferenciadas. O termo “competência”, mais amplo, está associado a capacidade, aptidão, resolução e, como tal, distingue-se do conhecimento memorizado de termos, factos e procedimentos básicos (ABRAHAM; COLLINS, 1998; KIRSCHNER et al., 1997; WESTERA, 2001). A definição de “competência” pressupõe a aquisição de um conjunto de conhecimentos e de processos que conduzam à compreensão, interpretação e resolução de problemas Necessidades de formação para o exercício profissional na perspectiva do treinador de Futebol em função da sua experiência e nível de formação Gabriel Barros da Cunha 1 Isabel Maria Ribeiro Mesquita 2 António Fernando Boleto Rosado 3 Tiago Sousa4 Pedro Pereira 4 Introdução Ao treinador desportivo é incumbido um conjunto alargado de funções, as quais requerem um corpo eclético de conhecimentos e competências para exercer cabalmente a sua atividade profissional (ABRAHAM; COLLINS; MARTINDALE, 2006; McCALLISTER; BLINDE; WEISS, 2000; VARGAS-TONSING, 2007). Tradicionalmente, tem-se vindo a atribuir idênticos significados aos conceitos de conhecimento e de competência, embora estes, G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira profissionais. Dentro dessa perspectiva, Kirschner et al. (1997) caracterizam competência como o corpo de conhecimentos e habilidades que um indivíduo possui e consegue utilizar de maneira eficaz num contexto específico. Os autores afirmam, ainda, que a competência transcende a acção, sendo uma forma de usar o conhecimento e a habilidade de maneira estratégica (KIRSCHNER et al., 1997). Deste modo, a competência deve ser interpretada como uma função do conhecimento, da habilidade, das atitudes e valores bem como da situação (KIRSCHNER et al., 1997; STEPHENSON; WEILL, 1992). planificação plurianual e anual; competências relacionadas com a orientação da prática e da competição e competências pessoais e de formação. No que se refere às competências dos treinadores, o projecto AEHESIS (DUFFY, 2008) e o programa Baccalaureate in Sport Intervention (DEMERS; WOODBURN; SAVARD, 2006) destacam como actividades profissionais dos treinadores: organizar, implementar e avaliar planos a curto, médio e longo prazo; conduzir e apoiar os praticantes no treino e nas competições; coordenar adjuntos e assistentes bem como outros membros da equipa técnica e participar na formação de jovens treinadores. Por outro lado, competências pessoais e sociais, como a capacidade de comunicação, de aprendizagem e de responsabilização (DUFFY, 2008) são descritas como a base da qualificação da interacção com os participantes e com os outros agentes desportivos e devem estar presentes nos programas de formação de treinadores (JONES; ARMOUR; POTRAC, 2002; SALMELA, 1996). Vargas-Tonsing (2007) relatou o interesse dos treinadores no desenvolvimento de competências de comunicação com os pais e com os atletas, considerando a comunicação uma parte essencial do treino (ABRAHAM; COLLINS; MARTINDALE, 2006; VARGAS-TONSING, 2007; WIERSMA; SHERMAN, 2005). Na realidade, o sucesso e o status social dos treinadores depende da sua habilidade em se tornarem credíveis aos olhos dos atletas, dos pais, dos directores, etc. Deste modo, competências pessoais e sociais bem como educativas são vistas como essenciais (SALMELA, 1996). Do mesmo modo, no seu estudo, Santos et al. Introdução (in press) destacam que os treinadores reconheceram que a comunicação era essencial para o sucesso, associando-a com a liderança e com as boas práticas educativas. Também Salmela (1996) sublinhou a importância dos valores morais e da sensibilidade social e cultural dos treinadores, particularmente importantes dado o longo período de tempo que os treinadores passam com os atletas e o poder que exercem sobre estes. Estudos recentes (CUSHION; JONES, 2006; JONES; ARMOUR; POTRAC, 2004; POTRAC; JONES, 2009) assinalam que as interações sociais constituem o centro do processo de treino, já que “os treinadores são seres sociais operando num meio social” (JONES; ARMOUR; POTRAC, 2002, p. 35). Contudo, Cushion e Jones (2006) relatam que a dinâmica social que fundamenta as relações entre os diversos agentes desportivos não está, ainda, suficientemente compreendida. Dessa forma, reconhecendo e acompanhando as tendências de valorização da actividade profissional do treinador desportivo (ABRAHAM; COLLINS; MARTINDALE, 2006; DEMERS; WOODBURN; SAVARD, 2006; DUFFY, 2008), o estudo das competências profissionais tem vindo a destacar-se, quer ao nível da organização das actividades profissionais quer na agenda científica, onde ainda é escasso o número de investigações, sendo amplamente reconhecida a relevância do tema para o desenvolvimento profissional dos treinadores desportivos. Entre o conjunto de competências profissionais que o treinador necessita possuir, destacam-se, no presente estudo, as competências centradas no treino e na competição, as competências de gestão de equipas técnicas e de apoio à formação de treinadores bem como as competências pessoais de suporte à atividade profissional, numa sistematização apoiada por estudos como os referenciados pelo projeto temático da AEHESIS (Alinhamento Europeu da Educação Superior nas actividades profissionais relacionadas com as Ciências do Desporto) (DUFFY, 2008), e por investigadores da área (ABRAHAM; COLLINS; MARTINDALE, 2006; DEMERS; WOODBURN; SAVARD, 2006). Demers, Woodburn e Savard (2006) e Abraham, Collins e Martindale (2006) consideram que para assegurar o desempenho profissional eficaz, o treinador terá de possuir competências para gerir, conduzir e formar atletas e integrantes da comissão técnica; competências para planear, conduzir e avaliar sessões de treinos, bem como competências para planear, implementar e avaliar as competições e as temporadas. Neste reduto, Santos et al. (in press) através de um estudo aplicado a 343 treinadores portugueses identificaram três grandes áreas de competências claramente diferenciadas, como relevantes, pelos treinadores; competências relacionadas com a 932 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. Introdução 2010 Necessidades de formação dos treinadores Necessidades de formação dos treinadores Neste alcance, o desenho curricular de programas de formação de treinadores deve estar em estreita relação com as necessidades e os requisitos das federações nacionais e internacionais, devendo os treinadores serem preparados de acordo com os diferentes papéis que vão desempenhar. As funções e as competências a desempenhar variam de contexto para contexto, nomeadamente, em função dos estádios de carreira desportiva dos praticantes (DUFFY, 2008) pelo que a formação e qualificação dos treinadores deverá potenciar a diversificação e a especialização baseada numa análise objectiva das necessidades de formação para cada nível de intervenção profissional. Sendo que a formação federativa configura níveis diferenciados de formação, mormente em Portugal quatro níveis, importa perceber se as necessidades de formação se alteram, e em que domínios, em função do nível de formação dos treinadores e destes em relação aos que não possuem qualquer tipo de formação. A análise de necessidades de formação é, hoje, vista como um processo partilhado entre os diversos intervenientes no processo desportivo, proporcionando um clima favorável a uma reflexão conjunta sobre as condições específicas como o trabalho é desenvolvido, a sua organização e condicionantes técnicas, levando, também, em conta as expectativas e anseios dos treinadores. A análise torna-se, deste modo, um verdadeiro ato formativo e gerador de um clima organizativo propiciador de mudanças qualitativas. Nesta senda, a formação resulta da análise da atividade profissional e do seu contexto, implicando neste processo todos os atores. O estudo das necessidades de formação dos treinadores permite-nos, por outro lado, a compreensão dos aspectos em que os treinadores acreditam ter maior necessidade de formação, o que facilita os processos de melhoria da sua própria formação. As investigações realizadas com o intuito de identificarem as necessidades de formação dos treinadores desportivos, em geral, priorizam a perspectiva do próprio treinador. Salmela (1996) refere que a crítica sobre a prática desportiva e seus modelos de formação deveria ser elaborada, não apenas por aqueles que observam externamente as realidades, permanecendo fora delas, mas, especialmente, por quem está no seu interior, de modo a que se formulem alternativas coerentes com as transformações qualitativas que se pretendem promover nessas realidades. Sob essa perspectiva, estudar as necessidades de formação dos treinadores desportivos, tendo em conta a formação federativa e a experiência profissional, aprofunda um debate sobre a formação dos treinadores desportivos, assim como potencia a eficácia da formação. Participantes p Participaram, no presente estudo, 81 treinadores, quase todos do sexo masculino (Masculino = 80; Feminino = 1), de Futebol, sendo 20 da região centro de Portugal (Aveiro e Viseu) e os restantes da região norte (Porto e Braga). Os treinadores são dominantemente jovens, apresentando uma média de idade de 27,9 anos (DP=5,25). São treinadores de escalões de formação, possuindo, na sua maioria (60,4%), o curso de nível I da Federação Portuguesa de Futebol. Um dos grupos é constituído por treinadores sem formação federativa (19,8%) e outro por treinadores com formação federativa de níveis II e III (19,8%). Estes dois últimos grupos foram agrupados uma vez que a diferença na formação federativa entre estes dois níveis é, em Portugal, muito ténue, não se distinguindo, substantivamente, no que se refere aos conteúdos da formação. Ainda sobre a formação dos treinadores, a maioria apresenta licenciatura em Educação Física e Desporto (80,3%). A experiência dos treinadores como atletas é muito diversificada, indo desde a quase ausência de experiência até 22 anos de prática desportiva (M=9,65; DP =5,01), acontecendo o mesmo com a experiência como treinador que se situa entre 1 e 20 anos (M=4,79; DP =3,78). Para a categorização da experiência profissional foi aplicada a classificação preconizada por Burden (1990), sendo os treinadores experientes aqueles que possuem cinco anos ou mais de experiência profissional (fase de estabilização) e os menos experientes aqueles que indicam tempo inferior a cinco anos de experiência profissional (fase de iniciação ou de aperfeiçoamento). No que se refere ao nível de formação federativa, os treinadores foram divididos em três grupos. O primeiro grupo foi composto por treinadores que não frequentaram cursos de formação federativa (n = 16; 19,7%). O segundo grupo foi constituído por treinadores que possuem o curso federativo de nível I (n = 49; 60,6%). Por fim, fizeram parte do terceiro grupo os treinadores com cursos federativos de níveis II e III (n = 16; 19,7%). O questionário foi dividido por domínio de competências: (a) Competências relacionadas com o Treino; (b) Competências relacionadas com a Competição; (c) Competências relacionadas com a Gestão Desportiva; (d) Competências relacionadas com o Papel de Formador; (e) Competências Pessoais. Introdução A percepção das necessidades de formação pode variar, também, de acordo com as características dos treinadores, nomeadamente, a sua experiência profissional. Na realidade, a experiência profissional é considerada uma importante fonte de conhecimento e competência profissional (GILBERT; TRUDEL, 2001; IRWIN; HANTON; KERWIN, 2004; JONES; ARMOUR; POTRAC, 2002; WRIGHT; TRUDEL; CULVER, 2007), sendo enquadrada, ainda, na visão de diversos pesquisadores (JONES; ARMOUR; POTRAC, 2003; VALLÉE; BLOOM, 2005) como um dos pré-requisitos para um treinador alcançar um elevado nível de mestria. De facto, durante a prática profissional, os treinadores são confrontados com dilemas e constrangimentos peculiares da actividade profissional, contribuindo, esses momentos, de maneira indubitável, para o desenvolvimento das capacidades e competências dos treinadores desportivos. Deste modo, um variado leque de competências é requerida para o exercício efectivo da actividade de treinador. Essas competências são geralmente obtidas a partir da formação concretizada por via dos cursos de formação de treinadores e da experiência profissional. As formações obtidas nos cursos promovidos pelas federações desportivas possuem um carácter obrigatório em diversos países para quem pretende tornar-se um treinador desportivo (CAMPBELL, 1993), sendo Portugal um dos países que adoptam esta política formativa. No entanto, é sabido que, mesmo fora do contexto legal, muitos treinadores, em Portugal, não possuem qualquer tipo de formação federativa. Este contexto evidencia a necessidade de aprofundar a qualidade da formação federativa dos treinadores desportivos. Considerando a indubitável pertinência das competências profissionais para o melhor exercício da função dos treinadores desportivos, bem como a relevância de um processo de formação complexo, eclético, contextualizado, objectivo e adequado às reais necessidades dos treinadores desportivo, parece-nos importante colocar as seguintes questões como problemas centrais de estudo: (1) Será que a formação federativa de diferentes níveis distingue os treinadores em relação às áreas de competências profissionais em que percepcionam maiores necessidades de formação? (2) É possível que os anos de experiência como treinador influenciem as percepções de necessidades de formação dos treinadores desportivos? 933 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira Com base nestas questões, o objectivo do presente estudo é identificar se a formação federativa e a experiência profissional diferenciam as necessidades de formação dos treinadores nas áreas relacionadas com o Treino, a Competição, a Gestão Desportiva, o Papel de Formador e as Competências Pessoais. Procedimentos Foi aplicado um questionário para avaliar as necessidades de formação dos treinadores em relação às suas competências profissionais para exercer a actividade profissional. O respectivo questionário foi construído a partir da literatura especializada na temática (ABRAHAM; COLLINS; MARTINDALE, 2006; CÔTÉ, 2006; DUFFY, 2008; SALMELA, 1996). Utilizaram-se três estratégias de desenvolvimento do questionário no sentido de preencher os requisitos de validade de construção e de conteúdo. Em primeiro lugar, o processo de geração dos itens e a construção da primeira versão do questionário baseou-se nos referenciais teóricos disponíveis sobre a temática (ABRAHAM; COLLINS; MARTINDALE, 2006; CÔTÉ; SALMELA; RUSSELL, 1995; DUFFY, 2008; KIRSCHNER et al., 1997). Num segundo momento, para a validação de conteúdo, recorreu-se ao método de peritagem, tendo sido consultados três doutorados em Ciências do Desporto e especialistas da área, que avaliaram se os itens representavam as competências profissionais definidas. Alguns itens foram removidos e outros modificados de acordo com o seu aconselhamento. Por fim, o questionário foi submetido a um estudo piloto com o objetivo de se aferir sobre a inteligibilidade, nível de compreensão, clareza e objetividade das questões. Dessa forma, foi aplicado a 30 treinadores de diferentes modalidades e escalões de prática e, ainda, com experiência profissional diferenciada. Após o estudo piloto, o questionário ficou composto por duas partes; a primeira destinada à caracterização demográfica do treinador e a segunda no reconhecimento das necessidades de formação pela apreciação de 26 itens através de uma escala de Likert que varia entre não necessito (1) a necessito muitíssimo (5). Procedimentos Estatísticos Após a análise dos resultados foi possível inferir que o grupo de treinadores com menor experiência reconheceu maiores necessidades de formação, quando comparado ao grupo de treinadores mais experientes. A tabela 1 apresenta os itens onde se revelaram as diferenças significativas sendo as necessidades de formação apreciadas, em média, como “necessárias” e “muito necessárias”. Para a análise estatística das variáveis do estudo, utilizou-se o programa estatístico Statistical Package for the Social Sciences (SPSS), versão 15.0. Recorremos ao estudo de normalidade com o teste de Komolgorov-Smirnov e ao estudo da homogeneidade das variâncias através do teste de Levene, tendo, ambos, preenchido os pressupostos exigidos para aplicação da estatística paramétrica. Para a comparação entre grupos recorremos ao teste Tabela1. Análise comparativa das necessidades de formação dos treinadores em relação as competências relacionadas com o Treino, Competição, Gestão Desportiva e Papel de Formador em função dos anos de experiência como treinador. Tabela1. Análise comparativa das necessidades de formação dos treinadores em relação as competências relacionadas com o Treino, Competição, Gestão Desportiva e Papel de Formador em função dos anos de experiência como treinador. MÉDIAS Competências relacionadas com o Treino ≤ 5 anos (n=49) > 5 anos (n=32) F Sig Realizar um plano de preparação plurianual 3,57 2,93 6,689 0,012 Organizar e implementar um planejamento plurianual 3,55 2,90 6,565 0,012 Competências relacionadas com a Competição Estabelecer um plano plurianual de participação na competição 3,42 2,84 5,797 0,018 Competências relacionadas com a Gestão Desportiva Gerir o desenvolvimento da carreira desportiva de atletas através da organização da minha actividade no treino e na competição 3,28 2,71 7,396 0,008 Assumir papel de treinador principal coordenando actividades de outros treinadores 3,22 2,81 4,370 0,040 Liderar uma organização envolvendo a coordenação da actividade de atletas, treinadores, técnicos, especialistas de Ciências do Desporto 3,36 2,71 9,481 0,003 Competências relacionadas com o Papel de Formador Orientar a formação de treinadores principiantes 3,36 2,78 6,497 0,013 Coordenar a formação de técnicos da modalidade 3,40 2,84 5,796 0,018 Diferença estatisticamente significativa (p≤0,05) As diferenças significativas verificadas nas áreas relacionadas com o Treino e com a Competição apontam no sentido de os treinadores menos experientes sentirem particular necessidade de formação nas dimensões plurianuais da planificação, quer do treino quer da competição; a experiência profissional associou- se a uma menor percepção da necessidade de formação no âmbito da orientação e gestão do treino, a médio e a longo prazo. Participantes A recolha de dados foi feita de maneira presencial nos programas de formação de treinadores organizados pelas Associações de Futebol, com excepção do Porto, durante a época desportiva de 2008/2009, onde foi realizado no dia de um seminário de treinadores de futebol organizado por uma entidade privada. Na entrega do questionário aos treinadores houve uma breve explicação sobre o inquérito por parte do pesquisador, sendo aberto um período para 934 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 Necessidades de formação dos treinadores paramétrico One-Way ANOVA (Análise de Variância). Já nas comparações entre mais de dois níveis da variável de agrupamento, foram estimados os Post-Hocs com recurso ao teste de Tukey para se identificar entre que grupos se encontravam as diferenças significativas. Para efeitos da interpretação e análise dos dados, o nível de confiança assumido foi de 0.05 (p≤0.05). responder a possíveis dúvidas dos treinadores. Depois de se assegurar a confidencialidade e o anonimato, os questionários foram entregues aos treinadores, que participaram voluntariamente, e devolvidos ao pesquisador durante os intervalos dos cursos de formação ou seminário. O preenchimento dos questionários foi realizado num ambiente calmo e reservado de forma a optimizar a atenção dos inquiridos, num espaço apropriado para esse fim, e teve a duração de 20 minutos em média. Procedimentos Estatísticos Na análise das competências profissionais em função dos anos de experiência como treinador desportivo, verificou-se que o domínio do Treino, o domínio da Competição, o domínio da Gestão Desportiva e o domínio do Papel de Formador apresentaram itens com diferenças estatisticamente significativas (em 22% dos itens, em 16% dos itens, em 100% dos itens e em 50% dos itens, respectivamente). Apenas no domínio das Competências Pessoais não se verificaram diferenças estatisticamente significativas. 935 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira atividades que envolvam outros agentes desportivos, bem como na participação no âmbito da formação de treinadores principiantes. Os treinadores com menor experiência percepcionaram, também, maior necessidade de formação nas áreas de gestão da carreira dos atletas e nos processos de coordenação de atividades que envolvam outros agentes desportivos, bem como na participação no âmbito da formação de treinadores principiantes. Os treinadores com menor experiência percepcionaram, também, maior necessidade de formação nas áreas de gestão da carreira dos atletas e nos processos de coordenação de Tabela 2. Análise comparativa das necessidades de formação dos treinadores em relação as competências relacionadas com o Treino, Competição, Gestão Desportiva e Papel de Formador em função dos níveis de formação federativa. Tabela 2. Análise comparativa das necessidades de formação dos treinadores em relação as competências relacionadas com o Treino, Competição, Gestão Desportiva e Papel de Formador em função dos níveis de formação federativa. Procedimentos Estatísticos Diferença estatisticamente significativa (p≤0,05) A = Treinadores sem formação federativa; B = Treinadores com formação federativa de nível I; C = Treinadores com formação federativa nível II e III MÉDIAS Competências relacionadas com o Treino A (n=16) B (n=49) C (n=16) F Sig Planejar a sessão de treino 3,81 3,28 2,56 1,858 A-C =0,005 Realizar o planejamento anual de treino 3,56 3,32 2,62 2,929 A-C =0,051 Realizar um plano de preparação plurianual 4,00 3,12 2,31 6,689 A-C = 0,041 B-C = 0,021 Organizar e conduzir a sessão de treino 3,62 3,10 2,56 1,501 A-B = 0,050 A-C = 0,001 Organizar e implementar o plano anual de treino 3,75 3,36 2,62 3,395 A-C = 0,022 Organizar e implementar um planejamento plurianual 3,68 2,97 2,37 6,565 A-C = 0,013 Avaliar e modificar o plano da sessão de treino 3,68 2,97 2,37 1,968 A-C = 0,011 Avaliar e modificar o plano anual de treino 3,68 3,18 2,68 2,590 A-C = 0,032 Competências relacionadas com a Competição Preparar um equipe/atleta para participar de forma digna em uma competição 3,68 2,95 2,25 1,864 A-C = 0,004 Preparar a competição de uma época estabelecendo objetivos adequados a equipe/atletas 3,62 3,12 2,43 3,003 A-C = 0,010 Dirigir uma equipe/atleta durante uma competição, nomeadamente nos aspectos disciplinares 3,50 2,95 2,37 1,107 A-C = 0,017 Competências relacionadas com o Papel de Formador Dar um contributo relevante para a formação desportiva dos atletas 3,81 3,10 2,62 0,758 A-C = 0,017 Orientar a formação de treinadores principiantes 3,87 3,08 2,56 6,497 A-B = 0,016 A-C = 0,001 Orientar a formação de pais de atletas e acompanhantes 3,68 3,06 2,68 3,374 A-B = 0,048 A-C = 0,007 Coordenar a formação de técnicos da modalidade 3,75 3,16 2,68 5,796 A-C = 0,012 Necessidades de formação e formação federativa treinadores de nível II e III existiram diferenças significativas em 88% dos itens e entre os treinadores sem formação e os treinadores de nível I em 11% dos itens. O mesmo percentual foi encontrado entre os treinadores de nível I e os de nível II e III nível, onde as diferenças encontradas se situaram nas competências de planificação e condução e avaliação do treino. Na análise dos resultados emergiu o maior reconhecimento das necessidades de formação por parte dos treinadores sem formação federativa quando comparados com treinadores com formação federativa de níveis II e III. Verificaram-se diferenças significativas em menor escala quando comparados os treinadores sem formação federativa e os treinadores de nível I em relação aos de nível II e III. A tabela 2 apresenta os itens onde se verificaram diferenças significativas. No domínio da Competição verificou-se, também, uma maior necessidade de formação por parte dos treinadores sem formação federativa comparativamente aos treinadores com formação de níveis II e III (50% dos itens), especificamente para estabelecer objetivos, preparar e gerir a equipe em situação de competição. No domínio do Treino, houve diferença estatisticamente significativa na comparação entre os três grupos de treinadores desportivos. Entre os treinadores sem formação e os 936 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 Necessidades de formação dos treinadores universitário, Baccalauearte in Sport Intervention (BIS), reconhece que a capacidade de fazer um planejamento a médio e longo prazo assume um papel importante para o desenvolvimento das capacidades dos atletas e de uma equipe, incluindo o desenvolvimento dessas competências no seu conteúdo programático da formação inicial de treinadores (DEMERS; WOODBURN; SAVARD, 2006). Os autores afirmam, consequentemente, que o desenvolvimento de planos anuais e plurianuais é um elemento decisivo de um programa de formação de treinadores. Já no domínio do Papel de Formador, foi possível identificar um maior reconhecimento de necessidades de formação por parte dos treinadores sem formação federativa quando comparados aos treinadores com formação federativa de nível I (50% dos itens) e aos treinadores de nível II e III (100% dos itens). È notória que nesta dimensão a percepção de necessidades de formação decresce à medida que a formação é mais avançada. Por fim, no domínio das Competências Pessoais e no domínio da Gestão Desportiva não se identificaram diferenças estatisticamente significativas em quaisquer dos itens. Necessidades de formação e formação federativa Neste estudo, como no estudo de Santos et al., (in press), os treinadores enfatizaram a competência de planeamento a longo prazo, considerando planos prospectivos e estratégicos como uma parte fundamental da sua competência profissional. A pesquisa tem, no entanto, dedicado pouca atenção aos aspectos do planeamento, nomeadamente, no que se refere ao planeamento a longo prazo. Côté e Sedgwick (2003) afirmam que os treinadores planeiam de forma proactiva, de acordo com preocupações de curto e de longo prazo, procurando preparar os seus atletas para situações não esperadas que podem ocorrer nos ambientes de treino e competição. G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira No que se refere ao papel de formador, os mais experientes reconheceram menores necessidades de formação para intervir no âmbito da formação de treinadores principiantes e de outros treinadores da modalidade. Nesse âmbito, a literatura especializada (ABRAHAM; COLLINS; MARTINDALE, 2006; CÔTÉ, 2006; DEMERS, WOODBURN; SAVARD 2006; DUFFY, 2008; GILBERT; TRUDEL, 2001; JONES; ARMOUR; POTRAC, 2004) reconhece a experiência como um importante fator de desenvolvimento profissional, tornando-se uma mais valia para o apoio ao desenvolvimento de acções associadas à formação de novas gerações de treinadores. Reforçando essa perspectiva, diversos autores reconhecem os anos de experiência profissional como um requisito indispensável para um treinador atingir o nível de expert (CUSHION; ARMOUR; JONES, 2003; JONES; ARMOUR; POTRAC, 2004) e sublinham a necessidade destes participarem activamente na formação de treinadores. No que se refere à influência da formação federativa na percepção das necessidades, os treinadores com maior nível de formação federativa (nível II e III) reconhecem menores necessidades em relação aos treinadores sem formação federativa, nomeadamente, nas áreas relacionadas com o Treino, a Competição e o Papel de Formador. No domínio do Treino verificou-se o maior número de itens (88%) com diferenças significativas entre os treinadores sem formação comparativamente aos treinadores com formação de níveis II e III. Estas diferenças de percepção no que se refere às necessidades de formação incidem sobre competências profissionais decisivas desde os mais elementares níveis de atividade como treinador. Neste contexto, Demers, Woodburn e Savard (2006) enfatizaram as competências de comunicação e de implementação das tarefas do treino, bem como o suporte e a gestão dos atletas em treino e competição, como competências profissionais de base. Por outro lado, os treinadores de elite entrevistados por Abraham, Collins, Martindale (2006) identificaram as competências de fornecimento de feedback e a aquisição de habilidades desportivas como os elementos - chave da pedagogia no treino. O estudo permite, assim, concluir que a experiência profissional interfere no reconhecimento das necessidades de formação dos treinadores desportivos. Discussão O presente estudo teve como objectivo identificar se o reconhecimento das necessidades de formação dos treinadores desportivos nas áreas relacionadas com o Treino, Competição, Papel de Formador, Gestão Desportiva e Competências Pessoais se diferencia em função da formação federativa e da experiência como treinador desportivo. Quando analisamos o reconhecimento das necessidades de formação de acordo com os anos de experiência do treinador, foi possível confirmar uma menor percepção de necessidades de formação por parte dos treinadores mais experientes num grupo muito significativo de aspectos, confirmando indicações de de Armour e Yelling (2004); Jones, Armour e Potrac (2003) e de Lemyre, Trudel e Durand-Bush (2007). De facto, o desenvolvimento de planos estratégicos facilita a clarificação dos micro e macro-planos e assegura a inscrição das acções particulares num mesmo plano geral, sendo apontada como uma peça muito importante da formação de treinadores (SANTOS et al., in press). A experiência profissional apresenta uma relação significativa com a competência de gestão de atletas e da comissão ou equipa técnica, tendo-se verificado resultados com diferenças estatisticamente significativas em 100% dos itens no domínio da Gestão Desportiva. A literatura especializada (ABRAHAM; COLLINS; MARTINDALE, 2006; DEMERS, WOODBURN; SAVARD 2006; DUFFY, 2008; VALLÉE; BLOOM, 2005) identifica como uma das principais funções do treinador a gestão de recursos humanos, nomeadamente, de atletas e da comissão técnica, enfatizando a capacidade de gestão como um elemento preponderante para um treinador atingir altos níveis de performance, reconhecendo a competência dos treinadores experts em gerir expectativas, conflitos, emoções e otimizar relacionamentos entre comissão técnica e atletas. As diferenças significativas verificadas nas áreas relacionadas com o Treino e com a Competição permitem-nos destacar que os treinadores com maior vivência profissional reconhecem menores necessidades de formação únicamente para o planejamento, organização e implementação de planos plurianuais de treinos e competições. A experiência profissional adquirida permite aos treinadores mais experientes percepcionarem um maior domínio de competências no âmbito da orientação e gestão do treino a médio e longo prazo. Essa capacidade de planejar e executar um processo de desenvolvimento a longo prazo parece apresentar um diferencial maior em função da experiência dos treinadores (CÔTÉ; SEDGWICK, 2003). Concordando com essa linha de pensamento, um programa canadense de formação de treinadores, conduzido no ambiente 937 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 Conclusões Este estudo considera as necessidades de formação dos treinadores na perspectiva dos próprios indivíduos, entendendo esta abordagem como fundamental na gestão da formação de treinadores. Os treinadores sem formação federativa reconheceram maiores necessidades de formação nos domínios do Treino, da Competição e do Papel de Formador comparativamente aos treinadores com formação de nível II e III; nomeadamente, no que se refere a questões de planificação anual e plurianual do treino, a competências relacionadas com a planificação e gestão das competições e a competências de orientação da formação de outros treinadores. Tendo em conta a experiência profissional, podemos concluir que esta exerce, também, um efeito diferenciador na percepção das necessidades de formação, principalmente no domínio da Gestão Desportiva, do Papel de Formador e no domínio do Treino, sendo que os treinadores menos experientes reconheceram menores necessidades de formação.. No entanto, se por um lado, emergiu dos resultados do presente estudo, uma diferença substancial de percepções de necessidades de formação entre os treinadores sem formação e os treinadores com formação federativa de nível II e III (57% dos itens), o mesmo não aconteceu no grupo de treinadores de nível I (11,5% dos itens) relativamente aos treinadores sem formação. Esse resultado indicia a necessidade de refletir sobre a qualidade dos cursos de formação federativa de nível I que parecem não afectar significativamente as percepções das necessidades de formação quando comparados com os treinadores sem formação. O facto de apresentarem programas curriculares muito elementares pode determinar a inviabilização de diferenças substanciais nas competências adquiridas pelos treinadores após cursar o programa, refletindo perfis de necessidades de formação muito idênticos aos dos treinadores sem formação. Os resultados revelam que a experiência profissional e a formação federativa desempenham um papel crucial na percepção das necessidades de formação dos treinadores, mostrando a necessidade de considerar estas variáveis aquando da planificação da formação de treinadores desportivos, na medida em que fornecem um contributo substancial, não só, para a valorização da experiência profissional e dos modos da sua construção como, também, para uma avaliação mais contextualizada e específica da formação do treinador de desporto no âmbito federativo. Os programas de formação representam o primeiro contato dos futuros treinadores com os conhecimentos e competências necessários para o exercício da função e, dessa forma, possuem elevado grau de importância pelo que a sua otimização é decisiva. outros estudos (CUSHION; ARMOUR; JONES, 2003; VARGAS-TONSING, 2007). BUSH, 2007) sugerem que os treinadores com reconhecida competência devem assumir o papel de formadores dos treinadores principiantes, sendo esta actividade vista como uma ferramenta para o desenvolvimento pessoal e profissional tanto dos formadores quanto dos formandos. Dessa forma, é natural que os treinadores que possuam maiores níveis de formação se percepcionem como mais preparados para actuar no papel de formadores. G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira Não estando em causa a pertinência que a experiência profissional assume na aquisição de um corpo de conhecimentos e competências profissionais, bem como no processo de crescimento e formação dos treinadores, é, contudo, importante que os treinadores estejam cientes de que apenas os anos de experiência não garantem um elevado nível de competência pois a construção do conhecimento através da experiência profissional está dependente de processos complexos associados à qualidade dessa experiência (GILBERT; TRUDEL, 1999). Aprender pela experiência é, indubitavelmente, um dos meios mais referidos de aprendizagem (FLEURENCE; COTTEAUX, 1999; JONES; ARMOUR; POTRAC, 2003; WRIGHT; TRUDEL; CULVER, 2007), bem como uma fonte de conhecimento (GILBERT; TRUDEL, 2001; JONES; ARMOUR; POTRAC, 2002; LEMYRE; TRUDEL, 2004; WRIGHT; TRUDEL; CULVER, 2007). Na realidade, os treinadores com mais experiência tendem a descrever-se como mais competentes (LEMYRE; TRUDEL, 2004), sendo apontada pelos próprios como o marcador principal para o desenvolvimento profissional (JONES; ARMOUR; POTRAC, 2004). Quando analisamos as competências relacionadas com a Competição, é possível verificar, novamente, diferenças significativas entre os treinadores sem formação e os treinadores de níveis II e III (50% dos itens), mais especificamente para estabelecer objectivos, planear e dirigir a equipe. Vallée e Bloom (2005) reconhecem que a aquisição dessas competências proporcionam aos treinadores a capacidade de promover maior desenvolvimento dos atletas, assumindo assim, um relevante papel no exercício da função de treinador. Dentro das competências relacionadas com o Papel de Formador, os treinadores sem formação federativa voltaram a admitir maior necessidade de formação comparativamente aos treinadores de níveis II e III (100% dos itens), nomeadamente, no que se refere à formação e educação dos atletas e seus acompanhantes, bem como de outros treinadores. Nessa linha de pensamento, valorizando o papel de formador do treinador, diferentes investigadores (ABRAHAM; COLLINS; MARTINDALE, 2006; DEMERS; WOODBURN; SAVARD, 2006; LEMYRE; TRUDEL; DURAND- 938 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 Necessidades de formação dos treinadores Science & Coaching. Brentwood, v. 1, n. 3, p. 217-222, 2006. In: PETRY, K.; FROBERG, K.; MADELLA, A.; TOKARSKY, W. (Org.). Higher Education in Sport in Europe: from labour market demand to training supply. Maidenhead: Meyer & Meyer Sport, 2008. p. 80-108. Science & Coaching. Brentwood, v. 1, n. 3, p. 217-222, 2006. determinando as funções ou grupo de funções a exercer, as condições e equipamentos necessários, assim como os saberes e competências requeridas. É também relevante a análise das necessidades de formação no nível organizacional procurando analisar quando e onde a formação pode e deve ser aplicada. Tal implica, entre outros aspectos, a análise dos recursos disponíveis, das condições específicas da organização, do sistema técnico, das relações de trabalho, entre outros. CÔTÉ, J.; SALMELA, J.; RUSSELL, S. The knowledge of high performance gymnastics coaches: competition and training considerations. The Sport Psychologist. Champaign, v. 9, n.1, p. 76-95, 1995. CÔTÉ, J.; SALMELA, J.; RUSSELL, S. The knowledge of high performance gymnastics coaches: competition and training considerations. The Sport Psychologist Champaign v 9 n 1 p CÔTÉ, J.; SEDGWICK, W. Effective behaviors of expert rowing coaches: a qualitative investigation of Canadian athletes and coaches. International Sports Journal. West Haven, v. 7, n. 1, p. 62-77, 2003. A investigação não deve omitir, também, a questão da construção social das necessidades de formação. Seguindo essa linha de pensamento, futuros estudos devem procurar identificar os elementos chaves de uma visão não meramente diagnóstica mas também prospectiva da atividade profissional e das necessidades de formação, ao descentrar-se da análise das condições do presente e equacionando os cenários futuros da atividade como treinador. Esta análise deve estender-se à visão dos treinadores sobre os conteúdos, itinerários e modalidades de formação, considerando, nomeadamente, as características pessoais, as fontes de conhecimentos e as áreas de conhecimentos e competências profissionais que os treinadores reconhecem como responsáveis pelo seu sucesso profissional. CUSHION, C.; ARMOUR, K.; JONES, R. Coach education and continuing professional development: experience and learning to coach. Quest. Champaign, v. 55, n. 3, p. 215-230, 2003. CUSHION, C.; JONES, R. Power, discourse, and symbolic violence in professional youth soccer: the case of albion football club. Sociology of Sport Journal. Champaign, v. 23, n. 2, p. 142-161, 2006. DEMERS, G.; WOODBURN, A.; SAVARD, C. The Development of an Undergraduate Competency- Based Coach Education Program. The Sport Psychologist. Champaign, v. 20, n. 2, p. 162-173, 2006. DUFFY, P. Implementation of the Bologna Process and Model Curriculum Development in Coaching. In: PETRY, K.; FROBERG, K.; MADELLA, A.; TOKARSKY, W. (Org.). Higher Education in Sport in Europe: from labour market demand to training supply. Maidenhead: Meyer & Meyer Sport, 2008. p. 80-108. DUFFY, P. Implementation of the Bologna Process and Model Curriculum Development in Coaching. Conclusões No caso da formação de treinadores é necessário ultrapassar a elaboração casuística de planos de formação, não precedida de diagnósticos sobre as necessidades de competências. Na elaboração desses diagnósticos, as percepções dos intervenientes são fundamentais e podem dirigir os planos de formação, garantindo que os programas de formação estejam adequados as reais necessidades dos treinadores. Neste caso, as indicações dos treinadores apontam no sentido de conferir maior atenção aos conhecimentos processuais e competências de cunho pedagógico-didático e de gestão, sem desvalorizar as questões técnicas específicas da modalidade confirmando indicações de diversos Mais importa que esta linha de investigação seja complementada pela realização de investigação mais aprofundada nesta área. Na realidade, a forma mais adequada de pensar a análise de necessidades de formação é inseri-la no âmbito de uma estratégia de melhoria da qualidade, numa organização ou comunidade. Ela não visa, apenas, detectar problemas pontuais, mas fornecer informações relevantes que permitam agir de uma forma estratégica, adaptando-se aos contextos muito diversos de exercício da actividade do treinador. Deste modo, importa que a investigação aprofunde a análise funcional da actividade de treinador, de 939 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 G, Cunha; I. Mesquita; A. Rosado; T. Sousa; P. Pereira Referências ABRAHAM, A.; COLLINS, D. Examining and extending research in coach development. Quest. Champaign, v. 50, p. 59-79, 1998. FLEURENCE, P.; COTTEAUX, V. Construction de l’expertise chez les entraîneurs sportifs d’athlètes de hau-niveau français. Avante. Ottawa, v. 5, p. 54-68, 1999. ABRAHAM, A.; COLLINS, D.; MARTINDALE, R. The coaching schematic: validation through expert coach consensus. Journal of Sport Sciences. Londres, v. 24, n. 6, p. 549-564, 2006. ABRAHAM, A.; COLLINS, D.; MARTINDALE, R. The coaching schematic: validation through expert coach consensus. Journal of Sport Sciences. Londres, v. 24, n. 6, p. 549-564, 2006. GILBERT, W.; TRUDEL, P. Framing the construction of coaching knowledge in experiential learning theory. SOSOL - Sociology of Sport Online. Otago, v. 2, n.1, 1999. ARMOUR, K. M.; YELLING, M. R. Continuing professional development for experienced physical education teachers: towards effective provision. Sport, Education and Society. Abingdon, v. 9, n. 1, p. 95-114, 2004. GILBERT, W.; TRUDEL, P. Learning to coach through experience: reflection in model youth sport coaches. Journal of Teaching in Physical Education. Champaign, v. 21, n. 1, p.16-34, 2001. BURDEN, P. R. Teacher development. In: HOUSTON W. R. (Org.). Handbook of research on teacher education. New York: Macmillan, 1990. p. 311 - 327. IRWIN, G.; HANTON, S.; KERWIN, D. Reflective practice and the origins of elite coaching knowledge. Reflective Practice. Abingdon, n. 5, p. 425-442, 2004. CAMPBELL, S. Coaching education around the world. Sport Science Review. Champaign, v. 2, n. 2, p. 62-74, 1993. CÔTÉ, J. The development of coaching knowledge. International Journal of Sport CAMPBELL, S. Coaching education around the world. Sport Science Review. Champaign, v. 2, n. 2, p. 62-74, 1993. Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 JONES, R.; ARMOUR, K.; POTRAC, P. Constructing expert knowledge: a case study of a top-level professional soccer coach. Sport JONES, R.; ARMOUR, K.; POTRAC, P. Constructing expert knowledge: a case study of a top-level professional soccer coach. Sport CÔTÉ, J. The development of coaching knowledge. International Journal of Sport 940 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010 Necessidades de formação dos treinadores Education and Society. Abingdon, v. 8, n. 2, p. 213-229, 2003. Education and Society. Abingdon, v. 8, n. 2, p. 213-229, 2003. WESTERA, W. Competences in education: a confusion of tongues. Journal of Curriculum Studies. Abingdon, v. 33, n. 1, p. 75-88, 2001. Education and Society. Abingdon, v. 8, n. 2, p. 213-229, 2003. JONES, R.; ARMOUR, K.; POTRAC, P. Referências Sports Coaching Cultures: from practice to theory. London: Routledge, 2004. WIERSMA, L.; SHERMAN, C. Volunteer Youth Sport Coaches' Perspectives of Coaching Education: certification and parental codes of conduct. Research Quarterly for Exercise & Sport. Reston, v. 76, n. 3, p. 324-338, 2005. JONES, R.; ARMOUR, K.; POTRAC, P. Understanding the coaching process: A framework for social analysis. Quest. Champaign, v. 54, n. 1, p. 34-48, 2002. WRIGHT, T.; TRUDEL, P.; CULVER, D. Learning how to coach: the different learning situations reported by youth ice hockey coaches. Physical Education & Sport Pedagogy. Abingdon, v. 12, n. 2, p.127-144, 2007. WRIGHT, T.; TRUDEL, P.; CULVER, D. Learning how to coach: the different learning situations reported by youth ice hockey coaches. Physical Education & Sport Pedagogy. Abingdon, v. 12, n. 2, p.127-144, 2007. KIRSCHNER, P. et al. The design of a study environment for acquiring academic and professional competence. Studies in Higher Education. Abingdon, v. 22, n. 2, p. 151-171, 1997. LEMYRE, F.; TRUDEL, P. Le parcours d’apprentissage au role d’entraineur bénévole. Avante. Ottawa, v.10, n. 3, p. 40-55, 2004. Agradecimentos Este artigo é parte de um projecto de investigação financiado pela Fundação para a Ciência e Tecnologia (FCT), Portugal, através do Programa Operacional Temático Factores de Competitividade (COMPETE) do Quadro Comunitário de apoio III, e comparticipado pelo Fundo da Comunidade Europeia (FEDER) com a seguinte referência: PTDC/DES/67190/2006 y FCOMP-01-0124-007177-FEDER. Este artigo é parte de um projecto de investigação financiado pela Fundação para a Ciência e Tecnologia (FCT), Portugal, através do Programa Operacional Temático Factores de Competitividade (COMPETE) do Quadro Comunitário de apoio III, e comparticipado pelo Fundo da Comunidade Europeia (FEDER) com a seguinte referência: PTDC/DES/67190/2006 y FCOMP-01-0124-007177-FEDER. LEMYRE, F.; TRUDEL, P.; DURAND-BUSH, N. How youth sport coaches learn to coach. The Sport Psychologist. Champaign. v. 21, n. 2, p. 191-209, 2007. McCALLISTER, S.G.; BLINDE, E.M.; WEISS, W.M. Teaching values and implementing philosophies: dilemmas of the youth sport coach. Physical Educator. Indianápolis, v. 57, n.1, p. 35- 45, 2000. POTRAC, P.; JONES, R. Power, conflict and co- operation: towards a micro-politics of coaching. Quest. Champaign, v. 61, p. 223-236, 2009. Endereço: Professora Doutora Isabel Mesquita R. Dr. Plácido Costa, nº 91, 4200. 450 Porto, Portugal. (351)225074776 e-mail: imesquita@fade.up.pt Endereço: Professora Doutora Isabel Mesquita R. Dr. Plácido Costa, nº 91, 4200. 450 Porto, Portugal. SALMELA, J. Great Job Coach!: getting the edge from proven winners. Ottawa: Potentiun, 1996. (351)225074776 e-mail: imesquita@fade.up.pt (351)225074776 e-mail: imesquita@fade.up.pt SANTOS, S. et al. Coaches’ perceptions of competence and acknowledgement of training needs related to professional competences. SANTOS, S. et al. Coaches’ perceptions of competence and acknowledgement of training needs related to professional competences. Journal of Sport Science and Medicine. Uludag, In press. Journal of Sport Science and Medicine. Uludag, In press. Recebido em: 11 de dezembro de 2009. Aceito em: 14 de junho de 2010. STEPHENSON, J.; WEILL, S. Quality in learning: a capability approach in higher education. London: Kogan Page, 1992. VALLÉE, C. N.; BLOOM, G. A. Building a successful university sport program: key and common elements of expert coaches. Journal of Applied Sport Psychology. Abingdon, v. 17, n. 3, p. 179-196, 2005. Motriz. Revista de Educação Física. UNESP, Rio Claro, SP, Brasil - eISSN: 1980-6574 - está licenciada sob Licença Creative Commons p cre ati ve co VARGAS-TONSING, T. Coaches' preferences for continuing coaching education. International Journal of Sports Science & Coaching. Brentwood, v. 2, n. 1, p. 25-35, 2007. 941 Motriz, Rio Claro, v.16, n.4, p.931-941, out./dez. 2010
https://openalex.org/W4249925518	https://www.researchsquare.com/article/rs-1147/v2.pdf?c=1631840466000	English	null	Linkage to Hepatitis C Care After Incarceration in Jail: A Prospective, Single Arm Clinical Trial	Research Square (Research Square)	2,019	cc-by	7,318	Linkage to Hepatitis C Care After Incarceration in Jail: A Prospective, Single Arm Clinical Trial Matthew Jiro Akiyama (  makiyama@montefiore.org ) Montefiore Medical Center/Albert Einstein College of Medicine https://orcid.org/0000-0003-4290-3083 Devin Columbus New York University School of Medicine Ross MacDonald New York City Health + Hospitals Alison O Jordan New York City Health + Hospitals Jessie Schwartz New York City Department of Health and Mental Hygiene Alain H Litwin Greenville Health System, University of South Caroline School of Medicine Benjamin Eckhardt New York University School of Medicine Ellie Carmody New York University School of Medicine Research article Keywords: HCV, jail, prison, linkage to care, PWID Posted Date: July 19th, 2019 DOI: https://doi.org/10.21203/rs.2.10021/v2 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License f f Linkage to Hepatitis C Care After Incarceration in Jail: A Prospective, Single Arm Clinical Trial Matthew Jiro Akiyama (  makiyama@montefiore.org ) Montefiore Medical Center/Albert Einstein College of Medicine https://orcid.org/0000-0003-4290-308 Devin Columbus New York University School of Medicine Ross MacDonald New York City Health + Hospitals Alison O Jordan New York City Health + Hospitals Jessie Schwartz New York City Department of Health and Mental Hygiene Alain H Litwin Greenville Health System, University of South Caroline School of Medicine Benjamin Eckhardt New York University School of Medicine Ellie Carmody New York University School of Medicine Research article Keywords: HCV, jail, prison, linkage to care, PWID Posted Date: July 19th, 2019 DOI: https://doi.org/10.21203/rs.2.10021/v2 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published on August 8th, 2019. See the published version at https://doi.org/10.1186/s12879-019-4344-1. Research article License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published on August 8th, 2019. See the published version at https://doi.org/10.1186/s12879-019-4344-1. Page 1/19 Page 1/19 Abstract Background: Hepatitis C virus (HCV) is a major public health problem in correctional settings. HCV treatment is often not possible in U.S. jails due to short lengths of stay. Linkage to care is crucial in these settings, but competing priorities complicate community healthcare engagement and retention after incarceration. Methods: We conducted a single arm clinical trial of a combined transitional care coordination (TCC) and patient navigation intervention and assessed the linkage rate and factors associated with linkage to HCV care after incarceration. Results: During the intervention, 84 participants returned to the community after their index incarceration. Most participants were male and Hispanic, with a history of mental illness and a mean age of 45 years. Of those who returned to the community, 26 (31%) linked to HCV care within a median of 20.5 days; 17 (20%) initiated HCV treatment, 15 (18%) completed treatment, 9 (11%) had a follow-up lab drawn to confirm sustained virologic response (SVR), and 7 (8%) had a documented SVR. Among those with follow-up labs the known SVR rate was (7/9) 78%. Expressing a preference to be linked to the participant’s existing health system, being on methadone prior to incarceration, and feeling that family or a loved one were concerned about the participant’s wellbeing were associated with linkage to HCV care. Reporting drinking alcohol to intoxication prior to incarceration was negatively associated with linkage to HCV care. Conclusion: We demonstrate that an integrated strategy with combined TCC and patient navigation may be effective in achieving timely linkage to HCV care. Additional multicomponent interventions aimed at treatment of substance use disorders and increasing social support could lead to further improvement. Introduction Lack of access to health insurance, lack of social support, exacerbation of mental health conditions, and unstable housing have all been documented as complicating linkage to care among PLWHIV.11–14 A variety of strategies have been evaluated for their impact on linkage to care among PLWHIV leaving jail and prison. Case management, which aims to coordinate medical, mental health, social, and other services has been the most extensively studied. Evidence suggests that case management plays a key role in facilitating linkage to HIV care. In one randomized controlled trial (RCT), case management was shown to be more effective than referral alone.15 Case management has also been associated with improved adherence to antiretroviral therapy (ART), HIV viral load suppression, retention in care, fewer emergency department visits, and less housing instability and food insecurity among PLWHIV.16,17 Transitional care coordination (TCC) is a form of case management that includes an intake assessment, discharge and transitional care plans, and appointment scheduling. TCC is now a nationally recognized evidence-informed intervention that plays a key role in facilitating linkage to HIV care after incarceration.18 Patient navigation is another strategy that may provide additional benefit. While patient navigation does not have a standard definition, programs incorporating this strategy tend to focus on advocacy, health education, and social support.19 In a recently published RCT study, patient navigation coupled with case management resulted in greater linkage to and retention in HIV care after incarceration in jail than standard case management alone.20 A variety of strategies have been evaluated for their impact on linkage to care among PLWHIV leaving jail and prison. Case management, which aims to coordinate medical, mental health, social, and other services has been the most extensively studied. Evidence suggests that case management plays a key role in facilitating linkage to HIV care. In one randomized controlled trial (RCT), case management was shown to be more effective than referral alone.15 Case management has also been associated with improved adherence to antiretroviral therapy (ART), HIV viral load suppression, retention in care, fewer emergency department visits, and less housing instability and food insecurity among PLWHIV.16,17 Strategies for linkage to HIV care may be applicable to other clinical and therapeutic areas including HCV. To date, models of care and their impact on linkage to HCV care after incarceration are poorly understood. There are also few data on predictors of linkage to HCV care after incarceration. Introduction As many as half a million people in correctional settings in the United States may be hepatitis C virus (HCV) antibody-positive.1 Due in part to the criminalization of individuals with substance use disorders known to increase HCV risk, HCV prevalence in correctional settings is in the order of 10–20 times that of the general population.2 It is estimated that one in three people living with HCV (PLWHCV) pass through the criminal justice system each year.3 Failure to effectively intervene for PLWHCV in HCV care within the criminal justice system may place PLWHCV at ongoing risk for progression of HCV-related liver disease (cirrhosis, end-stage liver disease, and hepatocellular carcinoma) and may facilitate ongoing transmission of HCV in the neighborhoods of greatest socio-economic disparities to which PLWHCV return after incarceration.4–6 The criminal justice system, being a high prevalence setting, is an important locus for HCV treatment with the goal of HCV elimination.7 For optimal operationalization of a response to HCV in this setting it is important to differentiate jails from prisons. People incarcerated in local jails are largely detained pre-trial and/or have short sentences of generally up to one year or less. People incarcerated in prisons are convicted of a crime and typically sentenced to incarceration for more than one year. While longer sentences more readily permit HCV treatment in prisons, treatment is complicated in jails where the median length of stay is 15 days.8 HCV regimens have been reduced to twelve weeks or less for the Page 2/19 Page 2/19 majority of individuals with direct acting antiviral (DAA) therapy making treatment feasible for some jail detainees.9 However, there are twelve million admissions to US jails per year, which is nineteen times the number in state and federal prisons.10 With the high volume of likely untreated detainees cycling through, U.S. jails are a critical site for linkage to HCV care after incarceration. Linkage to community care from correctional settings is complicated by competing priorities after incarceration. During incarceration, individuals experience social, structural, and medical disruptions that complicate the reentry period. To date, the majority of work on linkage to care from correctional settings has focused on people living with HIV (PLWHIV). Introduction We conducted a prospective, single arm clinical trial of a combined TCC and patient navigation intervention among HCV mono-infected individuals leaving the New York City jail system. The aim of this study was to measure linkage to and retention in care associated with this combined TCC and patient navigation model on the rate of linkage to and retention in HCV care. We also aimed to identify factors associated with linkage to HCV care and evaluate feasibility of operationalizing this care coordination strategy among justice-involved PLWHCV. Methods Study setting and participants Study setting and participants Study setting and participants Page 3/19 Page 3/19 The NYC jail system is the second largest in the U.S. with nearly fifty thousand admissions per year and an average daily census of over eight thousand (as of 2018).21 Health care service delivery in jails is the responsibility of Correctional Health Services, a division of NYC Health + Hospitals, the nation’s largest public health care system.21 We identified potential participants using clinical data reports identifying individuals with detectable HCV viral loads. Recruitment was conducted in three NYC jails and one hospital correctional health ward (Bellevue Hospital). We included consenting patients who were ≥18 years old at time of recruitment, were diagnosed with chronic HCV, and had a projected length of stay sufficient to anticipate community reentry within the study follow-up period. Exclusion criteria included HIV (due to existing TCC services on a population basis) and inability to obtain informed, signed consent. We conducted study enrollment and baseline interviews in private cubicles of jail health clinics. Participant tracking We tracked dates of incarceration and community return by querying public inmate lookup databases22,23 and communicating with jail health staff throughout the course of the study. Index incarceration was defined as the incarceration during which study enrollment and baseline interview took place. Reincarceration was defined as any incarceration of duration of greater than 24 hours occurring after the index incarceration. We collected detailed contact information for participants and participants’ next of kin to facilitate patient navigation upon reentry into the community. We also collected information on places and programs the participant may have frequented in the community (e.g., drug treatment centers; homeless shelters). HCV transitional care coordination & patient navigation HCV transitional care coordination & patient navigation We partnered with New York City Correctional Health Services teams that employ the well-established TCC intervention to facilitate linkages to care for PLWHIV after incarceration in the NYC jail system.18 The TCC intervention includes: intake and assessment services, discharge planning, health insurance assistance, court liaison to facilitate medical alternatives to incarceration, facilitating continuity of medication and patient navigation and coordination with service providers to facilitate linkages to care after incarceration.12 We developed a care coordination intervention employing a similar model for HCV mono-infected PLHCV with HCV education and community patient navigation performed by a master’s level study coordinator. Patient navigation included reminder calls, appointment rescheduling if needed, appointment accompaniment if requested, and free public transit passes for each successfully completed HCV appointment. Patient navigation did not include home visits or direct outreach due to resource limitations.We gave all participants contact information for key study team members and encouraged communication before and after return to the community. Study personnel offered participants assistance with scheduling HCV linkage appointments at their existing health system, if applicable, or one of two study-affiliated community clinical partners if they preferred. Appointment dates were targeted within two to four weeks of community return. Whether or not Page 4/19 Page 4/19 a community appointment was scheduled during the jail stay, we attempted to engage all participants within one week after incarceration using information from the enrollment interview. This included attempting to contact participants and/or next of kin via telephone, text, and/or email. Participants were followed for up to 180 days after the index incarceration during which time attempts were made at linkage to HCV care. a community appointment was scheduled during the jail stay, we attempted to engage all participants within one week after incarceration using information from the enrollment interview. This included attempting to contact participants and/or next of kin via telephone, text, and/or email. Participants were followed for up to 180 days after the index incarceration during which time attempts were made at linkage to HCV care. For participants contacted after incarceration, we made up to five attempts to schedule community HCV appointments within 180 days after the index incarceration before categorizing a participant as not linked. If the participant was reincarcerated prior to or after being linked to HCV care, we attempted to coordinate follow-up appointments after reincarceration. Primary and secondary outcomes Our primary outcome was linkage to HCV care, which was defined as a documented visit with an HCV treating provider within 180 days after the index incarceration. Once linked to care, clinicians assessed participants for HCV treatment eligibility according to standard of care. Secondary outcomes for treatment-eligible participants included treatment initiation, treatment completion, and documented sustained virologic response (SVR), or HCV cure.For those who linked to care we continued to assess for secondary outcomes beyond 180 days if applicable. HCV transitional care coordination & patient navigation For participants who linked to care, study personnel assisted with scheduling HCV clinical and laboratory follow-up visits. Data sources and covariates Following enrollment, we conducted participant interviews using a modified version of the Addiction Severity Index (ASI) to identify potential factors associated with linkage to HCV care. The ASI includes information on medical history, employment/financial support, drug/alcohol use (e.g., Have you used heroin in your lifetime; how many days did you use heroin in the 30 days prior to incarceration?), legal case status, family/social relationships, and psychiatric status. Individual participants’ answers were recorded by a study coordinator and entered into REDCap—a secure, web-based application designed to support data capture for research studies.24 Feasibility data were gathered from study staff logs detailing the number of attempted and successful contacts with participants/next of kins. Data on linkage to and retention in HCV care were obtained from participant medical records. HCV transitional care coordination & patient navigation The median duration of time between enrollment and community return after the index incarceration was 34.0 days (interquartile range, IQR = 71.8). Less than one half (43%) provided personal contact information, but almost all provided contact information for a next of kin (92%). On average, participants were engaged 1.2 (SD = 0.5) times by study staff during incarceration. In the 180 days after incarceration, an average of 3.8 (SD = 4.4) contact attempts were made resulting in 42 (50%) participants being successfully contacted, and an average of 2.7 (SD = 2.1) attempts to contact next of kin resulted in 53 (63%) successful contacts of participants’ next of kin. Statement of ethical standards Statement of ethical standards The study team obtained written informed consent from all participants enrolled in the study, and approval to conduct the study was obtained from New York University School of Medicine and New York City Department of Health and Mental Hygiene Institutional Review Boards (IRB). The New York University School of Medicine IRB approval process encompassed detailed review by the IRB prisoners’ advocate. Participants From May 2015 to April 2017 we identified and interviewed 185 people incarcerated in NYC jails for study participation; a total of 105 were enrolled and 5 were subsequently disenrolled as PLWHIV from the study group, leaving a cohort of 100 who were followed during and after their index incarceration. Of the 100 participants, 8 were transferred to state prison or another jurisdiction, 4 received HCV treatment in jail, 1 spontaneously cleared their HCV infection, 1 died prior to release, and 2 remained incarcerated at the completion of the study; leaving 84 who returned to the community after the index incarceration (Figure 1). Most participants were male (57%) and Hispanic (55%) with a mean age of 45 years (standard deviation, SD = 12). The majority (63%) had a reported history of mental illness. Over one quarter (26%) reported being homeless prior to incarceration. Most (81%) participants reported having health insurance in the community, but fewer than one quarter (24%) had an existing primary care provider and fewer than 10% had an existing HCV provider. Over two thirds (70%) reported a lifetime history of injection drug use and nearly two thirds (61%) reported injection drug use in the 30 days prior to incarceration. Most (89%) reported using heroin in their lifetime and over half reporting heroin use in the 30 days prior to incarceration. Nearly two thirds (60%) reported taking methadone prior to incarceration and nearly one third (29%) reported taking buprenorphine-naloxone (Table 1). Statistical analysis We conducted a bivariate analysis to examine associations with linkage to HCV care using Chi-square and Fisher’s exact tests for categorical variables. Significance was assessed at an alpha level of 0.05. Statistical analyses were conducted using SAS 9.2 (SAS Institute, Cary, NC). We also used EventFlow (University of Maryland, Human-Computer Interaction Lab; http://hcil. umd.edu/eventflow) to descriptively identify relevant temporal event sequences associated with the HCV corrections-community care cascade with an emphasis on linkage to HCV care and reincarceration. Page 5/19 Page 5/19 Associations with linkage to HCV care Expressing a preference to be linked to the participant’s existing health system, being on methadone prior to incarceration, and feeling that family or a loved one were concerned about the participant’s wellbeing were associated with linkage to HCV care. Reporting drinking alcohol to intoxication prior to incarceration was negatively associated with linkage to HCV care. Notably, homelessness was not significantly associated with linkage to HCV care and the proportion of those reincarcerated was similar in the group who were linked to care (8/26, 31%) and the group who were not linked to care (20/58, 36%, Chi-square p = 0.86) (Table 1). Linkage to & retention in HCV care Page 6/19 Of the participants who returned to the community after the index incarceration, 26 (31%) linked to HCV care within a median of 20.5 (IQR = 69.8) days. On average, 1.1 (SD = 1.1) linkage appointments were made among all participants who returned to the community. In terms of secondary outcomes, 17 (20%) participants initiated HCV treatment, 15 (18%) completed treatment, 9 (11%) had a follow-up lab drawn to confirm SVR,and 7 (8%) had a documented SVR (Figure 2). Therefore, the SVR rate among those with follow-up was (7/9) 78%. One participant who linked to care and initiated HCV treatment was reincarcerated mid-treatment. Our study coordinator received a phone call from a jail physician to confirm the participant was on treatment. The participant was continued on treatment and achieved SVR in jail. Reincarceration and loss to follow-up Overall, 28 (33%) participants were reincarcerated within a median of 88.0 (IQR = 85.8) days after their index incarceration. Reincarceration was an equally common initial event compared with linkage to HCV care with 23 (27%) participants being reincarcerated following community return. Five participants who were linked to HCV care were subsequently reincarcerated. Conversely, three participants who were reincarcerated subsequently linked to HCV care after incarceration. Nearly half (45%) of study participants were either not known to be linked to HCV care or reincarcerated in the 180 days after their index incarceration (Figure 3). Discussion To our knowledge, this is the first prospective trial to characterize the HCV care cascade in the transition from a correctional setting to the community. It is also the first cohort study to describe factors associated with linkage to HCV care after incarceration. These data provide preliminary evidence that a combined TCC and patient navigation strategy may be effective in achieving timely linkage to HCV care for a subset of PLWHCV after incarceration in jail. Linkage rates in our study are similar to those reported in other corrections-based public health programs and pilot studies from other jurisdictions. In a Rhode Island-based rapid HCV testing pilot study in which participants received a questionnaire, informational video, and a referral, 4/15 (27%) participants with confirmed viremia linked to HCV care.25 In a Massachusetts program in which participants with HCV received referrals to HCV care, 14/82 (17%) and 31/82 (38%) had laboratory evidence of linkage to HCV care within 6 and 12 months, respectively.26 In an HCV testing and linkage to care program in South and North Carolina, participants received HCV post-test counseling, a referral, appointment scheduling, and Page 7/19 Page 7/19 patient navigation (in North Carolina only). In South and North Carolina, respectively, 2/7 (29%) and 10/18 (56%) participants who were referred attended their first HCV appointment.27 All of these studies suggest there are substantial challenges associated with linkage to HCV care. Moreover, data from South Carolina, Massachusetts, and Rhode Island demonstrate these barriers are not overcome with education, referral, and appointment scheduling alone. While data from North Carolina suggest better linkage rates may be achievable with patient navigation, the small sample size limits the generalizability of the results and, still, only roughly half of participants linked to HCV care. Our intervention was modeled on the evidence-informed TCC intervention used in NYC jails to facilitate linkage to HIV care after incarceration. For PLWHIV, linkage rates associated with this program approach 75%.18 Yet, the linkage rate was substantially lower among monoinfected PLWHCV in our study. Several factors may account for the higher linkage rate among PLWHIV than PLWHCV. First, in our study we used telephone-based patient navigation with appointment accompaniment if requested as the community portion of the intervention. Discussion The TCC intervention for PLWHIV after incarceration in the NYC jails includes an extensive network of community resources, linkage agreements with community providers, dedicated community case managers funded through Ryan White Part A, and housing assistance.28 We were not able to provide such extensive services due to resource limitations and believe such services could have improved linkage rates.18 We hypothesize there are several other key differences leading to higher linkage rates among PLWHIV than PLWHCV. First, due to the earlier response of health care systems to the HIV epidemic at the federal, state, and local levels and relatively longstanding availability of effective ART, many PLWHIV are more informed about their HIV diagnoses and the effectiveness of ART. As an example, Loeliger et al. demonstrated that an HIV diagnosis of greater than one year predicted retention in HIV care among justice-involved PLWHIV.17 PLWHCV may be relatively unaware of the short- and long-term consequence of HCV infection and the effectiveness of DAA therapy. With increased collective HCV- and DAA-related knowledge, linkage rates may marginally improve over time. Additionally, relationships with medical providers prior to incarceration have been shown to result in higher linkage rates among PLWHIV.16,29,30 Our study mirrors this finding given a preference to be linked to one’s existing health system was associated with linkage to HCV care. Among PLWHIV who reported no prior history of HIV medical services, Molitor et al. identified a linkage rate of 29%.31 Very few of our study participants reported an existing relationship with an HCV provider prior to incarceration, and our linkage rate was similar. Taken together, these findings suggest increased familiarity with the healthcare system may facilitate linkage to care. Active substance use disorders also likely contribute to lower linkage rates among PLWHCV. Injection drug use is the number one risk factor for HCV in the U.S., and nearly two-thirds of our study participants reported injection drug use thirty days prior to incarceration. Substance use disorders are known to complicate linkage among PLWHIV.32 In our study, injection drug use was not associated with linkage to Page 8/19 Page 8/19 HCV care; however, taking methadone prior to incarceration was. Discussion Opioid agonist therapy is an evidence- based strategy leading to more stability from active drug use and higher linkage to care among PLWHIV.20 While we do not have follow-up data on continuation of methadone, we presume many of these participants continued on methadone as those who were on methadone prior to incarceration at the time of this study were generally maintained in the NYC jails unless they were expected to be transferred to prison based on available legal criteria. Given that the period after incarceration is associated with a high risk for active drug use and theoretical risk for HCV transmission,33–36 rapid linkages to HCV and substance use disorder treatment are a high priority. The importance of linkage interventions among people who inject drugs is further underscored by risk of overdose after incarceration.37 Reporting a history of drinking alcohol to intoxication prior to incarceration was negatively associated with linkage to HCV care. Data are limited on the impact of alcohol use on linkage to care following incarceration. However, alcohol has been shown to be negatively associated with DAA adherence among people who inject drugs.38 Drinking alcohol to intoxication in the 30 days prior to incarceration was not associated with linkage to care; however, it did show a similar trend. Therefore, screening and treatment of alcohol use disorders should also be considered an integral component of HCV linkage to care programs. Feeling family or a loved one were concerned about the participants’ wellbeing was also associated with linkage to HCV care in our study. The role of social support in promoting linkage to care after incarceration among PLWHCV is not known; however, there is evidence among PLWHIV that a lack of social support is a barrier. The use of patient or peer navigators with racial/ethnic concordance or shared life experience have been proposed as an important strategy.31,39,40 This may be even more important for those who lack support from family or a loved one. In our study, a navigator with a master’s level of education performed patient navigation that was mostly telephonic. It is possible peer navigation with active outreach might have resulted in a higher linkage rate. Feeling family or a loved one were concerned about the participants’ wellbeing was also associated with linkage to HCV care in our study. Discussion The role of social support in promoting linkage to care after incarceration among PLWHCV is not known; however, there is evidence among PLWHIV that a lack of social support is a barrier. The use of patient or peer navigators with racial/ethnic concordance or shared life experience have been proposed as an important strategy.31,39,40 This may be even more important for We did not identify a statistically significant relationship between reincarceration and linkage to HCV care. Reincarceration has been demonstrated to be a complicating factor in linkage to and retention in HIV care. However, since incarcerated persons have access to stable medical services, it can also be leveraged to improve HIV-related outcomes.17,41,42 In our study, reincarceration was more common in the 180 days after index incarceration than linkage to HCV care and was an equally common initial event to linkage to HCV care. If individuals are reincarcerated, communication can be interrupted and HCV linkage appointments may be missed. Conversely, reincarceration may be an opportunity to reengage individuals with HCV who have not yet linked to care, as was the case for three participants in this study. For those who initiate HCV treatment in the community and are at risk of ongoing justice-involvement, education should be provided to make jail healthcare staff aware to avoid HCV treatment interruption. This study has limitations. First, as a single arm trial, we are unable to determine if the observed linkage rate was directly attributable to the intervention. Moreover, the moderate sample size could lead to decreased statistical power to detect associations. However, we believe our study provides important Page 9/19 Page 9/19 Page 9/19 preliminary data on the rate and factors associated with linkage to HCV care after incarceration in jail following the implementation of a combined case management and patient navigation strategy. Second, the study took place in one large urban area, and community partners were able to schedule HCV appointments within 2–4 weeks after incarceration, which may limit generalizability. Third, our sample may not have been representative of PLWHCV in the jail population as a whole since the median duration of time between enrollment and community return was 34 days, and detainees with the shortest lengths of stay were not able to be recruited. Fourth, demographics and covariates like mental illness were obtained through self-report, which may under- or overestimate true rates. Discussion Lastly, follow-up for linkage to HCV care was conducted over 180 days after the index incarceration and participants who were lost to follow-up could have linked to non study-affiliated clinics so our linkage and retention outcomes may be underestimated. Despite these limitations, our study provides important real-world data on the rate and factors associated with linkage to HCV care after incarceration jail. Due to the modest linkage rates observed, future interventions should consider strengthening transitional care planning and community patient navigation among justice-involved PLWHCV. We believe that multicomponent intervention incorporating education, increased resources for TCC and community-based patient navigation, treatment of substance use disorders—specifically concomitant linkage to opioid agonist therapy, and increasing social support are needed. Declarations Ethics approval and consent to participate: The study team obtained written informed consent from all participants enrolled in the study, and approval to conduct the study was obtained from New York University School of Medicine and New York City Department of Health and Mental Hygiene Institutional Review Boards (IRB). The New York University School of Medicine IRB approval process encompassed detailed review by the IRB prisoners’ advocate. Consent for publication: Not applicable. Availability of data and material: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests: The authors declare that they have no competing interests Competing interests: The authors declare that they have no competing interests Funding: This work was funded by a Gilead Sciences Investigator Sponsored Research grant. Dr. Akiyama is also supported by a Pathway to Independence Award from the National Institute on Drug Abuse (K99DA043011). Authors’ contributions: MJA and DC drafted the manuscript. MJA, RM, and EC devised the study concept and design. Data acquisition was conducted by MJA, DC, JS. Data analysis was conducted by JS. All authors conducted critical revisions of the manuscript for important intellectual content. Page 10/19 Acknowledgements: Not applicable. References 1.Edlin BR, Eckhardt BJ, Shu MA, Holmberg SD, Swan T. Toward a more accurate estimate of the prevalence of hepatitis C in the United States. Hepatology (Baltimore, Md). 2015;62(5):1353–1363. 1.Edlin BR, Eckhardt BJ, Shu MA, Holmberg SD, Swan T. Toward a more accurate estimate of the prevalence of hepatitis C in the United States. Hepatology (Baltimore, Md). 2015;62(5):1353–1363. 2.Varan AK, Mercer DW, Stein MS, Spaulding AC. Hepatitis C seroprevalence among prison inmates since 2001: still high but declining. Public Health Rep. 2014;129(2):187–195. 3.Rich JD, Allen SA, Williams BA. Responding to hepatitis C through the criminal justice system. The New England journal of medicine. 2014;370(20):1871–1874. 4.Murray KF, Richardson LP, Morishima C, Owens JW, Gretch DR. Prevalence of hepatitis C virus infection and risk factors in an incarcerated juvenile population: a pilot study. Pediatrics. 2003;111(1):153–157. 5.Nyamathi A, Salem BE, Marlow E, Zhang S, Yadav K. Understanding correlates of hepatitis C virus infection among homeless recently paroled men. J Forensic Nurs. 2013;9(3):161–170. 6.Hudson AL, Nyamathi A, Bhattacharya D, et al. Impact of prison status on HIV-related risk behaviors. AIDS and behavior. 2011;15(2):340–346. 7.Cuadrado A, Llerena S, Cobo C, et al. Microenvironment Eradication of Hepatitis C: A Novel Treatment Paradigm. Am J Gastroenterol. 2018. 8.Spaulding AC, Perez SD, Seals RM, Hallman MA, Kavasery R, Weiss PS. Diversity of release patterns for jail detainees: implications for public health interventions. Am J Public Health. 2011;101 Suppl 1:S347– 352. 9.MacDonald R, Akiyama MJ, Kopolow A, et al. Feasibility of Treating Hepatitis C in a Transient Jail Population. Open Forum Infect Dis. 2017;4(3):ofx142. 10.Subramanian R, Delaney R, Roberts S, Fishman N, P.M. Incareration’s Front Door: The Misuse of Jails in America. New York, NY: VERA Institute of Justice;2015. 11.Iroh PA, Mayo H, Nijhawan AE. The HIV Care Cascade Before, During, and After Incarceration: A Systematic Review and Data Synthesis. Am J Public Health. 2015;105(7):e5–16. 12.Nunn A, Cornwall A, Fu J, Bazerman L, Loewenthal H, Beckwith C. Linking HIV-positive jail inmates to treatment, care, and social services after release: results from a qualitative assessment of the COMPASS Program. J Urban Health. 2010;87(6):954–968. 13.Rapp RC, Ciomcia R, Zaller N, Draine J, Ferguson A, Cagey R. The role of jails in engaging PLWHA in care: from jail to community. AIDS and behavior. 2013;17 Suppl 2:S89–99. Page 11/19 Page 11/19 Page 11/19 14.Avery A, Ciomica R, Gierlach M, Machekano R. Jail-Based Case Management Improves Retention in HIV Care 12 Months Post Release. References AIDS and behavior. 2018. 15.Gardner LI, Metsch LR, Anderson-Mahoney P, et al. Efficacy of a brief case management intervention to link recently diagnosed HIV-infected persons to care. AIDS. 2005;19(4):423–431. 16.Teixeira PA, Jordan AO, Zaller N, Shah D, Venters H. Health outcomes for HIV-infected persons released from the New York City jail system with a transitional care-coordination plan. Am J Public Health. 2015;105(2):351–357. 17.Loeliger KB, Meyer JP, Desai MM, Ciarleglio MM, Gallagher C, Altice FL. Retention in HIV care during the 3 years following release from incarceration: A cohort study. PLoS Med. 2018;15(10):e1002667. 18.Jordan AO, Cohen LR, Harriman G, Teixeira PA, Cruzado-Quinones J, Venters H. Transitional care coordination in New York City jails: facilitating linkages to care for people with HIV returning home from Rikers Island. AIDS and behavior. 2013;17 Suppl 2:S212–219. 19.Myers JJ, Koester KA, Kang Dufour MS, Jordan AO, Cruzado-Quinone J, Riker A. Patient navigators effectively support HIV-infected individuals returning to the community from jail settings. Int J Prison Health. 2017;13(3–4):213–218. 20.Myers JJ, Kang Dufour MS, Koester KA, et al. The Effect of Patient Navigation on the Likelihood of Engagement in Clinical Care for HIV-Infected Individuals Leaving Jail. Am J Public Health. 2018;108(3):385–392. 21.New York City Department of Corrections. DOC Statistics https://www1.nyc.gov/site/doc/about/doc- statistics.page. Accessed March 1st, 2018. 22.New York City Department of Corrections. New York City Department of Corrections Inmate Lookup Service http://a073-ils-web.nyc.gov/inmatelookup/pages/common/find.jsf. Accessed March 1st, 2018. 23.New York State Department of Corrections. New York State Department of Corrections Inmate Lookup http://nysdoccslookup.doccs.ny.gov Accessed March 1st, 2018. 24.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. 25.Beckwith CG, Kurth AE, Bazerman LB, et al. A pilot study of rapid hepatitis C virus testing in the Rhode Island Department of Corrections. J Public Health (Oxf). 2016;38(1):130–137. 25.Beckwith CG, Kurth AE, Bazerman LB, et al. A pilot study of rapid hepatitis C virus testing in the Rhode Island Department of Corrections. J Public Health (Oxf). 2016;38(1):130–137. 26.Cocoros N, Nettle E, Church D, et al. Screening for Hepatitis C as a Prevention Enhancement (SHAPE) for HIV: an integration pilot initiative in a Massachusetts County correctional facility. Public Health Rep. 2014;129 Suppl 1:5–11. Page 12/19 Page 12/19 27.Schoenbachler BT, Smith BD, Sena AC, et al. Hepatitis C Virus Testing and Linkage to Care in North Carolina and South Carolina Jails, 2012–2014. References Improving HIV Care Engagement in the South from the Patient and Provider Perspective: The Role of Stigma, Social Support, and Shared Decision-Making. AIDS Patient Care STDS. 2018;32(9):368–378. 41.Spaulding AC, Drobeniuc A, Frew PM, et al. Jail, an unappreciated medical home: Assessing the feasibility of a strengths-based case management intervention to improve the care retention of HIV- infected persons once released from jail. PloS one. 2018;13(3):e0191643. 42.Lim S, Nash D, Hollod L, Harris TG, Lennon MC, Thorpe LE. Influence of Jail Incarceration and Homelessness Patterns on Engagement in HIV Care and HIV Viral Suppression among New York City Adults Living with HIV/AIDS. PloS one. 2015;10(11):e0141912. 42.Lim S, Nash D, Hollod L, Harris TG, Lennon MC, Thorpe LE. Influence of Jail Incarceration and Homelessness Patterns on Engagement in HIV Care and HIV Viral Suppression among New York City Adults Living with HIV/AIDS. PloS one. 2015;10(11):e0141912. References Public Health Rep. 2016;131 Suppl 2:98–104. 28.New York City Human Resources Administration. HIV/AIDS Services. 2018; https://www1.nyc.gov/site/hra/help/hiv-aids-services.page. Accessed July 1st, 2018. 29.Ammon B, Iroh P, Tiruneh Y, et al. HIV Care After Jail: Low Rates of Engagement in a Vulnerable Population. J Urban Health. 2018;95(4):488–498. 30.Loeliger KB, Altice FL, Desai MM, Ciarleglio MM, Gallagher C, Meyer JP. Predictors of linkage to HIV care and viral suppression after release from jails and prisons: a retrospective cohort study. Lancet HIV. 2018;5(2):e96-e106. 31.Molitor F, Waltermeyer J, Mendoza M, et al. Locating and linking to medical care HIV-positive persons without a history of care: findings from the California Bridge Project. AIDS Care. 2006;18(5):456–459. 32.Molitor F, Kuenneth C, Waltermeyer J, et al. Linking HIV-infected persons of color and injection drug users to HIV medical and other services: the California Bridge Project. AIDS Patient Care STDS. 2005;19(6):406–412. 33.Merrall EL, Kariminia A, Binswanger IA, et al. Meta-analysis of drug-related deaths soon after release from prison. Addiction (Abingdon, England). 2010;105(9):1545–1554. 34.Adams J, Nowels C, Corsi K, Long J, Steiner JF, Binswanger IA. HIV risk after release from prison: a qualitative study of former inmates. J Acquir Immune Defic Syndr. 2011;57(5):429–434. 35.Bird SM, Hutchinson SJ. Male drugs-related deaths in the fortnight after release from prison: Scotland, 1996–99. Addiction (Abingdon, England). 2003;98(2):185–190. 36.Kariminia A, Butler T, Corben S, et al. Extreme cause-specific mortality in a cohort of adult prisoners— 1988 to 2002: a data-linkage study. Int J Epidemiol. 2007;36(2):310–316. 37.Binswanger IA, Stern MF, Deyo RA, et al. Release from prison—a high risk of death for former inmates. The New England journal of medicine. 2007;356(2):157–165. 38.Mason K, Dodd Z, Guyton M, et al. Understanding real-world adherence in the directly acting antiviral era: A prospective evaluation of adherence among people with a history of drug use at a community- based program in Toronto, Canada. Int J Drug Policy. 2017;47:202–208. 39.Koester KA, Morewitz M, Pearson C, et al. Patient navigation facilitates medical and social services engagement among HIV-infected individuals leaving jail and returning to the community. AIDS Patient Care STDS. 2014;28(2):82–90. Page 13/19 40.Taylor BS, Fornos L, Tarbutton J, et al. Improving HIV Care Engagement in the South from the Patient and Provider Perspective: The Role of Stigma, Social Support, and Shared Decision-Making. AIDS Patient Care STDS. 2018;32(9):368–378. 40.Taylor BS, Fornos L, Tarbutton J, et al. Table 1. Participant characteristics and associations with linkage to HCV care Table Table 1. Participant characteristics and associations with linkage to HCV care Table 1. Table Participant characteristics and associations with linkage to HCV care Page 14/19 Characteristics Linked to care N=26* (%) Unlinked to care N=58* (%) Total N = 84* (%) P value Age, mean (SD) 47 (11) 44 (12) 45 (12) 0.29 ≥45 <45 14 (54) 12 (46) 29 (50) 29 (50) 43 (51) 41 (49) 0.74 Sex Female Male 9 (35) 17 (65) 27 (47) 31 (53) 36 (43) 48 (57) 0.31 Race/ethnicity Hispanic NH Black NH White Other 15 (58) 5 (19) 5 (19) 1 (4) 31 (54) 10 (18) 12 (21) 4 (7) 46 (55) 15 (18) 17 (21) 4 (6) 0.94 Health insurance§ 22/26 (85) 45/57 (79) 67/83 (81) 0.54 Homeless§ 5 (19) 17 (29) 22 (26) 0.11 Graduated high school 14/25 (56) 34/55 (62) 48/80 (60) 0.62 Psychiatric diagnosis¶ 13 (50) 40 (69) 53 (63) 0.10 Care Coordination Personal contact info at enrollment 13 (50) 23 (40) 36 (43) 0.38 Next of kin contact info at enrollment 25 (96) 52 (90) 77 (92) 0.32 Known release date 14 (54) 34 (59) 48 (57) 0.68 HCV appointment scheduled pre-release 13 (50) 19 (33) 32 (38) 0.13 Existing HCV provider 3/24 (13) 3/54 (6) 6/78 (8) 0.29 Existing primary care provider 9/24 (38) 10/55 (18) 19/79 (24) 0.06 Prefer linkage to existing health system 8 (31) 7 (12) 15 (18) 0.04 Opioid agonist therapy§ Characteristics 5 (19) 19 (33) 24 (29) 0.21 5 (19) 19 (33) Suboxone Suboxone Page 16/19 Substance use Injection drug use 30 days prior 16 (62) 35 (60) 51 (61) 0.92 Lifetime 22 (85) 36 (63) 58 (70) 0.22 Heroin 30 days prior 13 (50) 34 (59) 47 (56) 0.46 Lifetime 25 (96) 50 (86) 75 (89) 0.17 Prescription opiates 30 days prior 4 (15) 2 (3) 6 (7) 0.90 Lifetime 14 (54) 23 (40) 37 (44) 0.22 Crack/Cocaine 30 days prior 9 (35) 23 (40) 32 (38) 0.66 Lifetime 21 (81) 48 (83) 69 (82) 0.83 Amphetamines 30 days prior 0 (0) 2 (3) 2 (2) 0.33 Lifetime 2 (8) 8 (14) 10 (12) 0.42 Marijuana 30 days prior 6 (23) 20 (34) 26 (31) 0.30 Lifetime 15 (58) 44 (76) 59 (70) 0.09 Alcohol to intoxication 30 days prior 2 (8) 12 (21) 14 (17) 0.14 Lifetime 5 (16) 26 (44) 31 (37) 0.03 Social support Feel supported socially 21/25 (84) 37/56 (66) 58/81 (72) 0.10 Feel family or a loved one is concerned about wellbeing 25/25 (100) 45/56 (64) 70/81 (86) 0.02** Reincarcerated ≥1 time within 180 days post-release 8 (31) 20 (36) 28 (33) 0.86 Data were obtained at enrollment in the jail setting. Table ¶ Includes depression, anxiety, bipolar, schizophrenia; § Refers to the period prior to incarceration; Unless denominator specified; *Statistically significant alpha level p<0.05. HCV = hepatitis C virus, NH = non-Hispanic, SD = standard deviation. Figures Figure 1 Figure 1 Participant flow chart. HCV = hepatitis C virus, HIV = human immunodeficiency virus, VL = viral load, Tx = treatment Page 17/19 Page 17/19 Figure 2 Hepatitis C virus (HCV) care cascade after incarceration in the New York City jails. In a cohort of 84 participants who returned to the community after incarceration, 26 (31%) linked to HCV care, 17 (20%) initiated HCV treatment, 15 (18%) completed treatment, 9 (11%) had a follow-up lab drawn to confirm sustained virologic response (SVR), and 7 (8%) had a documented SVR. Of the 9 participants with documented 12-week follow-up labs, this was a 78% SVR rate. Figure 2 Hepatitis C virus (HCV) care cascade after incarceration in the New York City jails. In a cohort of 84 participants who returned to the community after incarceration, 26 (31%) linked to HCV care, 17 (20%) initiated HCV treatment, 15 (18%) completed treatment, 9 (11%) had a follow-up lab drawn to confirm sustained virologic response (SVR), and 7 (8%) had a documented SVR. Of the 9 participants with documented 12-week follow-up labs, this was a 78% SVR rate. Page 18/19 Figure 3 Trajectories of people living with HCV after incarceration in the New York City Jails. Each row in an EventFlow figure represents one participant’s sequence of events during a period of time. The height of each bar is proportional to the number of records with that sequence, and its horizontal position is determined by the median time between events. Groups of sequences with the same preceding event a sorted by the number of records in each group. The sequence groups are shown from top to bottom in descending order of number of participants per group. For interval events such as returns to jail, the length of the intervals represents the mean duration of the grouped events. For point events such as linkage to HCV care, periods between the aggregated point events represent the mean length of time fr any previous point event. Figure 3 Figure 3 Trajectories of people living with HCV after incarceration in the New York City Jails. Each row in an EventFlow figure represents one participant’s sequence of events during a period of time. The height of each bar is proportional to the number of records with that sequence, and its horizontal position is determined by the median time between events. Groups of sequences with the same preceding event are sorted by the number of records in each group. The sequence groups are shown from top to bottom in descending order of number of participants per group. For interval events such as returns to jail, the length of the intervals represents the mean duration of the grouped events. For point events such as linkage to HCV care, periods between the aggregated point events represent the mean length of time from any previous point event. Page 19/19
https://openalex.org/W4362727849	https://e-journal.unmuhkupang.ac.id/index.php/jipend/article/download/686/429	Indonesian	null	ANALISIS PENGGUNAAN EJAAN DALAM KARANGAN NARASI SISWA KELAS VII SMP MUHAMMADIYAH KUPANG	Jurnal Ilmu Pendidikan/JIP (Jurnal Ilmu Pendidikan)	2,022	cc-by	5,889	Abstract This study aims to describe the use of capital letters in Muhammadiyah Middle School in Kupang, describing the use of punctuation marks (.) And commas (,) in narrative essays of students of class VII/B at Muhammadiyah Middle School in Kupang, and describing clitical writing you -, - me, you -, and in the narrative essay of Class VII / B Students of Muhammadiyah Middle School Kupang.The method used in this research is descriptive qualitative method. The data sources in the study were the results of the test writing a narr Kupang Academic Year 2017/2018, as many as 26 students. Data management techniques are carried out by means of; (a) read and understand students' essays. Data that has been obtained throug and understood one by one, (b) identifies elements of spelling errors, related to the use of capital letters, writing punctuation marks and commas, and clitics of you, me, you, and so, ( c) analyze and describe spelling mistakes in students' narrative essays, (d) discuss and present forms of spelling misuse; and (e) Conclude the results of data analysis in the form of language deviations found in students' essays.The results showed spelling errors in the VII grade students of Muhamma included: (a) Capital letter usage errors totaling 235 data cases of errors, (b) errors in the use of dot punctuation numbered 93 data cases of errors, (c ) errors in the use of comma pun of errors, (d) errors in writing clitical writingyou; Keywords: analysis, eyd, narrative essay. This study aims to describe the use of capital letters in narrative essays of Class VII/B Students of Muhammadiyah Middle School in Kupang, describing the use of punctuation marks (.) And commas (,) in narrative essays of students of class VII/B at Muhammadiyah Middle School in Kupang, and describing clitical , and in the narrative essay of Class VII / B Students of Muhammadiyah Middle School Kupang.The method used in this research is descriptive qualitative method. The data sources in the study were the results of the test writing a narrative essay of Class VII / B Students of Muhammadiyah Middle School Kupang Academic Year 2017/2018, as many as 26 students. Data management techniques are carried out by means of; (a) read and understand students' essays. Abstrak Penelitian ini bertujuan untuk mendeskripsikan VII/B SMP Muhammadiyah Kupang, Mendeskripsikan penggunaan tanda baca titik (.) dan tanda Koma (,) dalam karangan Narasi Siswa kelas VII/B SMP Muhammadiyah Kupang klitikakau-,-ku,-mu-, dan-nya dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang. Metode yang digunakan dalam penelitian ini adalah metode deskriptif kualitatif. Sumber data dalam penelitian hasil tes menulis karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran 2017/2018, sebanyak 26 siswa.Teknik pengolahan data dilakukan Data yang sudah diperoleh melalui tes, dibaca dan dipahami s kesalahan ejaan, terkait dengan penggunaan huruf kapital, penulisan tanda baca titik dan koma, serta klitikakau, ku, mu, dan nya, (c) menganalisis dan mendeskripsikan kesalahan penggunaan ejaan dalam karangan narasi siswa, (d) membahas dan menyajikan bentuk hasil analisis data berupa penyimpangan berbahasa yang terdapat dalam karangan siswa. menunjukkan kesalahan ejaan dalam karangan kasus kesalahan, yang meliputi: (a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, (b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, (c) kesalahan pemakaian t koma berjumlah 197 data kasus kesalahan, (d) kesalahan penulisan klitikakau; data kasus kesalahan. mendeskripsikan penggunaan huruf kapital dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang, Mendeskripsikan penggunaan tanda baca titik (.) dan tanda Koma (,) dalam karangan Narasi Siswa kelas VII/B SMP Muhammadiyah Kupang, danMendeskripsikan penulisan dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang. Metode yang digunakan dalam penelitian ini adalah metode deskriptif kualitatif. Sumber data dalam penelitian menulis karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran 2017/2018, sebanyak 26 siswa.Teknik pengolahan data dilakukan dengan cara; (a) membaca dan memahami karangan siswa. Data yang sudah diperoleh melalui tes, dibaca dan dipahami satu per satu, (b) mengidentifikasi unsur terkait dengan penggunaan huruf kapital, penulisan tanda baca titik dan koma, serta menganalisis dan mendeskripsikan kesalahan penggunaan ejaan dalam karangan narasi siswa, (d) membahas dan menyajikan bentuk-bentuk kesalahan penggunaan ejaan; dan (e) Menyimpulkan hasil analisis data berupa penyimpangan berbahasa yang terdapat dalam karangan siswa. kesalahan ejaan dalam karangan siswa kelas VII SMP Muhammadiyah kasus kesalahan, yang meliputi: (a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, (b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, (c) kesalahan pemakaian t koma berjumlah 197 data kasus kesalahan, (d) kesalahan penulisan klitikakau;-ku;- penggunaan huruf kapital dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang, Mendeskripsikan penggunaan tanda baca titik (.) dan tanda Koma (,) , danMendeskripsikan penulisan dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang. Abstrak Metode yang digunakan dalam penelitian ini adalah metode deskriptif kualitatif. Sumber data dalam penelitian adalah menulis karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran 2017/2018, membaca dan memahami karangan siswa. atu per satu, (b) mengidentifikasi unsur-unsur terkait dengan penggunaan huruf kapital, penulisan tanda baca titik dan koma, serta menganalisis dan mendeskripsikan kesalahan penggunaan ejaan dalam karangan bentuk kesalahan penggunaan ejaan; dan (e) Menyimpulkan hasil analisis data berupa penyimpangan berbahasa yang terdapat dalam karangan siswa. Hasil penelitian kelas VII SMP Muhammadiyah Kupang berjumlah 525 kasus kesalahan, yang meliputi: (a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, (b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, (c) kesalahan pemakaian tanda baca -mu;-dannya berjumlah 12 Kata kunci: Analisis, Ejaan, Karangan Narasi. Kata kunci: Analisis, Ejaan, Karangan Narasi. Kata kunci: Analisis, Ejaan, Karangan Narasi. Kata kunci: Analisis, Ejaan, Karangan Narasi. Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 ANALISIS PENGGUNAAN PENGGUNAAN EJAAN DALAM KARANGAN NARASI SISWA MUHAMMADIYAH KUPANG SISWA KELAS VII SMP Nurlailah, Abdul Syahril Muh. Program Studi Pendidikan Guru Sekolah Dasar Pos-el:abimnurlailah@mail.com A. PENDAHULUAN Menulis merupakan bagian yang tidak terpisahkan dalam sebuah proses belajar yang dialami siswa selama menuntut ilmu di sekolah. Menulis memerlukan keterampilan karena diperlukan latihan-latihan yang berkelanjutan. Dengan demikian, keterampilan menulis lebi banyak diperoleh dari pengalaman yang berulang-ulang melalui latihan terstruktur di sekolah. Halinisesuaideng (2016: 31)yang meny pembelajaran bahasa diha pesertadidik mengenal danbudaya orang lain, meng dan perasaan, berpartis masyarakatyang me bahasatersebut,dan menggunakan kemampuan imaginatif yang ada Pembelajaran keterampilan menulis memiliki berbagai macam bentuk. Salah satunya adalah keterampilan menulis karangan yang termasuk dalam menulis lanjutan. Ada lima bentuk karangan yakni; karangan narasi, persuasi, argumentasi, deskripsi, dan eksposisi. Menulis merupakan bagian yang tidak terpisahkan dalam sebuah proses belajar yang dialami siswa selama menuntut ilmu di sekolah. Menulis memerlukan keterampilan karena latihan yang berkelanjutan. Dengan demikian, keterampilan menulis lebih banyak diperoleh dari pengalaman yang ulang melalui latihan terstruktur di ganKurikulum2013 nyatakan bahwa sa diharapkanmembantu l dirinya,budayanya, gemukakan gagasan rpartisipasi dalam g menggunakan n menemukanserta n analitis dan da dalam dirinya. Pembelajaran keterampilan menulis memiliki bentuk. Salah satunya adalah keterampilan menulis karangan yang termasuk dalam menulis lanjutan. Ada lima bentuk karangan yakni; karangan narasi, persuasi, argumentasi, deskripsi, dan eksposisi. Menurut Arifin dan Tasai (2008:164) pemahaman akan tata bahasa perlu diperhatikan dalam menulis atau hasil tulisan lain yang bersifat ilmiah maupun non ilmiah. Dengan memperhatikan tata bahasa yang baik dan benar siswa dapat membiasakan bahwa hal te memanglah perlu dalam keterampilan berbahasa khususnya keterampilan menulis. Suparno dan Yunus (2012: 96) mengemukakan bahwa menulis harus menggunakan aturan terdapat dalam bahasa Indonesia. Sebagai pemakai bahasa, kita wajib mematuhi a baku berbahasa yang dinyatakan dalam ejaan yang disempurnakan atau yang sekarang dikenal dengan PUEBI. Menurut Kridalaksana dalam Kamus Linguistik edisi keempat (2008: 22) yang dimaksud dengan ejaan adalah keseluruhan peraturan bagaimana melambangkan ujaran dan bagaimana antar hubungan antara lambang-lambang itu (pemisahan dan penggabungannya dalam suatu bahasa). Secara teknis yang dimaksud dengan ejaan adalah penulisan huruf, penulisan kata, dan pemakaian tanda baca. Menurut Arifin dan Tasai (2008:164) pemahaman akan tata bahasa perlu diperhatikan dalam menulis atau hasil tulisan lain yang bersifat ilmiah maupun non ilmiah. Dengan memperhatikan tata bahasa yang baik dan benar siswa dapat membiasakan bahwa hal tersebut memanglah perlu dalam keterampilan berbahasa khususnya keterampilan menulis. Suparno dan Yunus (2012: 96) mengemukakan bahwa menulis harus menggunakan aturan-aturan yang terdapat dalam bahasa Indonesia. Sebagai pemakai bahasa, kita wajib mematuhi aturan baku berbahasa yang dinyatakan dalam ejaan yang disempurnakan atau yang sekarang dikenal dengan PUEBI. Abstract Data that has been obtained throug and understood one by one, (b) identifies elements of spelling errors, related to the use of capital letters, writing punctuation marks and commas, and clitics of you, me, you, and so, ( c) analyze and describe spelling mistakes students' narrative essays, (d) discuss and present forms of spelling misuse; and (e) Conclude the results of data analysis in the form of language deviations found in students' essays.The results showed spelling errors in the VII grade students of Muhammadiyah Middle School in Kupang amounted to 525 cases of errors, which included: (a) Capital letter usage errors totaling 235 data cases of errors, (b) errors in the use of dot punctuation numbered 93 data cases of errors, (c ) errors in the use of comma punctuation are 197 data in case of errors, (d) errors in writing clitical writingyou; -me; -you; and there are 11 data cases of errors. Keywords: analysis, eyd, narrative essay. narrative essays of Class VII/B Students of Muhammadiyah Middle School in Kupang, describing the use of punctuation marks (.) And commas (,) in narrative essays of students of class VII/B at Muhammadiyah Middle School in Kupang, and describing clitical , and in the narrative essay of Class VII / B Students of Muhammadiyah Middle School Kupang.The method used in this research is descriptive qualitative method. The data sources in the study were ative essay of Class VII / B Students of Muhammadiyah Middle School Kupang Academic Year 2017/2018, as many as 26 students. Data management techniques are carried out by means of; (a) read and understand students' essays. Data that has been obtained through composing tests, read and understood one by one, (b) identifies elements of spelling errors, related to the use of capital letters, writing punctuation marks and commas, and clitics of you, me, you, and so, ( c) analyze and describe spelling mistakes students' narrative essays, (d) discuss and present forms of spelling misuse; and (e) Conclude the results of data analysis in the form of language deviations found in students' essays.The results showed spelling errors in diyah Middle School in Kupang amounted to 525 cases of errors, which included: (a) Capital letter usage errors totaling 235 data cases of errors, (b) errors in the use of dot ctuation are 197 data in case you; and there are 11 data cases of errors. 1 1 Jurnal Ilmu Pendidikan (JIP), Vol. Abstract p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 yang telah terjadi?” Setiap orang pasti memiliki pengalaman. Dari sejumlah pengalaman itu tentu ada kesan atau hal yang menarik untuk diceritakan kepada orang lain. Di dalam kelas atau diluar kelas siswa sering bercerita dengan teman-teman sebayanya mengenai su Akan tetapi, cerita tersebut jika ditransformasikan dalam bentuk tulisan berupa karangan narasi, siswa masih mengalami kesulitan, salah-satu faktornya adalah Penguasaan ejaan siswa yang masih sangat rendah. Dikutip dari artikel “ literasi di Indonesia “ dari situs (diakses online pada 15Februari, 2018) adanya kecenderungan siswa kurang mengapresiasi keterampilan menulis itu, disebabkan minimnya penguasaan dan pemahaman siswa terkait ejaan dalam ranah tulis menulis. telah terjadi?” Setiap orang pasti memiliki pengalaman. Dari sejumlah pengalaman itu tentu ada kesan atau hal yang menarik untuk diceritakan kepada orang lain. Di dalam kelas atau diluar kelas siswa sering bercerita dengan teman sebayanya mengenai suatu hal. Akan tetapi, cerita tersebut jika ditransformasikan dalam bentuk tulisan berupa karangan narasi, siswa masih mengalami satu faktornya adalah Penguasaan ejaan siswa yang masih sangat rendah. Dikutip dari artikel “ Mundurnya budaya “ dari situs Portal Guru id. diakses online pada 15Februari, 2018) bahwa adanya kecenderungan siswa kurang mengapresiasi keterampilan menulis itu, disebabkan minimnya penguasaan dan pemahaman siswa terkait ejaan dalam ranah tulis- A. PENDAHULUAN Menurut Kridalaksana dalam Kamus Linguistik edisi keempat (2008: 22) yang dimaksud dengan ejaan adalah keseluruhan peraturan bagaimana melambangkan bunyi ujaran dan bagaimana antar hubungan antara lambang itu (pemisahan dan penggabungannya dalam suatu bahasa). Secara teknis yang dimaksud dengan ejaan adalah penulisan huruf, penulisan kata, dan pemakaian Menurut Keraf (2007: 134) dari kelima karangan tersebut karangan narasi lebih banyak diterapkan pada tahap awal menulis lanjutan, karena narasi bertujuan menggali hal dalam kehidupan siswa. Hal ini t Kurikulum 2013 kelas VII (2016: 43) Menengah Pertama kompetensi,yaitu;mengidentifikasi unsur teks narasi, yang dibaca dan didengar, menceritakan kembali isi teks narasi, menelaah struktur kebahasaan teks narasi, dan menyajikan gagasan kreatif dalam karangan narasi berupa cerita fantasi. Menurut Keraf (2007: 134) dari kelima gan narasi lebih banyak diterapkan pada tahap awal menulis lanjutan, karena narasi bertujuan menggali hal-hal aktual dalam kehidupan siswa. Hal ini tercantum dalam um 2013 kelas VII (2016: 43) Sekolah Menengah Pertama denganstandar ngidentifikasi unsur-unsur teks narasi, yang dibaca dan didengar, menceritakan kembali isi teks narasi, menelaah struktur kebahasaan teks narasi, dan menyajikan gagasan kreatif dalam karangan narasi berupa Narasi sebenarnya merupakan karangan yang mudah ditulis oleh siswa karena karangan ini dikembangkan melalui kegemaran siswa dalam mendengarkan cerita atau bercerita. Seperti yang dikemukakan oleh Keraf (2007: 136) bahwa narasi berusaha menjawab: “Apa 2 Narasi sebenarnya merupakan karangan yang mudah ditulis oleh siswa karena karangan ini dikembangkan melalui kegemaran siswa dalam mendengarkan cerita atau bercerita. Seperti yang dikemukakan oleh Keraf (2007: 136) bahwa narasi berusaha menjawab: “Apa 2 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 3)bagaimana penulisan klitikaku-;- karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran 2017/2018?. dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 3)bagaimana -mu-;dan-nya dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 3)bagaimana penulisan klitikaku-;- karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran 2017/2018?. dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 3)bagaimana -mu-;dan-nya dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang Tahun Ajaran Kepala Pusat Pembinaan dan Pengembangan Badan Bahasa Kemendikbud Prof. Dr. A. PENDAHULUAN Gufran Ali Ibrahim, M.Si dalam seminar yang bertemakan “ Meningkatkan Literasi Masyarakat’’ diselenggarakan di Jakarta, 27 Februari 2018 (diakses Online 15 Februari, 2018 berpendapat Minimnya pemahaman penguasaan ejaan bahasa Indonesia menyebabkan siswa tidak mampu menulis kata secara benar, tidak mampu menggunakan tanda baca dengan tepat, serta tidak mampu menuliskan huruf dengan baik dan benar. Sehingga karangan yang dihasil menjadi cacat atau tidak beraturan. Kepala Pusat Pembinaan dan Pengembangan Badan Bahasa Kemendikbud Prof. Dr. Gufran Ali Ibrahim, M.Si dalam seminar yang bertemakan “ Peran Guru dalam Meningkatkan Literasi Masyarakat’’ yang diselenggarakan di Jakarta, 27 Februari 2018 diakses Online 15 Februari, 2018). Beliau ndapat Minimnya pemahaman penguasaan ejaan bahasa Indonesia menyebabkan siswa tidak mampu menulis kata secara benar, tidak mampu menggunakan tanda baca dengan tepat, serta tidak mampu menuliskan huruf dengan baik dan benar. Sehingga karangan yang dihasilkan menjadi cacat atau tidak beraturan. tujuan penelitian ini adalah mengetahui dan mendeskripsikan penggunaan ejaan bahasa Indonesia khususnya pemakaian huruf kapital, tanda baca dan klitika dalam karangan narasi siswa kelas siswa SMP Muhammamdiyah Kupang. Adapun manfaatnya bagi guru dengan adanya penelitian ini diharapkan dapat bermanfaat bagi guru bahasa dan sastra Indonesia, sebagai bahan ulasan dan evaluasi. Untuk kedepannya memperhatikan dan memprioritaskan akan penggunaan ejaan dalam pembelajaran keterampilan menulis khususnya menulis karangan narasi, sehingg menghasilkan karangan yang baik dan benar, sedangkan untuk siswa penelitian ini diharapkan dapat memberikan pengetahuan dalam menulis karangan narasi sehingga keterampilan menulis karangan narasi dapat meningkat, serta menarik minat siswa agar senang dan menyukai kegiatan menulis. tujuan penelitian ini adalah untuk mengetahui dan mendeskripsikan penggunaan ejaan bahasa Indonesia khususnya pemakaian huruf kapital, tanda baca dan klitika dalam karangan narasi siswa kelas siswa SMP Muhammamdiyah Kupang. Adapun manfaatnya bagi guru dengan adanya penelitian ini harapkan dapat bermanfaat bagi guru bahasa dan sastra Indonesia, sebagai bahan ulasan dan evaluasi. Untuk kedepannya memperhatikan dan memprioritaskan akan penggunaan ejaan dalam pembelajaran keterampilan menulis khususnya menulis karangan narasi, sehingga siswa mampu menghasilkan karangan yang baik dan benar, penelitian ini diharapkan dapat memberikan pengetahuan dalam menulis karangan narasi sehingga keterampilan menulis karangan narasi dapat meningkat, serta menarik ar senang dan menyukai kegiatan Berdasarkan pengalaman dan pengamatan di kelas saat program pengalaman lapangan (PPL) di SMP Muhammadiyah Kupang, masih banyak ditemui terjadi kesalahan penulisan ejaan dalam tulisan siswa. Ini berarti kemampuan siswa dalam memahami ejaan masih tergolong rendah. Kesalahan penulisan ejaan yang sering ditemukan yakni penggunaan huruf kapital, tanda baca dan penulisan klitika. A. PENDAHULUAN Faktor penyebab pada umumnya adalah ketidaktahuan siswa dalam memahami penulisan ejaan yang benar, metode pembelajaran yang digunakan guru kurang efektif khususnya materi ejaan seperti penulisan huruf kapital, tanda baca dan klitika, dan mungkin guru kurang menekankan siswa untuk membiasakan menulis dengan memperhatikan kaidah-kaidah penulisan benar. Berdasarkan pengalaman dan pengamatan di kelas saat program pengalaman lapangan (PPL) di SMP Muhammadiyah Kupang, masih banyak ditemui terjadi kesalahan penulisan ejaan dalam tulisan siswa. Ini berarti siswa dalam memahami ejaan masih tergolong rendah. Kesalahan penulisan ejaan yang sering ditemukan yakni penggunaan huruf kapital, tanda baca dan penulisan klitika. Faktor penyebab pada umumnya adalah ketidaktahuan siswa dalam memahami penulisan ejaan yang benar, metode pembelajaran yang digunakan guru kurang efektif khususnya materi ejaan seperti penulisan huruf kapital, tanda baca dan klitika, dan mungkin guru kurang menekankan siswa untuk membiasakan menulis dengan kaidah penulisan yang B. METODE Penelitian ini dilakukan dengan menggunakan metode deskriptif kualitatif yang termasuk dalam penelitian kebahasaan. Boydan dan Taylor (dalam mengatakan bahwa pendekatan kualitatif merupakan prosedur penelitian yang menghasilkan data deskriptif berupa kata tertulis atau lisan dari orang yang dapat diamati. Sandjaja dan Heriyanto (2006: 4) mengemukakan b kualitatif memiliki tujuan utama yaitu mengumpulkan data deskriptif yang mendeskripsikan objek penelitiaan secara rinci dan mendalam dengan maksud mengembangkan konsep atau pemahaman dari suatu gejala. Pendekatan deskriptif kualitatif dipil Penelitian ini dilakukan dengan menggunakan metode deskriptif kualitatif yang termasuk dalam penelitian kebahasaan. Boydan Moleong, 2006: 4) mengatakan bahwa pendekatan kualitatif merupakan prosedur penelitian yang menghasilkan data deskriptif berupa kata-kata tertulis atau lisan dari orang-orang dan perilaku yang dapat diamati. Sandjaja dan Heriyanto (2006: 4) mengemukakan bahwa metodologi kualitatif memiliki tujuan utama yaitu mengumpulkan data deskriptif yang mendeskripsikan objek penelitiaan secara rinci dan mendalam dengan maksud mengembangkan konsep atau pemahaman dari suatu gejala. Pendekatan deskriptif kualitatif dipilih karena Untuk mengatasi hal dituntut untuk membina dan membekali siswa dengan pembelajaran atau materi kaidah kebahasaan seperti penguasaan ejaan yang tepat. Ini bertujuan memberi pe merangsang kreativitas siswa khususnya menulis karangan narasi. Berdasarkan latar belakang, maka rumusan masalah dirinci sebagai berikut; 1) bagaimana penggunaan huruf kapital dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 2)bagaimana penggunaan tanda baca titik dan tanda koma Untuk mengatasi hal-hal tersebut, guru dituntut untuk membina dan membekali siswa dengan pembelajaran atau materi kaidah-kaidah kebahasaan seperti penguasaan ejaan yang tepat. Ini bertujuan memberi pe ngetahuan dan merangsang kreativitas siswa dalam menulis, khususnya menulis karangan narasi. Berdasarkan latar belakang, maka rumusan masalah dirinci sebagai berikut; 1) bagaimana penggunaan huruf kapital dalam karangan narasi Siswa Kelas VII/B SMP Muhammadiyah Kupang?, 2)bagaimana baca titik dan tanda koma 3 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 dapat berupa pendapat, pengetahuan, pengelaman, keinginan, ataupun perasaan orang. Menulis tidak hanya mengungkapkan gagasan melalui bahasa tulis, tetapi meramu tulisan tersebut agar dapat dipahami pembaca. dapat berupa pendapat, pengetahuan, pengelaman, keinginan, ataupun perasaan orang. Menulis tidak hanya mengungkapkan gagasan melalui bahasa tulis, tetapi meramu tulisan tersebut agar dapat dipahami pembaca. masalah yang diteliti berupa data (karangan narasi siswa) yang lebih tepatnya dijelaskan dengan menggunakan kata langkah-langkah yang ditempuh yaitu: masalah yang diteliti berupa data (karangan narasi siswa) yang lebih tepatnya dijelaskan dengan menggunakan kata-kata. Adapun langkah yang ditempuh yaitu: a) Membaca dan memahami karangan siswa. 2. Karangan Narasi Keraf (1994: 136) menjelaskan bahwa narasi sebagai suatu bentuk wacana yang sasaran utamanya adalah tindakan dan dirangkaikan menjadi sebuah peristiwa yang terjadi dalam suatu kejadian waktu. Dapat juga dirumuskan dengan kata lain bahwa narasi adalah suatu bentuk wacana yang berusaha menggambarkan dengan sejelas pembaca suatu per terjadi..Nurudin (2010: 71) menerangkan bahwa narasi adalah bentuk tulisan yang menciptakan, mengisahkan, merangkaikan tindak-tanduk perbuatan manusia dalam sebuah peristiwa secara kronologis atau yang berlangsung dalam suatu kej Keraf (1994: 136) menjelaskan bahwa narasi sebagai suatu bentuk wacana yang sasaran utamanya adalah tindakan-tanduk yang dijalin dan dirangkaikan menjadi sebuah peristiwa yang terjadi dalam suatu kejadian waktu. Dapat juga dirumuskan dengan kata lain bahwa narasi adalah suatu bentuk wacana yang berusaha menggambarkan dengan sejelas-jelasnya kepada pembaca suatu peristiwa yang sudah Nurudin (2010: 71) menerangkan bahwa narasi adalah bentuk tulisan yang berusaha menciptakan, mengisahkan, merangkaikan tanduk perbuatan manusia dalam sebuah peristiwa secara kronologis atau yang berlangsung dalam suatu kejadian waktu tertentu. b) Mengidentifikasi unsur ejaan, terkait dengan penggunaan huruf kapital, penulisan tanda baca titik dan koma, serta klitikakau Mengidentifikasi unsur-unsur kesalahan terkait dengan penggunaan huruf kapital, penulisan tanda baca titik dan kau, ku, mu, dan nya; c) Menganalisis dan mendeskripsikan kesalahan penggunaan ejaan karangan narasi siswa; Menganalisis dan mendeskripsikan kesalahan penggunaan ejaan dalam karangan narasi siswa; d) Membahas dan menyajikan bentuk bentuk kesalahan penggunaan ejaan; dan Membahas dan menyajikan bentuk- bentuk kesalahan penggunaan ejaan; dan e) Menyimpulkan hasil analisis data berupa penyimpangan berbahasa yang terdapat dalam karangan siswa. Menyimpulkan hasil analisis data berupa penyimpangan berbahasa yang terdapat dalam karangan siswa. C. KAJIAN TEORI Narasi adalah karangan yang bersifat menceritakan sesuatu dari waktu kewaktu. Biasanya unsur waktu sangat menentukan dalam karangan narasi. Di sini, peristiwa disusun dengan cara tertentu untuk membentuk suatu cerita. Dari beberapa pendapat para ahli, maka narasi adalah suatu wacana atau karangan yang bertujuan untuk mengisahkan atau menceritakan suatu peristiwa atau kejadian dari waktu ke waktu. Biasanya digunakan oleh para penulis menurut urutan terjadinya (kronologis) agar pembaca dapat memetik hikmah dari cerita. Narasi adalah karangan yang bersifat menceritakan sesuatu dari waktu kewaktu. Biasanya unsur waktu sangat menentukan dalam karangan narasi. Di sini, peristiwa-peristiwa disusun dengan cara tertentu untuk membentuk beberapa pendapat para ahli, maka narasi adalah suatu wacana atau karangan yang bertujuan untuk mengisahkan atau menceritakan suatu peristiwa atau kejadian dari waktu ke waktu. Biasanya digunakan oleh para penulis menurut urutan terjadinya (kronologis) r pembaca dapat memetik hikmah dari cerita. B. METODE Data yang sudah diperoleh melalui tes mengarang, dibaca dan dipahami satu per satu; Membaca dan memahami karangan siswa. Data yang sudah diperoleh melalui g, dibaca dan dipahami satu 1. Keterampilan Menulis Menurut Tarigan (2008: 21) menulis ialah menurunkan atau melukiskan lambang lambang grafik yang menggambarkan suatu bahasa yang dipahamioleh seseorang sehingga orang lain dapat membaca lambang grafik tersebut kalau mereka memahami bahasa dan gambaran grafik itu. S (2008: 17 ) mengatakan bahwa menulis adalah mengungkapkan pikiran, perasaan, pengalaman, dan hasil bacaan dalam bentuk tulisan, bukan dalam bentuk tutur. Rosidi (2009: 10 mengemukakan bahwa, tulisan yang baik memiliki ciri-ciri; a) kesesuaian judul dengan isi tulisan, b) ketepatan penggunaan ejaan dan tanda baca, c), ketepatan dalam struktur kalimat, d) kesatuan, kepaduan, dan kelengkapan dalam setiap paragraf. 2008: 21) menulis ialah menurunkan atau melukiskan lambang- lambang grafik yang menggambarkan suatu bahasa yang dipahamioleh seseorang sehingga orang lain dapat membaca lambang-lambang grafik tersebut kalau mereka memahami bahasa dan gambaran grafik itu. Selanjutnya, Djibran (2008: 17 ) mengatakan bahwa menulis adalah mengungkapkan pikiran, perasaan, pengalaman, dan hasil bacaan dalam bentuk tulisan, bukan dalam bentuk tutur. Rosidi (2009: 10-11) mengemukakan bahwa, tulisan yang baik kesesuaian judul dengan isi tulisan, b) ketepatan penggunaan ejaan dan tanda baca, c), ketepatan dalam struktur kalimat, d) kesatuan, kepaduan, dan kelengkapan dalam 3. Pedoman Umum Ejaan Bahasa Indonesia Pada tahun 2016 berdasarkan Keputusan Menteri Pendidikan dan Kebudayaan, Dr. Anis Baswedan, Umum Ejaan Bahasa Indonesia yang Disempurnakan (PUEYD) diganti dengan nama Pedoman Umum Ejaan Bahasa Indonesia (PUEBI) yang penyempurnaan naskahnya disusun oleh Pusat Pengembangan dan Pedoman Umum Ejaan Bahasa Indonesia Pada tahun 2016 berdasarkan Keputusan Menteri Pendidikan dan Kebudayaan, Dr. Anis Baswedan, Pedoman Umum Ejaan Bahasa Indonesia yang (PUEYD) diganti dengan Pedoman Umum Ejaan Bahasa Indonesia (PUEBI) yang penyempurnaan naskahnya disusun oleh Pusat Pengembangan dan Dari beberapa pendapat diatas, jelas bahwa menulis adalah suatu kegiatan seseorang dalam mengungkapkan ide, gagasan, atau buah pikiran melalui tulisan. Buah Dari beberapa pendapat diatas, jelas bahwa menulis adalah suatu kegiatan seseorang dalam mengungkapkan ide, gagasan, atau buah- pikiran melalui tulisan. Buah-pikiran tersebut 4 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 Pelindungan, Badan Pengembangan dan Pembinaan Bahasa. Pelindungan, Badan Pengembangan dan Kesalahan penggunaan huruf kapital yang ditemukan dalam kar kelas VII/B SMP Muhammadiyah Kupang berjumlah 235 kesalahan. Kesalahan tersebut meliputi: Kesalahan penggunaan huruf kapital yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 235 kesalahan. b. Kesalahan Penggunaan Tanda Baca Koma b. Kesalahan Penggunaan Tanda Baca Koma gg Kesalahan penggunaan tanda baca koma yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 197 kesalahan. Kesalahan itu terjadi karena kurangnya kemampuan siswa membedakan penggunaan tanda titik dan tanda koma. Sehingga siswa seringkali susah membedakan mana yang harus dibubuhi tanda titik dan mana yang harus dibubuhi tanda koma. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penggunaan huruf tanda baca koma merupakan kesalahan terbanyak kedua yang terdapat dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. gg Kesalahan penggunaan tanda baca koma yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang lahan. Kesalahan itu terjadi karena kurangnya kemampuan siswa membedakan penggunaan tanda titik dan tanda koma. Sehingga siswa seringkali susah membedakan mana yang harus dibubuhi tanda titik dan mana yang harus dibubuhi tanda koma. ut, dapat diketahui bahwa kesalahan penggunaan huruf tanda baca koma merupakan kesalahan terbanyak kedua yang terdapat dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. c. Kesalahan Penggunaan Tanda Titik Kesalahan Penggunaan Tanda Titik c. Kesalahan Penggunaan Tanda Titik Kesalahan Penggunaan Tanda Titik 1. Hasil Penelitian Hasil penelitian yang akan disajikan adalah berupa kesalahan penggunaan ejaan pada karangan siswa kelas VII/B SMP Muhammadiyah Kupang. Data yang dikumpulkandalam penelitian ini sebanyak 26 karangan siswa. Berdasarkan hasil analisis kesalahan pada karangan narasi siswa kelas VII SMP Muhammadiyah Kupang diperoleh 525 data kesalahan yang meliputi; a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, c) kesalahan pemakaian tanda baca koma berjumlah 197 data kasus kesalahan, dan d)kesalahan penulisan klitika ku; berjumlah 12 data kasus kesalahan. Hasil penelitian yang akan disajikan adalah berupa kesalahan penggunaan ejaan pada karangan siswa kelas VII/B SMP Muhammadiyah Kupang. Data yang dikumpulkandalam penelitian ini sebanyak 26 karangan siswa. Berdasarkan hasil analisis narasi siswa kelas VII SMP Muhammadiyah Kupang diperoleh 525 data kesalahan yang meliputi; a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, c) akaian tanda baca koma berjumlah 197 data kasus kesalahan, dan d)kesalahan penulisan klitika ku;-mu;-dan nya data kasus kesalahan. 1. Keterampilan Menulis Kesalahan-kesalahan Kamus Besar Bahasa Indonesia (2008: 20) menyatakan bahwa ejaan a kaidah cara menggambarkan bunyi (kata dan kalimat dsb) dalam bentuk tulisan ( huruf ) serta penggunaan tanda baca. Kamus Besar Bahasa Indonesia (2008: 20) menyatakan bahwa ejaan adalah kaidah- kaidah cara menggambarkan bunyi (kata dan kalimat dsb) dalam bentuk tulisan ( huruf-huruf ) serta penggunaan tanda baca. 1. Kesalahan pemakaian huruf pertama pada awal kalimat. Kesalahan pemakaian huruf pertama pada awal kalimat. 2. Kesalahan pemakaian huruf pertama nama Tuhan dan kitab suci. Kesalahan pemakaian huruf pertama nama Tuhan dan kitab suci. Menurut Kridalaksana dalam Kamus Linguistik edisi keempat (2008: 22) “Ejaan adalah penggambaran bunyi bahasa dengan kaidah tulis-menulis yang distandarisasikan. Yang mempunyai tiga aspek yakni aspek fonologis, yang menyangkut penggambaran bunyi fonem dengan huruf, dan penyusunan abjad, aspek morfologis yang menyangkut penggambaran satuan-satuan morfemis, aspek sintaksis menyangkut penanda ujaran berupa tanda baca. Jadi, ejaan adalah keseluruhan peraturan yang resmi yang dijadikan padoman dalam setiap penulisan karya ilmiah untuk merujuk kepada bahasa Indonesiayang baik dan benar. Menurut Kridalaksana dalam Kamus Linguistik edisi keempat (2008: 22) “Ejaan adalah penggambaran bunyi bahasa dengan menulis yang distandarisasikan. Yang mempunyai tiga aspek yakni aspek fonologis, yang menyangkut penggambaran bunyi fonem dengan huruf, dan penyusunan abjad, aspek morfologis yang menyangkut satuan morfemis, aspek ngkut penanda ujaran berupa tanda baca. Jadi, ejaan adalah keseluruhan peraturan yang resmi yang dijadikan padoman dalam setiap penulisan karya ilmiah untuk merujuk kepada bahasa Indonesiayang baik dan 3. Kesalahan penulisan hur tempat/geografi. Kesalahan penulisan huruf pertama nama 4. Kesalahan pemakaian huruf pertama nama orang. Kesalahan pemakaian huruf pertama 5. Kesalahan pemakaian huruf pertama judul karangan. Kesalahan pemakaian huruf pertama 6. Kesalahan pemakaian huruf pertama kekerabatan. Kesalahan pemakaian huruf pertama Kesalahan tersebut terjadi karena kurangnya pengetahuan siswa tentang penjelasan dan contoh- kapital yang benar sehingga siswa seringkali menempatkan huruf kapital tidak berdasarkan kaidah-kaidahnya. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penggunaan huruf kapital merupakan kesalahan yang pali banyak terjadi dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. Kesalahan tersebut terjadi karena kurangnya pengetahuan siswa tentang -contoh penulisan huruf kapital yang benar sehingga siswa seringkali menempatkan huruf kapital tidak berdasarkan kaidahnya. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penggunaan huruf kapital merupakan kesalahan yang paling banyak terjadi dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. 2. Pembahasan a. Kesalahan Penggunaan Huruf Kapital a. Kesalahan Penggunaan Huruf Kapital a. Kesalahan Penggunaan Huruf Kapital a. Kesalahan Penggunaan Huruf Kapital 5 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 Dari 26 karangan siswa yang dianalisis rata-rata setiap karangan ditemukan 10 kesalahan ejaan. Melihat hal tersebut guru bahasa Indonesia hendaknya mencari pembelajaran baru yang menuntut siswa untuk menemukan sendiri kesalahan ejaan. Melihat banyak di temukannya kesalahan khususnya kesalahan penggunaan ejaan, hendaknya guru selalu memberikan perhatian tentang kesalahan ejaan yang dibuat siswa, dan melibatkan mareka dalam memperbaiki kesalahan tersebut. Dari 26 karangan siswa yang dianalisis rata setiap karangan ditemukan 10 kesalahan rsebut guru bahasa Indonesia hendaknya mencari pembelajaran baru yang menuntut siswa untuk menemukan sendiri kesalahan ejaan. Melihat banyak di temukannya kesalahan khususnya kesalahan penggunaan ejaan, hendaknya guru selalu memberikan salahan ejaan yang dibuat siswa, dan melibatkan mareka dalam memperbaiki kesalahan tersebut. Kesalahan penggunaan tanda baca titk yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 93 kesalahan. Kesalahan itu terjadi karena kurangnya kemampuan siswa membedakan penggunaan tanda titik dan tanda koma. Sehingga siswa seringkali susah membedakan mana yang harus dibubuhi tanda titik dan mana yang harus dibubuhi tanda koma. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penggunaan huruf tanda baca titik merupakan kesalahan terbanyak ketiga yang terdapat dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. Kesalahan penggunaan tanda baca titk yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 93 kesalahan. Kesalahan itu terjadi karena kurangnya kemampuan siswa membedakan penggunaan tanda titik dan tanda koma. Sehingga siswa seringkali susah yang harus dibubuhi tanda titik dan mana yang harus dibubuhi tanda koma. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penggunaan huruf tanda baca titik merupakan kesalahan terbanyak ketiga yang terdapat dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. F. DAFTAR PUSTAKA F. DAFTAR PUSTAKA Aristohadi, Sutopo. 2003. dan Flash. Yogyakarta: Graha Ilmu. Aristohadi, Sutopo. 2003. Multimedia Interaktif Yogyakarta: Graha Ilmu. d. Kesalahan Penulisan Klitika d. Kesalahan Penulisan Klitika Kesalahan penulisan klitika yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 11 kesalahan. Kesalahan tersebut terjadi disebabkan karena kurangnya pengetahuan tentang penjelasan dan contoh klitika yang benar sehingga siswa seringkali keliru menuliskan klitika tidak berdasarkan kaidah-kaidahnya. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penulisan klitika merupakan kesalahan dalam karangan siswa siswa kelas VII/B SMP Muhammadiyah Kupang. Kesalahan penulisan klitika yang ditemukan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 11 kesalahan. Kesalahan tersebut terjadi disebabkan karena kurangnya pengetahuan siswa tentang penjelasan dan contoh-contoh penulisan klitika yang benar sehingga siswa seringkali keliru menuliskan klitika tidak berdasarkan kaidahnya. Berdasarkan data tersebut, dapat diketahui bahwa kesalahan penulisan yang palang sedikit dalam karangan siswa siswa kelas VII/B SMP Arikunto, Suharsemi. 2006. Suatu Pendekatan Praktek. Rineka Cipta. Arikunto, Suharsemi. 2006. Prosedur Penelitian Pendekatan Praktek.Jakarta: Chaer, Abdul. 2006. Tata Bahasa Praktis Bahasa Indonesia. Jakarta Tata Bahasa Praktis Bahasa Indonesia. Jakarta: PT Rineka. Dendy, Sugono. 2010. Kamus Bahasa Indonesia Sekolah Dasar. Jakarta: PTGramedia Utama. Kamus Bahasa Indonesia . Jakarta: PTGramedia Djibran, Fahrudin. 2008. Yogyakarta: Juxtapose. Djibran, Fahrudin. 2008. Writing Is Amazing. Yogyakarta: Juxtapose. Enre, Fachrudin Ambo. 1998. Keterampilan Menulis. Enre, Fachrudin Ambo. 1998. Dasar-Dasar Keterampilan Menulis. Jakarta:Kanisius. E. Zaenal Arifin, dan S. Amran Tasai. 2008. Cermat Berbahasa Indonesia Jakarta:Akademika Pressindo. E. Zaenal Arifin, dan S. Amran Tasai. 2008. Cermat Berbahasa Indonesia. Jakarta:Akademika Pressindo. E. KESIMPULAN Berdasarkan hasil penelitian dan pembahsan maka dapat disimpulkan bahwa jumlah kesalahan penggunaan ejaan dalam karangan narasi siswa kelas VII/B SMP Muhammadiyah Kupang berjumlah 525 kesalahan. Bentuk ejaan dibedakan menjadi enam aspek meliputi; a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, c) kesalahan p tanda baca koma berjumlah 197 data kasus kesalahan, dan d) kesalahan penulisan klitika ku; mu;-dan nya berjumlah 11 data kasus kesalahan. Berdasarkan hasil penelitian dan pembahsan maka dapat disimpulkan bahwa jumlah kesalahan penggunaan ejaan dalam karangan narasi siswa kelas VII/B SMP ammadiyah Kupang berjumlah 525 kesalahan. Bentuk ejaan dibedakan menjadi enam aspek meliputi; a) kesalahan pemakaian huruf kapital berjumlah 235 data kasus kesalahan, b) kesalahan pemakaian tanda baca titik berjumlah 93 data kasus kesalahan, c) kesalahan pemakaian tanda baca koma berjumlah 197 data kasus kesalahan, dan d) kesalahan penulisan klitika ku;- dan nya berjumlah 11 data kasus kesalahan. Keraf, Goys. 1994. Komposisi Indah.2007. Argumentasi dan Narasi Jakarta: Gramedia Pustaka. Komposisi. Ende: Nusa . Argumentasi dan Narasi. Jakarta: Gramedia Pustaka. Kridalaksana, Harimurti. Bahasa Indonesia Kridalaksana, Harimurti. 1984. Kelas Kata dalam Bahasa Indonesia. Jakarta: Mahsun. 2012. Metode Penelitian Bahasa: Tahapan, Strategi, dan, Tekniknya Jakarta: Raja Grafindo. Metode Penelitian Bahasa: Tahapan, Strategi, dan, Tekniknya. Jakarta: Raja Grafindo. Moleong, Lexy. J. 2006. Kualitatif. Bandung: Rosda. Moleong, Lexy. J. 2006. Metodologi Penelitian . Bandung: Rosda. Moleong, Lexy. J. 2006. Kualitatif. Bandung: Rosda. Moleong, Lexy. J. 2006. Metodologi Penelitian . Bandung: Rosda. Moeliono, Anton. 1984. Balai Pustaka. Moeliono, Anton. 1984. Santun Bahasa. Jakarta: Moeliono, Anton. 1984. Balai Pustaka. Moeliono, Anton. 1984. Santun Bahasa. Jakarta: 6 Jurnal Ilmu Pendidikan (JIP), Vol. p-ISSN Jurnal Ilmu Pendidikan (JIP), Vol. 7, No.1, Edisi: ISSN 2477-1287 e-ISSN 2745-7516 , Edisi: April 2022 Nuruddin. 2010. Dasar Malang: UMM Press. Dasar-dasar Penulisan. Malang: UMM Press. Nuruddin. 2010. Dasar Malang: UMM Press. Nursisto. 1999. Penuntun Mengarang Yogyakarta: Adi Cita. Dasar-dasar Penulisan. Malang: UMM Press. Penuntun Mengarang. Yogyakarta: Adi Cita. g Nursisto. 1999. Penuntun Mengarang Yogyakarta: Adi Cita. g Penuntun Mengarang. Yogyakarta: Adi Cita. Pusat Bahasa Kemdiknas. 2008. Bahasa Indonesia. Jakarta: Gramedia. 2008. Kamus Besar . Jakarta: Gramedia. Rosidi, Imron. 2009. Menulis Siapa Takut Yogyakarta: Kanisius. Menulis Siapa Takut. Yogyakarta: Kanisius. Yogyakarta: Kanisius. Yogyakarta: Kanisius. Ramlan, Abdul Gani dan Mahmudah, amlan, Abdul Gani dan Mahmudah, Fitriyah. 2007. Pembinaan Bahasa Indonesia. Jakarta: Gramedia. Pembinaan Bahasa Indonesia. Sabarti , Akhadiah. 1993. Kemampuan Menulis Bahasa Indonesia Jakarta: Erlangga. Sabarti , Akhadiah. 1993. E. KESIMPULAN Pembinaan Kemampuan Menulis Bahasa Indonesia. Sandjaja, dan Heriyanto. 2006. Penelitian. Jakarta: Prasasti Pustaka. Sandjaja, dan Heriyanto. 2006. Panduan . Jakarta: Prasasti Pustaka. Semi, M. Atar. 1990. Menulis Efektif Angkasa Raya. Menulis Efektif. Padang: Suparno, dan M. Yunus. 2012. Keterampilan Berb Bandung: CV Karya Putra Darwati. Suparno, dan M. Yunus. 2012. Meningkatkan Keterampilan Berbahasa Indonesia. Bandung: CV Karya Putra Darwati. Tarigan, Hery Guntur. 1997. Berbahasa. Jakarta: Depdikbud. Tarigan, Hery Guntur. 1997. Analisis Kesalahan . Jakarta: Depdikbud. 7
https://openalex.org/W4280592701	https://journals.kozminski.edu.pl/system/files/Kwiatkowska_0.pdf	Polish	null	Cyfryzacja ochrony zdrowia a konieczność wprowadzania nowych rozwiązań prawnych	Krytyka Prawa	2,022	cc-by	7,022	Cyfryzacja ochrony zdrowia a konieczność wprowadzania nowych rozwiązań prawnych2 Wpłynął: 12.05.2021. Akceptacja: 20.02.2022 Wpłynął: 12.05.2021. Akceptacja: 20.02.2022 1 Dr Ewa M. Kwiatkowska – Akademia Leona Koźmińskiego (Polska); e-mail: ekwiatkowska@kozmin- ski.edu.pl; ORCID: 0000-0001-7576-1996. 2 Badania wykorzystane w artykule nie zostały sfinansowane przez żadną instytucję. „Krytyka Prawa”, tom 14, nr 1/2022, s. 154–170, ISSN 2080-1084, e-ISSN 2450-7938, © 2022 Author. This is an open access article distributed under the Creative Commons BY 4.0 license: http://creativecommons.org/licenses/by/4.0 „Krytyka Prawa”, tom 14, nr 1/2022, s. 154–170, ISSN 2080-1084, e-ISSN 2450-7938, © 2022 Author. This is an open access article distributed under the Creative Commons BY 4.0 license: http://creativecommons.org/licenses/by/4.0 „Krytyka Prawa”, tom 14, nr 1/2022, s. 154–170, ISSN 2080-1084, e-ISSN 2450-7938, © 2022 Author. This is an open access article distributed under the Creative Commons BY 4.0 license: http://creativecommons.org/licenses/by/4.0 EWA M. KWIATKOWSKA1 3 The research in this article has not been supported financially by any institution. Streszczenie Konieczność zapewnienia ochrony zdrowia obywatelom jest jedną z podstawo- wych zasad każdego państwa, regulowaną przepisami prawa zarówno krajowego, jak i unijnego czy międzynarodowego i uzupełnianą przez soft law. Zmiany cywi- lizacyjne i postęp technologiczny umożliwiają rozwój nowych narzędzi i usług z zakresu ochrony zdrowia (m.in. telemonitoringu, telekonsultacji, telerehabilita- cji oraz elektronicznej dokumentacji medycznej). Pandemia COVID-19 w znacznym stopniu podkreśliła potrzebę zapewnienia tego typu usług. Szerokie stosowanie e-zdrowia, które może być remedium na wiele problemów, z którymi boryka się sektor ochrony zdrowia, wymaga jednak zabezpieczenia praw osób, które są nie są włączone cyfrowo, szczególnie osób starszych. Powszechne występowanie różnych poziomów wykluczenia cyfrowego, nie tylko wśród seniorów, sprawia, że kwestie te są społecznie istotne i powinny być rozwiązywane przez prawo. Celem artykułu jest wstępna analiza zasadności rozwijania usług e-zdrowia w starzejącym się społeczeństwie w Polsce oraz przedstawienie propozycji dosto- sowania rozwiązań prawnych w zakresie wykorzystania narzędzi cyfrowych w ochronie zdrowia do zmieniającej się rzeczywistości. Słowa kluczowe: technologie informacyjno-komunikacyjne (ICT), ochrona zdrowia, osoby starsze, wykluczenie cyfrowe, telemedycyna. Słowa kluczowe: technologie informacyjno-komunikacyjne (ICT), ochrona zdrowia, osoby starsze, wykluczenie cyfrowe, telemedycyna. 1 Dr Ewa M. Kwiatkowska – Akademia Leona Koźmińskiego (Polska); e-mail: ekwiatkowska@kozmin- ski.edu.pl; ORCID: 0000-0001-7576-1996. 2 Badania wykorzystane w artykule nie zostały sfinansowane przez żadną instytucję. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 „Krytyka Prawa”, tom 14, nr 1/2022, s. 154–170, ISSN 2080-1084, e-ISSN 2450-7938, © 2022 Author. This is an open access article distributed under the Creative Commons BY 4.0 license: http://creativecommons.org/licenses/by/4.0 EWA M. KWIATKOWSKA EWA M. KWIATKOWSKA Abstract The necessity of providing healthcare to citizens is one of the basic principles of every state, regulated by provisions of national, EU or international law and comple mented by soft law. Civilisational changes and technological advancement make it possible to develop new healthcare tools and services (i.a. telemonitoring, telecon sulting, telerehabilitation and electronic medical documentation). The COVID-19 pandemic greatly emphasised the need to provide such services. The wide use of e-health, however, which may be a remedy for numerous problems that the health- care sector is wrestling with, requires securing the rights of those who are not digitally included, especially the elderly. The prevalence of various degrees of digital exclusion, not only among the elderly, makes these issues socially significant and they should be solved by law. The purpose of the article is an initial analysis of the validity of developing e-health services in the ageing society in Poland and a presentation of proposals for adjusting legal solutions as regards the use of digital tools in healthcare to the changing reality. Keywords: information and communications technologies (ICT), healthcare, the elderly, digital exclusion, telemedicine. DOI: 10.7206/kp.2080-1084.513 DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 156 Ewa M. Kwiatkowska 156 Ewa M. Kwiatkowska 4 W tekście używane są wymiennie: „senior”, „osoba starsza” i „osoba w podeszłym wieku” na określenie konstytucyjnego terminu „osoba w podeszłym wieku”. 5 Szerzej o wykluczeniu cyfrowym: A.P. Neto, M.B. Flynn, The Internet and Health in Brazil: Trends and Challenges, [w:] iidem (red.), The Internet and Health in Brazil: Challenges nad Trends, Springer 2019, s. 5 i n.; E.M. Kwiatkowska, Transformacja cyfrowa a osoby starsze w systemie ochrony zdrowia. O potrzebie implemen- towania nowych rozwiązań prawnych, „Transformacje” 2011, 108(1), s. 164 i n. 6 J.L.H. Birkland, Gerontechnology. Understanding Older Adult Information and Communication Technology Use, Emerald Publishing 2019, s. 4. 6 J.L.H. Birkland, Gerontechnology. Understanding Older Adult Information and Communication Technology Use, Emerald Publishing 2019, s. 4. 4 W tekście używane są wymiennie: „senior”, „osoba starsza” i „osoba w podeszłym wieku” na określenie konstytucyjnego terminu „osoba w podeszłym wieku”. Wprowadzenie Konieczność zapewnienia ochrony zdrowia obywatelom, a szczególnie osobom starszym4 zawarta jest w krajowych, unijnych oraz międzynarodowych przepisach prawa i niewiążących aktach normatywnych (soft law). Z uwagi na obserwowane nie tylko w Polsce zmiany demograficzne związane ze starzeniem się społeczeń- stwa jest to problematyka szczególnie istotna i uniwersalna. Projektując rozwią- zania z zakresu ochrony zdrowia, szczególnie rozwiązania cyfrowe, których celem jest usprawnienie funkcjonowania w systemie ochrony zdrowia zarówno świad- czeniodawcom, jak i świadczeniobiorcom, organy państwa powinny zwrócić szczególną uwagę na grupy obywateli najbardziej narażone na wykluczenie cyfrowe5. Uwaga ta powinna być skierowana przede wszystkim na osoby starsze, które częściej od innych korzystają z usług ochrony zdrowia. Rzadziej od młodszych korzystają natomiast z technologii informacyjno-komunikacyjnych, w tym kom- puterów i smartfonów6. Nowoczesne rozwiązania w sektorze ochrony zdrowia powinny być tak skonstruowane, aby rzeczywiście ułatwiać, a nie utrudniać oby- watelom funkcjonowanie w społeczeństwie i zaspokajanie podstawowych potrzeb, do których bezwzględnie należy właśnie ochrona zdrowia. Instrumenty te powinny zapewniać bezpieczeństwo pacjenta – oznaczające co najmniej zapewnienie pouf- ności dotyczących go danych wrażliwych – przy jednoczesnym dopuszczeniu świadczeniodawców do informacji niezbędnych w procesie leczenia. Projektowa- nie rozwiązań cyfrowych oraz regulujących je rozwiązań prawnych w sektorze ochrony zdrowia wymaga zagwarantowania poszanowania praw jednostki. Konieczne jest ich osadzenie w podlegającym nieustannym zmianom środowisku zgodnie z ideami zrównoważonego rozwoju. Występowanie, nie tylko wśród seniorów, różnych poziomów wykluczenia cyfrowego, sprawia, że problem ten jest społecznie istotny i powinien być rozwiązywany przez prawo. Należy przeciwdziałać DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 157 rosnącym nierównościom tworzącym się pomiędzy tymi, którzy posiadają i korzy- stają z technologii, a tzw. „technologiczną biedotą”, która może stać się pozbawioną dostępu do pewnych świadczeń7. Nie można dopuścić do sytuacji, w której kto- kolwiek byłby dyskryminowany m.in. z powodu braku umiejętności cyfrowych, czy też dostępu do Internetu8. Celem opracowania jest rozpoznanie i wskazanie konieczności dostosowania do zmieniającej się rzeczywistości regulacji prawnych w zakresie wykorzystania narzędzi cyfrowych w ochronie zdrowia w szczegól- ności do potrzeb osób starszych, często wykluczonych cyfrowo. Dokonana zostanie wieloaspektowa analiza sytuacji bieżącej, na którą niebagatelny wpływ ma panu- jąca pandemia COVID-19, oraz przedstawione zostaną propozycje dostosowania rozwiązań prawnych do zmian zachodzących we współczesnym świecie. 7 J. Bartlett, Ludzie przeciw technologii. Jak Internet zabija demokrację (i jak ją możemy ocalić), Katowice 2019, s. 185–186. 7 J. Bartlett, Ludzie przeciw technologii. Jak Internet zabija demokrację (i jak ją możemy ocalić), Katowice 2019, s. 185–186. 8 J. Janowski, Cyberkultura prawa. Współczesne problemy filozofii i informatyki prawa, Warszawa 2012, s. 326. 9 M. Balicki, Prezentacja koncepcji Centrum Zdrowia 75+, Konferencja „Centrum Zdrowia 75+. Zdążyć przed demograficznym tsunami”, Warszawa, 10.09.2019, https://www.prezydent.pl/aktualnosci/wy- darzenia/art,1530,prezydent-na-konferencji-centrum-zdrowia-75.html, (dostęp: 21.11.2020). 10 Centrum Zdrowia 75+, Zarys koncepcji opracowanej przez ekspertów z Narodowej Rady Rozwoju przy Prezy- dencie RP, Warszawa 2019, https://www.prezydent.pl/aktualnosci/wydarzenia/art,1530,prezydent-na- -konferencji-centrum-zdrowia-75.html (dostęp: 21.11.2020). 10 Centrum Zdrowia 75+, Zarys koncepcji opracowanej przez ekspertów z Narodowej Rady Rozwoju przy Prezy- dencie RP, Warszawa 2019, https://www.prezydent.pl/aktualnosci/wydarzenia/art,1530,prezydent-na- -konferencji-centrum-zdrowia-75.html (dostęp: 21.11.2020). J. Janowski, Cyberkultura prawa. Współczesne problemy filozofii i informatyki prawa, Warszawa 20 Zrównoważony rozwój w ochronie zdrowia W Polsce prawo do ochrony zdrowia wynika wprost z art. 68 Konstytucji RP. Konstytucja gwarantuje równy dostęp do świadczeń opieki zdrowotnej finansowa nych ze środków publicznych, nakładając na władze publiczne obowiązek zapew- nienia szczególnej opieki zdrowotnej dzieciom, kobietom ciężarnym, osobom niepełnosprawnym i osobom w podeszłym wieku. Istnieje zatem konstytucyjny obowiązek zapewnienia m.in. osobom, które są w podeszłym wieku rzeczywiście dostępnej opieki zdrowotnej, która będzie zorganizowana na innych zasadach niż opieka powszechna9. W chwili obecnej osobom starszym nie jest zapewniania taka szczególna opieka zdrowotna, choć podejmowane są prace nad określeniem założeń kompleksowej opieki nad seniorami, m.in. w ramach przygotowywanej przez ekspertów Narodowej Rady Rozwoju przy Prezydencie RP koncepcji Cen- trum Zdrowia 75+10. W przyjętej przez Radę Ministrów w 2017 r. Strategii na rzecz Odpowiedzial- nego Rozwoju do roku 2020 (z perspektywą do 2030 r.) podkreślono, że w związku z wyzwaniami współcześnie występującej ery globalizacji pojawiają się nie tylko DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 Ewa M. Kwiatkowska 158 szanse, ale i zagrożenia. Obecnie trwa czwarta rewolucja przemysłowa, co oznacza m.in. automatyzację produkcji, występowanie rewolucji cyfrowej, robotyzację oraz postęp technologiczny. W przyszłości zjawiska te mogą sprzyjać tworzeniu napięć na rynku pracy, a także występowaniu tarć społecznych11. Ten stan rzeczy nie pozostanie zapewne bez wpływu na sektor ochrony zdrowia, szczególnie ważny dla osób starszych, niejednokrotnie cierpiących na wielochorobowość, samotnych i w dużej mierze wykluczonych cyfrowo. Według danych zebranych w badaniu PolSenior2 w 2018–2019 r. blisko 57% osób w wieku 60+ nie używa Internetu12. Zgodnie z Agendą na rzecz zrównoważonego rozwoju 203013 przyjętą w 2015 r. na forum ONZ promocja zdrowia fizycznego, psychicznego, dobrobytu oraz dążenie do wydłużania oczekiwanej długości życia wymaga co najmniej zapewnienia wszyst- kim ludziom, bez pomijania kogokolwiek, powszechnej ochrony zdrowia i dostępu do opieki zdrowotnej wysokiej jakości14. W przyjętej w 2010 r. unijnej strategii na rzecz inteligentnego i zrównoważonego rozwoju sprzyjającego włączeniu społecz nemu15 określono zrównoważony rozwój jako wspieranie gospodarki coraz efek- tywniej wykorzystującej zasoby, bardziej konkurencyjnej i przyjaznej środowisku. Sprzyjać temu będzie m.in. przyspieszanie wprowadzania inteligentnych sieci teleinformatycznych, co umożliwi m.in. zapobieganie degradacji środowiska16. 16 Ibidem, s. 16. 11 M. Morawiecki, Słowo wstępne do Strategii na rzecz Odpowiedzialnego Rozwoju do roku 2020 (z perspektywą do 2030 r.), Warszawa 2017, s. 2. 12 E.M. Kwiatkowska, Wykluczenie cyfrowe, Konferencja ogólnopolska PolSenior2 – moduł 1, sytuacja socjalno-ekonomiczna osób starszych, 23.11.2020, https://konferencjapolsenior2.pl/ (dostęp: 12.05.2021). 13 ONZ, Przekształcamy nasz świat: Agenda na rzecz zrównoważonego rozwoju 2030, https://www.gov.pl/web/ rozwoj-praca-technologia/agenda-2030 (dostęp: 20.11.2020). 14 Ibidem, pkt. 26. 15 Komunikat Komisji: „Europa 2020. Strategia na rzecz inteligentnego i zrównoważonego rozwoju sprzyjającego włączeniu społecznemu”, Bruksela, 3.3.2010, KOM(2010) 2020 wersja ostateczna. 16 Ibidem, s. 16. 17 C. Marolla, Information and Communication Technology for Sustainable Development, Boca Raton–London– New York 2019, s. 94; S. Smith, The Internet of Risky Things. Trusting the Devices That Surround Us, O’Reilly Media 2017, s. 211. 15 Komunikat Komisji: „Europa 2020. Strategia na rzecz inteligentnego i zrównoważonego rozwoju sprzyjającego włączeniu społecznemu”, Bruksela, 3.3.2010, KOM(2010) 2020 wersja ostateczna. 11 M. Morawiecki, Słowo wstępne do Strategii na rzecz Odpowiedzialnego Rozwoju do roku 2020 (z perspektywą do 2030 r.), Warszawa 2017, s. 2. 17 C. Marolla, Information and Communication Technology for Sustainable Development, Boca Raton–London– New York 2019, s. 94; S. Smith, The Internet of Risky Things. Trusting the Devices That Surround Us, O’Reilly Media 2017, s. 211. 13 ONZ, Przekształcamy nasz świat: Agenda na rzecz zrównoważonego rozwoju 2030, https://www.gov.pl/web/ rozwoj-praca-technologia/agenda-2030 (dostęp: 20.11.2020). 22 Komunikat Komisji do Parlamentu Europejskiego, Rady, Europejskiego Komitetu Ekonomiczno-Spo- łecznego i Komitetu Regionów w sprawie umożliwienia transformacji cyfrowej opieki zdrowotnej i społecznej na jednolitym rynku cyfrowym; wzmacniania pozycji obywateli i budowania zdrowszego społeczeństwa, Bruksela, 25.04.2018, COM(2018) 233 final, s. 1. Zrównoważony rozwój w ochronie zdrowia W odniesieniu do ochrony zdrowia wykorzystywanie technologii informacyjno- -komunikacyjnych (ICT), a ściślej usług telemedycznych, takich jak telekonsulta- cje, telemonitoring czy telerehabilitacja, może nie tylko zdecydowanie ułatwić życie pacjentom, którzy w wielu wypadkach nie będą musieli opuszczać swojego miejsca zamieszkania i dojeżdżać w celu wykonania podstawowych procedur medycz- nych, ale również sprawić, że będą czuli się bezpieczniej, zwiększy się jakość ich życia, a także będą niezależni od osób trzecich17. Jest to szczególnie istotne dla osób starszych, którym powinno zapewnić się przede wszystkim opiekę zdrowotną w ich miejscu zamieszkania, zaś jedynie w niezbędnych przypadkach kierować 17 C. Marolla, Information and Communication Technology for Sustainable Development, Boca Raton–London– New York 2019, s. 94; S. Smith, The Internet of Risky Things. Trusting the Devices That Surround Us, O’Reilly Media 2017, s. 211. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 159 na leczenie specjalistyczne i szpitalne18. Świadczenia telemedyczne, w tym zdalne monitorowanie – dostępne dzięki narzędziom Internetu Rzeczy, pozwala osobom starszym i schorowanym pozostawać dłużej we własnym domu czy mieszkaniu19. Należy przy tym zadbać, aby brak umiejętności korzystania z ICT nie wykluczał nikogo z dostępu do usług zdrowotnych. Zgodnie z komunikatem Komisji Euro- pejskiej z 2017 r. zamierza ona w ramach wsparcia celów zrównoważonego rozwoju upowszechnić wykorzystywanie ICT20. Mogą one być pomocne w poprawie stanu zdrowia ludności, a także wspomóc systemowe wyzwania związane z opieką zdrowotną. Mogą dostarczać kosztowo opłacalne narzędzia pozwalające na prze- chodzenie od systemów szpitalnej opieki zdrowotnej do pacjentocentrycznej, dostępnej opieki zintegrowanej21. W efekcie takie działania zgodnie z zasadami zrównoważonego rozwoju mają pozwolić pacjentom na pozostawanie w swoim środowisku tak długo, jak to tylko możliwe. 18 A. Malarewicz-Jakubów, Wsparcie prawne osób starszych, Warszawa 2017, s. 112. 19 R.H. Weber, R. Weber, Internet of Things. Legal Perspectives, Springer 2010, s. 123. 20 Komunikat Komisji do Parlamentu Europejskiego, Rady, Europejskiego Komitetu Ekonomiczno-Spo- łecznego i Komitetu Regionów w sprawie przeglądu śródokresowego realizacji strategii jednolitego rynku cyfrowego. Połączony jednolity rynek cyfrowy dla wszystkich, Bruksela, 10.05.2017, COM(2017) 228 final, s. 28. 21 Ibidem, s. 22. Nowe technologie remedium na problem starzenia się społeczeństwa Przed europejskimi systemami opieki zdrowotnej oraz społecznej stoją poważne wyzwania obejmujące m.in. starzenie się społeczeństw, wielochorobowość oraz braki kadrowe wśród pracowników ochrony zdrowia. Odpowiedzią na te wyzwa- nia może być wprowadzenie rozwiązań cyfrowych przyczyniających się do poprawy dobrostanu obywateli oraz radykalnie zmieniających sposób świadczenia usług zdrowotnych i opiekuńczych22. Cyfryzacja pozwala przekształcić systemy opieki zdrowotnej z opartych w głównej mierze na opiece szpitalnej na bardziej środowiskowe i pozostające blisko pacjenta. Wykorzystanie ICT może spowodować zmniejszenie zapotrzebowania na konsultacje w przychodniach czy szpitalach na rzecz tych świadczonych telefonicznie. 22 Komunikat Komisji do Parlamentu Europejskiego, Rady, Europejskiego Komitetu Ekonomiczno-Spo- łecznego i Komitetu Regionów w sprawie umożliwienia transformacji cyfrowej opieki zdrowotnej i społecznej na jednolitym rynku cyfrowym; wzmacniania pozycji obywateli i budowania zdrowszego społeczeństwa, Bruksela, 25.04.2018, COM(2018) 233 final, s. 1. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 160 Ewa M. Kwiatkowska Jednym z celów polskiej polityki społecznej dotyczącej seniorów jest podnosze nie jakości ich życia przez stworzenie możliwości pozostawania możliwie jak najdłużej samodzielnymi, a także zapewnienia bezpieczeństwa23. Realizację tego celu umożliwia m.in. wykorzystywanie nowoczesnych technologii, w tym zdalne monitorowanie stanu zdrowia pacjenta, stosowanie instrumentów telemedycyny i teleopieki24. Takie działanie będzie nie tylko tańsze i bardziej efektywne z punktu widzenia systemu opieki zdrowotnej, ale także bardziej pożądane społecznie z punktu widzenia jakości życia seniorów, którzy najlepiej czują się w znanym sobie środowisku, a więc przede wszystkim we własnym domu. Dlatego też tworząc modelowy system wsparcia seniorów należy pamiętać o ich ograniczeniach, wyni- kających m.in. z braku umiejętności cyfrowych. Planowane i wdrażane rozwiązania muszą być tak skonstruowane, aby nie pogłębiać ich izolacji i niepewności. Nie można dopuścić do sytuacji, w której informacje lub usługi będą dostępne tylko dla osób posługujących się siecią, zaś dostępne w ograniczonym zakresie lub w ogóle niedostępne dla e-wykluczonych. Istotne jest, aby informacje, czyli treści przeka- zywane nadawcy przez odbiorcę, mogły zostać przez każdego adresata odebrane25. y y p g y p g Zapewnienie powszechnego dostępu do usług zdrowotnych ma kluczowe znaczenie w redukowaniu nierówności w zdrowiu. Usługi te muszą być dostępne zawsze i wszędzie tam gdzie ludzie ich potrzebują, także w małych miejscowoś ciach i na obszarach wiejskich, gdzie szczególnie widoczne są braki kadrowe wśród lekarzy. Systemy opieki zdrowotnej powinny być projektowane w sposób elastyczny, tak aby można było skutecznie reagować na zmieniające się potrzeby w zakresie opieki zdrowotnej wynikające ze zmian demograficznych i lepiej wykorzystywać potencjał nowych technologii cyfrowych w celu wzmocnienia profilaktyki i opieki zdrowotnej26. 23 Uchwała nr 161 Rady Ministrów z dnia 26 października 2018 r. w sprawie przyjęcia dokumentu Polityka społeczna wobec osób starszych 2030. Bezpieczeństwo – Uczestnictwo – Solidarność, Warszawa 30.11.2018, M.P. poz. 1169, s. 5. 24 Ibidem, s. 12. 25 T. Pszczołowski, Organizacja od dołu i od góry, Warszawa 1978, s. 285–286 i 290. 26 OECD/EU, Health at a Glance: Europe 2018: State of Health in the EU Cycle, DOI: 10.1787/health_ glance_eur-2018-en, s. 13. Nowe technologie remedium na problem starzenia się społeczeństwa Pandemia COVID-19 pokazała, że rozwiązania telemedyczne są nie tylko moż- liwe do wprowadzenia, ale wręcz konieczne. O ile w Europie wprowadzanie roz- wiązań z zakresu telemedycyny przed pandemią było ograniczone, o tyle podczas jej trwania rozwiązania te w wielu państwach były szeroko stosowane. Tam, gdzie świadczenie tego typu usług było już dozwolone przed pandemią (np. Polska i Fran- cja) ułatwiono świadczeniodawcom i świadczeniobiorcom korzystanie z konsul- tacji zdalnych. W Polsce w trakcie trwania pierwszej fali pandemii około 80% Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 161 konsultacji było świadczonych zdalnie27. Na mocy nowelizacji dokonanej ustawą z 9 października 2015 r. o zmianie ustawy o systemie informacji w ochronie zdro- wia oraz niektórych innych ustaw28, umożliwiono lekarzom świadczenie usług zdrowotnych na odległość za pośrednictwem systemów teleinformatycznych29. y g p y y y Wyzwaniem nie tylko dla Polski, ale i większości państw europejskich jest starzenie się społeczeństw, powodujące konieczność zwiększania nakładów na opiekę zdrowotną30. Według danych Eurostatu w 2019 r. udział osób w wieku 65+ w UE wynosił 20%, zaś osób w wieku 80+ 5,7%31. W Polsce wg danych z 30.06.2020 r. udział seniorów (osób w wieku 65+) w populacji wyniósł 18,4% (tzw. wskaźnik starości), a osoby w wieku 80+ (tzw. grupa osób w sędziwym wieku) stanowiły 4,44% społeczeństwa32. Jak widać, wskaźniki te dla Polski są niższe niż średnie dla państw członkowskich. Jednak od 2000 r. nastąpił ponad dwukrotny wzrost liczby osób najstarszych (w 2000 r. osoby w wieku 80+ stanowiły 2% ludności)33. Prognozy wskazują, że w 2070 r. w UE udział osób w wieku 65+ ma już być bliski 29%, zaś 80+ bliski 13%. W Polsce natomiast w 2070 r. ma być ponad 33% osób w wieku 65+ i ponad 16% w wieku 80+34. Należy podkreślić, że Polska spośród krajów UE jest najszybciej starzejącym się krajem35. Co więcej, według prognoz GUS ponad 53,3% jednoosobowych gospodarstw ma być prowadzonych przez osoby w wieku 65+ (ponad 2,7 mln gospodarstw), zaś blisko 17,3% przez seniorów w wieku 80+ (blisko 0,89 mln gospodarstw)36. Starzenie się społeczeństwa charak- 28 Dz.U. poz. 1991 ze zm. 29 Szerzej: E.M. Kwiatkowska, Internet of Things – wykorzystanie nowych technologii w ochronie zdrowia, [w:] Ł. Sułkowski, D. Kaczorowska-Spychalska (red.), Internet of Things. Nowy paradygmat rynku, Warszawa 2018, s. 185–186. 30 Strategia na rzecz Odpowiedzialnego Rozwoju do roku 2020 (z perspektywą do 2030 r.), https://www. gov.pl/documents/33377/436740/SOR.pdf (dostęp: 20.11.2020), s. 16. 27 OECD/European Union, Health at a Glance: Europe 2020: State of Health in the EU Cycle, DOI: 10.1787/82129230-en, s. 64. 29 Szerzej: E.M. Kwiatkowska, Internet of Things – wykorzystanie nowych technologii w ochronie zdrowia, [w:] Ł. Sułkowski, D. Kaczorowska-Spychalska (red.), Internet of Things. Nowy paradygmat rynku, Warszawa 2018, s. 185–186. 28 Dz.U. poz. 1991 ze zm. Nowe technologie remedium na problem starzenia się społeczeństwa 31 Eurostat, https://appsso.eurostat.ec.europa.eu/nui/submitViewTableAction.do (dostęp: 2 32 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2020 r. (Stan w dniu 30 czerwca 2020), Tablica 1. Ludność wg płci i wieku w 2020 r. Stan w dniu 30 czerwca, https://stat.gov.pl/obszary-tema- tyczne/ludnosc/ludnosc/ludnosc-stan-i-struktura-ludnosci-oraz-ruch-naturalny-w-przekroju- terytorialnym-stan-w-dniu-31-12-2020,6,29.html (dostęp: 12.05.2021). 33 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r. Stan w dniu 31.XII., Warszawa 2020, s. 18. 34 European Commission, The 2018 Ageing Report. Economic & Budgetary Projections for the 28 EU Member States (2016–2070), s. 23, https://ec.europa.eu/info/sites/info/files/economy-finance/ip079_en.pdf (dostęp: 20.11.2020). 35 Strategia…, s. 149. 35 Strategia…, s. 149. 36 GUS, Prognoza gospodarstw domowych według województw na lata 2003–2030, Tablica 3. Prognoza gospodarstw domowych według liczby osób i wieku głowy gospodarstwa (stan w dniu 31.XII), https://stat.gov.pl/obszary- 36 GUS, Prognoza gospodarstw domowych według województw na lata 2003–2030, Tablica 3. Prognoza gospodarstw domowych według liczby osób i wieku głowy gospodarstwa (stan w dniu 31.XII), https://stat.gov.pl/obszary- Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 DOI: 10.7206/kp.2080-1084.513 Ewa M. Kwiatkowska 162 teryzuje się więc w Polsce singularyzacją starości. W najbliższych latach nakłady na opiekę zdrowotną muszą więc wzrosnąć, przy czym w szczególności wzrost ten powinien dotyczyć wydatków na długotrwałą opiekę nad seniorami. Już teraz widoczne są niedobory w usługach opieki długoterminowej, a usługi zdrowotne świadczone w domu pacjenta są słabo rozwinięte. Z powodu niedofinansowania tego segmentu są one uzależnione od opieki nieformalnej (m.in. opiekunowie rodzinni) i leczenia szpitalnego. Formalna opieka długoterminowa nie jest odpo- wiednio rozwinięta. Wydatki na ten cel stanowią tylko 6% wydatków bieżących, co wskazuje na przepaść w porównaniu ze średnią UE wynoszącą powyżej 16%37. We wszystkich grupach społecznych obserwowane jest zjawisko rozwarstwienia społecznego przejawiającego się w nierównym dostępie m.in. do usług zdrowotnych czy też usług opieki nad seniorami38. teryzuje się więc w Polsce singularyzacją starości. W najbliższych latach nakłady na opiekę zdrowotną muszą więc wzrosnąć, przy czym w szczególności wzrost ten powinien dotyczyć wydatków na długotrwałą opiekę nad seniorami. Już teraz widoczne są niedobory w usługach opieki długoterminowej, a usługi zdrowotne świadczone w domu pacjenta są słabo rozwinięte. Z powodu niedofinansowania tego segmentu są one uzależnione od opieki nieformalnej (m.in. opiekunowie rodzinni) i leczenia szpitalnego. Formalna opieka długoterminowa nie jest odpo- wiednio rozwinięta. Wydatki na ten cel stanowią tylko 6% wydatków bieżących, co wskazuje na przepaść w porównaniu ze średnią UE wynoszącą powyżej 16%37. We wszystkich grupach społecznych obserwowane jest zjawisko rozwarstwienia społecznego przejawiającego się w nierównym dostępie m.in. -tematyczne/ludnosc/prognoza-ludnosci/prognoza-gospodarstw-domowych-wedlug-wojewodztw- na-lata-2003-2030,9,3.html (dostęp: 12.05.2021). 37 OECD/European Observatory on Health Systems and Policies, State of Health in the EU, Polska: Profil systemu ochrony zdrowia 2019, DOI: 10.1787/c7cfb688-pl, s. 10, 15 i 20. 38 Strategia…, s. 153. 39 Ibidem, s. 154. 40 Ibidem, s. 21–22 i 270. 37 OECD/European Observatory on Health Systems and Policies, State of Health in the EU systemu ochrony zdrowia 2019, DOI: 10.1787/c7cfb688-pl, s. 10, 15 i 20. -tematyczne/ludnosc/prognoza-ludnosci/prognoza-gospodarstw-domowych-wedlug-w na-lata-2003-2030,9,3.html (dostęp: 12.05.2021). -tematyczne/ludnosc/prognoza-ludnosci/prognoza-gospodarstw-domowych-wedlug-wojewodztw- na-lata-2003-2030,9,3.html (dostęp: 12.05.2021). 37 OECD/European Observatory on Health Systems and Policies, State of Health in the EU, Polska: Profil Nowe technologie remedium na problem starzenia się społeczeństwa do usług zdrowotnych czy też usług opieki nad seniorami38. Coraz większym problemem jest zapewnienie opieki nad tymi seniorami, którzy mogliby samodzielnie funkcjonować przy wsparciu np. w postaci zdalnej opieki medycznej lub społecznej. Niejednokrotnie ich najbliżsi muszą się decydować na zapewnienie tzw. nieformalno-rodzinnej opieki nad bliskimi. Wiąże się to z dużym obciążeniem finansowym przy prywatnym zatrudnieniu opiekunów, a także z ryzy- kiem wykluczenia zawodowego i społecznego tych, którzy poświęcają czas na opiekę nad seniorem39. Obserwowane potrzeby stanowią więc bodziec dla rozwoju systemu usług opiekuńczych, w tym przede wszystkim realizowanych z wykorzy- staniem ICT w opiece nad osobami starszymi40. Przykładem takiego rozwiązania mogą być np. przyciski życia pozwalające seniorom w prosty sposób wezwać pomoc, co może umożliwiać im samodzielne zamieszkiwanie. Sama świadomość posia- dania takiego urządzenia umożliwiającego uzyskanie całodobowej pomocy często poprawia jakość i bezpieczeństwo życia seniorów. Rozwiązania takie powinno się wprowadzać powszechnie, a nie tylko w ramach programów pilotażowych. Wdro- żenie tego typu zdalnych usług byłoby właściwym krokiem w kierunku rzeczy- wistej realizacji konstytucyjnego nakazu zapewnienia seniorom szczególnej opieki zdrowotnej. Według badania ankietowego przeprowadzonego w 2018 r. przez Centrum Systemów Informacyjnych Ochrony Zdrowia (CSIOZ, obecnie Centrum e-Zdrowia) 40 Ibidem, s. 21–22 i 270. Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 163 blisko 84% podmiotów wykonujących działalność leczniczą nie wykorzystywała rozwiązań telemedycznych. Największy udział podmiotów je wykorzystujących dotyczył szpitali (ponad 40%), a najniższy – ambulatoryjnej opieki zdrowotnej (ok. 10% podmiotów)41. Paradoksalnie, badanie z 2019 r. wskazało na wzrost tego odsetka – już ponad 88% podmiotów wykonujących działalność leczniczą nie wyko- rzystywało telemedycyny42. Co więcej, kontrola NIK-u wykazała, że w Polsce, w porównaniu do innych państw UE, istnieją opóźnienia w implementacji nowo- czesnych technologii w leczeniu wielu schorzeń43.l Jednym z flagowych projektów zaprezentowanych w 2017 r. przez Radę Mini- strów jest projekt Telemedycyna, który dzięki wykorzystaniu technologii komu- nikacyjnych ma stymulować rozwój innowacyjnych usług i produktów medycznych służących zwiększeniu dostępności do specjalistycznych usług medycznych44. Wprowadzenie nowoczesnych technologii może pozwolić zmienić tradycyjny, kapi- tałochłonny sposób zapewnienia opieki zdrowotnej w nowoczesne e-zdrowie, poprawiając jakość usług oraz wprowadzając oszczędności. Zastosowanie powinny znaleźć m.in. technologie telemedyczne pozwalające np. monitorować i sterować rehabilitacją pacjenta w jego domu oraz urządzenia do noszenia przesyłające informacje o stanie zdrowia (wearables), a także technologie telemedyczne stosowane w diagnostyce i terapii wspierające opiekę skoordynowaną (pozwalające m.in. na konsultacje na odległość)45. 41 CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 3 wyd., Warszawa 2018, s. 38. 48 OECD/European Observatory on Health Systems and Policies, op. cit., s. 22. 43 NIK, Informacja o wynikach kontroli. Raport: System ochrony zdrowia w Polsce – stan obecny i pożądane kierunki zmian, Warszawa 2019, s. 8. 42 CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 4 wyd., Warszawa 2019, s. 21. 41 CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 3 wyd., Warszawa 2018, s. 38. 42 CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 4 wyd., Warszawa 2019, s. 21. 43 NIK, Informacja o wynikach kontroli. Raport: System ochrony zdrowia w Polsce – stan obecny i pożądane kierunki zmian, Warszawa 2019, s. 8. 44 Strategia s 78 p f y j p y ją y ą y 2018, s. 38. 42 CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 4 wyd., Warszawa 2019, s. 21. 43 NIK, Informacja o wynikach kontroli. Raport: System ochrony zdrowia w Polsce – stan obecny i pożądane kierunki zmian, Warszawa 2019, s. 8. 47 NIK, op. cit., s. 21. 44 Strategia…, s. 78. 57 European Commission, Benchmarking Deployment of eHealth among General Practitioners (2018) – Final report, Luxemburg 2018, s. 31. 58 Centrum e-Zdrowia, Biuletyn Statystyczny 2020, Warszawa 2020, s. 25. 49 NIK, op. cit., s. 8. 50 OECD/European Observatory on Health Systems and Policies, op. cit., s. 22. 51 OECD/EU, op. cit., s. 11. 52 Uchwała nr 161…, s. 42–43. 53 Strategia…, s. 154. 54 OECD/European Observatory on Health Systems and Policies, op. cit., s. 3, 9–10. 55 Ibidem, s. 3 i 15. Nowe technologie remedium na problem starzenia się społeczeństwa Rozwój telemedycyny i teleopieki doprowadzić powi- nien do zwiększenia poziomu bezpieczeństwa seniorów pozostających w swoich domach, a także zwiększyć dostępność do świadczeń medycznych oraz usprawnić funkcjonowanie systemu ochrony zdrowia46. Takie rozwiązania mogłyby sprzyjać zmniejszeniu liczby pacjentów konsultowanych w odległych często ośrodkach47. Spośród państw europejskich w Polsce w przypadku schorzeń przewlekłych liczba hospitalizacji, których można by uniknąć, należy do jednej z najwyższych48. Koszty leczenia szpitalnego stanowią ponad połowę kosztów świadczeń zdrowot- 44 Strategia…, s. 78. 44 Strategia…, s. 78. 45 Ibidem, s. 281–282. 45 Ibidem, s. 281–282. 46 Uchwała nr 161…, s. 45. 47 NIK, op. cit., s. 21. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 Ewa M. Kwiatkowska 164 nych finansowanych przez NFZ49, podczas gdy leczenie wielu chorób przewlekłych, generujących wysokie wskaźniki hospitalizacji możliwe jest ambulatoryjnie50. W różnych państwach członkowskich nawet 20% wydatków ponoszonych na opiekę zdrowotną jest marnotrawione i mogłoby być lepiej wykorzystane51. Roz- wój i zwiększenie dostępności telemedycznych konsultacji skutkować powinno zmniejszeniem liczby niepotrzebnych hospitalizacji, a w konsekwencji spadkiem kosztów całego systemu ochrony zdrowia. Telemedycyna może mieć szczególnie istotne znaczenie na obszarach wiejskich, gdzie dostęp do usług medycznych nie jest tak powszechny jak w miastach52. Udział w PKB wydatków na ochronę zdrowia plasuje Polskę na jednym z ostat- nich miejsc spośród państw OECD53, oraz UE. Średni udział wydatków na opiekę zdrowotną w UE wynosił w 2017 r. 9,8% PKB, zaś w Polsce kształtował się on na poziomie 6,5% PKB. W przeliczeniu na mieszkańca w 2017 r. wydatki w Polsce na opiekę zdrowotną wynosiły jedynie 1507 Euro (jest to szósty najgorszy wynik w UE), stanowiąc około połowę średniej UE. Co więcej, wydatki ze środków publicznych w Polsce są na niższym poziomie niż średnia UE (odpowiednio 70% i 79%)54. Tak niski udział w wydatkach środków publicznych może budzić obawy co do dostęp- ności tych usług dla wszystkich obywateli. Poziom niezaspokojonych potrzeb medycznych związany z wysokimi kosztami usług i towarów zdrowotnych oraz czasem oczekiwania na dostępność świadczeń medycznych jest wyższy niż średnia unijna55. Dostęp do świadczeń uzależniony jest w coraz większej liczbie przypadków od sytuacji finansowej pacjenta56. Dodatkowo w Polsce zauważalny jest niedobór pracowników ochrony zdrowia. Według danych z 2017 r. liczba praktykujących lekarzy na 1000 mieszkańców równa była 2,4 i była najniższa spośród państw OECD. Na obszarach wiejskich dostęp do lekarzy jest gorszy niż w miastach57. Dodatkowo według stanu z 31.12.2019 r. blisko co czwarty uprawniony do wyko- nywana zawodu lekarz przekroczył 65 rok życia58. Nowe technologie remedium na problem starzenia się społeczeństwa W najbliższej przyszłości mogą p 57 European Commission, Benchmarking Deployment of eHealth among General Practitioners (2018) – Final report, Luxemburg 2018, s. 31. 58 Centrum e-Zdrowia, Biuletyn Statystyczny 2020, Warszawa 2020, s. 25. Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 165 więc wzrastać trudności związane z dostępnością lekarzy, zwłaszcza, że z powo- du starzenia się społeczeństwa potrzeba konsultacji medycznych będzie rosła. W 2018 r. seniorzy w wieku 65+, których udział w populacji wynosił 17,5%59, skorzystali z blisko 30% wszystkich udzielonych porad ambulatoryjnych. Korzy- stali oni z konsultacji lekarskich ponad dwukrotnie częściej niż osoby w wieku poni- żej 65 lat60. Średnie wydatki publicznych środków na opiekę zdrowotną seniorów w wieku 65+ są około trzykrotnie wyższe w porównaniu do osób młodszych61. ą y y p y Z uwagi na proces starzenia się społeczeństwa wzrasta konieczność zapewnie- nia specjalistycznej opieki najstarszym pacjentom, często cierpiącym na wielocho- robowość. Zapewnienie im fragmentarycznej, składającej się z konsultacji u wielu specjalistów opieki nie jest pożądane, podobnie jak odsyłanie na wizyty do lekarzy geriatrów, przyjmujących niejednokrotnie w oddalonych przychodniach. Rozwią- zaniem tego problemu mogłyby być specjalistyczne telekonsylia geriatryczne, które mogą być kontraktowane od 2015 r.62. Niestety rozwiązanie to, będące pew- nym remedium na brak specjalistów w zakresie geriatrii i ich nierównomierne rozmieszczenie (głównie w ośrodkach miejskich), pomimo ponad pięcioletniego okresu obowiązywania nie jest właściwie w praktyce wykorzystywane63. 63 Według stanu z 27.01.2021 r. umowy na świadczenie telekonsyliów geriatrycznych zostały podpisane jedynie z 2 podmiotami z województwa małopolskiego. Co więcej od 2016 r. są to ci sami świadcze- niodawcy. Jedynie w 2018 r. dodatkowo zawarta była 1 umowa z podmiotem z województwa kujaw- sko-pomorskiego (https://aplikacje.nfz.gov.pl/umowy/Provider/Search?Year=2021&Produc- t=11.1060.000.02&Branch=06, 27.01.2021). Widoczny jest więc brak zainteresowania świadczeniem tych usług. 62 Zarządzenie Nr 63/2015/DSOZ Prezesa NFZ z dnia 30.09.2015 r. zmieniające zarządzenie w sprawie określenia warunków zawierania i realizacji umów w rodzaju świadczenia zdrowotne kontraktowane odrębnie, http://www.nfz.gov.pl/zarzadzenia-prezesa/zarzadzenia-prezesa-nfz/zarzadzenie-nr- -632015dsoz,6410.html (dostęp: 12.05.2021). 59 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r.…, s. 16. 60 GUS, Sytuacja osób starszych w Polsce w 2018 r., Warszawa–Białystok 2020, s. 46–47. 61 NIK, op. cit., s. 13. 62 Zarządzenie Nr 63/2015/DSOZ Prezesa NFZ z dnia 30.09.2015 r. zmieniające zarządzenie w sprawie określenia warunków zawierania i realizacji umów w rodzaju świadczenia zdrowotne kontraktowane odrębnie, http://www.nfz.gov.pl/zarzadzenia-prezesa/zarzadzenia-prezesa-nfz/zarzadzenie-nr- -632015dsoz,6410.html (dostęp: 12.05.2021). 63 Według stanu z 27.01.2021 r. umowy na świadczenie telekonsyliów geriatrycznych zostały podpisane jedynie z 2 podmiotami z województwa małopolskiego. Co więcej od 2016 r. są to ci sami świadcze- niodawcy. Jedynie w 2018 r. dodatkowo zawarta była 1 umowa z podmiotem z województwa kujaw- sko-pomorskiego (https://aplikacje.nfz.gov.pl/umowy/Provider/Search?Year=2021&Produc- t=11.1060.000.02&Branch=06, 27.01.2021). Widoczny jest więc brak zainteresowania świadczeniem tych usług. 59 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r.…, s. 16 60 GUS, Sytuacja osób starszych w Polsce w 2018 r., Warszawa–Białystok 2020, s. 46–47. 61 NIK, op. cit., s. 13. 59 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r.…, s. 16. 60 GUS, Sytuacja osób starszych w Polsce w 2018 r., Warszawa–Białystok 2020, s. 46–47. 61 NIK, op. cit., s. 13. Wnioski Starzenie się społeczeństwa oznacza wzrost zapotrzebowania nie tylko na usługi ochrony zdrowia, ale także na świadczenia z zakresu opieki długoterminowej. Oczekuje się, że wydatki na opiekę długoterminową będą rosły szybciej niż wydatki na opiekę zdrowotną. Technologie informacyjno-komunikacyjne oferują realne możliwości promowania zdrowszego starzenia się oraz bardziej wydajnej i pacjento 62 Zarządzenie Nr 63/2015/DSOZ Prezesa NFZ z dnia 30.09.2015 r. zmieniające zarządzenie w sprawie określenia warunków zawierania i realizacji umów w rodzaju świadczenia zdrowotne kontraktowane odrębnie, http://www.nfz.gov.pl/zarzadzenia-prezesa/zarzadzenia-prezesa-nfz/zarzadzenie-nr- -632015dsoz,6410.html (dostęp: 12.05.2021). 63 Według stanu z 27.01.2021 r. umowy na świadczenie telekonsyliów geriatrycznych zostały podpisane jedynie z 2 podmiotami z województwa małopolskiego. Co więcej od 2016 r. są to ci sami świadcze- niodawcy. Jedynie w 2018 r. dodatkowo zawarta była 1 umowa z podmiotem z województwa kujaw- sko-pomorskiego (https://aplikacje.nfz.gov.pl/umowy/Provider/Search?Year=2021&Produc- t=11.1060.000.02&Branch=06, 27.01.2021). Widoczny jest więc brak zainteresowania świadczeniem tych usług. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 Ewa M. Kwiatkowska 166 centrycznej opieki64. Niestety w Polsce takie zmiany nie do końca są widoczne. Opieka długoterminowa w wielu wypadkach pozostaje domeną najbliższych, którzy nie mogą liczyć na systemowe wsparcie. Zrównoważony rozwój pozwala na wykorzystywanie postępu technologicz- nego w służbie człowiekowi. Jedną z dziedzin, w której wykorzystywanie nowych technologii może mieć immanentny wpływ na życie i zdrowie człowieka, jest ochrona zdrowia. Wykorzystywanie nauki do wdrażania nowych rozwiązań technologicz- nych – łatwo dostępnych i adaptowalnych do zmieniających się potrzeb – kształ- tuje przyszłość w nowy, dotychczas nieosiągalny sposób. Postęp technologiczny powoduje, że podstawowe świadczenia z zakresu ochrony zdrowia realizowane dzięki rozwojowi usług cyfrowych mogą stawać się coraz tańsze i dostępne dla szer- szego grona odbiorców. Cyfrowe zdrowie, w skład którego wchodzą m.in. świad- czenia realizowane na odległość (telemedycyna), jest odpowiedzią na rozwój cywi- lizacyjny i problemy epidemiologiczne, których przykładem jest panująca pandemia COVID-19. Jednak aby zmiany mogły być realizowane zgodnie z zasadami zrówno ważonego rozwoju, obok procesów cyfryzacji ochrony zdrowia musi być zapewniona możliwość normalnej egzystencji osób wykluczonych cyfrowo. Jeśli świadczenia byłyby realizowane wyłącznie elektronicznie, wymaga to zabezpieczenia interesów grupy społecznej wykluczonej cyfrowo – z powodu braku umiejętności cyfrowych lub braku dostępu do sprzętu lub usług zapewniających dostęp do Internetu. Ochrona przed dyskryminacją obywateli należących do tej grupy pociąga za sobą konieczność wdrożenia rozwiązań takich jak, na przykład, wprowadzenie instytucji asystenta medycznego, społecznego czy w inny sposób rozszerzenia zakresu działania opieki społecznej. Wdrażając zdalne usługi cyfrowe, nie wolno zaniedbywać kreowania i rozwoju równoległych rozwiązań dla osób, które nie mogą, nie chcą lub nie potrafią korzystać z nowych technologii. 64 OECD/EU, op. cit., s. 14. Tom 14, nr 1/2022 Wnioski Przykładem takiego działania, zapewnia jącego poszanowanie praw pacjenta i niewykluczanie kogokolwiek z zakresu świadczeń, może być proces zapisów na szczepienia przeciwko COVID-19. W ramach tej procedury uruchomiono równolegle kilka kanałów komunikacji, umożliwiając zarówno pacjentom wykluczonym cyfrowo (kontakt bezpośrednio w placówce lub ew. telefoniczny), jak i tym włączonym cyfrowo zapisywanie się na szczepienia. Nie wolno także zaniedbywać rozwoju tradycyjnych metod leczenia, bowiem tylko wzajemna interakcja i zrównoważenie metod tradycyjnych oraz tych wykorzy- stujących nowe technologie, w tym możliwość kontaktu elektronicznego w czasie rzeczywistym, pozwalają na nawiązanie i rozwój prawidłowej relacji między pacjentem a lekarzem. Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… Kwiatkowska 168 poradni geriatrycznych, których istnienie pozwalałoby lekarzom POZ na konsul- towanie starszych, cierpiących na wielochorobowość pacjentów. poradni geriatrycznych, których istnienie pozwalałoby lekarzom POZ na konsul- towanie starszych, cierpiących na wielochorobowość pacjentów. Wdrażanie nowych rozwiązań technologicznych, szczególnie w tak wrażliwym z punktu widzenia praw jednostki sektorze, jakim jest sektor ochrony zdrowia, wiąże się z koniecznością wprowadzania rozwiązań prawnych, które nie zniweczą pozytywnych aspektów wdrażania nowych technologii, a pozwolą na pełne wykorzystanie ich zalet. Tylko w takiej sytuacji można mówić o zrównoważonym rozwoju. CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 167 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… Kwestie związane z zapewnieniem dostępu do publicznych świadczeń z zakresu e-zdrowia powinny być domeną państwa. Instrumenty prawne powinny być skon- struowane tak, aby zagwarantować prawa do wszelkich świadczeń wszystkim grupom społecznym, pamiętając szczególnie o narażonych na wykluczenie i cyfrowo wykluczonych. Związane z tym obowiązki powinny obciążać podmioty ochrony zdrowia lub opieki społecznej. Rozsądnym rozwiązaniem byłoby zapew- nienie dostępu do sprzętu pozwalającego na korzystanie z Internetu, a tym wyklu- czonym cyfrowo lub po prostu nieradzącym sobie z ciągłymi zmianami pomoc tzw. asystenta cyfrowego, np. osoby z opieki społecznej lub wolontariusza. Należy przy tym pamiętać, że wszelkie naruszenia prywatności i wynikające z nich skutki powodować będą w zasadzie natychmiastową utratę zaufania użytkowników systemu – zarówno świadczeniobiorców, jak i świadczeniodawców, a odbudowa zaufania – o ile w ogóle możliwa – byłaby procesem długotrwałym i kosztownym. Usługi cyfrowe, szczególnie z zakresu telemedycyny, eliminują konieczność dojazdu do daleko położonych ośrodków ochrony zdrowia, co przekłada się nie tylko na dostępność specjalistycznych usług dla szerszego grona odbiorców, ale również powoduje skrócenie czasu oczekiwania i załatwienia sprawy, a tym samym zwięk- szoną dostępność dla pacjentów. Ponieważ jednak bezpośredni kontakt między pacjentem a lekarzem jest niezwykle istotny, stąd dobrym rozwiązaniem wydaje się być wprowadzanie rozwiązań hybrydowych, łączących elektroniczny kontakt na odległość pacjenta z lekarzem specjalistą – w obecności lekarza POZ, który znając pacjenta i jego problemy zdrowotne, może zwrócić specjaliście uwagę na symptomy nierozpoznawalne przez będącego zwykle laikiem pacjenta. Takim rozwiązaniem są kontraktowane przez NFZ telekonsultacje geriatryczne i kardio- logiczne. Niestety ich nikłe wykorzystywanie świadczy o niedopasowaniu szczegó- łów tych rozwiązań do realiów. Ich zorganizowanie nie jest odpowiednie. Są one drogie, wymagające czasu, którym lekarz POZ nie dysponuje. Niejednokrotnie potrzebuje on konsultacji w danym momencie, a nie takiej, na którą pacjent zapisuje się z wyprzedzeniem. Wartym rozważenia, możliwym dzięki ICT rozwiązaniem mogłoby być stworzenie publicznego, ogólnokrajowego systemu dyżurów specja- listów on-line, dzięki któremu każdy lekarz POZ miałby możliwość konsultowania konkretnego przypadku w czasie rzeczywistym. W skali kraju takie np. 12-godzinne konsultacje on-line wymagałyby zaangażowania minimalnego czasu ze strony każdego specjalisty, a mogłyby ułatwić diagnozowanie i pozwolić na uniknięcie niepotrzebnego kierowania pacjentów na specjalistyczne wizyty, czy też ich zbęd- nego hospitalizowania. Jednocześnie takie rozwiązanie odciążyłoby lekarzy POZ, na których obecnie koncentruje się opieka nad wszystkimi pacjentami. Innym, niewymagającym wykorzystania rozwiązań cyfrowych wsparciem byłoby kontrak towanie przez NFZ przynajmniej w każdym powiecie świadczeń specjalistycznych Tom 14, nr 1/2022 DOI: 10.7206/kp.2080-1084.513 Ewa M. Bibliografia Balicki M., Prezentacja koncepcji Centrum Zdrowia 75+, Konferencja „Centrum Zdrowia 75+. Zdążyć przed demograficznym tsunami”, Warszawa, 10.09.2019, https://www. prezydent.pl/aktualnosci/wydarzenia/art,1530,prezydent-na-konferencji-centrum- -zdrowia-75.html (dostęp: 21.11.2020). Bartlett J., Ludzie przeciw technologii. Jak Internet zabija demokrację (i jak ją możemy ocalić), Katowice 2019. Bartlett J., Ludzie przeciw technologii. Jak Internet zabija demokrację (i jak ją możemy ocalić), Katowice 2019. Bartlett J., Ludzie przeciw technologii. Jak Internet zabija demokrację (i jak ją możemy ocalić), Katowice 2019. Birkland J.L.H., Gerontechnology. Understanding Older Adult Information and Communica- tion Technology Use, Emerald Publishing 2019. Birkland J.L.H., Gerontechnology. Understanding Older Adult Information and Communica- tion Technology Use, Emerald Publishing 2019. Centrum e-Zdrowia, Biuletyn Statystyczny 2020, Warszawa 2020. Centrum Zdrowia 75+, Zarys koncepcji opracowanej przez ekspertów z Narodowej Rady Rozwoju przy Prezydencie RP, Warszawa 2019, https://www.prezydent.pl/aktualnosci/ wydarzenia/art,1530,prezydent-na-konferencji-centrum-zdrowia-75.html (dostęp: 21.11.2020). CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 4 wyd., Warszawa 2019. CSIOZ, Badanie stopnia informatyzacji podmiotów wykonujących działalność leczniczą, 3 wyd., Warszawa 2018. European Commission, Benchmarking Deployment of eHealth among General Practi tioners (2018) – Final report, Luxemburg 2018. European Commission, The 2018 Ageing Report. Economic & Budgetary Projections for the 28 EU Member States (2016-2070), https://ec.europa.eu/info/sites/info/files/ economy-finance/ip079_en.pdf (20.11.2020). Eurostat, https://appsso.eurostat.ec.europa.eu/nui/submitViewTableAction.do (dostęp: 25.11.2021). GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2020 r. (Stan w dniu 30 czerwca 2020), Tablica 1. Ludność wg płci i wieku w 2020 r. Stan w dniu 30 czerwca, https://stat.gov.pl/obszary-tematyczne/ludnosc/ludnosc/ludnosc-stan-i-struktura-lu- dnosci-oraz-ruch-naturalny-w-przekroju-terytorialnym-stan-w-dniu-31-12-2020,6,29. html (dostęp: 12.05.2021). DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… 169 GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r. Stan w dniu 31.XII., Warszawa 2020. GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r. Stan w dniu 31.XII., Warszawa 2020. GUS, Ludność. Stan i struktura oraz ruch naturalny w przekroju terytorialnym w 2019 r. Stan w dniu 31.XII., Warszawa 2020. GUS, Sytuacja osób starszych w Polsce w 2018 r., Warszawa–Białystok 2020. GUS, Prognoza gospodarstw domowych według województw na lata 2003-2030, Tablica 3. Prog noza gospodarstw domowych według liczby osób i wieku głowy gospodarstwa (stan w dniu 31.XII), 1.08.2003, https://stat.gov.pl/obszary-tematyczne/ludnosc/prognoza-ludno- sci/prognoza-gospodarstw-domowych-wedlug-wojewodztw-na-lata-2003-2030,9,3. html (dostęp: 12.05.2021). Janowski J., Cyberkultura prawa. Współczesne problemy filozofii i informatyki prawa, Warszawa 2012. Komunikat Komisji do Parlamentu Europejskiego, Rady, Europejskiego Komitetu Eko- nomiczno-Społecznego i Komitetu Regionów w sprawie przeglądu śródokresowego realizacji strategii jednolitego rynku cyfrowego. Połączony jednolity rynek cyfrowy dla wszystkich, Bruksela, 10.05.2017, COM(2017) 228 final. Komunikat Komisji do Parlamentu Europejskiego, Rady, Europejskiego Komitetu Eko- nomiczno-Społecznego i Komitetu Regionów w sprawie umożliwienia transformacji cyfrowej opieki zdrowotnej i społecznej na jednolitym rynku cyfrowym, wzmac- niania pozycji obywateli i budowania zdrowszego społeczeństwa, Bruksela, 25.04.2018, COM(2018) 233 final. Komunikat Komisji „Europa 2020 Strategia na rzecz inteligentnego i zrównoważonego rozwoju sprzyjającego włączeniu społecznemu”, Bruksela, 3.03.2010, KOM(2010) 2020 wersja ostateczna. Kwiatkowska E.M., Wykluczenie cyfrowe, Konferencja ogólnopolska PolSenior2 – moduł 1, sytuacja socjalno-ekonomiczna osób starszych, 23.11.2020, https://konferencjapol- senior2.pl/ (dostęp: 12.05.2021). Kwiatkowska E.M., Internet of Things – wykorzystanie nowych technologii w ochronie zdrowia, [w:] Ł. Sułkowski, D. Kaczorowska-Spychalska (red.), Internet of Things. Nowy paradyg- mat rynku, Warszawa 2018, s. 180–199. Kwiatkowska E.M., Transformacja cyfrowa a osoby starsze w systemie ochrony zdrowia. O potrzebie implementowania nowych rozwiązań prawnych, „Transformacje” 2011, 108(1), s. 163–174. Malarewicz-Jakubów A., Wsparcie prawne osób starszych, Warszawa 2017. Marolla C., Information and Communication Technology for Sustainable Development, Boca Raton–London–New York 2019. Morawiecki M., Słowo wstępne do Strategii na rzecz Odpowiedzialnego Rozwoju do roku 2020 (z perspektywą do 2030 r.), Warszawa 2017. Neto A.P., Flynn M.B., The Internet and Health in Brazil: Trends and Challenges, [w:] A.P. Neto, M.B. Flynn (red.), The Internet and Health in Brazil: Challenges and Trends, Springer 2019, s. 1–11. NIK, Informacja o wynikach kontroli. Raport: System ochrony zdrowia w Polsce – stan obecny i pożądane kierunki zmian, Warszawa 2019. DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 Ewa M. Kwiatkowska 170 OECD/EU, Health at a Glance, Europe 2018. DOI: 10.7206/kp.2080-1084.513 CYFRYZACJA OCHRONY ZDROWIA A KONIECZNOŚĆ WPROWADZANIA… State of Health in the EU Cycle, DOI: 10.1787/ health_glance_eur-2018-en. OECD/EU, Health at a Glance, Europe 2018. State of Health in the EU Cycle, DOI: 10.1787/ health_glance_eur-2018-en. OECD/European Observatory on Health Systems and Policies, State of Health in the EU, Polska: Profil systemu ochrony zdrowia 2019, DOI: 10.1787/c7cfb688-pl. OECD/European Union, Health at a Glance, Europe 2020: State of Health in the EU Cycle, DOI: 10.1787/82129230-en. ONZ, Przekształcamy nasz świat: Agenda na rzecz zrównoważonego rozwoju 2030, https://www. gov.pl/web/rozwoj-praca-technologia/agenda-2030 (dostęp: 20.11.2020). Pszczołowski T., Organizacja od dołu i od góry, Warszawa 1978. Rada Ministrów, Uchwała nr 161 Rady Ministrów z dnia 26 października 2018 r. w spra- wie przyjęcia dokumentu Polityka społeczna wobec osób starszych 2030. Bezpie- czeństwo-Uczestnictwo-Solidarność, Warszawa 30.11.2018, M.P. poz. 1169. Smith S., The Internet of Risky Things. Trusting the Devices That Surround Us, O’Reilly Media 2017. Strategia na rzecz Odpowiedzialnego Rozwoju do roku 2020 (z perspektywą do 2030 r.), https:// www.gov.pl/documents/33377/436740/SOR.pdf (20.11.2020). Weber R.H., Weber R., Internet of Things. Legal Perspectives, Springer 2010. Zarządzenie Nr 63/2015/DSOZ Prezesa NFZ z dnia 30.09.2015 r. zmieniające zarządze- nie w sprawie określenia warunków zawierania i realizacji umów w rodzaju świad- czenia zdrowotne kontraktowane odrębnie, http://www.nfz.gov.pl/zarzadzenia- -prezesa/zarzadzenia-prezesa-nfz/zarzadzenie-nr-632015dsoz,6410.html (dostęp: 12.05.2021). DOI: 10.7206/kp.2080-1084.513 Tom 14, nr 1/2022 Tom 14, nr 1/2022
https://openalex.org/W4200425120	https://www.biorxiv.org/content/biorxiv/early/2021/04/12/2021.04.12.439478.full.pdf	English	null	Epitope profiling using computational structural modelling demonstrated on coronavirus-binding antibodies	PLOS computational biology/PLoS computational biology	2,021	cc-by	14,191	Epitope proﬁling of coronavirus-binding antibodies using computational structural modelling The properties (including sequence and, where pos- sible, structure) of these antibodies are documented in the Coronavirus Antibody Database (CoV-AbDab), which tracks patents and the academic literature on a weekly basis (1). One way to use this collated data is to look for similari- ties between binders. For instance, when a novel antibody is antigen-baited out of SARS-CoV-2 response serum, or is identiﬁed as an expanded clonal lineage post-SARS-CoV- 2 infection, one can assess whether it bears resemblance to any other antibody previously reported to bind a coronavirus. This resemblance can then be used to predict functional prop- erties of the newly-isolated antibody, such as its site of en- gagement with the antigen (the ‘epitope’). to identify such binders is ‘serum baiting’, where an extra- cellular coronavirus antigen is used to pan donated blood serum directly for complementary antibodies (2, 3). Another increasingly-used method is deep sequencing of the SARS- CoV-2 convalescent B-cell receptor (BCR) repertoire, which can implicate particular expanded antibody lineages as im- portant to adaptive immunity without biasing towards a cho- sen antigen bait (4–6). Other discovery methods have in- cluded mining surface display libraries, challenging and har- vesting transgenic animals, and antibody engineering (1). g g , y g g ( ) As of 11th March 2021, over 2,400 SARS-CoV-2 binding antibodies and nanobodies had been identiﬁed, of which just under one-third show neutralisation activity against the virus. The properties (including sequence and, where pos- sible, structure) of these antibodies are documented in the Coronavirus Antibody Database (CoV-AbDab), which tracks patents and the academic literature on a weekly basis (1). One way to use this collated data is to look for similari- ties between binders. For instance, when a novel antibody is antigen-baited out of SARS-CoV-2 response serum, or is identiﬁed as an expanded clonal lineage post-SARS-CoV- 2 infection, one can assess whether it bears resemblance to any other antibody previously reported to bind a coronavirus. This resemblance can then be used to predict functional prop- erties of the newly-isolated antibody, such as its site of en- gagement with the antigen (the ‘epitope’). A common way to cluster antibodies into such functional groupings is ‘clonotyping’, a form of clonal lineage cluster- ing. This can be performed in several different ways (7). Epitope proﬁling of coronavirus-binding antibodies using computational structural modelling For example, strict Fv-clonotyping maps both VH and VL anti- body chains to their closest immunoglobulin V- and J-gene and subsequently clusters identical gene mappings by their CDRH3 and CDRL3 lengths and sequence identities (using a threshold close to 100% per CDR3 region). This approach typically yields tight and functionally-signiﬁcant clustering, but is severely limited by its ability to bring together all an- tibodies able to engage a particular epitope (8–10). As a re- sult, leniency is often introduced to the clonotyping protocol, by lowering the sequence identity threshold to 80% (11), ig- noring J-gene annotations, and/or only considering the heavy chain (VH-only clonotyping) (4). structural clustering \| coronavirus antibody \| structure-function \| repertoire proﬁling These authors contributed equally to this work. These authors contributed equally to this work. Correspondence: deane@stats.ox.ac.uk Correspondence: deane@stats.ox.ac.uk Epitope proﬁling of coronavirus-binding antibodies using computational structural modelling Sarah A. Robinson1, Matthew I. J. Raybould1, Constantin Schneider1, Wing Ki Wong1, Claire Marks1, and Charlotte M. Deane1, 1Oxford Protein Informatics Group, Department of Statistics, University of Oxford, UK 1Oxford Protein Informatics Group, Department of Statistics, University of Oxford, UK Identifying the epitope of an antibody is a key step in under- standing its function and its potential as a therapeutic. It is well- established in the literature that sequence-based clonal cluster- ing can identify antibodies with similar epitope complementar- ity. However, there is growing evidence that antibodies from markedly different lineages but with similar structures can en- gage the same epitope with near-identical binding modes. Here, we describe a novel computational method for epitope proﬁl- ing based on structural modelling and clustering, and show how it can identify sequence-dissimilar antibodies that engage the same epitope. We start by searching for evidence of structural conservation across the latest solved SARS-CoV-2—binding an- tibody crystal structures. Despite the relatively small number of solved structures, we ﬁnd numerous examples of sequence- diverse but structurally-similar coronavirus-binding antibodies engaging the same epitope. We therefore developed a high- throughput structural modeling and clustering method to iden- tify functionally-similar antibodies across the set of thousands of coronavirus-binding antibody sequences in the Coronavirus Antibody Database (CoV-AbDab). In the resulting multiple- occupancy structural clusters, 92% bind to consistent domains based on CoV-AbDab metadata. Our approach functionally links antibodies with distinct genetic lineages, species origins, and coronavirus speciﬁcities. This indicates greater convergence exists in the immune responses to coronaviruses than would be suggested by sequence-based approaches. Our results show that applying structural analytics to large class-speciﬁc anti- body databases will enable high conﬁdence structure-function relationships to be drawn, yielding new opportunities to identify functional convergence hitherto missed by sequence-only analy- sis. to identify such binders is ‘serum baiting’, where an extra- cellular coronavirus antigen is used to pan donated blood serum directly for complementary antibodies (2, 3). Another increasingly-used method is deep sequencing of the SARS- CoV-2 convalescent B-cell receptor (BCR) repertoire, which can implicate particular expanded antibody lineages as im- portant to adaptive immunity without biasing towards a cho- sen antigen bait (4–6). Other discovery methods have in- cluded mining surface display libraries, challenging and har- vesting transgenic animals, and antibody engineering (1). As of 11th March 2021, over 2,400 SARS-CoV-2 binding antibodies and nanobodies had been identiﬁed, of which just under one-third show neutralisation activity against the virus. . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint Introduction The COVID-19 pandemic has generated worldwide efforts to isolate and characterise antibodies able to confer protection against SARS-CoV-2. Hundreds of studies have now released data on diverse antibodies and nanobodies capable of binding at least one coronavirus antigen (1). Due to the escalating number of individuals infected by SARS-CoV-2, most of the reported coronavirus-binding an- tibodies to date have been sourced directly from the blood of convalescent human patients. The primary technique used Convergent lenient VH-only clonotypes have been identiﬁed between multiple SARS-CoV-2 infected or convalescent in- dividuals (12–18) and across different studies, for example Robinson and Raybould et al. \| bioRχiv \| April 12, 2021 \| 1–14 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint the overlap between the clonotypes found by Galson et al. (4) and Nielsen et al. (19). Several papers have compared BCR sequences from individuals to veriﬁed SARS-CoV-2 binders in CoV-Ab-Dab and identiﬁed clonal similarities (e.g. (4, 15)). antibodies with high variable domain (Fv) structural simi- larity are able to exhibit functional commonality. We then model and structurally cluster the thousands of antibody Fv sequences in CoV-AbDab and show that 92% of multiple- occupancy structural clusters bin together antibodies that bind to consistent coronavirus antigens/domains. We also show that, in accordance with our analysis of the SARS-CoV- 2 X-ray co-crystal structures, the antibodies within these structural clusters frequently transcend clonal lineages. Results Sequences and structures in CoV-Ab-Dab. The growth of coronavirus-binding antibody and nanobody data in CoV- AbDab since its public release on 7th May 2020 is shown in Figure 1. The antibody plot indicates how the availability of sequence data rose much more rapidly than structural data at the start of the pandemic, stabilising at a level roughly an or- der of magnitude higher. However, the availability of solved antibody structures increased markedly in October 2020 and has continued to grow at an even faster rate throughout 2021. Analysis of SARS-CoV-2—antibody structural com- plexes. Experimentally-solved structures allow us to analyse the diversity of antibody geometries and paratopes that en- gage coronavirus antigens (28–30). At the time of this study (11st March 2021), CoV-AbDab (1) referenced 2,304 anti- bodies and 420 nanobodies able to bind to a coronavirus, with 132 having at least one solved structure. 111 of these solved structures are binders to SARS-CoV-2 (Figure 1), of which 91 were solved in complex with the cognate antigen. Structural comparisons offer a way to analyse antibody data over and above clonotype-based approaches. Analysis of epi- tope regions using solved structures and competition assays of SARS-CoV-2—binding antibodies has already revealed discrete antibody binding sites (28, 29). However, such as- sessments are very biased towards RBD-binding, neutralising antibodies, whose therapeutic potential renders them worth the expense and effort of structure determination. The vast majority of datapoints in CoV-AbDab do not have solved structures and must instead be structurally modeled. While homology models are provided alongside each structurally- unsolved CoV-AbDab entry, no studies have yet harnessed this data for functional annotation. A total of 48 antibodies and 12 nanobodies had at least one published solved X-ray crystal structure in complex with SARS-CoV-2 (12, 14, 20, 28, 31–54), all binding to the spike receptor binding domain (RBD, see Table S1 and Table S2 for names and PDB codes), while a further 31 antibodies and nanobodies were solely structurally characterised by cryo- EM (1). In our analysis we have focused on the 60 crystal structures, in order to determine more precise antibody bind- ing site topologies and paratope proﬁles. Epitope Binning. Inspecting the 48 antibodies solved by X- ray crystallography in complex with the RBD, 46 appear to fall cleanly into the binding regions previously deﬁned by Dejnirattisai et al. (28) (only approximate as the original clustering was performed via competition assays). Introduction This not only demonstrates that our computational structural anal- ysis pipeline provides orthogonal information to clonotyping to improve antibody functional proﬁling, but also that anti- body immune responses to SARS-CoV-2 are likely to be even more convergent than currently understood. Despite this, the clonotypes found to be enriched/to bear sim- ilarity to CoV-AbDab antibodies post SARS-CoV-2 exposure often differ across studies (20–22). This may be partly due to the small sample sizes used in individual studies, and the intrinsic biases in individual VJ gene usage; Xiang et al. found a larger variation between individuals within a cohort (healthy, or three different severities of COVID symptoms) than between cohorts (22). Another contributing factor is lin- eage clustering itself. Levels of functional convergence may in fact be higher than implied even by lenient clonotyping, as antibodies that derive from different lineages can engage the same epitope. The evidence for this phenomenon has been growing over recent years (9, 10, 23–25). For exam- ple, solved structures of antibody-antigen complexes reveal pairs of antibodies with different genetic origins but sufﬁ- ciently similar binding site geometry and paratope similarity that they bind to the same antigen with near-identical bind- ing modes (9, 10). Furthermore, given that individuals’ naive repertoires typically share very few clonotypes (11, 26) and yet are often found to respond to similar ‘immunodominant’ epitopes, it follows that multiple evolutionary routes may lead from low-moderate afﬁnity naive BCRs to high afﬁnity antibodies against the same antigen surface region. This is supported by statistical arguments showing the implausabil- ity of a purely “random repertoire” for an efﬁcient immune response (25, 27). Epitope immunodominance could be ra- tionalised via the existence of a more ‘public’ set of backbone structures in the BCR repertoire and the concept that BCRs with similar topologies and sufﬁcient chemical complemen- tarity engage the same epitope (24, 25). 2 \| bioRχiv Robinson and Raybould et al. \| Results The two remaining antibodies spanned the left and right shoulder re- gions (see Table S1). Interestingly, most of the 12 nanobodies could also be assigned to these predeﬁned regions (9/12, see Table S2). In this analysis, we examine how structurally analysing CoV-AbDab can enhance our functional understanding of coronavirus-binding antibodies. We ﬁrst analyse all X-ray crystal structures of antibodies/nanobodies bound to SARS- CoV-2 antigens, showing both that structure is conserved more often than clonality across same-epitope binders, and that paratope proﬁles typically involve multiple regions of the antibody across both chains. This provides direct evi- dence that relatively sequence dissimilar coronavirus-binding Structural alignment of the RBD of all complexes reveals that the 22 antibodies that bind to the ‘neck’ cluster (as termed by Dejnirattisai et al. (28)) have high structural conserva- . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Fig. 1. Comparing the quantity and growth of sequence (red) vs. structural (blue) data referenced by CoV-AbDab for antibodies (left-hand plot) and nanobodies (right-hand plot) against any coronavirus antigen. Structures are classed as solved if evaluated at least once either by X-ray crystallography or cryo-electron microscopy. The x-axis measures the database timestamp after the initial public release of CoV-AbDab on 7th May 2020 (Day 0), up until 11th March 2021 (Day 308). CDR3 sequences are often released ahead of full sequences to protect intellectual property during peer review and/or patent ﬁling. Fig. 1. Comparing the quantity and growth of sequence (red) vs. structural (blue) data referenced by CoV-AbDab for antibodies (left-hand plot) and nanobodies (right-hand plot) against any coronavirus antigen. Structures are classed as solved if evaluated at least once either by X-ray crystallography or cryo-electron microscopy. The x-axis measures the database timestamp after the initial public release of CoV-AbDab on 7th May 2020 (Day 0), up until 11th March 2021 (Day 308). Results CDR3 sequences are often released ahead of full sequences to protect intellectual property during peer review and/or patent ﬁling. Fig. 2. A) A cartoon representation of the 22 antibodies in the RBD ‘neck’ cluster binding to the SARS-CoV-2-RBD (salmon) [PDB code 6XC2] at a similar site to ACE-2. See Table S1 for PDB codes of the 22 antibodies. B) A cartoon representation of ACE-2 (green) binding to SARS-CoV-2 RBD (salmon) [PDB code 6VW1, ACE-2 chain A, SARS-CoV-2-RBD chain E]. C) The CDRH3 sequences represented across the 22 RBD ‘neck’-binding antibodies. Lenient VH-clonotypes are separated with solid lines, with the cluster representative highlighted in bold font. Fig. 2. A) A cartoon representation of the 22 antibodies in the RBD ‘neck’ cluster binding to the SARS-CoV-2-RBD (salmon) [PDB code 6XC2] at a similar site to ACE-2. See Table S1 for PDB codes of the 22 antibodies. B) A cartoon representation of ACE-2 (green) binding to SARS-CoV-2 RBD (salmon) [PDB code 6VW1, ACE-2 chain A, SARS-CoV-2-RBD chain E]. C) The CDRH3 sequences represented across the 22 RBD ‘neck’-binding antibodies. Lenient VH-clonotypes are separated with solid lines, with the cluster representative highlighted in bold font. align closest to the IGHV3-53 gene, while the remaining 3 align closest to IGHV3-66. tion (Figure 2A). These antibodies all compete for the ACE- 2 receptor binding site (Figure 2B). Dejnirattisai et al. iden- tiﬁed 13 IGHV3-53/IGHV3-66-derived antibodies engaging this binding site. Even from relatively early in the pandemic, it was clear that a disproportionate number of antibodies re- ported as able to block ACE-2 binding exploited the IGHV3- 53/IGHV3-66 genes (1, 28, 30); they appeared signiﬁcantly more often as binders of this region than would be expected by their abundance in healthy antibody repertoires. Banach et al. realised that many of the coronavirus-binding antibodies deriving from the IGHV3-53/IGHV3-66 genes possess a con- served set of structural motifs that enable complementarity to a SARS-CoV-2 RBD epitope (55). Our updated analysis reit- erates the importance of these V gene origins in engaging this highly conserved binding site: 19/22 (86%) of the antibodies tion (Figure 2A). These antibodies all compete for the ACE- 2 receptor binding site (Figure 2B). Dejnirattisai et al. iden- tiﬁed 13 IGHV3-53/IGHV3-66-derived antibodies engaging this binding site. Robinson and Raybould et al. \| Results Even from relatively early in the pandemic, it was clear that a disproportionate number of antibodies re- ported as able to block ACE-2 binding exploited the IGHV3- 53/IGHV3-66 genes (1, 28, 30); they appeared signiﬁcantly more often as binders of this region than would be expected by their abundance in healthy antibody repertoires. Banach et al. realised that many of the coronavirus-binding antibodies deriving from the IGHV3-53/IGHV3-66 genes possess a con- served set of structural motifs that enable complementarity to a SARS-CoV-2 RBD epitope (55). Our updated analysis reit- erates the importance of these V gene origins in engaging this highly conserved binding site: 19/22 (86%) of the antibodies Despite the highly similar V genes, these 22 tightly structurally-clustered antibodies represent 15 different ‘le- nient VH-only clonotypes’ (Figure 3 C) (clustered antibod- ies require the same CDRH3 length, 80% CDRH3 sequence identity, and an identical heavy V gene, see Methods). This corresponds to 18 ‘lenient Fv clonotypes’ when the light chain is also considered (clustered antibodies must also have the same CDRL3 length, 80% CDRL3 sequence identity, and an identical light V gene, see Methods). The analysis of the co-complex structures of these 22 antibodies suggests highly similar functionality, which cannot be wholly identi- ﬁed through clonotyping. Even using lenient clonotype def- initions, the antibodies would not be grouped together, so bioRχiv \| 3 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Fig. 3. A) The CDRH3 loops of nanobodies and antibodies of two binding regions, the ‘neck’ and ‘left ﬂank’, binding to SARS-CoV-2 RBD (salmon). The ‘neck’ cluster includes 22 antibodies (red). The ‘left ﬂank’ region includes ﬁve antibodies (the four antibodies identiﬁed by (28) are shown in orange, antibody MW06 is shown in magenta) and four nanobodies (shown in blue). MW06 and the nanobodies were not included in the analysis by (28). Results As expected, structures with longer CDRH3 regions tend to exhibit more CDRH3-dominated binding; the paratopes of the nine antibodies with CDRH3 length ≥19 on average comprised 41% CDRH3 residues. For the 22 highly structurally-conserved antibodies in the RBD neck epitope region, an average of 23% of the paratope residues originate from CDRH3 (range 8-34%). Dejnirattisai et al. (28) also highlighted an antibody binding cluster located away from the ACE-2 binding site, termed the ‘left ﬂank’. This binding region contains four antibodies, EY6A, CR3022, S304 and COVA1-16, (see Methods and Ta- ble S1 for antibody naming conventions). As shown in Figure 3A, these antibodies appear to bind to slightly different areas. In our updated analysis, we identify one antibody (MW06) and four nanobodies (VHH_U, VHH_V, VHH_W and VHH- 72) all able bind to the ‘left ﬂank’ binding region (Figure 3A). We highlight the antibody pair of EY6A and S304, which are structurally similar and adopt a common binding mode to the same RBD epitope (29), but share only 43% CDRH3 se- quence identity (Figure 3B) so could not have been identiﬁed as binding to the same site by sequence data alone. These RBD neck-binding antibodies exhibit high levels of paratope residue conservation across the CDRH1 and CDRH2 (Figure 4), which is also seen at the sequence simi- larity level. This explains the predominance of the IGHV3- 53/IGHV3-66 genes, as the residues and topologies pre- encoded by this germline play a key role in neck epitope complementarity. Paratope conservation is considerably less consistent across the CDRH3 region, accounting for the 18 unique CDRH3 sequences seen across the 22 antibodies. For such epitopes, a clonotyping framework (which conditions on high CDRH3 sequence identity being a pre-requisite for same-epitope binding), will clearly fail to capture the func- tional similarity of the spectrum of cognate antibodies. Even across the relatively small number of solved SARS- CoV-2—antibody structures, we can see numerous exam- ples of functionally and structurally similar antibodies that would not be grouped by sequence clustering alone. Group- ing coronavirus-binding antibodies into sets that have similar structures therefore represents an orthogonal and promising approach by which to highlight the potential functional com- monalities of sequence-dissimilar antibodies. Paratope analysis. 4 \| bioRχiv Robinson and Raybould et al. \| Results See Table S1 and Table S2 for antibody and nanobody PDB codes respectively. B) A ribbbon representation of sequence dissimilar antibodies S304 (blue) [PDB code 7L0N] and EY6A (green) [PDB code 6ZER, chain A] binding to the SARS-CoV-2-RBD (salmon) [PDB code 6ZER, chain B]. The CDRH1, CDRH2 and CDRH3 loops are illustrated in cartoon. The CDRH3 sequences of the two antibodies, S304 and EY6A, are shown, with dissimilar residues indicated in bold. Fig. 3. A) The CDRH3 loops of nanobodies and antibodies of two binding regions, the ‘neck’ and ‘left ﬂank’, binding to SARS-CoV-2 RBD (salmon). The ‘neck’ cluster includes 22 antibodies (red). The ‘left ﬂank’ region includes ﬁve antibodies (the four antibodies identiﬁed by (28) are shown in orange, antibody MW06 is shown in magenta) and four nanobodies (shown in blue). MW06 and the nanobodies were not included in the analysis by (28). See Table S1 and Table S2 for antibody and nanobody PDB codes respectively. B) A ribbbon representation of sequence dissimilar antibodies S304 (blue) [PDB code 7L0N] and EY6A (green) [PDB code 6ZER, chain A] binding to the SARS-CoV-2-RBD (salmon) [PDB code 6ZER, chain B]. The CDRH1, CDRH2 and CDRH3 loops are illustrated in cartoon. The CDRH3 sequences of the two antibodies, S304 and EY6A, are shown, with dissimilar residues indicated in bold. their similar binding mode and functional similarity would be missed. conservation over particular regions is typically sufﬁcient. p g yp y An average of ~67% of paratope residues were found to lie on the heavy (VH) chain while ~33% reside on the light chain (VL). The percentage of paratope residues donated by the most hypervariable region, the CDRH3 loop, varies from just 9% up to 58%. As expected, structures with longer CDRH3 regions tend to exhibit more CDRH3-dominated binding; the paratopes of the nine antibodies with CDRH3 length ≥19 on average comprised 41% CDRH3 residues. For the 22 highly structurally-conserved antibodies in the RBD neck epitope region, an average of 23% of the paratope residues originate from CDRH3 (range 8-34%). An average of ~67% of paratope residues were found to lie on the heavy (VH) chain while ~33% reside on the light chain (VL). The percentage of paratope residues donated by the most hypervariable region, the CDRH3 loop, varies from just 9% up to 58%. Robinson and Raybould et al. \| Results We then analysed the binding interfaces across the set of 48 antibodies, to investigate whether struc- ture needs to be conserved across the whole Fv, or whether For those antibody paratopes that use CDRH3 sparsely, the paratope is mostly drawn from the CDRH1, CDRH2, CDRL1 and CDRL3 residues, with occasional amino acids provided . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CDR paratope conservation and divergence across the 22 antibodies in the ‘neck’ cluster. Frequency indicates the number of times an amino acid was se deﬁned CDR position (56). Paratope residues are coloured by side chain chemistry (black = hydrophobic, green = polar neutral, red = acidic, blue = bas ). Grey indicates amino acids present at the positions but not in the paratope (within 4.5Å of the SARS-CoV-2 antigen). CDRL2 has not been shown, as i contain paratope residues. Produced using Logomaker (57). he framework regions (FWRs). In particular, a serine due at position 83 and/or glycine at position 84 of the light n (part of the formal FWRL3, IMGT numbering) are dis- portionately involved in binding. These residues were nd in 18/48 (38%) of crystallised-antibody paratopes, of which 15 are from the ‘neck’ structural cluster. Using gio (58) to identify the type of binding interactio FWRL3 residues were involved in, 9 paratopes conta drogen bonds between L83 and the RBD, almost al RBD residue 498. . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is Fig. 4. CDR paratope conservation and divergence across the 22 antibodies in the ‘neck’ cluster. Frequency indicates the number of times an amino acid was seen at each IMGT-deﬁned CDR position (56). Results ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint in vitro, antibodies with similar Fv structures should have a high chance of binding to the same surface region and there- fore being complementary to the same epitope, as demon- strated in our analysis of solved structures. Overall, we conclude that structural conservation is impor- tant across the entire Fv, to ensure that the paratope residues, which in most coronavirus-binding antibodies are spread evenly throughout the CDRs, are held in equivalent topolog- ical positions. y Nonetheless, inaccuracy in CDR structure prediction and/or the lack of consideration for paratope residues can lead to misleading clusters of antibodies that have markedly differ- ent functionalities. Therefore, to assess (in the context of this large disease-speciﬁc antibody database) how predictive be- longing to the same structural cluster is of engaging the same epitope, we estimated a pseudo-‘true positive’/‘false positive’ ratio based on the consistency of the CoV-AbDab epitope metadata across antibodies grouped into the same structural cluster. ‘Domain-consistent’ structural clusters were classed as those only containing antibodies reported as binding to internally-consistent antigen domains (see Methods). We now perform a comprehensive structure-based analysis on CoV-AbDab, where 95% of SARS-CoV-2 binding anti- bodies have no solved structure and 27% completely lack binding-domain annotation. Structural Convergence across CoV-AbDab antibod- ies. As of 11th March, just ~5% (113/2,304) of the antibod- ies in CoV-AbDab had at least one solved X-ray or cryo- EM structure, while ~90% (2,063/2,304) of the antibodies had full Fv amino acid sequences (Figure 1). We used high- throughput homology modelling approaches to approximate and analyse the geometries of this much broader set of neu- tralising and non-neutralising antibodies able to bind to mul- tiple coronaviruses, antigens, and domains. A total of 184/200 (92%) of our multiple-occupancy struc- tural clusters were domain-consistent, indicating that struc- turally clustering with another member of CoV-AbDab is likely to be highly predictive of function. p g We used ABodyBuilder (59) to homology model the 2,063 antibody entries in CoV-AbDab with full Fv sequences. This resulted in a total of 1,500 models in which every loop was entirely FREAD-modellable without any need for ab initio loop modelling or backbone adjustment (see Methods); we focus on this subset of models as we have the highest conﬁ- dence in their accuracy (23, 24, 59). Results Paratope residues are coloured by side chain chemistry (black = hydrophobic, green = polar neutral, red = acidic, blue = basic, purple = amide). Grey indicates amino acids present at the positions but not in the paratope (within 4.5Å of the SARS-CoV-2 antigen). CDRL2 has not been shown, as it was found not to contain paratope residues. Produced using Logomaker (57). g. 4. CDR paratope conservation and divergence across the 22 antibodies in the ‘neck’ cluster. Frequency indicates the number of times an amino acid was seen at e Fig. 4. CDR paratope conservation and divergence across the 22 antibodies in the ‘neck’ cluster. Frequency indicates the number of times an amino acid was seen at each IMGT-deﬁned CDR position (56). Paratope residues are coloured by side chain chemistry (black = hydrophobic, green = polar neutral, red = acidic, blue = basic, purple = amide). Grey indicates amino acids present at the positions but not in the paratope (within 4.5Å of the SARS-CoV-2 antigen). CDRL2 has not been shown, as it was found not to contain paratope residues. Produced using Logomaker (57). which 15 are from the ‘neck’ structural cluster. Using Arpeg- gio (58) to identify the type of binding interactions these FWRL3 residues were involved in, 9 paratopes contained hy- drogen bonds between L83 and the RBD, almost always to RBD residue 498. by the framework regions (FWRs). In particular, a serine residue at position 83 and/or glycine at position 84 of the light chain (part of the formal FWRL3, IMGT numbering) are dis- proportionately involved in binding. These residues were found in 18/48 (38%) of crystallised-antibody paratopes, of bioRχiv \| 5 bioRχiv \| 5 Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. Results This represents 72.7% modellability across the set of Fv sequences, a remarkably high percentage relative to recent studies on both healthy and disease-related natural antibody datasets (23, 24). Typically, only ~40% of randomly-sampled human CDRH3s can be ho- mology modelled by FREAD. The increase in modellability is likely to be related to the scientiﬁc effort that has gone into solving a large number of SARS-CoV-2-binding anti- body structures within the ﬁrst year of the pandemic. It also hints at a high degree of underlying structural convergence across the reported coronavirus-binding antibodies in CoV- AbDab. Amongst the 16 structural clusters considered to be ‘false positives’, the domain-inconsistent antibody within four structural clusters (SC2, SC84, SC96, SC136) bore signiﬁ- cant similarities to at least one other antibody in the cluster, suggesting that some experimentally-deduced epitope labels may be inaccurate. The original papers detailing the bind- ing of the antibodies support this possibility. For example, we classify SC2 as domain-inconsistent, as it contains 8 an- tibodies that bind to spike outside the RBD (60, 61), and one (COVA2-32 (62)) labelled as binding to the RBD. How- ever, COVA2-32 only marginally met the threshold AUC to be classiﬁed as an RBD binder (see Figure 4 of Brouwer et al. (62)). The remaining twelve ‘false positive’ structural clusters are likely to result from a combination of inaccurate structural modelling and/or the fact that bearing the same binding site structure does not guarantee functional commonality. We fo- cus the remainder of our analysis on the 92% of domain- consistent structural clusters. After modelling, we performed structural clustering. Brieﬂy, Fvs with 6 modellable CDRs are ﬁrst clustered according to their combination of six CDR lengths, and are then fur- ther structurally grouped by a greedy clustering algorithm that considers the pairwise structural root-mean square de- viation (RMSD) between the selected FREAD template for each CDR region (see Methods for a full description). The result is a set of predicted ‘structural clusters’, each adopted by at least one Fv sequence. Epitope binning. Some members of a structural cluster can have a lower resolution of functional characterisation than others. In these cases, functional properties of the less well- characterised antibodies can be inferred from other antibod- ies predicted to adopt the same structure. Robinson and Raybould et al. \| 6 \| bioRχiv Results Thirty-one antibodies experimentally shown only to bind to the whole spike protein, or to bind the spike protein but not the RBD, can be localised to a more precise epitope using our structural clusters. For example, three of the antibodies as- signed to SC11 that were shown to bind the full-length spike protein, but not a soluble RBD protein (60), can be inferred to bind to the S2 domain in the same way as cluster members DH1147 and DH1149 (61). Similarly, CC12.24, previously shown only to bind to the whole SARS-CoV-2 spike protein (36), can be localised to the same binding site in the RBD as C139 (12) and COVOX-45 (28) (SC57). The 1,500 homology modelled Fv regions fell into 1,159 structural clusters, of which 200 were adopted by more than one Fv sequence (‘multiple-occupancy’ structural clusters). In total, 541/1500 (36.1%) of the antibodies belonged to a multiple-occupancy structural cluster. For a full breakdown of each multiple-occupancy structural cluster, labelled SC0- SC199, see SI Dataset 2. When applied to an entire antibody repertoire, the number of antibodies with similar structures but different functionalities is likely to be signiﬁcant (24). However, on ‘cleaner’ datasets such as CoV-AbDab, where every antibody has been shown to bind a coronavirus antigen (and often a particular domain) A further 62 antibodies fall into 19 structural clusters for which no antibody has been resolved as binding to a par- . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Fig. 5. SARS-CoV-2 RBD-binding antibodies with similar predicted structure that span multiple clonotypes. A representative of structural cluster 3 (SC3, cyan), structural cluster 21 (SC21, orange), and structural cluster 24 (SC24, green) are aligned in the context of the COVOX-253:RBD co-crystal complex (RBD in salmon). The CDRH3 and CDRL3 structures are highlighted in cartoon representation. Results All 100% sequence non-redundant CDRH3 and CDRL3 sequences across the three structures are listed, grouped by lenient VH- or VL-clonotype, with the cluster representative in bold font. The alignment shows the various CDRH3 and CDRL3 structures are likely to be topologically compatible with this RBD epitope. Residues E484 and N501, commonly mutated in SARS-CoV-2 variants of concern, are highlighted as sticks and coloured by default atom type. Closest heavy-atom distances between the functional group (carboyxlate/amide heteroatom) and the different structural classes of antibody are shown. Fig. 5. SARS-CoV-2 RBD-binding antibodies with similar predicted structure that span multiple clonotypes. A representative of structural cluster 3 (SC3, cyan), structural cluster 21 (SC21, orange), and structural cluster 24 (SC24, green) are aligned in the context of the COVOX-253:RBD co-crystal complex (RBD in salmon). The CDRH3 and CDRL3 structures are highlighted in cartoon representation. All 100% sequence non-redundant CDRH3 and CDRL3 sequences across the three structures are listed, grouped by lenient VH- or VL-clonotype, with the cluster representative in bold font. The alignment shows the various CDRH3 and CDRL3 structures are likely to be topologically compatible with this RBD epitope. Residues E484 and N501, commonly mutated in SARS-CoV-2 variants of concern, are highlighted as sticks and coloured by default atom type. Closest heavy-atom distances between the functional group (carboyxlate/amide heteroatom) and the different structural classes of antibody are shown. coronavirus neutralising antibodies, and could be exploited in epitope-focused vaccine design strategies (64) to achieve a more broadly neutralising response. ticular domain. For some structures, the selected FREAD templates could offer an indication of epitope speciﬁcity: the PDB structure ‘6nb8’ (63) is used to model all three light chain CDRs in at least one antibody assigned to SC6, SC15, and SC71, while ‘7ben’ (28) is used to model the CDRH3 in all four antibodies assigned to SC16. The antibody in both of these PDB structures engages the SARS-CoV-2 RBD, which could imply that the cluster groups RBD-complementary an- tibodies. Even ignoring these indications, just 19 binding characterisation experiments could lend functional annota- tions to over three-times the number of CoV-AbDab entries (62). The epitope and corresponding paratope residues within these binding sites will differ between lineages, meaning that sequence-based clustering approaches would struggle to spot their functional commonality. Structural clustering can see beyond paratope proﬁles to capture broader epitope topology conservation via the geometries of their cognate antibodies. Epitopes targetable by multiple species. Robinson and Raybould et al. \| Structural clusters frequently span multiple clonal lineages. Structural clusters frequently span multiple clonal lineages. We analysed each of the 184 domain-consistent structural clusters to determine how often the antibodies clustered to- gether belonged to multiple lenient Fv clonotypes. A total of 88 (47.8%) contained at least one pair of antibodies from dif- ferent lenient Fv clonotypes and 73 (39.7%) of the structural clusters contain at least two lenient VH-only clonotypes. It is clear that antibodies with both heavy and light chains of dif- fering clonality can frequently co-exist within our structural clusters. Many structural clusters contain at least one pair of antibod- ies from the same clonotype; this is unsurprising since the ‘near-identical sequence, similar function’ assumption under- pinning clonotyping experiments is often correct. However, the high frequency with which we group antibodies spanning several clonotypes into the same structural cluster recapitu- lates the ﬁndings of other papers (8–10), and our earlier anal- ysis on solved structures, that clonotyping cannot group to- gether all antibodies capable of same-epitope engagement. In most cases where multiple clonotypes are found in the same structural cluster, it is due to signiﬁcant dif- ferences in the CDRH3 sequence. However, some clus- ters such as SC134 (which pools COV2-2490 (60) with H712061+K711727 (61)), align closest to different heavy V (IGHV3-7 vs. IGHV3-30) and light V (IGKV1-5 vs. IGKV1D-16) genes. g Soon after we identiﬁed this broad structural cluster, a preprint was released by Schmitz et al. (70) showing that many IGHV1-58-encoded SARS-CoV-2 binding antibodies have highly similar residues at equivalent paratope positions (deﬁned by the S2-E12 crystal structure (71)). On this occa- sion, the other CDRH3 sequence-diverse antibodies hypoth- esised to have similar function were shortlisted through in- spection of their sequences; they all derive from IGHV1-58 germline and bear the -C(X)4C- motif within their CDRH3 loop (X4 representing four non-cysteine residues). We have shown that structural modelling and clustering supports the theory that these antibodies are functionally similar, and of- fers a systematic route to the identiﬁcation of other sequence- diverse clusters of functionally-common antibodies that do not bear such clearly conserved motifs. ‘Public’ Response Antibodies. ‘Public’ antibodies, those that are raised independently across multiple individuals against an immunodominant epitope, are of high interest to several ﬁelds of research, from vaccinology to drug discovery (67– 69). Results It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint IGHV1-58/IGKV2-30 gene transcripts but have sequence- diverse CDRH3s all containing a common intra-loop disul- ﬁde bridge. of Ab_511E7). Even more remarkably, SC174 pools anti- bodies from different species conﬁrmed to bind to different coronaviruses; in SC174, a human antibody (C131 (12)) that binds the SARS-CoV-2 RBD is grouped with a murine an- tibody (F26G18 (66)) shown to bind the full-length SARS- CoV-1 spike protein. Should F26G18 be conﬁrmed to engage the RBD, and the SC62 antibodies compete for the same epi- tope, this would show the ability of structural clustering to identify cross-coronavirus epitopes targetable by multiple or- ganisms. The SC21 and SC24 clusters also map to these genes and contain a disulﬁde bridge; antibodies assigned to SC21 and SC3 have identical CDR lengths but are predicted to have different CDR structures, while SC24 is necessarily classed as a separate structural cluster to SC3 as its CDRL3 loop is of a different length (10 residues rather than 9). g ( ) When aligned to the solved COVOX-253:RBD co-crystal structure, SC3, SC21, and SC24 appear to all be topolog- ically complimentary to the same RBD epitope (Figure 5). The length-10, more protruding CDRL3 loop of SC24 is ac- commodated by the small G485 residue on the RBD, while the CDRH3s across all three structural clusters protrude to a similar extent towards residues F456-N460 on the RBD. The antibodies mapped to SC21 and SC24 (from an ad- ditional three independent sources) comprise an additional three lenient Fv-clonotypes, making a total of eight lenient Fv-clonotypes with potential same-epitope complementarity. This set of similar structures has been observed across eight independent studies indicating that the corresponding epitope is immunodominant. Moreover, none of the antibodies di- rectly engage the carboxylate group of residue E484 (5Å for length-10 CDRL3s or 11Å for length-9 CDRL3s, with too acute an angle for hydrogen bonding) nor the amide group of N501 (≥12Å for all antibody topologies). Structural clusters frequently span multiple clonal lineages. Several studies have already identiﬁed public SARS-CoV-2 response antibodies based on convergence towards particular clonotype lineages (12–18), but none have yet considered the fact that antibodies from different lineages can exert similar functions. As demonstrated above, structural clustering en- ables us to group together clonally-distinct antibodies with a high chance of engaging the same epitope. We therefore ex- amined our structural clusters to reveal functionally similar groups of antibodies from different genetic lineages that have been independently isolated across several different studies (i.e. “public structures”). 8 \| bioRχiv Robinson and Raybould et al. \| Results This should make them of particular interest as clones that might neutralise both wildtype SARS-CoV-2 and the more recent E484K/N501Y- containing variants of concern. Results A recent study has shown that mouse and human antibodies use a different distri- bution of CDRH3 structures, which also varies by B-cell mat- uration stage (23). This can be rationalised by their species- speciﬁc gene loci having different predetermined structural biases and the fact that negative selection occurs against dif- ferent self-epitopes. Nevertheless, assuming that each CDR loop can only adopt a ﬁnite set of geometries imposed by the loop closure criterion, and that many of the same antigens would be considered pathogenic to both species, there ought to be some structural overlap between human and mouse an- tibodies and therefore the potential for some epitopes to be targetable by both species. These would be extremely hard to identify by sequence alone, as human and murine gene loci are highly sequence dissimilar. Evolutionarily-conserved epitope topologies across coro- naviruses. Evolutionarily-conserved epitope topologies are implied by our structural clusters that contain antibodies able to bind multiple coronaviruses. As an extreme example, SC0 pools together a set of 13 IGHV1-69-derived antibodies, of which at least one which has been shown to engage each of SARS-CoV-1, SARS-CoV-2, HKU1, 229E, NL63, OC43, and MERS-CoV. More broadly, 69/184 (37.5%) of our struc- tural clusters contain an antibody shown on the current lev- els of data to have cross-coronavirus binding potential. This number may be an underestimate, as several antibodies have only been tested against a single coronavirus strain in vitro. These epitopes could represent sites of particular vulnera We identiﬁed SC62, which groups human antibody Ab_511E7 (65) alongside the two murine antibodies DK4 & DK7 (patent CN111978395A), all of which have been shown to bind the SARS-CoV-2 RBD (albeit only weakly in the case These epitopes could represent sites of particular vulnera- bility across coronaviruses; antigen regions whose structure must be preserved for viral function. They therefore reﬂect the most promising regions against which to design pan- bioRχiv \| 7 bioRχiv \| 7 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. Discussion ple, a cluster of 22 antibodies from 15 different lenient VH clonotypes all approach the same SARS-CoV-2 RBD ‘neck’ epitope with a closely-related binding mode. The paratopes of these antibodies are highly conserved across the CDRH1 and CDRH2 regions (accounting for the strong bias towards IGHV3-53/IGHV3-66 gene origins) while the CDRH3 se- quence can diverge substantially in sequence identity. The same phenomenon was observed for two antibodies solved engaging the RBD ‘left ﬂank’ with a near-identical binding mode but just 43% CDRH3 sequence identity. These scenar- ios represent a problem for the functional interpretation of sequence-based clustering approaches such as clonotyping; antibodies that are functionally similar would be binned into different clusters. Inspired by the structural similarity of an- tibodies that bind to the same epitope, we predicted and clus- tered the structures of the broad set of antibodies documented in CoV-AbDab (1). Overall, our results show that structural information via computational modeling enhances the picture of disease- characteristic convergence across SARS-CoV-2 response an- tibodies. It is clear that while clonotyping COVID-19 anti- body repertoires can offer an indication for which epitopes are public in the response, they risk understating the true lev- els of same-epitope reactivity across individuals. Accurately capturing the functions of the antibodies raised during the im- mune response is critical when evaluating vaccine efﬁcacy, both against the injected viral strain and against new vari- ants of concern that may arise and alter the immunogenicity of certain epitopes. Structure prediction and clustering has a crucial role to play alongside clonotyping to yield the max- imum functional inference from the vast amount of disease- speciﬁc antibody data available. We found that this structural clustering is likely to achieve a very high accuracy of epitope binning. Up to 92% of multiple-occupancy structural clusters grouped antibodies re- ported to bind to consistent domains, based on the current levels of metadata in CoV-AbDab. This suggests that we can use this method to predict the epitopes of many as-yet un- characterised coronavirus binders, as well as prospectively to predict the epitopes of newly-isolated SARS-CoV-2 binding antibodies. These structural clusters also offer a unique perspective on the data that can not be identiﬁed through standard, sequence- based clonotyping. First, they can functionally-associate antibodies that derive from highly distinct clonal lineages; around 40% of our structural clusters contain at least two antibodies from different lenient VH clonotypes. Discussion They also functionally connect disease-response antibodies that origi- nate from different species, of interest in the study of func- tional crossovers between immune repertoires that exploit different gene loci. Moreover, they can reveal which epi- tope topologies are likely to be conserved across coronavirus strains, helping to co-ordinate efforts to design prophylactics towards more fruitful sites for pan-coronavirus neutralisation. Robinson and Raybould et al. \| Discussion It is clear that while clonotyping COVID-19 anti- body repertoires can offer an indication for which epitopes are public in the response, they risk understating the true lev- els of same-epitope reactivity across individuals. Accurately capturing the functions of the antibodies raised during the im- mune response is critical when evaluating vaccine efﬁcacy, both against the injected viral strain and against new vari- ants of concern that may arise and alter the immunogenicity of certain epitopes. Structure prediction and clustering has a crucial role to play alongside clonotyping to yield the max- imum functional inference from the vast amount of disease- speciﬁc antibody data available. ple, a cluster of 22 antibodies from 15 different lenient VH clonotypes all approach the same SARS-CoV-2 RBD ‘neck’ epitope with a closely-related binding mode. The paratopes of these antibodies are highly conserved across the CDRH1 and CDRH2 regions (accounting for the strong bias towards IGHV3-53/IGHV3-66 gene origins) while the CDRH3 se- quence can diverge substantially in sequence identity. The same phenomenon was observed for two antibodies solved engaging the RBD ‘left ﬂank’ with a near-identical binding mode but just 43% CDRH3 sequence identity. These scenar- ios represent a problem for the functional interpretation of sequence-based clustering approaches such as clonotyping; antibodies that are functionally similar would be binned into different clusters. Inspired by the structural similarity of an- tibodies that bind to the same epitope, we predicted and clus- tered the structures of the broad set of antibodies documented in CoV-AbDab (1). As more coronavirus-binding antibody structures continue to be released to the PDB, the coverage and expected accu- racy of the CoV-AbDab homology models will increase ac- cordingly, likely further improving the accuracy of our epi- tope binning over time. However, solved antibody-pandemic virus structures are not a prerequisite of meaningful epi- tope binning via predicted structure. Forty-eight of our domain-consistent structural clusters currently connect anti- bodies that bind outside of the RBD, despite the fact no high- quality X-ray structures of antibodies binding outside the SARS-CoV-2 RBD had been solved at the time of this anal- ysis (≤2.5Å resolution, a requirement for use as an ABody- Builder template). This indicates that structural clustering is able to draw functional connections between antibodies iso- lated at the start of a pandemic, even if the number of solved antibody-pandemic virus structures is very low. Discussion Here, we have analysed the solved X-ray crystal structures of antibodies and nanobodies bound to SARS-CoV-2 from the perspective of their structural and paratope conservation, and performed the ﬁrst analysis of thousands of structural models of coronavirus-binding antibodies reported in over 100 inde- pendent literature sources. We have updated the previously-reported sets of antibod- ies shown to bind to speciﬁc regions of the RBD (28) and demonstrated that the antibodies within these clusters are of- ten structurally similar but sequence dissimilar. For exam- A striking example of a public structure that spans multi- ple clonotypes is SC3 (Figure 5). SC3 contains nine an- tibodies from ﬁve independent sources, spanning ﬁve le- nient Fv-clonotypes. All antibodies align closest to the Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint As more coronavirus-binding antibody structures continue to be released to the PDB, the coverage and expected accu- racy of the CoV-AbDab homology models will increase ac- cordingly, likely further improving the accuracy of our epi- tope binning over time. However, solved antibody-pandemic virus structures are not a prerequisite of meaningful epi- tope binning via predicted structure. Forty-eight of our domain-consistent structural clusters currently connect anti- bodies that bind outside of the RBD, despite the fact no high- quality X-ray structures of antibodies binding outside the SARS-CoV-2 RBD had been solved at the time of this anal- ysis (≤2.5Å resolution, a requirement for use as an ABody- Builder template). This indicates that structural clustering is able to draw functional connections between antibodies iso- lated at the start of a pandemic, even if the number of solved antibody-pandemic virus structures is very low. Overall, our results show that structural information via computational modeling enhances the picture of disease- characteristic convergence across SARS-CoV-2 response an- tibodies. Methods CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: oRxiv preprint SARS-CoV-2-RBD chain using PyMOL functions. Paratope residues were deﬁned as any antibody residues with a heavy atom with 4.5Å of an antigen heavy atom. We refer to an- tibodies by name as referenced in CoV-AbDab, using the nomenclature as set out in the literature from which the anti- body was sourced. See Table S1 and Table S2 for the names and corresponding PDB codes of antibodies. SARS-CoV-2-RBD chain using PyMOL functions. Paratope residues were deﬁned as any antibody residues with a heavy atom with 4.5Å of an antigen heavy atom. We refer to an- tibodies by name as referenced in CoV-AbDab, using the nomenclature as set out in the literature from which the anti- body was sourced. See Table S1 and Table S2 for the names and corresponding PDB codes of antibodies. 2. a structural cluster that only contains antibodies char- acterised as binding to internally-consistent domains (e.g. some antibodies labeled as spike N-Terminal Domain (NTD) binders and others labeled as S1 non-RBD binders; where the S NTD is a subdomain of S1 non-RBD). 3. a structural cluster that contains some antibodies that are characterised as binding to the same domain, and others that bind to the same antigen without domain-level resolution (e.g. 4 antibodies shown to bind the spike RBD, and 2 antibodies shown to bind to the full-length spike protein). Structural Modeling and Analysis. The 2,063 full vari- able domain (Fv) sequences in CoV-AbDab were submit- ted to the ABodyBuilder antibody modelling tool (59) with default Environment Substitution Score cutoffs. In the ﬁrst instance, ABodyBuilder seeks to model antibody CDR re- gions entirely by homology; i.e. Methods to use the FREAD software (72, 73) to identify a CDR structural ‘template’ likely to be adopted by each of the submitted antibody’s CDR sequences, considering backbone dihedral angle compatibility and loop graftability onto the framework template. If no suitable struc- tural template can be found for a CDR sequence, ab initio or hybrid homology/ab initio approaches must be used to predict the loop structure, adding uncertainty to model qual- ity. To ensure high model quality, only the 1,500 models for which ABodyBuilder used FREAD to homology model all six CDR loops were carried forward for structural clustering (59) (these models are available as SI Dataset 1). 4. a structural cluster that only contains antibodies char- acterised to bind to the same antigen, but no antibody has domain-level resolution (e.g. 5 antibodies all shown to bind to the full-length spike protein, but none are localised to a particular domain). The following structural clusters would both be considered ‘domain-inconsistent’: 1. a structural cluster that contains antibodies shown to bind to different antigens. 1. a structural cluster that contains antibodies shown to bind to different antigens. 2. a structural cluster that contains antibodies that bind to the same antigen, but to inconsistent domains (e.g. 3 antibodies that have been shown to bind the spike RBD and 1 shown to bind the spike S2 domain). Structural Clustering Algorithm. These 1,500 Fvs were ﬁrst split by their combination of six CDR lengths. For each unique CDR length combination, the ﬁrst antibody in the list was selected as a cluster centre and every subsequent anti- body is fed into the following equation: Clonotyping. Clonotyping was performed using an in-house script. Our lenient VH-clonotyping protocol groups Fvs with matching IGHV genes, the same length CDRH3, and ≥80% CDRH3 sequence identity, while our lenient Fv-clonotyping protocol additionally requires cluster members to have a matching IG[K/L]V gene, the same length CDRL3, and ≥ 80% CDRL3 sequence identity. These are lenient clonotyp- ing threshold conditions by community standards (7), as the CDR3 sequence identity threshold is set to its typical lower bound and there is no requirement for cluster members to map to the same IGHJ/IG[K/L]J gene. AUTHOR CONTRIBUTIONS All authors contributed to the design of the research; SAR and MIJR performed the research; SAR, MIJR, CS, and WKW compiled datasets; SAR, MIJR, and CMD wrote the original manuscript; all authors contributed to reviewing and editing the manuscript. ACKNOWLEDGEMENTS This work was supported by a Engineering and Physical Sciences Research Coun- cil and Medical Research Council Grant [EP/L016044/1] and further research fund- ing from AstraZeneca and UCB Pharma. Methods v u u t P(H1−H3,L1−L3) X D2 X12LX P(H1−H3,L1−L3) X LX where the sum over X refers to each of the six CDRs, LX is the length of North CDRX, and DX12 is the Cα RMSD be- tween the template used to model CDRX in Fv 1 and Fv 2 respectively. If this formula equates to ≤0.75Å, the Fv is clustered with the ﬁrst cluster centre, otherwise it is held out for the next round of clustering. Once all the Fvs have been considered relative to the ﬁrst cluster centre, the algorithm progresses in a greedy fashion to select the next unclustered Fv region as the second cluster centre. The result is a set of structural cluster centres and associated antibody Fv se- quences, where each structural cluster only contains antibod- ies with six identical CDR lengths. This algorithm is adapted from the ﬁnal step of Repertoire Structural Proﬁling (24). Bibliography 1. Matthew IJ Raybould, Aleksandr Kovaltsuk, Claire Marks, and Charlotte M Deane. CoV- AbDab: the Coronavirus Antibody Database. Bioinformatics, page btaa739, 2020. doi: 10.1093/bioinformatics/btaa739. ‘Domain-Consistent’ Structural Clusters. The multiple- occupancy structural clusters were each classiﬁed as ‘domain-consistent/inconsistent’ based on the CoV-AbDab binding metadata of their mapped antibody sequences. For instance, the following examples of structural clusters would each be considered as ‘domain-consistent’: 2. Renhong Yan, Yuanyuan Zhang, Yaning Li, Lu Xia, Yingying Guo, et al. Structural basis for the recognition of SARS-CoV-2 by full-length human ACE2. Science, 367(6485):1444– 1448, 2020. doi: 10.1126/science.abb2762. 3. Jinsung Yang, Simon JL Petitjean, Melanie Koehler, Qingrong Zhang, Andra C Dumitru, et al. Molecular interaction and inhibition of SARS-CoV-2 binding to the ACE2 receptor. Nat Commun, 11:4541, 2020. doi: 10.1038/s41467-020-18319-6. 4. Jacob D Galson, Sebastian Schaetzle, Rachael JM Bashford-Rogers, Matthew IJ Raybould, Aleksandr Kovaltsuk, et al. Deep sequencing of B cell receptor repertoires from COVID-19 patients reveals strong convergent immune signatures. Front Immunol, 11:605170, 2020. doi: 10.3389/ﬁmmu.2020.605170. 1. a structural cluster that only contains antibodies charac- terised as binding to the same antigen and the same domain (e.g. all shown to bind the RBD of the spike protein). 1. a structural cluster that only contains antibodies charac- terised as binding to the same antigen and the same domain (e.g. all shown to bind the RBD of the spike protein). 5. Rafael R De Assis, Aarti Jain, Rie Nakajima, Algis Jasinskas, Jiin Felgner, et al. Analysis of SARS-CoV-2 antibodies in COVID-19 convalescent blood using a coronavirus antigen microarray. Nat Commun, 12:6, 2021. doi: 10.1038/s41467-020-20095-2. 10 \| bioRχiv Robinson and Raybould et al. \| Methods Database Preparation. The version of CoV-AbDab (1) used throughout this analysis was timestamped to the 11th March 2021. The framework and 6 FREAD (72, 73) CDR loop databases, which were used during structural modeling to ﬁnd suitable homologous templates for each antibody re- gion, were also timestamped to contain the quality-ﬁltered contents of SAbDab (74) on 11th March 2021. Quality ﬁlter- ing restricts templates to those solved by X-ray crystallogra- phy, with a resolution ≤2.5Å and a B-factor < 80. Such powerful structure-function relationships are likely only possible due to the collation of clean, high conﬁdence bind- ing data. Our work demonstrates the clear value of building large class-speciﬁc databases of antibody-binders against ex- tracellular disease-associated antigens. Numbering Scheme and Region Deﬁnitions. IMGT numbering (56) is used throughout the manuscript. IMGT CDR region deﬁnitions are used to analyse the solved SARS- CoV-2 structures. ABodyBuilder uses North CDR deﬁnitions in template selection. The North-deﬁned and IMGT-deﬁned CDR3 region lies between IMGT residue numbers 105 and 117 in both the heavy and light chains, meaning clonotype deﬁnitions are consistent regardless of region deﬁnition (see Clonotyping). In addition to offering functional annotations to as-yet un- characterised antibodies, these databases could also be used to identify important gaps in structural space. This would en- sure that time-consuming experimental structure determina- tion efforts are targeted towards sets of antibodies that yield maximal functional insight. For structural clusters with lim- ited existing epitope knowledge, a central antibody could be selected for structural evaluation with the cognate antigen, enabling the functional annotation of many other antibodies in the same structural cluster. Similarly, structurally mod- elling the entire database reveals which antibodies cannot currently be accurately modeled and should therefore be pri- oritised for experimental structure characterisation. Solved Co-crystal Structure Analysis. Sixty solved X- ray co-crystal structures of antibodies and nanobodies bound to SARS-CoV-2 were downloaded from SAbDab. All anti- bodies were aligned based on the coordinates of the cognate bioRχiv \| 9 bioRχiv \| 9 . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint . Robinson and Raybould et al. \| Robinson and Raybould et al. \| Lewis, Jiye Shi, James Snowden, Bruck Taddese, and Charlotte M. Deane. Ab-Ligity: Identify- ing sequence-dissimilar antibodies that bind to the same epitope. mAbs, 13(1):1873478, 2021. doi: 10.1101/2020.03.24.004051. 38. Nicholas C Wu, Meng Yuan, Hejun Liu, Chang-Chun D Lee, Xueyong Zhu, Sandhya Ban- garu, Jonathan L Torres, Tom G Caniels, Philip JM Brouwer, Marit J Van Gils, et al. An alternative binding mode of IGHV3-53 antibodies to the SARS-CoV-2 receptor binding do- main. Cell Rep, 33(3):108274, 2020. doi: 10.1016/j.celrep.2020.108274. 11. Cinque Soto, Robin G. Bombardi, Andre Branchizio, Nurgun Kose, Pranathi Matta, et al. High frequency of shared clonotypes in human B cell receptor repertoires. Nature, 566: 398–402, 2019. doi: 10.1038/s41586-019-0934-8. 39. Nicholas C Wu, Meng Yuan, Sandhya Bangaru, Deli Huang, Xueyong Zhu, et al. A nat- ural mutation between SARS-CoV-2 and SARS-CoV determines neutralization by a cross- reactive antibody. PLoS Pathog, 16(12):e1009089, 2020. doi: 10.1371/journal.ppat.1009089. 12. Davide F Robbiani, Christian Gaebler, Frauke Muecksch, Julio CC Lorenzi, Zijun Wang, et al. Convergent antibody responses to SARS-CoV-2 in convalescent individuals. Nature, 584(7821):437–442, 2020. doi: 10.1038/s41586-020-2456-9. 40. Nicholas K Hurlburt, Emilie Seydoux, Yu-Hsin Wan, Venkata Viswanadh Edara, Andrew B Stuart, et al. Structural basis for potent neutralization of SARS-CoV-2 and role of antibody afﬁnity maturation. Nat Commun, 11(1):5413, 2020. doi: 10.1038/s41467-020-19231-9. 13. Fan Zhang, Rui Gan, Ziqi Zhen, Xiaoli Hu, Xiang Li, et al. Adaptive immune responses to SARS-CoV-2 infection in severe versus mild individuals. Signal Transduct Target Ther, 5: 156, 2020. doi: 10.1038/s41392-020-00263-y. 41. Daming Zhou, Helen ME Duyvesteyn, Cheng-Pin Chen, Chung-Guei Huang, Ting-Hua Chen, et al. Structural basis for the neutralization of SARS-CoV-2 by an antibody from a convalescent patient. Nat Struct Mol Biol, 27(10):950–958, 2020. doi: 10.1038/ s41594-020-0480-y. 14. Jakob Kreye, S Momsen Reincke, Hans-Christian Kornau, Elisa Sánchez-Sendin, Victor M Corman, et al. A SARS-CoV-2 neutralizing antibody protects from lung pathology in a COVID-19 hamster model. Cell, 183(4):1058–1069, 2020. doi: 10.1016/j.cell.2020.09.049. 42. Edurne Rujas, Iga Kucharska, Yong Zi Tan, Samir Benlekbir, Hong Cui, et al. Multivalency transforms SARS-CoV-2 antibodies into broad and ultrapotent neutralizers. bioRxiv, 2020. doi: 10.1101/2020.10.15.341636. 15. Georgia Bullen, Jacob D Galson, Pedro Villar, Lien Moreels, Line Ledsgaard, et al. Deep mining of early antibody response in COVID-19 patients yields potent neutralisers and re- veals high level of convergence. bioRxiv, 2020. doi: 10.1101/2020.12.29.424711. 43. Yu Guo, Lisu Huang, Guangshun Zhang, Yanfeng Yao, He Zhou, Shu Shen, Bingqing Shen, Bo Li, Xin Li, Mingjie Chen, et al. Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint 33. Shuo Du, Yunlong Cao, Qinyu Zhu, Guopeng Wang, Xiaoxia Du, et al. Structures of potent and convergent neutralizing antibodies bound to the SARS-CoV-2 spike unveil a unique epitope responsible for exceptional potency. bioRxiv, 2020. doi: 10.1101/2020.07.09.195263. 6. Ping Zhao, Jie Cao, Lan-Juan Zhao, Zhao-Lin Qin, Jin-Shan Ke, et al. Immune responses against SARS-coronavirus nucleocapsid protein induced by DNA vaccine. Virology, 331(1): 128–135, 2005. doi: 10.1016/j.virol.2004.10.016. 33. Shuo Du, Yunlong Cao, Qinyu Zhu, Guopeng Wang, Xiaoxia Du, et al. Structures of potent and convergent neutralizing antibodies bound to the SARS-CoV-2 spike unveil a unique epitope responsible for exceptional potency. bioRxiv, 2020. doi: 10.1101/2020.07.09.195263. 7. Laura López-Santibáñez-Jácome, S. Eréndira Avendaño-Vázquez, and Carlos Fabián Flores-Jasso. The Pipeline Repertoire for Ig-Seq Analysis. Front Immunol, page 899, 2019. doi: 10.3389/ﬁmmu.2019.00899. 34. Sarah Ashley Clark, Lars Eric Clark, Junhua Pan, Adrian Coscia, Lindsay GA McKay, et al. Molecular basis for a germline-biased neutralizing antibody response to SARS-CoV-2. bioRxiv, 2020. doi: 10.1101/2020.11.13.381533. 35. Rui Shi, Chao Shan, Xiaomin Duan, Zhihai Chen, Peipei Liu, et al. A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2. Nature, 584(7819):120–124, 2020. doi: 10.1038/s41586-020-2381-y. 8. Aleksandr Kovaltsuk, Konrad Krawczyk, Jacob D. Galson, Dominic F. Kelly, Charlotte M. Deane, and Johannes Trück. How B-Cell Receptor Repertoire Sequencing Can Be Enriched with Structural Antibody Data. Front Immunol, 8:1753, 2017. doi: 10.3389/ﬁmmu.2017. 01753. y 36. Thomas F Rogers, Fangzhu Zhao, Deli Huang, Nathan Beutler, Alison Burns, et al. Isolation of potent SARS-CoV-2 neutralizing antibodies and protection from disease in a small animal model. Science, 369(6506):956–963, 2020. doi: 10.1126/science.abc7520. 9. Eve Richardson, Jacob D Galson, Paul Kellam, Dominic F Kelly, Sarah E Smith, et al. A computational method for immune repertoire mining that identiﬁes novel binders from different clonotypes, demonstrated by identifying anti-Pertussis toxoid antibodies. mAbs, 13 (1):1869406, 2021. doi: 10.1080/19420862.2020.1869406. 37. Hejun Liu, Nicholas C Wu, Meng Yuan, Sandhya Bangaru, Jonathan L Torres, et al. Cross- neutralization of a SARS-CoV-2 antibody to a functionally conserved site is mediated by avidity. Immunity, 53(6):1272–1280, 2020. doi: 10.1016/j.immuni.2020.10.023. 10. Wing Ki Wong, Sarah A. Robinson, Alexander Bujotzek, Guy Georges, Alan P. Robinson and Raybould et al. \| A SARS-CoV-2 neutralizing antibody with exceptional spike binding coverage and optimized therapeutic potentials. Research Square (Nature Preprint), 2020. doi: 10.21203/rs.3.rs-78945/v1. 16. Aurélien Sokal, Pascal Chappert, Giovanna Barba-Spaeth, Anais Roeser, Slim Fourati, et al. Maturation and persistence of the anti-SARS-CoV-2 memory B cell response. Cell, 184(5): 1201–1213, 2021. doi: 10.1016/j.cell.2021.01.050. 44. Cheolmin Kim, Dong-Kyun Ryu, Jihun Lee, Young-Il Kim, Ji-Min Seo, et al. A therapeutic neutralizing antibody targeting receptor binding domain of SARS-CoV-2 spike protein. Nat Commun, 12(1):288, 2021. doi: 10.1038/s41467-020-20602-5. 17. Christoph Schultheiß, Lisa Paschold, Donjete Simnica, Malte Mohme, Edith Willscher, et al. Next-generation sequencing of T and B cell receptor repertoires from COVID-19 patients showed signatures associated with severity of disease. Immunity, 53(2):442–455, 2020. doi: 10.1016/j.immuni.2020.06.024. 45. Luca Piccoli, Young-Jun Park, M Alejandra Tortorici, Nadine Czudnochowski, Alexandra C Walls, et al. Mapping neutralizing and immunodominant sites on the SARS-CoV-2 spike receptor-binding domain by structure-guided high-resolution serology. Cell, 183(4):1024– 1042, 2020. doi: 10.1016/j.cell.2020.09.037. 18. Zachary Montague, Huibin Lv, Jakub Otwinowski, William S DeWitt, Giulio Isacchini, et al. Dynamics of B-cell repertoires and emergence of cross-reactive responses in COVID-19 patients with different disease severity. arXiv, 2020. doi: arXiv:2007.06762. 46. Federico Bertoglio, Viola Fühner, Maximilian Ruschig, Philip Alexander Heine, Ulfert Rand, et al. A SARS-CoV-2 neutralizing antibody selected from COVID-19 patients by phage dis- play is binding to the ACE2-RBD interface and is tolerant to known RBD mutations. bioRxiv, 2020. doi: 10.1101/2020.12.03.409318. 19. Sandra C. A. Nielsen, Fan Yang, Ramona A. Hoh, Katherine J. L. Jackson, Katharina Roelt- gen, et al. Human B Cell Clonal Expansion and Convergent Antibody Responses to SARS- CoV-2. Cell Host Microbe, 28(4):516–525, 2020. doi: 10.1016/j.chom.2020.09.002. 20. Bin Ju, Qi Zhang, Jiwan Ge, Ruoke Wang, Jing Sun, et al. Human neutralizing antibod- ies elicited by SARS-CoV-2 infection. Nature, 584(7819):115–119, 2020. doi: 10.1038/ s41586-020-2380-z. 47. Jiangdong Huo, Audrey Le Bas, Reinis R Ruza, Helen ME Duyvesteyn, Halina Mikolajek, et al. Neutralizing nanobodies bind SARS-CoV-2 spike RBD and block interaction with ACE2. Nat Struct Mol Biol, 27(9):846–854, 2020. doi: 10.1038/s41594-020-0469-6. 21. Meng Yuan, Hejun Liu, Nicholas C Wu, Chang-Chun D Lee, Xueyong Zhu, et al. Struc- tural basis of a shared antibody response to SARS-CoV-2. Science, 369(6507):1119–1123, 2020. doi: 10.1126/science.abd2321. 48. Hebang Yao, Hongmin Cai, Tingting Li, Bingjie Zhou, Wenming Qin, et al. A high-afﬁnity RBD-targeting nanobody improves fusion partner’s potency against SARS-CoV-2. PLoS Pathog, 17(3):e1009328, 2021. doi: 10.1371/journal.ppat.1009328. 22. Robinson and Raybould et al. \| Haitao Xiang, Yingze Zhao, Xinyang Li, Peipei Liu, Longlong Wang, et al. Landscapes and dynamic diversiﬁcations of B-cell receptor repertoires in COVID-19 patients. bioRxiv, 2020. doi: 10.1101/2020.12.28.424622. 49. Yufei Xiang, Sham Nambulli, Zhengyun Xiao, Heng Liu, Zhe Sang, et al. Versatile and multivalent nanobodies efﬁciently neutralize SARS-CoV-2. Science, 370(6523):1479–1484, 2020. doi: 10.1126/science.abe4747. 23. Aleksandr Kovaltsuk, Matthew IJ Raybould, Wing Ki Wong, Claire Marks, Sebastian Kelm, et al. Structural diversity of B-cell receptor repertoires along the B-cell differentiation axis in humans and mice. PLOS Computational Biology, 16(2):e1007636, 2020. doi: 10.1371/ journal.pcbi.1007636. 50. Daniel Wrapp, Dorien De Vlieger, Kizzmekia S Corbett, Gretel M Torres, Nianshuang Wang, et al. Structural basis for potent neutralization of betacoronaviruses by single-domain camelid antibodies. Cell, 181(5):1004–1015, 2020. doi: 10.1016/j.cell.2020.04.031. 24. Matthew I. J. Raybould, Claire Marks, Aleksandr Kovaltsuk, Alan P. Lewis, Jiye Shi, and Charlotte M. Deane. Public Baseline and Shared Response Structures Support the Theory of Antibody Repertoire Functional Commonality. PLoS Computat Biol, 17(3):e1008781, 2021. doi: 10.1371/journal.pcbi.1008781. 51. Paul-Albert Koenig, Hrishikesh Das, Hejun Liu, Beate M Kümmerer, Florian N Gohr, et al. Structure-guided multivalent nanobodies block SARS-CoV-2 infection and suppress muta- tional escape. Science, 371(6530):eabe6230, 2021. doi: 10.1126/science.abe6230. 52. Tingting Li, Hongmin Cai, Hebang Yao, Bingjie Zhou, Yapei Zhao, et al. A potent synthetic nanobody targets RBD and protects mice from SARS-CoV-2 infection. BioRxiv, 2020. doi: 10.1101/2020.06.09.143438. 25. Anthony R Rees. Understanding the human antibody repertoire. MAbs, 12(1):1729683, 2020. doi: 10.1080/19420862.2020.1729683. 26. Bryan Briney, Anne Inderbitzin, Collin Joyce, and Dennis R. Burton. Commonality despite exceptional diversity in the baseline human antibody repertoire. Nature, 566(7744):393– 397, 2019. doi: 10.1038/s41586-019-0879-y. 53. Micah Rapp, Yicheng Guo, Eswar Reddy Reddem, Lihong Liu, Pengfei Wang, et al. Modular basis for potent SARS-CoV-2 neutralization by a prevalent VH1-2-derived antibody class. Cell Rep, 35(1):108950, 2021. doi: 10.1016/j.celrep.2021.108950. y 27. Rohit Arora and Ramy Arnaout. Private Antibody Repertoires Are Public. bioRxiv, 2020. doi: 10.1101/2020.06.18.159699. 54. Emma C Thomson, Laura E Rosen, James G Shepherd, Roberto Spreaﬁco, Ana da Silva Filipe, et al. Circulating SARS-CoV-2 spike N439K variants maintain ﬁtness while evading antibody-mediated immunity. Cell, 184(5):1171–1187, 2021. doi: 10.1016/j.cell. 2021.01.037. 28. Wanwisa Dejnirattisai, Daming Zhou, Helen M Ginn, Helen ME Duyvesteyn, Piyada Supasa, et al. The antigenic anatomy of SARS-CoV-2 receptor binding domain. Cell, 2021. doi: 10.1016/j.cell.2021.02.032. 55. Bailey B. Banach, Gabriele Cerutti, Ahmed S. Fahad, Chen-Hsiang Shen, Matheus Oliveira de Souza, et al. Robinson and Raybould et al. \| bioRxiv, 2021. doi: 10.1101/2020.12.31.424729. in mice and nonhuman primates. bioRxiv, 2021. doi: 10.1101/2020.12.31.424 62. Philip JM Brouwer, Tom G Caniels, Karlijn van der Straten, Jonne L Snitselaar, Yoann Al- don, et al. Potent neutralizing antibodies from COVID-19 patients deﬁne multiple targets of vulnerability. Science, 369(6504):643–650, 2020. doi: 10.1126/science.abc5902. vulnerability. Science, 369(6504):643–650, 2020. doi: 10.1126/science.abc590 63. Dora Pinto, Young-Jun Park, Martina Beltramello, Alexandra C Walls, M Alejandra Tortorici, et al. Cross-neutralization of SARS-CoV-2 by a human monoclonal SARS-CoV antibody. Nature, 583(7815):290–295, 2020. doi: 10.1038/s41586-020-2349-y. 64. Bruno E. Correia, John T. Bates, Rebecca J. Loomis, Gretchen Baneyx, Chris Carrico, et al. Proof of principle for epitope-focused vaccine design. Nature, 507:201–206, 2014. doi: 10.1038/nature12966. 65. Xiaojian Han, Yingming Wang, Shenglong Li, Chao Hu, Tingting Li, et al. A rapid and efﬁcient screening system for neutralizing antibodies and its application for the discovery of potent neutralizing antibodies to SARS-CoV-2 S-RBD. bioRxiv, 2020. doi: 10.1101/2020.08. 19.253369. 66. Michael J. Gubbins, Frank A. Plummer, Xin Y. Yuan, Darrell Johnstone, Mike Drebot, et al. Molecular characterization of a panel of murine monoclonal antibodies speciﬁc for the SARS-coronavirus. Mol Immunol, 42(1):125–136, 2005. doi: https://doi.org/10.1016/j. molimm.2004.06.032. 67. Ian Setliff, Wyatt J. McDonnell, Nagarajan Raju, Robin G. Bombardi, Amyn A. Murji, et al. Multi-Donor Longitudinal Antibody Repertoire Sequencing Reveals the Existence of Public Antibody Clonotypes in HIV-1 Infection. Cell Host Microbe, 26(3):845–854, 2018. doi: 10.1016/j.chom.2018.05.001. 68. David Forgacs, Rodrigo B. Abreu, Giuseppe A. Sautto, Greg A. Kirchenbaum, Elliott Drabek, et al. Convergent antibody evolution and clonotype expansion following inﬂuenza virus vaccination. PLoS ONE, 16(2):e0247253, 2021. doi: 10.1371/journal.pone.0247253. 69. Claire Marks and Charlotte M. Deane. How antibody repertoire data are changing antibody science. J Biol Chem, 295(25):9823–9837, 2021. doi: 10.1074/jbc.REV120.010181. 69. Claire Marks and Charlotte M. Deane. How antibody repertoire data are ch science. J Biol Chem, 295(25):9823–9837, 2021. doi: 10.1074/jbc.REV120 70. Aaron J. Schmitz, Jackson S. Turner, Zhuoming Liu, Ishmael D. Aziati, Rita E. Chen, et al. A public vaccine-induced human antibody protects against SARS-CoV-2 and emerging vari- ants. bioRxiv, 2021. doi: 10.1101/2021.03.24.436864. 71. M. Alejandra Tortorici, Martina Beltramello, Florian A. Lempp, Dora Pinto, Ha V. Dang, et al. Ultrapotent human antibodies protect against SARS-CoV-2 challenge via multiple mecha- nisms. Science, 370(6519):950–957, 2020. doi: 10.1126/science.abe3354. 72. Yoonjoo Choi and Charlotte M. Deane. FREAD revisited: Accurate loop structure prediction using a database search algorithm. Proteins, 78(6):1431–1440, 2010. doi: 10.1002/prot. 22658. 73. Yoonjoo Choi and Charlotte M. Deane. Robinson and Raybould et al. \| Paired heavy and light chain signatures contribute to potent SARS-CoV-2 neutralization in public antibody responses. bioRxiv, 2021. doi: 10.1101/2020.12.31.424987. 29. Rui Yin, Johnathan D. Guest, Ghazaleh Taherzadeh, Ragul Gowthaman, Ipsa Mittra, et al. Structural and energetic proﬁling of SARS-CoV-2 antibody recognition and the impact of circulating variants. bioRxiv, 2021. doi: 10.1101/2021.03.21.436311. 56. M-P. Lefranc, Christelle Pommié, Manuel Ruiz, V Giuducelli, Elodie Foulquier, et al. IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfam- ily V-like domains. Dev Comp Immunol, 27(1):55–77, 2003. doi: 10.1016/S0145-305X(02) 00039-3. 30. Christopher O Barnes, Claudia A Jette, Morgan E Abernathy, Kim-Marie A Dam, Shannon R Esswein, et al. SARS-CoV-2 neutralizing antibody structures inform therapeutic strategies. Nature, 588:682–687, 2020. doi: 10.1038/s41586-020-2852-1. 31. Yan Wu, Feiran Wang, Chenguang Shen, Weiyu Peng, Delin Li, et al. A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding to its receptor ACE2. Science, 368(6496):1274–1278, 2020. doi: 10.1126/science.abc2241. 57. Ammar Tareen and Justin B Kinney. Logomaker: beautiful sequence logos in Python. Bioin- formatics, 36(7):2272–2274, 2020. doi: 10.1093/bioinformatics/btz921. 58. Harry C Jubb, Alicia P Higueruelo, Bernardo Ochoa-Montaño, Will R Pitt, David B Ascher, et al. Arpeggio: a web server for calculating and visualising interatomic interactions in protein structures. J Mol Biol, 429(3):365–371, 2017. doi: 10.1016/j.jmb.2016.12.004. 32. Bryan E Jones, Patricia L Brown-Augsburger, Kizzmekia S Corbett, Kathryn Westerndorf, Julian Davies, et al. The neutralizing antibody, LY-CoV555, protects against SARS-CoV- 2 infection in non-human primates. Sci Transl Med, page eabf1906, 2021. doi: 10.1126/ scitranslmed.abf1906. 59. Jinwoo Leem, James Dunbar, Guy Georges, Jiye Shi, and Charlotte M. Deane. ABody- Builder: Automated antibody structure prediction with data–driven accuracy estimation. bioRχiv \| 11 Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint 60. Seth J. Zost, Pavlo Gilchuk, Rita E. Chen, James Brett Case, Joseph X. Reidy, et al. Rapid isolation and proﬁling of a diverse panel of human monoclonal antibodies tar- geting the SARS-CoV-2 spike protein. Nat Med, 26:1422–1427, 2020. doi: 10.1038/ s41591-020-0998-x. 61. Dapeng Li, Robert J Edwards, Kartik Manne, David R. Martinez, Alexandra Schäfer, et al. The functions of SARS-CoV-2 neutralizing and infection-enhancing antibodies in vitro and in mice and nonhuman primates. Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Robinson and Raybould et al. \| Antibodies whose epitope cluster is labelled with a Ab Name PDB Code Citation Epitope Cluster B38 7BZ5 (31) Neck Bamlanivimab (LY-CoV555) 7KMG (32) Left Shoulder BD-236 7CHB (33) Neck BD-368-2 7CHC (33) Left Shoulder BD-604 7CHF (33) Neck BD-629 7CH5 (33) Neck C102 7K8M (12) Neck C1A-B3 7KFW (34) Neck C1A-B12 7KFV (34) Neck C1A-C2 7KFX (34) Neck C1A-F10 7KFY (34) Neck CB6 7C01 (35) Neck CC12.1 6XC2 (36) Neck CC12.3 6XC4 (36) Neck COVA1-16 7JMW (37, 62) Left Flank COVA2-04 7JMO (38, 62) Neck COVA2-39 7JMP (38, 62) Right Shoulder COVOX-45 7BEL (28) Right Flank COVOX-75 7BEO (28) Right Shoulder COVOX-88 7BEL (28) Left Shoulder COVOX-150 7BEI (28) Neck COVOX-158 7BEJ (28) Neck COVOX-253 7BEN (28) Left Shoulder COVOX-253H55L 7BEO (28) Left Shoulder COVOX-269 7BEM (28) Neck COVOX-316 7BEH (28) Left Shoulder COVOX-384 7BEP (28) Left Shoulder CR3022 7JN5 (39, 75) Left Flank CV07-250 6XKQ (14) Right Shoulder CV07-270 6XKP (14) Left Shoulder CV30 6XE1 (40, 76) Neck EY6A 6ZER (41) Left Flank Fab_15033-7 7KLH (77) Right Shoulder Fab-52 7K9Z (42) Left Shoulder Fab-298 7K9Z (42) Right Shoulder Fab2-15 7L5B (53, 78) Shoulder LY-CoV481 7KMI (32) Neck LY-CoV488 7KMH (32) Neck MW06 7DPM (79) Left Flank P4A1 7CJF (43) Neck P2B-2F6 7BWJ (20) Left Shoulder P2C-1F11 7CDI (80) Neck P2C-1A3 7CDJ (81) Shoulder Regdanvimab 7CM4 (44) Right Shoulder S2H14 7JX3 (45) Right Shoulder S304 7L0N (45, 63) Left Flank S309 7JX3 (45, 63) Right Flank STE90-C11 7B30 (46) Neck Robinson and Raybould et al. \| ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Ab Name PDB Code Citation Epitope Cluster B38 7BZ5 (31) Neck Bamlanivimab (LY-CoV555) 7KMG (32) Left Shoulder BD-236 7CHB (33) Neck BD-368-2 7CHC (33) Left Shoulder BD-604 7CHF (33) Neck BD-629 7CH5 (33) Neck C102 7K8M (12) Neck C1A-B3 7KFW (34) Neck C1A-B12 7KFV (34) Neck C1A-C2 7KFX (34) Neck C1A-F10 7KFY (34) Neck CB6 7C01 (35) Neck CC12.1 6XC2 (36) Neck CC12.3 6XC4 (36) Neck COVA1-16 7JMW (37, 62) Left Flank COVA2-04 7JMO (38, 62) Neck COVA2-39 7JMP (38, 62) Right Shoulder COVOX-45 7BEL (28) Right Flank COVOX-75 7BEO (28) Right Shoulder COVOX-88 7BEL (28) Left Shoulder COVOX-150 7BEI (28) Neck COVOX-158 7BEJ (28) Neck COVOX-253 7BEN (28) Left Shoulder COVOX-253H55L 7BEO (28) Left Shoulder COVOX-269 7BEM (28) Neck COVOX-316 7BEH (28) Left Shoulder COVOX-384 7BEP (28) Left Shoulder CR3022 7JN5 (39, 75) Left Flank CV07-250 6XKQ (14) Right Shoulder CV07-270 6XKP (14) Left Shoulder CV30 6XE1 (40, 76) Neck EY6A 6ZER (41) Left Flank Fab_15033-7 7KLH (77) Right Shoulder Fab-52 7K9Z (42) Left Shoulder Fab-298 7K9Z (42) Right Shoulder Fab2-15 7L5B (53, 78) Shoulder LY-CoV481 7KMI (32) Neck LY-CoV488 7KMH (32) Neck MW06 7DPM (79) Left Flank P4A1 7CJF (43) Neck P2B-2F6 7BWJ (20) Left Shoulder P2C-1F11 7CDI (80) Neck P2C-1A3 7CDJ (81) Shoulder Regdanvimab 7CM4 (44) Right Shoulder S2H14 7JX3 (45) Right Shoulder S304 7L0N (45, 63) Left Flank S309 7JX3 (45, 63) Right Flank STE90-C11 7B30 (46) Neck Tables ble S1. Antibodies with a solved crystal structures in complex with a SARS-CoV-2 antigen. Antibodies were assi ir nearest epitope cluster, as deﬁned by Dejnirattisai et al. (28). Robinson and Raybould et al. \| Predicting antibody complementarity determining region structures without classiﬁcation. Mol BioSyst., 12(7):3327–3334, 2011. doi: 10.1039/ c1mb05223c. 74. James Dunbar, Konrad Krawczyk, Jinwoo Leem, Terry Baker, Angelika Fuchs, et al. SAbDab: The structural antibody database. Nucleic Acids Research, 42(D1), 2014. doi: 10.1093/nar/gkt1043. 75. Jan Ter Meulen, Edward N Van Den Brink, Leo LM Poon, Wilfred E Marissen, Cynthia SW Leung, et al. Human monoclonal antibody combination against SARS coronavirus: synergy and coverage of escape mutants. PLoS Med, 3(7):e237, 2006. doi: 10.1371/journal.pmed. 0030237. 76. Emilie Seydoux, Leah J Homad, Anna J MacCamy, K Rachael Parks, Nicholas K Hurlburt, et al. Characterization of neutralizing antibodies from a SARS-CoV-2 infected individual. bioRxiv, 2020. doi: 10.1101/2020.05.12.091298. 77. Shane Miersch, Zhijie Li, Reza Saberianfar, Mart Ustav, James Brett Case, et al. Tetravalent SARS-CoV-2 Neutralizing Antibodies Show Enhanced Potency and Resistance to Escape Mutations. bioRxiv, 2020. doi: 10.1101/2020.10.31.362848. 78. Lihong Liu, Pengfei Wang, Manoj S Nair, Jian Yu, Micah Rapp, et al. Potent neutralizing antibodies against multiple epitopes on SARS-CoV-2 spike. Nature, 584(7821):450–456, 2020. doi: 10.1038/s41586-020-2571-7. 79. J. Wang, S. Jiao, R. Wang, J. Zhang, M. Zhang, et al. Crystal structure of SARS-CoV-2 Spike RBD in complex with MW06 Fab. 2020. doi: 10.2210/pdb7DPM/pdb. 80. X. Wang, L. Zhang, J. Ge, and R. Wang. Crystal structure of SARS-CoV-2 antibody P2C- 1F11 and RBD. 2020. doi: 10.2210/pdb7CDI/pdb. 1F11 and RBD. 2020. doi: 10.2210/pdb7CDI/pdb. 81. X. Wang and J. Ge. Crystal structure of SARS-CoV-2 antibody P2C-1A3 and RBD. 2020. doi: 10.2210/pdb7CDJ/pdb. 12 \| bioRχiv Robinson and Raybould et al. \| Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. Robinson and Raybould et al. \| Table S1. Antibodies with a solved crystal structures in complex with a SARS-CoV-2 antigen. Antibodies were assigned to their nearest epitope cluster, as deﬁned by Dejnirattisai et al. (28). Antibodies whose epitope cluster is labelled with a star () were not included in the previous analysis by Dejnirattisai et al. (28). The new cluster label ‘Shoulder’ indicates that antibody’s epitope lies between the left and right shoulder clusters. SI Tables Table S1. Antibodies with a solved crystal structures in complex with a SARS-CoV-2 antigen. Antibodies were assigned to their nearest epitope cluster, as deﬁned by Dejnirattisai et al. (28). Antibodies whose epitope cluster is labelled with a star () were not included in the previous analysis by Dejnirattisai et al. (28). The new cluster label ‘Shoulder’ indicates that antibody’s epitope lies between the left and right shoulder clusters. bioRχiv \| 13 bioRχiv \| 13 Robinson and Raybould et al. \| . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 12, 2021. ; https://doi.org/10.1101/2021.04.12.439478 doi: bioRxiv preprint Nb Name PDB Code Citation Epitope Cluster H11-D4 6YZ5 (47) Left Shoulder H11-H4 6ZBP (47) Left Shoulder MR17 7C8W (48, 52) Shoulder MR17_K99Y 7CAN (52) Shoulder Nb20 7JVB (49) Shoulder SR4 7C8V (52) Left Shoulder SR31 7D2Z (52) Left Shoulder VHH-72 6WAQ (50) Left Flank VHH_E 7KN5 (51) Left Shoulder VHH_U 7KN5 (51) Left Flank VHH_V 7KN6 (51) Left Flank VHH_W 7KN7 (51) Left Flank Table S2. Nanobodies with a solved crystal structures in complex with a SARS-CoV-2 antigen. nanobodies were assigned to their nearest epitope cluster, as deﬁned by Dejnirattisai et al. (28). Nanobodies whose epitope cluster is labelled with a star () were not included in the previous analysis by Dejnirattisai et al. (28). The new cluster label ‘Shoulder’ indicates that antibody’s epitope lies between the left and right shoulder clusters. 14 \| bioRχiv Robinson and Raybould et al. \| Robinson and Raybould et al. \|
https://openalex.org/W3213414501	https://radjapublika.com/index.php/IJERLAS/article/download/34/37	English	null	ANALYSIS OF THE EFFECT OF TRANSFORMATIONAL LEADERSHIP AND COMMUNICATION ON EMPLOYEE PERFORMANCE (CASE STUDY AT THE NATIONAL HIGHER FOUNDATION Dr. WAHIDIN SUDIROHUSODO MEDAN)	International Journal of Educational Review, Law And Social Sciences	2,021	cc-by	3,191	Volumes 1 No 1 (2021) Volumes 1 No 1 (2021) Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Analysis of the Effect of Transformational Leadership and Communication on Employee Performance S h Si b l ABSTRACT This study aims to examine the effect of transformational leadership and communication on employee performance at the Dr Wahidin Sudirohusodo National Higher Education Foundation, Medan. The number of samples in this study were 76 respondents using simple random sampling method. This type of research is quantitative research using path analysis techniques and questionnaire data. Path analysis hypothesis test is done by using multiple linear analysis, classical assumption test and intervention test. The transformational leadership and communication variables partially show a significant effect on employee performance. Transformational leadership variables and communication simultaneously affect employee performance with a coefficient of determination of 87.8% which means transformational leadership variables, communication can explain employee performance as much as 87.8% and the remaining 12.2% is explained by other variables such as discipline, commitment, compensation and others. Transformational leadership and communication variables simultaneously affect employee performance with a coefficient of determination of 87.8% which means that transformational leadership and communication variables can explain employee performance as much as 87.8% and the remaining 12.2% is explained by other variables such as discipline, commitment. , compensation and others. Keywords : Transformational Leadership, Employee Performance, Communication Keywords : Transformational Leadership, Employee Performance, Communication ANALYSIS OF THE EFFECT OF TRANSFORMATIONAL LEADERSHIP AND COMMUNICATION ON EMPLOYEE PERFORMANCE (CASE STUDY AT THE NATIONAL HIGHER FOUNDATION Dr. WAHIDIN SUDIROHUSODO MEDAN) Sahat Simbolon Institut Bisnis IT&B Sumatera Utara, Indonesia E-mail: sahats_ Simbolon@yahoo.com 1. INTRODUCTION Human resources are the main component in an organization which is an active planner and actor in every organizational activity. They have heterogeneous thoughts, feelings, desires, status and educational background, age, gender which are brought to an organization to collaborate to achieve goals. The strength of an organization lies in its human resources, not in its systems, technology, procedures, or funding sources. So, the functioning of the parts in the organization depends on the ability of the people in the organization concerned to move them towards the achievement of the goals that have been set. Quality human resources will develop high performance for the organization. Employee performance is the result of work achieved by someone in carrying out tasks in accordance with their responsibilities. Jobs that match the job description are expected to be completed on time and meet quality and quantity standards. Employees can work well if they have high performance so that they can produce good work. With high employee competence, it is expected that organizational goals can be achieved. Transformational leadership is considered the most appropriate model of the many existing leadership models because transformational leaders better understand the needs of subordinates that must be met so that they can support the improvement of subordinates' performance. Transformational leadership is a leader who stimulates and inspires (changes) subordinates to achieve extraordinary results. A leader with this leadership style will be able to motivate his employees to always work optimally. A transformational leader can bring about major changes in International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS both followers and organizations. It is not only leadership that needs to be considered, but the communication that exists between good leaders and subordinates is also something that must be considered. both followers and organizations. It is not only leadership that needs to be considered, but the communication that exists between good leaders and subordinates is also something that must be considered. Communication in an organization is an important factor in undergoing interaction with each other, if there is no communication, all individuals in the organization cannot know what they should do for the organization, leaders cannot receive information input and providers cannot give instructions. NS National Higher Education Foundation Dr. Wahidin Sudirohusodo concerned in the field of education, especially primary and secondary education, both junior high and general. This foundation has been around for a long time and is a private foundation. In the course of its activities, the Foundation has a vision and mission to be achieved in the future. Based on the problems above, the authors are interested in analyzing the effect of transformational leadership and communication on employee performance (Study at the Dr. Wahidin Sudirohusodo College Foundation). E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS International Journal of Educational Review, Law And Social Sciences \|IJERLAS 2. RESEARCH METHOD The method used in this study was done through data collection by questionnaires and testing using path analysis. This study uses primary data. Where the primary data in the form of questionnaires and interviews conducted to 76 respondents. This research was conducted from April to July 2018 at the Dr. National Higher Education Foundation. Wahidin Sudirohusodo Medan. 3. RESULTS AND DISCUSSION 3. RESULTS AND DISCUSSION 3.1. Results of Observation Using Questionnaire The Effect of Transformational Leadership and Communication on Employee Performance Results of Observation Using Questionnaire The Effect of Transformational Leadership and Communication on Employee Performance 3.1. Results of Observation Using Questionnaire The Effect of Transformational Leader 1. Results of Observation Using Questionnaire The Effect of Transformational Leadershi and Communication on Employee Performance This research was conducted on employees of the National Higher Education Foundation D Wahidin Sudirohusodo Medan with a research sample of 76 people. Table 1. Descriptive Analysis Results No Penyataan Alternatif Jawaban SS S R TS STS F % F % F % F % F % 1 Ketepatan waktu 52 68,42 15 19,74 9 11,84 0 0,00 0 0,00 2 Ketepatan 20 26,32 36 47,37 19 25.00 1 1,32 0 0,00 3 Terampil 55 72,37 10 13,16 11 14,47 0 0,00 0 0,00 4 kemampuan kerja 24 31,58 36 47,37 14 18,42 2 2,63 0 0,00 5 Total Produksi 23 30,26 43 56,58 10 13,16 0 0,00 0 0,00 6 Luas Pengetahuan 43 56,58 22 28,95 10 13,16 1 1,32 0 0,00 7 Memiliki keterampilan 21 27,63 43 56,58 12 15,79 0 0,00 0 0,00 8 Keaktifan dalam menyampaikan pendapat 45 59,21 24 31,58 5 6,58 2 2,63 0 0,00 9 Kegiatan desain 49 64,47 16 21,05 8 10,53 3 3,95 0 0,00 Total 436,84 322,37 128,95 11,84 0,00 Rata-rata 48,54 35,82 14,33 1,32 0,00 Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Based on table 1, it can be seen that the average respondents answered strongly agree (SS) as much as 48.54%, answered agree (S) at 35.82%, in doubt (R) as much as 14.33% and disagree (TS) by 1.32% and strongly disagree (STS) as much as 0.00%. Of all respondents' answers, there are 84.36% of respondents who chose the answer strongly agree and the answer agrees where the percentage of answers between an assessment score of 80 to 100 with a very good category. This explains that employee participation in expressing opinions and ideas at the meeting is quite good and employees are quite skilled in completing the work given, while from all respondents' answers there are 1.32% of respondents who answered disagree. This shows that there are still weaknesses related to these problems, such as employees in designing work activities before implementing them, and the leadership must encourage employees to get optimal results. Table 2. Results of Multiple Linear Regression Analysis Model Koefisien Tidak Standar Koefisien Standar T Tanda tangan. B Std. Kesalahan Beta (Konstan) -9,412 2.113 -4.455 ,000 Kepemimpinan Transformasional ,077 ,036 ,107 2.154 ,035 Komunikasi ,347 ,063 ,338 5.479 ,000 A. Variabel Dependen: Kinerja Karyawan Sumber: Data diolah dari hasil penelitian, 2019 Table 2. Results of Multiple Linear Regression Analysis Based on table 2, it can be seen that the constant value of -9.412 means that if the transformational leadership (X1) and communication (X2) is 0 (zero), there is a tendency for the employee performance value (Y) to decrease by -9.412. The regression coefficient for transformational leadership is 0.077 and is positive, meaning that if transformational leadership has increased by one unit, the employee's performance value will increase by 0.077, cateris paribus. The communication regression coefficient is 0.347 and has a positive direction, meaning that if communication increases by one unit, the employee performance value will increase by 0.347, cateris paribus. Table 3. Test Results of the Effect of Transformational Leadership, Communication on Employee Performance Model Koefisien Tidak Standar Koefisien Standar T Tanda tangan . B Std. Kesalahan Beta (Konstan) -9,412 2.113 -4.455 ,000 Kepemimpinan Transformasional ,077 ,036 ,107 2.154 ,035 Komunikasi ,347 ,063 ,338 5.479 ,000 A. Variabel Dependen: Kinerja Karyawan Sumber: Data diolah dari hasil penelitian, 2019 able 3. Test Results of the Effect of Transformational Leadership, Communication o mployee Performance Results of the Effect of Transformational Leadership, Communication on formance A. Volumes 1 No 1 (2021) Volumes 1 No 1 (2021) Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Variabel Dependen: Kinerja Karyawan Sumber: Data diolah dari hasil penelitian, 2019 Based on table 3, it can be seen that the tcount value of each independent variable in this study. The t-count value of each independent variable will be compared with the t-table value, with a confidence interval of 95% or 0.05, so that the t-table value = 1.666. Table 3 also shows that transformational leadership has a tcount of 2.154 > ttable 1.666, which means that transformational leadership has a positive and significant effect on employee performance. The communication International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS 90 variable has a tcount of 5.479 > ttable of 1.666, meaning that communication has a positive and significant effect on employee performance. 3.2. The Effect of Transformational Leadership on Employee Performance From this research, the value of tcount is 2.154 > ttable 1.666 and a significance value of 0.035 <0.05. The regression coefficient (ß) and tcount use a significant level of 0.05. Then Ha is accepted and H0 is rejected, so it can be concluded that the results of the transformational leadership variable have a positive and significant effect on the employee performance variable. These results are similar to research conducted by Agustina Ritawati (2013). Agustina stated that the transformational leadership style has a positive effect on the performance of insurance employees by taking an interpersonal approach to subordinates, so that subordinates feel happy with the way the leader provides direction to achieve higher performance. variable has a tcount of 5.479 > ttable of 1.666, meaning that communication has a positive and significant effect on employee performance. variable has a tcount of 5.479 > ttable of 1.666, meaning that communication has a positive and significant effect on employee performance. 3.3. The Effect of Communication on Employee Performance p y The results of the analysis showed that the tcount value was 5,479 > ttable 1,666 and the significance value was 0.000 <0.05. The regression coefficient (ß) and tcount use a significant level of 0.05. This means that Ha is accepted and H0 is rejected, so it can be concluded that the results of the communication variable have a positive and significant effect on the employee performance variable. National Higher Education Foundation Dr. Wahidin Sudirohusodo Medan has provided information to members of the organization about the goals and policies of the organization, which has been carried out by organizational communication between superiors and subordinates. They can interact well starting from information, messages and advice and suggestions. So that organizational communication at the National Higher Education Foundation Dr. Wahidin Sudirohusodo Medan. International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS 4. CONCLUSION From the research conducted, it was found that the variables of transformational leadership and employee performance were in the very good category, while communication was in the good category at the Dr Wahidin Sudirohusodo National Higher Education Foundation, Medan. Based on the results of the partial hypothesis test (t test), the transformational leadership variable has a positive and significant effect on employee performance. The communication variable has a positive and significant effect on employee performance. Simultaneously or the F test, transformational leadership and communication have a positive and significant effect on employee performance at the Dr. National Higher Education Foundation. Wahidin Sudirohusodo Medan. Based on the results of path analysis, it was found that transformational leadership variables had more influence on employee performance than the direct influence of transformational leadership on employee performance. Communication is also more influential on employee performance than direct communication has an effect on employee performance, with a coefficient of determination (R2) of 87.8% which means that the variables of transformational leadership, communication and work motivation can explain the performance of 87 respondents. Leaders are able to convey the company's vision to employees by inspiring them to always be optimistic in achieving goals and providing personal support to train employees in carrying out their duties and good cooperation between colleagues and superiors and fully responsible for the work desk given. Each employee designs a work plan in advance so that every job can be done on time and with the support of the leadership so that the results obtained are more optimal. Volumes 1 No 1 (2021) Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon on the dependent variable simultaneously, so that employee performance has increased significantly. Wahidin Sudirohusodo Medan can use excellent communication according to indicators such as providing information about assignments, performance results to subordinates, subordinates easily providing information as input and consulting with superiors if there are obstacles, overcoming problems and equality in communication between work groups. All independent variables run well on the dependent variable simultaneously, so that employee performance has increased significantly. on the dependent variable simultaneously, so that employee performance has increased significantly. Wahidin Sudirohusodo Medan can use excellent communication according to indicators such as providing information about assignments, performance results to subordinates, subordinates easily providing information as input and consulting with superiors if there are obstacles, overcoming problems and equality in communication between work groups. All independent variables run well on the dependent variable simultaneously, so that employee performance has increased significantly. Performance Based on the results of the analysis, Fcount is 181.552, while Ftable (3;73) from the distribution table is 2.72 and the significance value is 0.000. This means that Fcount > Ftable or 181.552 > 2.72 and the significant value is less than the alpha value of 0.05, significant < (0.000 < 0.05). So Ha is accepted and H0 is rejected. This means that transformational leadership and communication variables have a simultaneous effect on employee performance variables. This highlights that the employees of the National Higher Education Foundation Dr. Wahidin Sudirohusodo has a leader who has a reaction to critical things, example, charisma, inspirational personality, intellectual stimulation and employee trust with attention both individually and in groups. As for the communication variable, it shows that communication has a simultaneous effect on employee performance. This explains that the staff of the National Higher Education Foundation, Dr. Wahidin Sudirohusodo Medan can use excellent communication according to indicators such as providing information about assignments, performance results to subordinates, subordinates easily providing information as input and consulting with superiors if there are obstacles, overcoming problems and equality in communication between work groups. All independent variables run well on the dependent variable simultaneously, so that employee performance has increased significantly. Wahidin Sudirohusodo Medan can use excellent communication according to indicators such as providing information about assignments, performance results to subordinates, subordinates easily providing information as input and consulting with superiors if there are obstacles, overcoming problems and equality in communication between work groups. All independent variables run well REFERENCE Brigham, Eugene, F. 2011. Manajemen Keuangan. Erlangga, Jakarta Brigham, Eugene, F. 2011. Manajemen Keuangan. Erlangga, Jakarta Danim, Sudarwan, 2012. Kepemimpinan Pendidikan, Alfabeta, Bandung Danim, Sudarwan, 2012. Kepemimpinan Pendidikan, Alfabeta, Bandung ______. 2013, Menjadi Komunitas Belajar: Transformasional Kepemimpinan dalam Komunitas Organisasi Pembelajaran, Bumi Aksara, Jakarta Dessler, Gary, 2011. Manajemen Sumber Daya Manusia, Edisi 10, Indeks, Jakarta Emad, dkk. 2014. Pentingnya Kepemimpinan dalam Peningkatan Produktivitas Konstruksi. Jurnal Penelitian Lanjutan Global Studi Manajemen dan Bisnis, Vol. 3, no.3, hal: 114-125 Erriani K. 2018. Hubungan Motivasi Kerja dengan Kinerja Pustakawan Bidang Pertanian di Kementerian Pertanian. Skripsi Institut Pertanian Bogor. Jurnal Penyuluhan, Maret 2018 No. 14 Vol. 1 Fattah, Nanang, 2014. Yayasan Manajemen Pendidikan, PT Remaja Rosdakarya, Bandung. International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS Febryantahanuji, 2017, Pengaruh Kepuasan Kerja dan Komunikasi Terhadap Komitmen Organisasi dengan Kepemimpinan Sebagai Variabel Moderasi Pada Stekom Semarang. Jurnal Nusamba, Vol.2, No.1 Haryani. 2010. Analisis Pengaruh Komunikasi, Motivasi, dan Kepemimpinan Terhadap Kepuasan Kerja Dalam Meningkatkan Kinerja Karyawan PT Arisa Mandiri Pratama,Jurnal Manajemen dan Kewirausahaan, volume 13. Hasibuan, Melayu. 2008. Dasar Manajemen, Pengertian, dan Masalah. PT Bumi Aksara, Jakarta Kuswana, Wowa S, 2013. Manajemen Kinerja, PT. Raja Grafindo Persada, Jakarta Mangkunegara, AP 2010, Manajemen Sumber Daya Manusia, Edisi Rev., Penerbit Bina Aksara, Jakarta ______. 2012, Manajemen Sumber Daya Manusia Perusahaan, Remaja Rosdakarya, Bandung Marshall Sashkin & Molly G. Sashkin, 2011. Prinsip Kepemimpinan, Erlangga, Jakarta Nyowani. Tri Wahyuni, 2014, Pengaruh Motivasi dan Pelatihan Kerja Terhadap Kinerja Pegawai pada BPRS Karya Mugi Sentosa Trenggalek dan BPRS Tanmiya Artha Kediri IAIN Tulungagung. Richard L. Hughes, dkk, 2012, Kepemimpinan: Memperkaya Pelajaran dari Pengalaman, Edisi 7, Salemba Humanika, Jakarta Rivai, Veitzal & Ella Jauvani Sagala. 2011. Manajemen Sumber Daya Manusia Perusahaan. PT. Raja Grafindo Persada. Jakarta. Robbins, Stephen P, & Timothy A Hakim. 2014. Perilaku Organisasi, Edisi Ketigabelas. Pearson Prentice Hall, New Jersey Sandy Marta, Muhammad. 2015. Karakteristik Pekerjaan dan Kinerja Dosen Berprestasi di UIN Sunan Gunung Djati Bandung: Komitmen Organisasi Sebagai Variabel Moderating. Skripsi di Universitas Widayatama, Bandung: Tidak Dipublikasikan. Santoso, S.2010. Statistik Multivariat; Konsep dan Aplikasi dengan SPSS. Jakarta: PT Elex Media Komputindo Sedarmayanti. 2011. Manajemen Sumber Daya Manusia, Reformasi Birokrasi dan Manajemen Pegawai Negeri Sipil (cetakan kelima). PT Refika, Bandung mamora, Wawan, 2016. Manajemen Sumber Daya Manusia, STIE YKPN, Yogyakarta. ugiyono. 2017. Statistik untuk Penelitian. CV Alfabeta, Bandung Sunyoto, Danang. 2012.Sumber daya manusia Manajemen. Pusat Layanan Penerbitan Akademik (CAPS), Yogyakarta. ______. 2013, Teori, Kuesioner, dan Analisis Data Sumber Daya Manusia (Praktek Penelitian), Pusat Layanan Penerbitan Akademik (CAPS), Yogyakarta. Supardi, dan Anwar Syaiful, 2012, Dasar-dasar Perilaku Organisasi, UII Pres, Yogyakarta. Sutrisno, Edy. 2011.Sumber daya manusia Manajemen Kencana, Jakarta utrisno, Edy. 2011.Sumber daya manusia Manajemen Kencana, Jakarta Sutrisno, Edy. 2011.Sumber daya manusia Manajemen Kencana, Jakarta International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS 93 94 International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Analysis of the Effect of Transformational Leadership and Communication on Employee Performance Sahat Simbolon Suyadi. 2013. Strategi Pembelajaran Pendidikan Karakter. Remaja Rosdakarya. Bandung mam, Khaerul, 2012. Perilaku Organisasi, Pustaka Setia, Bandung Veithzal Rivai. 2014.Manajemen Sumber Daya Manusia untuk Perusahaan, Edisi 6, PT. Raja Grafindo Persada, Depok Veithzal Rivai. 2014.Manajemen Sumber Daya Manusia untuk Perusahaan, Edisi 6, PT. Raja Grafindo Persada, Depok Yukl, Gary, 2015. Kepemimpinan dan Organisasi, Edisi 5, Indeks , Jakarta 94 International Journal of Educational Review, Law And Social Sciences \|IJERLAS E-ISSN: 2808-487X \| https://radjapublika.com/index.php/IJERLAS 94
https://openalex.org/W2763676075	https://www.nature.com/articles/s41598-017-13463-4.pdf	English	null	Sedum alfredii SaNramp6 Metal Transporter Contributes to Cadmium Accumulation in Transgenic Arabidopsis thaliana	Scientific reports	2,017	cc-by	9,583	Sedum alfredii SaNramp6 Metal Transporter Contributes to Cadmium Accumulation in Transgenic Arabidopsis thaliana Shuangshuang Chen1,2, Xiaojiao Han1,2, Jie Fang1,2,3, Zhuchou Lu4, Wenmin Qiu1,2, Mingying Liu1,2, Jian Sang1,2, Jing Jiang1,2 & Renying Zhuo 1,2 Received: 3 January 2017 Accepted: 25 September 2017 Published: xx xx xxxx Received: 3 January 2017 Accepted: 25 September 2017 Published: xx xx xxxx The plant natural resistance-associated macrophage protein (Nramp) family plays an important role in tolerance to heavy metal stress. However, few Nramps have been functionally characterized in the heavy metal-accumulating plant Sedum alfredii. Here, Nramp6 was cloned and identified from S. alfredii and its function analyzed in transgenic Arabidopsis thaliana. SaNramp6 cDNA contains an open reading frame of 1, 638 bp encoding 545 amino acids. SaNramp6′s expression can be induced by cadmium (Cd) stress, and, after treatment, it peaked at one week and 12 h in the roots and leaves, respectively. SaNramp6 localized to the plasma membrane in protoplasts isolated from A. thaliana, Nicotiana benthamiana lower leaf and onion (Allium cepa) epidermal cells. The heterologous expression of SaNramp6 in the Δycf1 yeast mutant increased the Cd content in yeast cells. SaNramp6 also rescued the low Cd accumulation of the A. thaliana nramp1 mutant. Transgenic A. thaliana expressing SaNramp6 exhibited high Cd accumulation levels, as determined by a statistical analysis of the Cd concentration, translocation factors and net Cd2+ fluxes under Cd stress. Thus, SaNramp6 may play a significant role in improving Cd accumulation, and the gene may be useful for the biotechnological development of transgenic plants for phytoremediation. A well-balanced cellular concentration of essential metals such as iron (Fe), copper (Cu) and manganese (Mn), plays a fundamental role in the normal growth and development of plants1. However, the absorption of heavy metals such as lead (Pb), cadmium (Cd) and arsenic (As), can upset the normal metabolism within plant cells and also cause damage to human and animal health through the cumulative effects of the food chain. For exam- ple, Cd is a carcinogenic factor closely related to the generation of breast and kidney cancer2,3, and high levels of Pb toxicity can lead to irreversible damage to the central nervous system4. The ever-increasing worldwide con- tamination of soil and water by heavy metals is a problem that demands a prompt solution5. Phytoremediation is presently regarded as an eco-friendly and cost-effective strategy to clean heavy metal-polluted soils with the help of hyperaccumulating plants6. Among the more than 400 naturally hyperaccumulating plants, an ecotype of Sedum alfredii that co-hyperaccumulates Cd, Zn and Pb was first found in China7,8. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 3 January 2017 Accepted: 25 September 2017 Published: xx xx xxxx Sedum alfredii SaNramp6 Metal Transporter Contributes to Cadmium Accumulation in Transgenic Arabidopsis thaliana Shuangshuang Chen1,2, Xiaojiao Han1,2, Jie Fang1,2,3, Zhuchou Lu4, Wenmin Qiu1,2, Mingying Liu1,2, Jian Sang1,2, Jing Jiang1,2 & Renying Zhuo 1,2 Previous physiological studies suggested that this ecotype is a promising hyper-accumulator for the decontamination of polluted soils, because it can accumulate up to nine g of Cd per kg of leaf dry weight (DW)9–11. However, the detailed molecular mechanism underlying its hyperaccumulation and tolerance of heavy metals is still unclear. Taking advantage of this genetic resource for the breeding of future phytoremediation-associated plants requires a functional analysis of potential heavy metal-responsive genes in the hyperaccumulating ecotype of S. alfredii.f p y p g yp g yp f Metal transporters are essential for the maintenance of appropriate metal ions concentrations within different cellular compartments12,13. Among the identified metal transporters, natural resistance-associated macrophage protein genes (Nramps) are considered to play potentially important roles mediating metal ion homeostasis at 1State Key Laboratory of Forest Genetics and Breeding, Xiangshan Road, Beijing, 100091, P.R. China. 2Key Laboratory of Tree Breeding of Zhejiang Province, The Research Institute of Subtropical of Forestry, Chinese Academy of Forestry, Hangzhou, 311400, China. 3Chemical Biology Center, Lishui Institute of Agricultural Sciences, Lishui, Zhejiang Province, 323000, China. 4Biotechnology Research Center of China, Three Gorges University, Yichang, Hubei, 443002, China. Shuangshuang Chen, Xiaojiao Han and Jie Fang contributed equally to this work. Correspondence and requests for materials should be addressed to R.Z. (email: zhuory@gmail.com) Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 1 1 www.nature.com/scientificreports/ multiple cellular levels in plants. First cloned in mouse, Nramp gene family members are relatively evolutionarily conserved throughout organisms, including plants, animals, yeast and bacteria14. The Nramp genes comprise a small family represented by six members in Arabidopsis thaliana15, 12 members in rice (Oryza sativa; http:// www.ncbi.nlm.nih.gov/gene/? term = Nramp + Oryza + sativa), eight members in soybean (Glycine max; http:// www.phytozome.net/soybean) and six members in poplar (Populus trichocarpa)16. Several NRAMP members have been experimentally characterized in A. thaliana and are involved in the uptake, intracellular transport, translocation and detoxification of metals14,17,18. They are all membrane spanning proteins, with the 10–12 hydro- phobic transmembrane domains characteristic of metal transporters19. When overexpressed in yeast, AtNramp1, AtNramp3 and AtNramp4 show high affinities for Fe, Mn and Cd, whereas AtNramp6 can transport Cd, but not Fe or Mn20–23. In rice, OsNramp1 shows transport activity for Cd and Fe, but not Mn. OsNramp4 is the first transporter identified for the trivalent aluminium ion, and the knockout of OsNramp5 results in a significantly reduced Cd uptake18,24,25. Results l i 2b, SaNramp6 shows 80% sequence similar- ity to Nramp6 from A. thaliana, 79% to Nramp6 from Theobroma cacao, 78% to Nramp1 from Populus trichocarpa, and 71% to Nramp1 from Nicotiana tabacum. A phylogenetic analysis revealed that the SaNramp6 was most closely related to AtNramp6 (Fig. 2b). Based on this, we designated this gene as SaNramp6 (GenBank accession no. KF887490). Expression profiles of SaNramp6 under CdCl2 stress. To examine transcriptional changes under CdCl2 stress, the expression of SaNramp6 was monitored at different Cd-stress treatment times in leaves, stems, and roots. Without heavy metal treatment, SaNramp6 was highly expressed in roots and leaves (Fig. 3a). However, the relative expression levels of SaNramp6 varied greatly in different tissues under Cd treatment period progressed. Despite starting at a higher level, SaNramp6′s expression was not induced in leaves and in fact was reduced during treatment, reaching only a maximum of less than one-fold of the initial level at 12 h (Fig. 3d). In stems, SaNramp6′s expression increased gradually before 12 h, and then declined (Fig. 3c). By contrast, SaNramp6′s transcript accumulation was highly induced in roots (Fig. 3b). It began to increase within 12 h of treatment and peaked at around one week (14-fold). SaNramp6′s expression enhances Cd2+ sensitivity and increases Cd2+ content in yeast. To investigate the cellular function of SaNramp6, the protein was expressed in Saccharomyces cerevisiae yeast mutant (∆ycf1) susceptible to Cd excess. SaNramp6 and empty vector-complemented ∆ycf1 cells were grown in SG-U medium overnight. Cells grown overnight were used for spotting on SG-U agar plates supplemented with 0, 15 and 20 µM CdCl2 at indicated dilutions. The Cd supplementation of the medium caused more considerable growth inhibition in yeast cells expressing SaNramp6 than in the control (Fig. 4a). We also analyzed the relative growth in liquid media in the presence of Cd in yeast cells. The growth of ∆ycf1 cells expressing SaNramp6 were lower than cells transformed with the empty vector (Fig. 4b). The growth inhibition due to the functional SaNramp6 in ∆ycf1 suggested that SaNramp6 may facilitate the import of Cd inside the yeast. p y f gg p y p y To test our hypothesis that SaNramp6 may mediate the Cd uptake, the metal content was measured in yeast cells expressing SaNramp6 or the vector that were grown in the presence of Cd. Results l i Isolation and sequence analysis of SaNramp6. To identify the function of SaNramp6 from S. alfredii, a full-length cDNA sequence of 2, 055 nucleotides was isolated, comprising a 1, 638-bp open reading frame, and 95-bp 5′and 322-bp 3′-untranslated regions. The specific primers SaNramp6-F and SaNramp6-R were used to amplify the sequence of the SaNramp6 from genomic DNA to investigate the genomic structure of SaNramp6. The genomic sequence spanned 3, 587 bp including 10 introns and 11 exons (Fig. 1a). A sequence comparison revealed that SaNramp6 is similar to members of group I from A. thaliana (Fig. 1b).h p g p ( g ) The predicted protein encoded by SaNramp6 contained 545 amino acid residues with a putative molecular weight of 58.44 kD and an isoelectric point of 7.97. The deduced amino acid sequence was not predicted to have a signal peptide by SignalP software (ExPASy). Based on analyses using the CELLO and SOSUI programs, we hypothesized that this protein is located at the plasma membrane and has 11 transmembrane domains (Fig. 1c). Multiple sequence alignments with SaNramp6 revealed high levels of similarity to the Nramps of other species (Fig. 2a). To investigate the evolutionary relationships among Nramps from different species, a phylogenetic analy- sis was performed based on the amino acid sequences. As shown in Fig. 2b, SaNramp6 shows 80% sequence similar- ity to Nramp6 from A. thaliana, 79% to Nramp6 from Theobroma cacao, 78% to Nramp1 from Populus trichocarpa, and 71% to Nramp1 from Nicotiana tabacum. A phylogenetic analysis revealed that the SaNramp6 was most closely related to AtNramp6 (Fig. 2b). Based on this, we designated this gene as SaNramp6 (GenBank accession no. KF887490). p g p ( g ) The predicted protein encoded by SaNramp6 contained 545 amino acid residues with a putative molecular weight of 58.44 kD and an isoelectric point of 7.97. The deduced amino acid sequence was not predicted to have a signal peptide by SignalP software (ExPASy). Based on analyses using the CELLO and SOSUI programs, we hypothesized that this protein is located at the plasma membrane and has 11 transmembrane domains (Fig. 1c). yp p p g Multiple sequence alignments with SaNramp6 revealed high levels of similarity to the Nramps of other species (Fig. 2a). To investigate the evolutionary relationships among Nramps from different species, a phylogenetic analy- sis was performed based on the amino acid sequences. As shown in Fig. Sedum alfredii SaNramp6 Metal Transporter Contributes to Cadmium Accumulation in Transgenic Arabidopsis thaliana Shuangshuang Chen1,2, Xiaojiao Han1,2, Jie Fang1,2,3, Zhuchou Lu4, Wenmin Qiu1,2, Mingying Liu1,2, Jian Sang1,2, Jing Jiang1,2 & Renying Zhuo 1,2 Nramp genes have also been cloned and characterized from other plants, such as tomato (Solanum lycopersicum)26, soybean27 and some metal-hyperaccumulating species. A better understanding of the mechanisms used by metal transporters will provide insights into the detoxification and accumulation of toxic heavy metals in plants. y p Although Nramp genes have been cloned and analyzed in other plant species, few studies have been reported regarding Nramps in the hyperaccumulating ecotype of S. alfredii. The transcriptome of S. alfredii under Cd stress indicated that an Nramp gene was greatly up–regulated after CdCl2 treatment28. The gene has an 80% homology with AtNramp6. Here, we described the isolation and characterization of the Nramp gene SaNramp6 from S. alfredii. A subcellular localization analysis indicated that SaNramp6 is a plasma membrane transporter. Moreover, the overexpression of SaNramp6 in A. thaliana increased the uptake and accumulation of Cd. Thus, SaNramp6 may be a potentially important heavy metal-responsive gene that could be useful for phytoremediation. This work will aid in understanding heavy metal hyperaccumulation and tolerance in S. alfredii. Results l i A significantly enhanced accumu- lation of Cd was observed in yeast cells expressing SaNramp6 compared with the control (Fig. 4c) Subcellular localization of SaNramp6. Bioinformatics analysis using the CELLO v2.5 program software predicted that SaNramp6 is localized in plasma membrane. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 2 www.nature.com/scientificreports/ Figure 1. SaNramp6 gene structure. (a) Genomic organization of SaNramp6. Black boxes and lines denote exons and introns, respectively. The numbers refer to the position between the exons and introns, (b) Comparison of the genomic DNA structure of SaNramp6 and several Nramp genes of Arabidopsis available in GenBank. The white boxes represent the introns, and the grey boxes represent the exons. The numbers indicate the length of the sequence. І and II indicate the groups of Nramps in A. thaliana, (c) Transmembrane domains predicted by the SOSUI program. (AtNramp1: AT1G80830; AtNramp2: AT1G47240; AtNramp3: AT2G23150; AtNramp4: AT5G67330; AtNramp5: AT4G18790; AtNramp6: AT1G15960). Figure 1. SaNramp6 gene structure. (a) Genomic organization of SaNramp6. Black boxes and lines denote exons and introns, respectively. The numbers refer to the position between the exons and introns, (b) Comparison of the genomic DNA structure of SaNramp6 and several Nramp genes of Arabidopsis available in GenBank. The white boxes represent the introns, and the grey boxes represent the exons. The numbers indicate the length of the sequence. І and II indicate the groups of Nramps in A. thaliana, (c) Transmembrane domains predicted by the SOSUI program. (AtNramp1: AT1G80830; AtNramp2: AT1G47240; AtNramp3: AT2G23150; AtNramp4: AT5G67330; AtNramp5: AT4G18790; AtNramp6: AT1G15960). To test the prediction, the subcellular localization of SaNramp6 was analyzed by transiently expressing the SaNramp6-GFP fusion protein in protoplasts isolated from A. thaliana, onion epidermal cells and N. benthami- ana epidermal cells, respectively. As shown in Fig. 5, visualized fluorescence indicated that the SaNramp6-GFP signal was localized at the plasma membrane, whereas the green fluorescent signal in the GFP control vector was distributed throughout the cytosol taken chlorophyll as control in protoplasts of A. thaliana. g y p y p p The plasma membrane localization of SaNramp6 was further confirmed by the transient expression of SaNramp6-GFP in onion epidermal cells and N. benthamiana epidermal cells. The fusion protein was found to be targeted to the plasma membrane by colocalization with FM4–64 within 5 min of the onset of staining (Supplementary Figure S1). Results l i These results indicate that SaNramp6 is localized at the plasma membrane, consistent with the prediction by the CELLO software. SaNramp6 participates in oxidative damage in transgenic Arabidopsis. The production of reac- tive oxygen species (ROS) in the different lines was analyzed using H2O2 and O2 − accumulation. As shown in Fig. 6, the contents of H2O2 and O2 − in the transgenic lines (OE 2 and OE 3) were markedly increased and both were nearly 30% higher than those in WT line. However, they were decreased or slightly increased in the mutant nramp1 (Atnr) and rescue of nramp1 lines (Atnr-N24 and Atnr-N28), respectively (Fig. 6a,b,e,f). Thus, upon Cd stress, the overexpression of SaNramp6 could result in a high level of H2O2 and O2 − accumulation.h We next examined the scavenging ability of ROS by determining CAT and POD activities. The concentra- tion of CAT and POD in the different lines had no difference in the control. Nevertheless, the CAT activity was dramatically increased in transgenic lines as was the POD activity under Cd treatment (Fig. 6c,d,g,h). Thus, the overexpression of SaNramp6 caused more damage and enhanced CAT and POD activities under Cd treatment.f p p g In addition, the four lines (WT, overexpression lines, mutant line, rescue lines) had no obvious differences in roots without Cd treatment (Supplementary Figure S2). However, the root length of transgenic lines (OE 2 and OE 3) was longer than that in the other lines in two weeks after the Cd treatment (Supplementary Figure S2). Overexpression of SaNramp6 resulted in an increased Cd concentration. Time-dependent Cd-uptake experiments using aerial parts and roots were conducted to evaluate the differences in Cd-uptake abil- ities by the different organs of the four lines. The time-dependent experiment on the four lines (WT, overexpres- sion lines, mutant line, rescue lines) showed that Cd concentration increased as the treatment period progressed Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 3 www.nature.com/scientificreports/ Figure 2. Comparison of SaNramp6 to Nramps of other species based on Nramp amino acid sequences from different species. Accession numbers for sequences used are listed in Table 2. Figure 2. Comparison of SaNramp6 to Nramps of other species based on Nramp amino acid sequences from different species. Accession numbers for sequences used are listed in Table 2. Results l i Primers Sequence (5′-3′) Description SaN1-F ATGGCATCAACTGTCGGAAACGC qRT-PCR SaN1-R ACATGCCAATTCCACAGCGA qRT-PCR SaNramp6-F ATGGCATCAACTGTCGGAAACGC Gene specific amplification SaNramp6-R CTACTCTAAGACAGCTCTGCGTTGCGG Gene specific amplification SaNramp6-GF CACCATGGCATCAACTGTCGGAAACGC Gene specific amplification SaNramp6-RT-F TGTTTGGCGATTGTGCCAAG qRT-PCR SaNramp6-RT-R ACATGCCAATTCCACAGCGA qRT-PCR UBC9-F TGGCGTCGAAAAGGATTCTGA qRT-PCR UBC9-R CCTTCGGTGGCTTGAATGGATA qRT-PCR AtActin-F GCACCCTGTTCTTCTTACCG qRT-PCR AtActin-R AACCCTCGTAGATTGGCACA qRT-PCR APL AAGCAGTGGTATCAACGCAGAGTACGC(G)10 5′-RACE adapter primer APS AAGCAGTGGTATCAACGCAGAGT 5′-RACE universal primer B 26 GACTCTAGACGACATCGA(T)18 3′-RACE adapter primer B 25 GACTCTAGACGACATCGA 3′-RACE universal primer 5P1 ATGCAATTGAAACAAGAAAACCAGG Reverse primer for 5′-RACE 5P2 ATGCCATCTATCGATCAAACTGTT Reverse primer for 5′-RACE 3P1 AGGCTGGCGTGGTTGATACATGTG Forward primer for 3′-RACE 3P2 TGTGCGAATCGGATCAAGTTT Forward primer for 3′-RACE Table 1. Degenerate and specific primers used in this work. Table 1. Degenerate and specific primers used in this work. and the pattern of Cd uptake by roots displayed an initial slower stage during the first eight hours, followed by a second, rapid stage over the subsequent two weeks and it was significant lower in OE 3 than in WT, Atnr and Atnr-N24 (Fig. 7a). The Cd concentrations in the roots of all of the lines increased remarkably under Cd-stress conditions for two weeks. However, compared with the other lines, the OE 3 had significant higher concentra- tion of Cd in its aerial parts. Cd was transferred to aboveground parts began in 8 h and increased steadily in the following two weeks. The translocation factor of the transgenic lines was markedly higher than that of the other Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 4 www.nature.com/scientificreports/ Figure 3. Expression patterns of SaNramp6 in S. alfredii under Cd stress. (a) Different tissue without any heavy metal treatments, (b) Root, (c) Stem, (d) Leaf. The normalized mRNA levels without treatment (y-axis “Relative mRNA expression”) were set arbitrarily to 1. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. Different letters on the bars indicate significant difference between the treatments. P-values of the two-way ANOVAs of Time, Cd (Cd treatment) and their interaction (Time × Cd) are indicated. P < 0.05; P < 0.01. Figure 3. Expression patterns of SaNramp6 in S. alfredii under Cd stress. (a) Different tissue without any heavy metal treatments, (b) Root, (c) Stem, (d) Leaf. The normalized mRNA levels without treatment (y-axis “Relative mRNA expression”) were set arbitrarily to 1. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. Different letters on the bars indicate significant difference between the treatments. Results l i Names and accession numbers of the Nramp protein family members. Species Name Accession numbers Arabidopsis thaliana AtNramp1 NP_178198.1 Arabidopsis thaliana AtNramp6 NP_173048.3 Populus trichocarpa PtNramp1 XP_006368514.1 Morus notabilis MnNramp6 EXB50420.1 Triticum urartu TuNramp3 EMS65084.1 Nicotiana tabacum NtNramp1 BAH66919.1 Selaginella moellendorffii SmNramp XP_002966634.1 Solanum melongena SmNramp1 BAM34953.1 Solanum torvum StNramp1 BAM34952.1 Brassica juncea BjNramp ACR16683.1 Medicago truncatula MtNramp XP_003602053.1 Arachis hypogaea AhNramp1 AFQ37304.1 Theobroma cacao TcNramp6 XP_007023419.1 Oryza sativa Japonica OsNramp5 NP_001059312.1 Table 2. Names and accession numbers of the Nramp protein family members. Table 2. Names and accession numbers of the Nramp protein family members. lines after Cd exposure for two weeks (Fig. 7b). These results indicated that the transgenic lines may have a better absorption capacity for Cd. Thus, SaNramp6 may influence the accumulation ability of Cd in S. alfredii. To decipher the phenomenon of Cd accumulation in the four different lines, Cd2+ was measured in 30 d after 30 μM Cd treatment. Significantly higher Cd2+-influx rates were identified in overexpression lines (OE 2 and OE 3) d ith WT li (Fi 8 b) d k dl l i At li (Fi 8 d) h th diff lines after Cd exposure for two weeks (Fig. 7b). These results indicated that the transgenic lines may have a better absorption capacity for Cd. Thus, SaNramp6 may influence the accumulation ability of Cd in S. alfredii.ft lines after Cd exposure for two weeks (Fig. 7b). These results indicated that the transgenic lines may have a better absorption capacity for Cd. Thus, SaNramp6 may influence the accumulation ability of Cd in S. alfredii.ft hl To decipher the phenomenon of Cd accumulation in the four different lines, Cd2+ was measured in 30 d after 30 μM Cd treatment. Significantly higher Cd2+-influx rates were identified in overexpression lines (OE 2 and OE 3) compared with WT line (Fig. 8a,b) and markedly lower in Atnr line (Fig. 8c,d); however, there were no differences between rescue lines (Atnr-N24 and Atnr-N28) and WT line (Fig. 8c,d). Results l i P-values of the two-way ANOVAs of Time, Cd (Cd treatment) and their interaction (Time × Cd) are indicated. P < 0.05; *P < 0.01. Figure 4. SaNramp6 expression increases Cd2+ sensitivity and Cd2+ content in yeast. (a) Growth of ∆ycf1 yeast cells expressing SaNramp6 on plates containing SG-U without CdCl2. (Left) or supplemented with 15 μM CdCl2. (Middle) and 20 μM CdCl2. (Right), (b) Time-dependent growth of yeast strains in SG-U liquid medium supplemented with 5 μM CdCl2, (c) Cd content of ∆ycf1 yeast cells expressing SaNramp6 grown for 48 h in liquid SG-U supplemented with 5 μM CdCl2. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from the ∆ycf1 + EV. Figure 4. SaNramp6 expression increases Cd2+ sensitivity and Cd2+ content in yeast. (a) Growth of ∆ycf1 yeast cells expressing SaNramp6 on plates containing SG-U without CdCl2. (Left) or supplemented with 15 μM CdCl2. (Middle) and 20 μM CdCl2. (Right), (b) Time-dependent growth of yeast strains in SG-U liquid medium supplemented with 5 μM CdCl2, (c) Cd content of ∆ycf1 yeast cells expressing SaNramp6 grown for 48 h in liquid SG-U supplemented with 5 μM CdCl2. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from the ∆ycf1 + EV. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 5 www.nature.com/scientificreports/ Figure 5. Subcellular localization of SaNramp6. Control, Non-transformed protoplast; GFP-vector, protoplast transformed with p35S-GFP vector; SaNramp6-GFP, protoplast transformed with SaNramp6-GFP fusion. Scale bar = 7.5 μm. Figure 5. Subcellular localization of SaNramp6. Control, Non-transformed protoplast; GFP-vector, protoplast transformed with p35S-GFP vector; SaNramp6-GFP, protoplast transformed with SaNramp6-GFP fusion. Scale bar = 7.5 μm. Figure 5. Subcellular localization of SaNramp6. Control, Non-transformed protoplast; GFP-vector, protoplast transformed with p35S-GFP vector; SaNramp6-GFP, protoplast transformed with SaNramp6-GFP fusion. Scale bar = 7.5 μm. Species Name Accession numbers Arabidopsis thaliana AtNramp1 NP_178198.1 Arabidopsis thaliana AtNramp6 NP_173048.3 Populus trichocarpa PtNramp1 XP_006368514.1 Morus notabilis MnNramp6 EXB50420.1 Triticum urartu TuNramp3 EMS65084.1 Nicotiana tabacum NtNramp1 BAH66919.1 Selaginella moellendorffii SmNramp XP_002966634.1 Solanum melongena SmNramp1 BAM34953.1 Solanum torvum StNramp1 BAM34952.1 Brassica juncea BjNramp ACR16683.1 Medicago truncatula MtNramp XP_003602053.1 Arachis hypogaea AhNramp1 AFQ37304.1 Theobroma cacao TcNramp6 XP_007023419.1 Oryza sativa Japonica OsNramp5 NP_001059312.1 Table 2. Discussion AtNramp6 is located in a vesicular-shaped endomembrane compartment and works as an intracellular Cd transporter23. Similarly, OsNramp1 was localized to the plasma membrane in onion epidermal cells, and the overexpression of OsNramp1 results in a Cd accumulation in the leaves33. The soybean Nramp homologue, GmDmt, is located on the peribacteroid membrane of root nodules and mediates ferrous iron uptake in yeast27. In Thlaspi japonicum H., a nickel (Ni) hyperaccumulator, TjNramp4 could specifically transport Ni and increase Ni concentrations34. The deduced amino acid sequence of SaNramp6 shares an 80% identity with AtNramp6, and the phylogenetic tree also indicated that SaNramp6 was most similar to AtNramp6. In addition, our subcellular localization analysis showed that SaNramp6 was located in the plasma membrane. Thus, SaNramp6 could function as a metal transporter in the plasma membrane (Fig. 5). h g An plasma membrane localization is consistent with SaNramp6 conferring Cd uptake by increasing the Cd content in plant tissues. AtNramp3 is involved in increased metal tolerance or accumulation. However, AtNramp6 leads to Cd hypersensitivity when overexpressed in Arabidopsis, and Arabidopsis plants lacking AtNramp6 are more resistant to Cd than WT lines23. In rice, OsNramp1 participates in cellular Cd uptake or transport and the overexpression of OsNramp1 enhances tolerance to Cd and increases Cd accumulation in shoots24. OsNramp5 is a major transporter for Cd uptake, influencing Cd absorption in both solution and soil cultures25,33. Hence, compared with the control, SaNramp6′s overexpression in A. thaliana could accumulate more ROS of roots when exposed to CdCl2. We believe that understanding its functions in plants will facilitate the development of Cd-accumulating plants.h g p The growth of the Cd sensitive yeast strain (∆ycf1) transformed with the empty vector was inhibited by Cd, and yeast cells harbouring the SaNramp6-expression vector exhibited weaker growth activities. The results indi- cated that SaNramp6 cannot complement the Cd sensitivity or rescue the Cd-sensitivity phenotype in the mutant yeast strain. However, the Cd concentration of the SaNramp6-expression strain was 10% higher than that of the empty vector strain (Fig. 4). The induction of SaNramp6 expression by CdCl2 suggested that the gene might be involved in responding to heavy metal stress and is a transporter for Cd uptake in S. alfredii. Consistent results have been reported using a similar approach. Discussion Here, we cloned an Nramp family member from a heavy metal-accumulating ecotype of S. alfredii and the results showed that it conferred the ability to accumulate Cd in overexpression transgenic A. thaliana. Cd is a strongly toxic heavy metal transported across plant membranes by physiological metal transporters29. To date, various gene families related to the transport of Cd have been reported, such as the P-type ATPase superfamily30, ABC transporters31 and the CE family32. Among these metal transporters, the Nramp family is widely distributed in mammals, fungi and bacteria. 6 Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 www.nature.com/scientificreports/ Figure 6. ROS accumulation responses to Cd stress and physiological indicators in four different lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,e) H2O2, (b,f) O2 −, (c,g) CAT activity, (d,h) POD activity. Control, without Cd treatment; Cd treatment, 30 µM Cd treatment for two weeks. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from wild type. Figure 6. ROS accumulation responses to Cd stress and physiological indicators in four different lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,e) H2O2, (b,f) O2 −, (c,g) CAT activity, (d,h) POD activity. Control, without Cd treatment; Cd treatment, 30 µM Cd treatment for two weeks. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from wild type. Figure 6. ROS accumulation responses to Cd stress and physiological indicators in four different lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,e) H2O2, (b,f) O2 −, (c,g) CAT activity, (d,h) POD activity. Control, without Cd treatment; Cd treatment, 30 µM Cd treatment for two weeks. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from wild type. The determination of SaNramp6′s subcellular localization is important for understanding its potential roles in the process of accumulating metals. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 Discussion WT (wild type); OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 (rescue lines). Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. Different letters on the bars indicate significant difference at same time within treatments (WT, overexpression lines, mutant line, rescue lines). P-values of the two-way ANOVAs of Time, Cd (Cd treatment) and their interaction (Time × Cd) are indicated. P < 0.05; *P < 0.01. Figure 7. Time-dependent Cd-uptake experiments (a) and Cd-translocation factors (Tf) (b) in four lines. WT (wild type); OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 (rescue lines). Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. Different letters on the bars indicate significant difference at same time within treatments (WT, overexpression lines, mutant line, rescue lines). P-values of the two-way ANOVAs of Time, Cd (Cd treatment) and their interaction (Time × Cd) are indicated. P < 0.05; *P < 0.01. The improved Cd uptake caused by SaNramp6 may be due to the exertion of direct effects on several major pathways or may work in cooperation with other genes participating heavy metal uptake, transport, sequestra- tion and detoxification. A similar case was reported recently, the uptake of Fe in roots by NRAMP1 requires the partnership of another transporter, IRT1, in A. thaliana36. Although the function of SaNramp6 is still unclear, the gene appears to be related to the hyperaccumulator characteristic of S. alfredii. These findings will contribute to understanding the function of Nramp genes and provide experimental evidence and theoretical guidance for further studies. Discussion Thomine et al.22 found that the growth of transgenic yeast expressing AtNramp1, AtNramp3, AtNramp4 was strongly reduced in liquid cultures supplemented with 3 µM CdCl2 com- pared with the control, and these genes increased the Cd content in yeast. However, AtNramp3-OE in Arabidopsis were found to be hypersensitive to Cd. TcNRAMP3′s-expression increased Cd sensitivity and the Cd content in yeast, and TcNRAMP3-OE in tobacco resulted in a slight sensitivity of root growth to Cd35. The growth of yeast strain Δycf1 was affected by OsNRAMP5, which is involved in Cd transport25. Therefore, these data showed that some NRAMP members could increase Cd sensitivity and Cd concentration. Here, our results from SaNramp6 in transgenic yeast with Cd-sensitivity phenotype and Cd concentration (Fig. 4) and that in transgenic A. thaliana were consistent with the previous results (Fig. 7).h The capacity to reduce Cd-associated oxidation may be an important mechanism contributing to Cd uptake and transport. To test the role of SaNramp6 in heavy metal-stress tolerance, a functional analysis was carried out by overexpressing SaNramp6 in A. thaliana and rescuing the Arabidopsis mutant nramp1. In the physiological assays (Fig. 6), the root lengths of overexpression transgenic plants were markedly longer and the contents of H2O2, O2 −, CAT and POD were also higher than those in WT lines, which suggested that the overexpression of SaNramp6 enhanced the Cd uptake and accumulation in transgenic plants. p p g p Taken together, we have functioned SaNramp6 in transgenic yeast and A. thaliana and the data presented in this study suggested that SaNramp6 may be a critical Cd transporter responsible for Cd accumulation in S. alfredii. It was hard to place the functions of SaNramp6 into specific categories such as uptake or translocation. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 7 www.nature.com/scientificreports/ Figure 7. Time-dependent Cd-uptake experiments (a) and Cd-translocation factors (Tf) (b) in four lines. WT (wild type); OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 (rescue lines). Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. Different letters on the bars indicate significant difference at same time within treatments (WT, overexpression lines, mutant line, rescue lines). P-values of the two-way ANOVAs of Time, Cd (Cd treatment) and their interaction (Time × Cd) are indicated. P < 0.05; *P < 0.01. Figure 7. Time-dependent Cd-uptake experiments (a) and Cd-translocation factors (Tf) (b) in four lines. Methods Plant materials and growth conditions. A hyperaccumulating ecotype of S. alfredii was collected from the area of an old Pb/Zn mine in Quzhou City, Zhejiang Province, P. R. China. The plants were water-cultivated in an artificial climate chamber at 25 °C with a 16 h light/8 h dark cycle. The S. alfredii seedlings used for the stress treatment were asexual propagated to ensure consistency and grown in half-strength Hoagland-Arnon solution for about two weeks until relatively vigorous roots grew. For the expression analyses of target genes, plants were treated with 400 μM CdCl2 for 0 h, 0.5 h, 6 h, 12 h, 24 h, 48 h, 72 h, 96 h, 1 week and 2 week. Each treatment was replicated three times. All samples were quickly frozen in liquid nitrogen followed by storage at −80 °C until use. 8 Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 www.nature.com/scientificreports/ Figure 8. Comparison of Cd concentrations and net Cd2+-influx rates in four lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,c) Cd2+ flux rates with Cd treatment for 24 h. (b,d) Mean flow rates of Cd2+. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from wild type. Figure 8. Comparison of Cd concentrations and net Cd2+-influx rates in four lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,c) Cd2+ flux rates with Cd treatment for 24 h. (b,d) Mean flow rates of Cd2+. Bars indicate means ± standard deviations (SDs) Figure 8. Comparison of Cd concentrations and net Cd2+-influx rates in four lines - WT (wild type); OE 2 and OE 3 (overexpression lines); Atnr (mutant line); Atnr-N24 and Atnr-N28 (rescue lines). (a,c) Cd2+ flux rates with Cd treatment for 24 h. (b,d) Mean flow rates of Cd2+. Bars indicate means ± standard deviations (SDs) of at least three independent biological experiments. One or two asterisks indicate a significant difference at P < 0.05 or P < 0.01 from wild type. A. thaliana (ecotype Columbia) was grown in a controlled environmental chamber at 22 °C under a long-day cycle (16 h light, 8 h dark), with a white light intensity of approximately 125 mmol·m−2·s−1 and 70% relative humidity. Methods CELLO v.2.5: subCELlular LOcalization predictor (http://cello.life.nctu.edu.tw/) and SOSUI version 1.11 (http://bp.nuap.nagoya-u.ac.jp/sosui/) were used to predict subcellular localization and transmem- brane domains, respectively. For the multiple sequence alignment, Clustal Omega (http://www.ebi.ac.uk/Tools/ msa/) was performed to align amino acid sequences first, and subsequently, the results were edited by GeneDoc. Additionally, a phylogenetic tree was constructed by MEGA 5.2 software using the Neighbour-joining method with 1,000 replicates based on amino acid sequences of the NRAMP proteins. The known NRAMP protein sequences from NCBI GenBank are shown in Table 2. Expression pattern analysis. SYBR-based quantitative real-time PCR (qRT-PCR) reactions (SYBR pre- mix EX Tag reagent, TaKaRa, Da Lian, China) were carried out in triplicate on a 7300 Real-Time PCR System (Applied Biosystems, CA, USA) according to the manufacturer’s instructions. Relative gene expression was esti- mated based on the 2−ΔΔCt method, applying the geometric mean of two reference genes: UBC9 and TUB39,40. All of the primers for RT-qPCR are listed in Table 1. Expression vector construction. The open reading frame of SaNramp6 was amplified by PCR using High Fidelity KOD-Plus DNA Polymerase (Toyobo, Japan) from the cDNA of S. alfredii using the specific primers SaNramp6-GF and SaNramp6-R (Table 1). The yeast expression vector pYES2.1 -SaNramp6 was generated using pYES2.1 TOPO® TA Expression Kit (Invitrogen, Carlsbad, USA). For subcellular location and plant expression vector, the purified PCR products were then cloned into the Gateway entry vector pENTR/D-Topo (Invitrogen, Carlsbad, USA) and positive clones were further sequenced to verify the direction and sequence accuracy. The sequence-verified plasmid was then recombined in pK7WGF2.0 and pH2GW7.041 to generate pK7W- GF2.0-SaNramp6 and pH2GW7.0-SaNramp6, respectively. Subcellular localization of SaNramp6. The correct plasmid pK7WGF2.0-SaNramp6 fused to the green fluorescent protein (GFP) was extracted by Plasmid Maxprep Kit (Vigorous, Beijing, China). Free vector p35S-GFP was used as control. A. thaliana protoplast isolation and transfection were performed as previously described42. The subcellular location of SaNramp6 was further investigated by transient expression in onion epi- dermal cells and Nicotiana benthamiana lower leaf epidermal cells as described by Liu et al.43 and Zheng et al.44, respectively. A LSM510 confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany) was used to observe the signals. Heterologous expression of SaNramp6 in yeast. The Saccharomyces cerevisiae strain BY4742 ∆ycf1 (MATα; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0; YDR135c::kanMX4) was a Cd-sensitive mutant, which lacked the ability to compartmentalize Cd into vacuoles45, was used to assess the Cd tolerance of SaNramp6. Methods Overexpression lines, mutant and rescue lines, were selected for physiological assays. The seeds were surface sterilized and germinated on 1/2 Murashige and Skoog (MS) agar plates containing 25 mg·L−1 hygromy- cin. Whereafter, 30 d-old Arabidopsis seedlings were soaked in Hoagland-Arnon solution with or without 50 µM CdCl2 for 24 h and them used to measure the Cd2+ flux. For physiological assays, 30 d-old homozygous transgenic seedlings were soaked in Hoagland-Arnon solution with or without 30 µM CdCl2 for two weeks, and treatments for 0 h, 10 min, 20 min, 30 min, 1 h, 2 h, 4 h, 8 h, 12 h, 24 h, 48 h, 96 h, 1 week and 2 week used for Cd-uptake assay. RNA preparation, cDNA synthesis and DNA extraction. Leaves, stems and roots were harvested after each treatment, and all of the samples were frozen in liquid nitrogen and stored at −80 °C for analysis. Total RNA was isolated from the tissues using the Total RNA Purification Kit (NORGEN, Thorold, Canada). First-strand cDNA was then synthesized from 2 µg of total RNA by the Superscript RT III first-strand cDNA synthesis kit fol- lowed by RNase H (Invitrogen, Carlsbad, USA) treatment. Genomic DNA was isolated from seedling leaves using cetyltrimethyl ammonium bromide (CTAB) method as described by Murray and Thompson37. Cloning of SaNramp6 gene. The full-length SaNramp6 cDNA was amplified by reverse transcription-PCR RT-PCR) and rapid amplification of cDNA ends-PCR (RACE-PCR).hi (R CR) a d ap d a pi cat o o c N e ds CR (R C CR). The internal fragment of SaNramp6 was isolated from S. alfredii using the specific primers SaNramp6-F and SaNramp6-R, which were designed according to transcriptome data28. To obtain the 3′-end cDNA and 5′-ready cDNA, four gene specific primers 3P1, 3P2, 5P1 and 5P2, were designed and synthesized based on the sequence of the cloned internal fragment. The cloning was performed as described by Wang et al.38. Additionally, the genomic sequence of SaNramp6 was amplified by PCR using genomic DNA as the template with primers SaNramp6-F and SaNramp6-R. All of the primers are listed in Table 1. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 9 www.nature.com/scientificreports/ Bioinformatics analysis of SaNramp6. To compare the genomic structure, the genomic sequences of A. thaliana Nramps from GenBank were searched and the intron-exon structure was analyzed. p y Translation and protein analyses of SaNramp6 were initially performed using ExPASy tools (http://www. expasy.org/tools/). Methods thaliana (Atnr) and the rescue of the Arabidopsis mutants (Atnr-N24 and Atnr-N28), as well as wild type, were used for abiotic stress-related physiological analyses, including root length, peroxidase Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 10 www.nature.com/scientificreports/ (POD) activity, catalase (CAT) content, H2O2 and superoxide anion accumulations, and measurements of the Cd2+ flux. All of the experiments were independently carried out three times. (POD) activity, catalase (CAT) content, H2O2 and superoxide anion accumulations, and measurements of the Cd2+ flux. All of the experiments were independently carried out three times. l p p y As for analyzing peroxidase (POD) activity, catalase (CAT) content, H2O2 and superoxide anion accumula- tions, approximately 0.1 g of root tissue was ground in liquid nitrogen and placed it in 2-mL tubes. The extraction of these physiological indices used the appropriate kits according to the instruction manual (Comin, Suzhou, China).hl The net Cd2+ fluxes in the roots of Arabidopsis were measured noninvasively by the Younger USA NMT Service Centre (Xuyue, Beijing) using the NMT system (NMT100 Series, Younger, USA LLC, Amherst, MA, USA). Prior to the flux measurement, the roots were equilibrated for 15 min in testing liquid (0.05 mM CdCl2, 0.1 mM KCl, 0.02 mM CaCl2, 0.02 mM MgCl2, 0.5 mM NaCl, 0.1 mM Na2SO4 and 0.3 mM MES, pH 5.7). Then, the transmembrane Cd2+ flux in roots was measured of different lines (120 µm to root apex) for 15 min by a Cd2+-selective microelectrode. All of the measurements were repeated at least six times independently. Cd concentration assay. To test the characteristic of SaNramp6′s Cd accumulation in A. thaliana, SaNramp6-OE A. thaliana (OE 3), homozygous mutant A. thaliana (Atnr), the rescue of the Arabidopsis mutant (Atnr-N28) and wild type were used in this experiment. yp p Roots and aerial parts were harvested individually for the Cd concentration analysis, and roots were resorbed by dipping in 1 mM EDTA for 30 min, and then washed three times with distilled water. All of the samples con- taining yeast cells for Cd determination were dried at 105 °C for 30 min, and then placed at 70 °C until they reached a constant weight. The dried samples were digested with a concentrated acid mixture of HNO3, HClO4, and H2SO4 (volume ratio = 4:1:0.5) at 250 °C for 8 h. References References 1. Fox, T. C. & Guerinot, M. L. Molecular biology of cation transport in plants. Plant Biology 49, 669–696 (1998). , , gy p p gy , ( ) 2. Satarug, S., Garrett, S. H., Sens, M. A. & Sens, D. A. Cadmium, environmental exposure, and health outcomes. Ciência & Saúde Coletiva 16, 2587–2602 (2010). gy p p gy ( ) 2. Satarug, S., Garrett, S. H., Sens, M. A. & Sens, D. A. Cadmium, environmental exposure, and health outcomes. Ciência & Saúde Coletiva 16, 2587–2602 (2010). 3. Nawrot, T. et al. Environmental exposure to cadmium and risk of cancer: a prospective population-based study. Lancet Oncology 7 119–126 (2006).f ( ) 4. Mason, L. H., Harp, J. P. & Han, D. Y. Pb neurotoxicity: neuropsychological effects of lead toxicity. Biomed Research International 2014 (2014). ( ) 4. Mason, L. H., Harp, J. P. & Han, D. Y. Pb neurotoxicity: neuropsychological effects of lead toxicity. Biomed Research International 2014 (2014). 0 ( 0 ). 5. Mcgrath, S. P., Zhao, F. J. & Lombi, E. Plant and rhizosphere processes involved in phytoremediation of metal-contaminated soils. Plant and Soil 232, 207–214 (2001). ( ) 5. Mcgrath, S. P., Zhao, F. J. & Lombi, E. Plant and rhizosphere processes involved in phytoremediation of metal-contaminated soils. Plant and Soil 232, 207–214 (2001). , ( ) 6. Mcgrath, S. P. & Zhao, F. J. Phytoextraction of metals and metalloids from contaminated soils. Current Opinion in Biotechnology 14 277–282 (2003).i 7. Yang, X. E., Long, X. X., Ni, W. Z. & Fu, C. X. Sedum alfredii H: A new Zn hyperaccumulating plant first found in China. Science Bulletin 47, 1634–1637 (2002).f 7. Yang, X. E., Long, X. X., Ni, W. Z. & Fu, C. X. Sedum alfredii H: A new Zn hyperaccumulating plant first found in China. Science Bulletin 47, 1634–1637 (2002).f ( ) 8. Yang, X. E. & Stoffella, P. J. Cadmium tolerance and hyperaccumulation in a new Zn-hyperaccumulating plant species (Sedum alfredii Hance). Plant and Soil 259, 181–189 (2004). f ) ( ) 9. Lu, L. et al. Enhanced root-to-shoot translocation of cadmium in the hyperaccumulating ecotype of Sedum alfredii. Journal o Experimental Botany 59, 3203–3213 (2008). 10. Yang, X. et al. Zinc compartmentation in root, transport into xylem, and absorption into leaf cells in the hyperaccumulating species of Sedum alfredii Hance. Planta 224, 185–195 (2006). 11. Tian, S. & Brown, P. Methods The metal concentration in the digested solution was deter- mined by atomic absorption spectrometry (M6; SOLLAR) and an inductively coupled plasma-mass spectrometer (ICP-MS; NexION 300; PerkinElmer) after dilution. Data processing. Data were exhibited as the means ± standard deviations (SDs) of at least three independent biological experiments. Statistical analysis was performed using SPSS 17.0 statistics software. To test significant changes in mRNA relative expression and Cd concentration, time and Cd treatment were regarded as the main factors. Tukey- HSD method was used to correct all P-values of these multiple comparisons. In addition, one asterisk () or two asterisk (**), significantly different from control at P = 0.05, 0.01, respectively. if The translocation factor for Cd within a plant was expressed by the concentration in the aerial parts (µg·g−1DW)/the concentration in the roots (µg·g−1DW), which showed the Cd-translocation properties from roots to aerial parts49. Methods The yeast transformation was performed using the lithium acetate method46. Yeast ∆ycf1 cells transformed with the empty pYES2.0 vector were used as controls. The transformed yeast cells were selected on synthetic defined medium lacking uracil. For complementation assays, a series of three 1:10 dilutions from each culture was spotted onto synthetic-galactose-uracil (SG-U) agar plates supplemented with 0, 15 and 20 µM CdCl2 and incubated at 30 °C for three days. The relative growth of transformants was determined by measuring the OD600 at 6 h intervals. For the Cd-uptake assay, yeast cells transformed with the empty or SaNramp6 vector were grown for 48 h at 30 °C on SG-U supplemented with 5 µM CdCl2, then measured the Cd content. Detection of the Arabidopsis Atnr mutant by Atnramp1. To understand the functions of SaNramp6, we obtained mutant alleles from the SALK collection of sequence-indexed T-DNA insertions47. However, the mutant alleles of AtNramp6 were not found. Therefore, a single insertion line (SALK_053236; nramp1-1) was con- firmed for SaNramp6 because AtNramp6 and AtNramp1 have similar genomic structure. A homozygous mutant was detected by PCR using the primers (LP/RP and universal primers BP) designed based on the T-DNA web- site (http://signal.salk.edu/tdnaprimers.2.html) (data not shown). The collected homozygous mutant seeds were air-dried and stored at 4 °C. Generation of transgenic A. thaliana. The recombinant plasmid pH2GW7.0-SaNramp6 was introduced into Agrobacterium tumefaciens strain EHA105. A. thaliana ecotype Columbia plants were transfected by the floral dip method48. Positive transformants were selected based on hygromycin (Hyg, 20 μg·mL−1) resistance and confirmed by PCR and RT-PCR using the primers described above, AtActin (Table 1) was the internal control. Homozygous lines were identified by screening for non-segregation from each independent transformant (T3 generation). Physiological analysis of SaNramp6 transgenic, mutant and rescue of mutant lines. Six overex- pression lines (designated OE) with high transcriptional levels of SaNramp6 and 26 rescue of Arabidopsis mutant lines were obtained in this study. Among them, the OE 2, OE 3, Atnr-N24 and Atnr-N28 lines were selected in the following study owing to their phenotypes.f g y g p yp To investigate the potential effects of SaNramp6 in A. thaliana, SaNramp6-OE A. thaliana (OE 2 and OE 3), homozygous mutant A. www.nature.com/scientificreports/ ff y 1. Karimi, M., Inzé, D. & Depicker, A. GATEWAY vectors for Agrobacterium-mediated plant transformation. Trends in Plant Science 7, 193–195 (2002). 42. Yoo, S. D., Cho, Y. H. & Sheen, J. Arabidopsis mesophyll protoplasts: a versatile cell system for transient gene expression analysis. Nature Protocols 2, 1565–1572 (2007). 3. Liu, M. et al. Functional characterization of a gene in sedum alfredii hance resembling rubber elongation factor endowed with functions associated with cadmium tolerance. Frontiers in Plant Science 7 (2016).fi 44. Zheng, H. et al. A Rab-E GTPase mutant acts downstream of the Rab-D subclass in biosynthetic membrane traffic to the pl membrane in tobacco leaf epidermis. Plant Cell 17, 2020–2036 (2005). p ( ) 45. Li, Z. S. & Rea, P. A. A new pathway for vacuolar cadmium sequestration in Saccharomyces cerevisiae: YCF1-catalyzed transpo bis (glutathionato) cadmium. Proceedings of the National Academy of Sciences of the United States of America 94, 42–47 (1997)hi 46. Szczypka, M. S., Wemmie, J. A., Moye-Rowley, W. S. & Thiele, D. J. A yeast metal resistance protein similar to human cystic fibrosis transmembrane conductance regulator (CFTR) and multidrug resistance-associated protein. Journal of Biological Chemistry 269, 22853–22857 (1994). ( ) 47. Alonso, J. M. & Ecker, J. R. Genome-wide insertional mutagenesis of Arabidopsis thaliana. Science 301, 653–657 (2003).i M. & Ecker, J. R. Genome-wide insertional mutagenesis of Arabidop l l d l fi d h d f b . Alonso, J. M. & Ecker, J. R. Genome-wide insertional mutagenesi l h l l d l fi d h d f g p 48. Clough, S. J. & Bent, A. F. Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant Journal 16, 735–743 (1998). 9. Stoltz, E. & Greger, M. Accumulation properties of As, Cd, Cu, Pb and Zn by four wetland plant species growing on submerged mine tailings. Environmental & Experimental Botany 47, 271–280 (2002). Acknowledgements We thank the anonymous referees and the editor for the comments and suggestions that helped improve the manuscript. This work was supported by Basic Scientific Research Project of Nonprofit Central Research Institutions (No. CAFYBB2012040 and TGB2013008), National High-tech R&D Program of China No. 2013AA102701-3, the National Natural Science Foundation of China (No. 31200465), the National Key R&D Program of China (No. 2016YFD0800801) and the National Nonprofit Institute Research Grant of CAF (No. CAFYBB2012040, No. RISF2014010). www.nature.com/scientificreports/ www.nature.com/scientificreports/ 22. Thomine, S. & Schroeder, J. I. Cadmium and iron transport by members of a plant metal transporter family in Arabidopsis with homology to Nramp genes. Proceedings of the National Academy of Sciences of the United States of America 97, 4991–4996 (2000). 23 C illi R L B B i J F M i S & C i C Th NRAMP6 l ib d i i i gy p g g f y f f f 3. Cailliatte, R., Lapeyre, B., Briat, J. F., Mari, S. & Curie, C. The NRAMP6 metal transporter contributes to cadmium toxicity Biochemical Journal 422, 217–228 (2009).h 4. Takahashi, R. et al. The OsNRAMP1 iron transporter is involved in Cd accumulation in rice. Journal of Experimental Botany 62 4843–4850 (2011). 5. Sasaki, A., Yamaji, N., Yokosho, K. & Ma, J. F. Nramp5 is a major transporter responsible for manganese and cadmium uptake in rice Plant Cell 24, 2155–2167 (2012).f ( ) 6. Bereczky, Z., Wang, H. Y., Schubert, V., Ganal, M. & Bauer, P. Differential regulation of Nramp and Irt metal transporter genes in wild t d i t k t t f t t J l f Bi l i l Ch i t 278 24697 24704 (2003) 26. Bereczky, Z., Wang, H. Y., Schubert, V., Ganal, M. & Bauer, P. Differential regulation of Nramp and Irt metal transporter genes in wild type and iron uptake mutants of tomato. Journal of Biological Chemistry 278, 24697–24704 (2003). yp p J f g y , ( ) 27. Kaiser, B. N. et al. The soybean NRAMP homologue, GmDMT1, is a symbiotic divalent metal transporter capable of ferrous iron transport. Plant Journal for Cell & Molecular Biology 35, 295–304 (2003). 28. Han, X. et al. Integration of small RNAs, degradome and transcriptome sequencing in hyperaccumulator Sedum alfredii uncovers a complex regulatory network and provides insights into cadmium phytoremediation. Plant Biotechnology Journal 33, 552–575(524) (2016). 29. Clemens, S., Antosiewicz, D. M., Ward, J. M., Schachtman, D. P. & Schroeder, J. I. The plant cDNA LCT1 mediates the uptake of calcium and cadmium in yeast. Proceedings of the National Academy of Sciences of the United States of America 95, 12043–12048 (1998). 0. Alvarez, M. E. et al. Reactive Oxygen intermediates mediate a systemic signal network in the establishment of plant immunity. Cel 92, 773–784 (1998).i 31. Rea, P. A., Li, Z. S., Lu, Y. P., Drozdowicz, Y. M. www.nature.com/scientificreports/ & Martinoia, E. From vacuolar gs-x pumps to multispecific ABC transporters. Plant Biology 49, 727–760 (1998).h gy 32. Paulsen, I. T. & Saier, M. H. A novel family of ubiquitous heavy metal ion transport proteins. The Journal of Membrane Biology 156, 99–103 (1997). 3. Ishimaru, Y. et al. Characterizing the role of rice NRAMP5 in manganese, iron and cadmium transport. Scientific Reports 2, 989–993 (2012).h 34. Mizuno, T. et al. Cloning of three ZIP/Nramp transporter genes from a Ni hyperaccumulator plant Thlaspi japonicum and their Ni2+-transport abilities. Plant Physiology & Biochemistry Ppb 43, 793–801 (2005). 35. Wei, W. et al. The Thlaspi caerulescens NRAMP homologue TcNRAMP3 is capable of divalent cation transport. Molecular Biotechnology 41, 15–21 (2009).hfi gy 6. Castaings, L., Caquot, A., Loubet, S. & Curie, C. The high-affinity metal transporters NRAMP1 and IRT1 team up to take up iron under sufficient metal provision. Scientific Reports 6 (2016).h fii 37. Murray, M. G. & Thompson, W. F. Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8, 4321–4325 (1980). 38 Wang M Zhang Y Wang J W X & G o X A no el MAP kinase gene in cotton (Gossypium hirsutum L ) GhMAPK is in ol ed fii 37. Murray, M. G. & Thompson, W. F. Rapid isolation of high molec fii 37. Murray, M. G. & Thompson, W. F. Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8, 4321–4325 (1980). 38. Wang, M., Zhang, Y., Wang, J., Wu, X. & Guo, X. A novel MAP kinase gene in cotton (Gossypium hirsutum L.), GhMAPK, is involved in response to diverse environmental stresses Journal of Biochemistry & Molecular Biology 40 325–332 (2007) 37. Murray, M. G. & Thompson, W. F. Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research 8, 4321–4325 (1 38. Wang, M., Zhang, Y., Wang, J., Wu, X. & Guo, X. A novel MAP kinase gene in cotton (Gossypium hirsutum L.), GhMAPK, is invo in response to diverse environmental stresses. Journal of Biochemistry & Molecular Biology 40, 325–332 (2007). 9. Vandesompele, J. et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple interna control genes. Genome Biology 3, 1–11 (2002). g gy 40. Sang, J. et al. Selection and validation of reference genes for real−time quantitative PCR in hyperaccumulating ecotype of Sedum alfredii under different heavy metals stresses. Plos One 8, e82927–e82927 (2013). Author Contributions S.C., X.H., J.F., R.Z. planned and designed the research. S.C., X.H., and J.F. performed the experiments. Z.L., W.M., M.L., J.S. and J.J. contributed analytical tools, and S.C., X.H. wrote the manuscript and coordinated its revision. R.Z. contributed reagents/materials/funds support. All authors read and provided helpful discussions, and approved the final version. References Cellular sequestration of cadmium in the hyperaccumulator plant species Sedum alfredii. Plant Physiology 157, 1914–1925 (2011). 2. Page, M. D. & Merchant, S. S. Two chlamydomonas CTR copper transporters with a novel cys-met motif are localized to the plasma membrane and function in copper assimilation. Plant Cell 21, 928–943 (2009). pp 13. Vert, G., Briat, J. F. & Curie, C. Arabidopsis IRT2 gene encodes a root-periphery iron transporter. Plant Journal 26, 181–189 (2001).h 13. Vert, G., Briat, J. F. & Curie, C. Arabidopsis IRT2 gene encodes a root-periphery iron transporter. Plant Journal 26, 181–189 (2001). 14. Nevo, Y. & Nelson, N. The NRAMP family of metal-ion transporters. Biochimica Et Biophysica Acta Molecular Cell Research 1763, 609–620 (2006). 13. Vert, G., Briat, J. F. & Curie, C. Arabidopsis IRT2 gene encode 15. Segond, D. et al. NRAMP genes function in Arabidopsis thaliana resistance to Erwinia chrysanthemi infection. Plant Journal for Cell & Molecular Biology 58, 195–207 (2009). gy 6. Migeon, A. et al. Genome-wide analysis of plant metal transporters, with an emphasis on poplar. Cellular & Molecular Life Science Cmls 67, 3763–3784 (2010). 7. Colangelo, E. P. & Guerinot, M. L. Put the metal to the petal: metal uptake and transport throughout plants. Current Opinion in Plan Biology 9, 322–330 (2006). gy 18. Xia, J., Yamaji, N., Kasai, T. & Ma, J. F. Plasma membrane-localized transporter for aluminum in rice. Proceedings of the National Academy of Sciences of the United States of America 107, 18381–18385 (2010).i y f f f 9. Cellier, M. et al. Nramp defines a family of membrane proteins. Proceedings of the National Academy of Sciences 92, 10089–10093 (1995). ( ) 20. Curie, C., Alonso, J. M., Le, J. M., Ecker, J. R. & Briat, J. F. Involvement of NRAMP1 from Arabidopsis thaliana in iron transport. Biochemical Journal 347(Pt 3), S34–36 (2000). J ( ), ( ) 21. Lanquar, V. et al. Mobilization of vacuolar iron by AtNRAMP3 and AtNRAMP4 is essential for seed germination on low iron. Embo Journal 24, 4041–4051 (2006). Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 11 Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-017-13463-4. Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps an institutional affiliations. Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 12 nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © Th A th ( ) 2017 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2017 © The Author(s) 2017 Scientific Reports \| 7: 13318 \| DOI:10.1038/s41598-017-13463-4 13
https://openalex.org/W4234047965	https://www.researchsquare.com/article/rs-417270/latest.pdf	English	null	Association Between Body Mass Index, Its Change And Cognitive Impairment Among Chinese Older Adults: A Community-Based, 9-Year Prospective Cohort Study	Research Square (Research Square)	2,021	cc-by	6,847	Association Between Body Mass Index, Its Change And Cognitive Impairment Among Chinese Older Adults: A Community-Based, 9-Year Prospective Cohort Study Page 1/21 y Shanshan Wu Capital Medical University Xiaozhen Lv Peking University Institute of Mental Health: Peking University Sixth Hospital Jie Shen (  shenjie_96@163.com ) Zhejiang University - School of Medicine - School of Public Health Hui Chen Zhejiang University - School of Medicine - School of Public Health Yuan Ma Harvard University T H Chan School of Public Health Xurui Jin MindRank AI ltd. Jiaxi Yang Harvard University T H Chan School of Public Health Yaying Cao Shanghai Institutes of Nutrition and Health CAS: Chinese Academy of Sciences Shanghai Institutes of Nutrition and Health Geng Zong Chinese Academy of Sciences Shanghai Institutes of Nutrition and Health Huali Wang Peking University Institute of Mental Health: Peking University Sixth Hospital Changzheng Yuan Zhejiang University - School of Medicine - School of Public Health Research Article Keywords: BMI, BMI change, cognitive impairment, older adults, Chinese, cohort Posted Date: April 26th, 2021 DOI: https://doi.org/10.21203/rs.3.rs-417270/v1 Jiaxi Yang Harvard University T H Chan School of Public Health Research Article Page 1/21 Keywords: BMI, BMI change, cognitive impairment, older adults, Chinese, cohort Posted Date: April 26th, 2021 DOI: https://doi.org/10.21203/rs.3.rs-417270/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published at European Journal of Epidemiology on August 9th, 2021. See the published version at https://doi.org/10.1007/s10654-021-00792-y. Page 2/21 Abstract Objective To examine the association of baseline body mass index (BMI) and BMI change with cognitive impairment among older adults in China. Methods The study included data from the Chinese Longitudinal Healthy Longevity Study, a national community-based prospective cohort study from 2002-2018. Baseline BMI and BMI change measurements were available for 12,027 adults aged older than 65 years. Cognitive impairment was defined as Chinese version of the Mini Mental State Examination score less than 18. Multivariable Cox proportional hazard model was used. Results Among 12027 participants (mean age was 81.23 years old and 47.48% were male), the proportion of underweight, normal, overweight and obese at baseline was 33.87%, 51.39%, 11.39% and 3.34%, respectively. During an average of 5.9 years’ follow-up, 3086 participants (4.35 per 100 person-years) with incident cognitive impairment were identified. Compared with normal weight group, adjusted hazard ratio (AHR) for cognitive impairment was 0.86 (95% CI 0.75-0.99) among overweight group, whereas corresponding AHR was 1.02 (95% CI 0.94-1.10) in underweight and 1.01 (95% CI 0.80-1.28) in obese. Large weight loss (<-10%) was significantly associated with an increased risk of cognitive impairment (AHR, 1.42, 95% CI 1.29-1.56), compared to stable weight status group (-5%~5%). In the restricted cubic spline models, BMI change showed a L-shaped association with cognitive impairment. Conclusions BMI-defined overweight is associated with a reduced risk of cognitive impairment among Chinese older adults, while large weight loss is associated with increased risk. More attention should be paid to older adults with significant weight loss. Study population This study is based on the Chinese Longitudinal Healthy Longevity Study (CLHLS), which is an ongoing, prospective cohort study of community-dwelling Chinese older people from 1998, covering approximately 85% of China’s population with representative data to investigate determinants of longevity[12-14]. New participants are enrolled during every 2-4 years of follow-up in order to reduce the attrition due to death and loss to follow-up. The surveys are conducted through face-to-face interviews in participants’ homes by trained interviewers with a structured questionnaire. Each interviewer was accompanied by a local doctor, a nurse, or a medical college student. All participants or their proxy respondents signed written consent forms to participate in the baseline and follow-up surveys. The study was approved by the Biomedical Ethics Committee of Peking University (IRB00001052-13074). The present study was based on three successive 9 to 10-year cohorts (the 2002-2011, 2005-2014 and 2008-2018 cohorts) within the CLHLS. Four waves were performed in each cohort (2002, 2005, 2008 and 2011 in the 2002-2011 cohort; 2005, 2008, 2011 and 2014 in the 2005-2014 cohort; and 2008, 2011, 2014 and 2018 in the 2008-2018 cohort). Participants who were with normal cognition (score≥18 on the Chinese version of Mini-Mental State Examination, MMSE) at baseline and completed at least once cognitive function assessments during follow-up surveys were included as study population. Participants were excluded due to poor cognitive function at baseline (MMSE score<18), a history of stroke or dementia at baseline, without baseline BMI or weight at the first follow-up, dead during the first year and lost to follow- up at the first 3-year survey. For participants who were enrolled in two or more cohorts, we only chose the record with the longest follow-up period. After excluding 5402 duplicated participants in three cohorts, a total of 12,027 participants were included in the final analysis, with 5285 from the 2002-2011 cohort, 2088 from the 2005-2014 cohort and 4654 from the 2008-2018 cohort (Figure 1). Introduction With the rapid growing of elderly population, cognitive impairment and dementia have become a major public health concern worldwide, particularly in China[1]. Epidemiological studies estimated the prevalence of mild cognitive impairment as 15.5% among Chinese elderly over 60 years old, representing approximately 38 million cases[2]. Currently, there is limited effective therapy on dementia, the critical prognosis of cognitive impairment. Therefore, it is critically important to identify the potential risk factors associated with cognitive impairment, particularly modifiable risk factors, to prevent or delay cognitive impairment effectively. Given the high prevalence of obesity, measured by body mass index (BMI), and its biological plausibility, increasing attention is being paid to investigate its relationship with cognitive impairment in older adults. However, the results are paradoxical. Some studies reported higher BMI associated with poor cognitive function in late life[1, 3, 4], whereas others[5–7] reported the protective effect of higher BMI on cognitive function. Discrepancies may be partly explained by differences in sample size, age distributions (mean age < 70 years old or above), and follow-up time (< 5 years or more). Also, limited studies have adopted prospective methods on this topic among Chinese elderly population, despite some cross-sectional Page 3/21 Page 3/21 studies[6, 8]. To our knowledge, the association between BMI status and cognitive impairment has not been thoroughly evaluated among a national cohort of Chinese older adults. studies[6, 8]. To our knowledge, the association between BMI status and cognitive impairment has not been thoroughly evaluated among a national cohort of Chinese older adults. Besides static weight status, among older individuals, weight changes often reflect declines in muscle mass and bone density[9], whereas few studies investigated the relationship between elderly weight/BMI change and cognitive impairment[7, 10, 11]. Also, limited period of follow up and suboptimal selection of weight loss measures hindered these studies from reflecting the long-term prospective relationship. Therefore, the aim of our study was to examine the association between BMI and its change with cognitive impairment based on Chinese Longitudinal Healthy Longevity Study (CLHLS), a national community-based prospective cohort of oldest adults in China with long-term follow up. Measurement of baseline BMI and BMI change Body weight and height were measured by trained medical staff according to the standardized protocol. Body weight was measured when individuals wearing light clothing in each wave. Height was measured as knee height (vertical distance from sole of the foot to the upper surface of the knee, with knee and ankle Page 4/21 each flexed to a 90° angle) in 2002 wave, and direct height in 2005 and 2008 waves. A validated equation[15] was used to calculate height at baseline for participants in 2002-2011 cohort (men, height=67.78+2.01knee height; women, height=74.08+1.81knee height). Baseline BMI was classified into 4 categories according to the guideline for Chinese[16], including underweight (BMI <18.5kg/m2), normal weight (18.5≤ BMI <23.9 kg/m2), overweight (24≤ BMI <27.9 kg/m2) and obese (BMI ≥28 kg/m2). BMI change during the first-3 year was calculated through the difference between BMI at the third year and baseline BMI divided by baseline BMI, and classified into 5 groups, including large weight loss (BMI change <-10%), small weight loss (-10%≤ BMI change <-5%), stable weight status (-5%≤ BMI change <5%), small weight gain (5%≤ BMI change <10%) and large weight gain (BMI change>10%). Assessment of cognitive function Cognitive function was measured by the Chinese version of MMSE scale during each survey through a homebased interview. The validity and reliability of the Chinese MMSE has been verified in several studies[12, 13, 17]. Based on the literature[18, 19], we considered responses of “unable to answer” as“wrong”. The MMSE score ranged from 0 to 30, with a higher score indicating better cognitive function. Since 54.9% of participants in the cohort were illiterate, a relatively low cutoff with MMSE score less than 18 was defined as cognitive impairment[20]. Covariates Sociodemographic characteristics, health behaviors and diet habits at baseline were adjusted as covariates in the model, which was selected as a prior based on the literature[18, 20]. Potential confounders included age (as a continuous variable), sex (male or female), type of residence (urban or rural), marital status (married or not), education (illiterate, defined as receiving<1 year of any formal education; literate, defined as receiving≥1 year of any formal education), living arrangement (with family member or alone/in nursing home), smoking status (never smoking, current smoking, former smoking), drinking status(never drinking, current drinking, former drinking), regular exercise (yes or no), vegetables intake (always, defined as eat vegetables almost every day; not always, defined as eat vegetables except winter, occasionally, or rarely/never) and fruit intake (always, defined as eat fruit almost every day; not always, defined as eat fruit except winter, occasionally, or rarely/never). Statistical analysis Cox proportional hazard model was performed to investigate the association of baseline BMI and BMI change with cognitive impairment. The endpoint was the first occurrence of cognitive impairment. The follow-up period started from baseline to the date of the first occurrence of cognitive impairment, or censored at the end of the study (the fourth wave), the date of death or lost-to-follow-up for those participants who did not develop cognitive impairment. Considering the very small percentage of missing values for covariates (0.017% of fruit consuming and 0.042% of vegetable consuming variables were missing), missing indicators were used to handle the analysis. Proportional hazard assumption was Page 5/21 Page 5/21 ascertained and satisfied by Kaplan-Meier curves for categorized variables and testing linear regression of scaled Schoenfeld residuals on functions of time for continuous variables. ascertained and satisfied by Kaplan-Meier curves for categorized variables and testing linear regression of scaled Schoenfeld residuals on functions of time for continuous variables. For both baseline BMI and BMI change, the adjustment was accomplished via three models: (1) model 1, univariable analysis; (2) model 2, adjusted for age and sex; (3) model 3, additionally adjusted type of residence, marital status, education, living arrangement, smoking status, drinking status, regular exercise, vegetables and fruit intake. For the association of BMI change and cognitive impairment, baseline BMI was additionally adjusted in model 3. Moreover, restricted cubic spline analysis was conducted to examine the shape of the association of baseline BMI and BMI change with cognitive impairment after adjusting covariates in model 3, respectively, with knots placed at 10th, 50th and 90th percentiles and median value of baseline BMI (19.84 kg/m2) and BMI change (0%) as reference point. Additionally, subgroup analysis was performed to investigate whether the association with baseline BMI and BMI change varied by age (65-79 years, ≥80 years), sex (male, female), educational (illiterate or not), smoking status (current drinker or not), alcohol drinking status (current drinker or not), regular exercise (yes, no). Effect modification was also detected by adding interaction terms of BMI and the abovementioned variables in the multivariable model, respectively. Further stratified multivariable analysis was conducted to explore the impact of BMI change on cognitive impairment in participants with different baseline BMI level (underweight, normal, overweight and obese). Statistical analysis Since participants who were lost to follow-up might be more likely to develop cognitive impairment, sensitivity analysis was conducted to assess the robustness of the results by excluding participants who were lost to follow-up during the second 3 years. In addition, another sensitivity analysis was conducted by using the MMSE score less than 24 as the definition of cognitive impairment. History of multiple major chronic diseases (as at least one of the following diseases: diabetes, hypertension, heart disease and cancer) were not included in our multivariable model, because it may lie within the causal pathway of BMI and cognitive impairment. However, we also conducted sensitivity analyses after further adjustment of these variables. A two-tailed P value <0.05 was considered to be statistically significant. All analyses were conducted using SAS software Version 9.4 and R version 4.0.2 (ggplot2 and forestplot package). Association of baseline BMI and BMI change with cognitive impairment Cox proportional hazard regression model with restricted cubic spline indicated that baseline BMI (as a continuous variable) was linearly associated with risk of cognitive impairment, with a negative and monotonic association (P=0.512, Figure 2A). Table 2 showed the association of baseline BMI with cognitive impairment. Compared with normal baseline BMI group, participants with overweight showed a decreased risk of cognitive impairment (adjusted HR=0.86, 95%CI: 0.75-0.99) according to multivariable adjusted model, whereas those with underweight and obese both had a similar risk of cognitive impairment as those with normal baseline BMI. Association of BMI change with cognitive impairment The median BMI change (25th, 75th percentile) of 12,027 participants during the first 3 years was 0% (-10.0%, 9.4%). The proportion of large weight loss, small weight loss, stable weight status, small weight gain and large weight gain was 24.21%, 12.56%, 29.87%, 9.86% and 23.50%, respectively. The incidence of cognitive impairment was 6.08, 3.56, 3.55, 3.65 and 4.51 per 100 person-years for those with large weight loss, small weight loss, stable weight status, small weight gain and large weight gain, respectively. According to the results of Cox proportional hazard analysis with restricted cubic spline, there was a non- linear (U-shaped) association between BMI change and cognitive impairment (P<0.001, Figure 2B). Table 3 showed the association of BMI change with cognitive impairment. In comparison with participants under stable weight status during the first 3 years, participants with large weight loss were more likely to develop cognitive impairment (adjusted HR=1.42, 95%CI: 1.29-1.56). No significant risk of cognitive impairment was detected in those with small weight loss, small weight gain and large weight gain compared with participants under stable weight status. Participant characteristics Among the 12,027 participants, 47.48% were male. The mean (standard deviation, SD) age was 81.23 (10.72) years old at baseline. Approximately 55% of the participants were illiterate, and 44.93% were married. The median (25th, 75th percentile) MMSE score was 28(25, 29) and the mean (SD) baseline BMI was 20.28 (3.86) kg/m2 at baseline. The proportion of underweight, normal, overweight and obese at baseline was 33.87%, 51.40%, 11.39% and 3.34%, respectively. As shown in Table 1, the participants with underweight were more likely to be older, be female, live in rural, be illiterate, be not married, live alone or in nursing home, have less regular exercise, and not always consuming vegetables or fruits. Page 6/21 Page 6/21 The average (SD) length of follow-up period was 5.9 (2.8) years (range, 2.1-11.2 years). A total of 3086 (25.7%) of 12,027 participants with cognitive impairment were identified. During the 70,936 person-years of follow-up, the incidence of cognitive impairment was 4.35 per 100 person-years (5.98, 3.91, 2.52 and 3.0 per 100 person-years for those with underweight, normal, overweight, and obesity, respectively). Subgroup analysis For both baseline BMI and BMI change, similar findings were observed across age and gender subgroups (Figure 3). In particular, compared with normal BMI group, similar risk was detected for overweight participants with 65-79 years (adjusted HR=0.89, 95%CI: 0.67-1.17) and ≥80 years (adjusted HR=0.89, 95% CI: 0.76-1.06) (Figure 3A). As shown in Figure 3B, the adjusted HR associated with large weight loss versus stable weight status equal to 1.43 (95% CI: 1.13-1.81) for participants with 65-79 years and 1.36 (95% CI: 1.22-1.51) for those with ≥80 years. Moreover, no significant modification was observed by age, gender, education level, major lifestyle factors and status of multiple major chronic diseases (data not shown). Furthermore, the associations of BMI change and cognitive impairment were non-significantly different across subgroups of baseline BMI status (P for interaction=0.388, Table S1). Compared with stable weight Page 7/21 Page 7/21 status, the increased risk of cognitive impairment associated with large weight loss was not only detected in participants with underweight, but also in participants with normal weight and overweight. In those participants with obese (N=402), large weight loss was associated with increased risk of cognitive impairment but not statistically significant, whereas significant higher risk of cognitive impairment associated with small weight gain was detected. Sensitivity analysis The results of sensitivity analysis were similar to the main analysis, when excluding participants who were lost to follow-up during the second 3 years, using a cutoff of MMSE score equals to 24 as the definition of cognitive impairment, or further adjustment of major chronic diseases in the multivariable model, which indicating the robustness of our results (Table S2). Discussion Another cohort based on Women's Health Initiative Study of Cognitive Aging with average 5.4 years of follow-up also demonstrated women aged 65 to 79 years old with weight loss ≥ 5% had a significantly lower global cognitive function score[11]. One possible mechanism may be that sarcopenia, a syndrome with generalized loss of skeletal muscle mass and strength, which could lead to low physical activity and further contribute to cognitive decline[33-35]. In addition, a L-shape relationship between BMI change and cognitive impairment was detected in our study, suggesting that large weight loss could be associated with greater risk of cognitive impairment. However, we observed a significant higher risk of cognitive impairment associated with small weight gain among participants with obese. Considering the limited sample of obese elderly in our study, the results need to be treated with caution and confirmed in further research. In clinical practice, this finding emphasizes the importance of considering history of weight loss as a potential predictor of cognitive dysfunction among elderly patients. A major strength of our study is the use of a well-designed, large-scaled, national representative, prospective cohort of older adults in China[12-14]. Also, the long-term and repeated follow up data allowed us to investigate the association of BMI dynamic change with cognitive impairment. Moreover, to our knowledge, this is the first study evaluating the relationship between late life BMI and its change with cognitive impairment in a national cohort of Chinese older people. Additionally, the average age of the participants in this study was 81 years old, a population with high risk of dementia, and our findings may provide crucial evidence to reduce dementia drastically. Furthermore, we conducted a variety of sensitivity analyses and verified the robustness of our results. Our study also has limitations. Firstly, some potential covariates, either unmeasured (such as medical treatment) or unknown, may confound the association between BMI and its change with cognitive impairment owing to the observational design, although we carefully controlled for numerous potential confounders. Secondly, some confounding factors based on self-reported data, such as smoking and drinking, without accurate number of consuming cigarettes and alcohol. Thus, it may lead to recall bias although we already adjusted them in the analysis. Thirdly, the average age of our population was 81 years old, and the generalizability of our results in younger individuals needed to be further validated. Discussion In this community-dwelling prospective cohort study, we observed a lower risk of cognitive impairment among elderly participants who were overweight, but not obese, at baseline, after adjusting for demographic and major lifestyle factors. In addition, individuals with large weight loss within a 3-year period had a greater risk of cognitive impairment compared with those in stable weight status. The protective effect of overweight in late life in our study was consistent with several previous studies[5-7, 21-27]. A 9-year longitudinal cohort study showed overweight was associated with a 25% decreased risk of dementia versus normal weight in individuals ≥75 years old[5]. Another 3-year cohort study also reported the inverse association between baseline BMI and cognitive impairment in participants ≥65 years old, with 3% risk reduction of cognitive impairment per 1 kg/m2 increase of baseline BMI[7]. Another Korean cohort study[25] with an average of 3.7 years follow up demonstrated overweight or obese older adults showed slower cognitive decline. Similarly, a cross-sectional study conducted in 1100 Chinese individuals greater than 60 years old demonstrated a decreased risk of cognitive impairment associated with overweight with adjusted odds ratio 0.46 (95%CI: 0.30, 0.70)[6]. Several potential mechanisms might contribute to this protective effect of overweight in elderly participants. Firstly, it is well-known that excess body adiposity tends to accrue during early and middle adulthood for most people. Among older individuals, however, low BMI often reflect underlying illness, decline in muscle mass and bone density[10], leading to the decreased validity of BMI as a measure of adiposity among older persons[28, 29]. Therefore, compared to older adults with lower BMI who may have worse health status, those with higher BMI may instead have a better late-life cognitive function. Secondly, some hormonal factors, such as high estrogen level in overweight elderly women secreted by extragonadal tissue[30] and leptin secreted by adipose tissue in both women and men, may play an important role in improving cognitive function[31, 32]. In terms of BMI change, the current findings suggested that large weight loss in older adults is associated with higher risk of cognitive impairment, regardless of baseline BMI, which was in line with other studies[7, 10, 11]. A 3-year follow up cohort study conducted in 5239 older participants aged over 65 years old in the Page 8/21 United States also showed the risk of cognitive decline would increase 98% in participants with BMI decrease greater than 10%[7]. Discussion Finally, the observed associations may be subject to potential reverse causation, which means that weight loss may indicate a prodromal phase of dementia. Declarations We are grateful to all cooperating organizations and their staff in CLHLS whose hard work made this study possible. We thank the interviewees and their families for their voluntary participation in the CLHLS study. We are grateful to all cooperating organizations and their staff in CLHLS whose hard work made this study possible. We thank the interviewees and their families for their voluntary participation in the CLHLS study. Funding: This work was funded by the National Natural Science Foundation of China (No. 82003539 and 71490732), Capital's Funds for Health Improvement and Research (CFH2020-3-4114). Funding: This work was funded by the National Natural Science Foundation of China (No. 82003539 and 71490732), Capital's Funds for Health Improvement and Research (CFH2020-3-4114). Conflicts of interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare no potential conflicts of interest relevant to this article. Conflicts of interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare no potential conflicts of interest relevant to this article. Data and materials availability: The CLHLS questionnaires are available at https://sites.duke.edu/centerforaging/ programs/chinese-longitudinal-healthy-longevity-survey- clhls/survey-documentation/questionnaires/. The full datasets used in this analysis are available from the corresponding author upon reasonable request. y q https://sites.duke.edu/centerforaging/ programs/chinese-longitudinal-healthy-longevity-survey- clhls/survey-documentation/questionnaires/. The full datasets used in this analysis are available from the corresponding author upon reasonable request. clhls/survey-documentation/questionnaires/. The full datasets used in this analysis are available from the corresponding author upon reasonable request. Author Contributions: SSW, XZL and CZY designed the study. SSW and JS drafted the manuscript. SSW and XZL analyzed the data. XRJ validated the analysis. HC, YM, XZL, JXY, YYC, GZ, LHW and CZY revised the manuscript. SSW, XZL and CZY interpreted the results, incorporated comments for the co-authors and finalized the manuscript. All authors approved the final version of the paper. Financial Disclosures: All authors have nothing to disclose. Financial Disclosures: All authors have nothing to disclose. Financial Disclosures: All authors have nothing to disclose. Ethical approval and Consent to participate: The CLHLS study was approved by the Research Ethics Committee of Peking University (IRB00001052-13074), and all participants or their proxy respondents provided written informed consent. Consent for publication: All authors are agree with the plant to submit to European Journal of Epidemiology. R f Conclusions Among a national representative cohort of Chinese older adults, overweight (defined as 24≤ BMI <27.9 kg/m) was associated with a lower risk of cognitive impairment, while large weight loss was associated with an increased risk of cognitive impairment. These findings support a potential role for great weight loss independent of static body mass index in the development of cognitive impairment. History of body weight change is an easy-to-recall measure among elderly adults, future public health recommendation and clinical practice on body weight management should take this into consideration for the prevention of cognitive impairment among elderly adults. Page 9/21 Page 9/21 References [1] Livingston G, Sommerlad A, Orgeta V, et al. Dementia prevention, intervention, and care. Lancet. 2017;390(10113):2673-734. https://doi.org/10.1016/s0140-6736(17)31363-6 [1] Livingston G, Sommerlad A, Orgeta V, et al. Dementia prevention, intervention, and care. Lancet. 2017;390(10113):2673-734. https://doi.org/10.1016/s0140-6736(17)31363-6 [1] Livingston G, Sommerlad A, Orgeta V, et al. Dementia prevention, intervention, and care. Lancet. 2017;390(10113):2673-734. https://doi.org/10.1016/s0140-6736(17)31363-6 [2] Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health. 2020;5(12):e661-e71. https://doi.org/10.1016/s2468-2667(20)30185-7 [2] Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health. 2020;5(12):e661-e71. https://doi.org/10.1016/s2468-2667(20)30185-7 [2] Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health. 2020;5(12):e661-e71. https://doi.org/10.1016/s2468-2667(20)30185-7 [3] Elias MF, Elias PK, Sullivan LM, Wolf PA, D'Agostino RB. Obesity, diabetes and cognitive deficit: The Framingham Heart Study. Neurobiol Aging. 2005;26 Suppl 1:11-6. https://doi.org/10.1016/j.neurobiolaging.2005.08.019 [3] Elias MF, Elias PK, Sullivan LM, Wolf PA, D'Agostino RB. Obesity, diabetes and cognitive deficit: The Framingham Heart Study. Neurobiol Aging. 2005;26 Suppl 1:11-6. https://doi.org/10.1016/j.neurobiolaging.2005.08.019 [4] Gunstad J, Lhotsky A, Wendell CR, Ferrucci L, Zonderman AB. Longitudinal examination of obesity and cognitive function: results from the Baltimore longitudinal study of aging. Neuroepidemiology. 2010;34(4):222-9. https://doi.org/10.1159/000297742 Page 10/21 Page 10/21 [5] Atti AR, Palmer K, Volpato S, Winblad B, De Ronchi D, Fratiglioni L. Late-life body mass index and dementia incidence: nine-year follow-up data from the Kungsholmen Project. J Am Geriatr Soc. 2008;56(1):111-6. https://doi.org/10.1111/j.1532-5415.2007.01458.x [5] Atti AR, Palmer K, Volpato S, Winblad B, De Ronchi D, Fratiglioni L. Late-life body mass index and dementia incidence: nine-year follow-up data from the Kungsholmen Project. J Am Geriatr Soc. 2008;56(1):111-6. https://doi.org/10.1111/j.1532-5415.2007.01458.x [6] Hou Q, Guan Y, Yu W, et al. Associations between obesity and cognitive impairment in the Chinese elderly: an observational study. Clin Interv Aging. 2019;14:367-73. https://doi.org/10.2147/cia.S192050 [6] Hou Q, Guan Y, Yu W, et al. Associations between obesity and cognitive impairment in the Chinese elderly: an observational study. Clin Interv Aging. 2019;14:367-73. https://doi.org/10.2147/cia.S192050 [7] Rodríguez-Fernández JM, Danies E, Martínez-Ortega J, Chen WC. Cognitive Decline, Body Mass Index, and Waist Circumference in Community-Dwelling Elderly Participants. J Geriatr Psychiatry Neurol. 2017;30(2):67-76. References https://doi.org/10.1177/0891988716686832 [8] Zhang T, Yan R, Chen Q, et al. Body mass index, waist-to-hip ratio and cognitive function among Chinese elderly: a cross-sectional study. BMJ Open. 2018;8(10):e022055. https://doi.org/10.1136/bmjopen-2018- 022055 [9] Manson JE, Bassuk SS, Hu FB, Stampfer MJ, Colditz GA, Willett WC. Estimating the number of deaths due to obesity: can the divergent findings be reconciled? J Womens Health (Larchmt). 2007;16(2):168-76. https://doi.org/10.1089/jwh.2006.0080 [10] Power BD, Alfonso H, Flicker L, Hankey GJ, Yeap BB, Almeida OP. Changes in body mass in later life and incident dementia. Int Psychogeriatr. 2013;25(3):467-78. https://doi.org/10.1017/s1041610212001834 [11] Driscoll I, Espeland MA, Wassertheil-Smoller S, et al. Weight change and cognitive function: findings from the Women's Health Initiative Study of Cognitive Aging. Obesity (Silver Spring). 2011;19(8):1595-600. https://doi.org/10.1038/oby.2011.23 [12] Lv YB, Gao X, Yin ZX, et al. Revisiting the association of blood pressure with mortality in oldest old people in China: community based, longitudinal prospective study. Bmj. 2018;361:k2158. https://doi.org/10.1136/bmj.k2158 [13] Zeng Y, Feng Q, Hesketh T, Christensen K, Vaupel JW. Survival, disabilities in activities of daily living, and physical and cognitive functioning among the oldest-old in China: a cohort study. Lancet. 2017;389(10079):1619-29. https://doi.org/10.1016/s0140-6736(17)30548-2 [14] Zeng Y. Introduction to the Chinese Longitudinal Healthy Longevity Survey (CLHLS). In: Zeng Y, Dudley LP, Denese AV, Danan G, editors. Healthy Longevity in China: demographic, socioeconomic, and psychological dimensions.: Springer, Dordrecht; 2008. p. p. 23–38. [15] Zhang H, Hsu-Hage BH, Wahlqvist ML. The use of knee height to estimate maximum stature in elderly Chinese. J Nutr Health Aging. 1998;2(2):84-7. [15] Zhang H, Hsu-Hage BH, Wahlqvist ML. The use of knee height to estimate maximum stature in elderly Chinese. J Nutr Health Aging. 1998;2(2):84-7. [16] Chen C, Lu FC. The guidelines for prevention and control of overweight and obesity in Chinese adults. Biomed Environ Sci. 2004;17 Suppl:1-36. Page 11/21 Page 11/21 [17] An R, Liu GG. Cognitive impairment and mortality among the oldest-old Chinese. Int J Geriatr Psychiatry. 2016;31(12):1345-53. https://doi.org/10.1002/gps.4442 [18] Lv X, Li W, Ma Y, et al. Cognitive decline and mortality among community-dwelling Chinese older people. BMC Med. 2019;17(1):63. https://doi.org/10.1186/s12916-019-1295-8 [19] Zhang Z, Gu D, Hayward MD. Childhood nutritional deprivation and cognitive impairment among older Chinese people. Soc Sci Med. 2010;71(5):941-9. https://doi.org/10.1016/j.socscimed.2010.05.013 [20] Wang J, Li T, Lv Y, et al. Fine Particulate Matter and Poor Cognitive Function among Chinese Older Adults: Evidence from a Community-Based, 12-Year Prospective Cohort Study. Environ Health Perspect. 2020;128(6):67013. https://doi.org/10.1289/ehp5304 [21] Ghaderpanahi M, Fakhrzadeh H, Sharifi F, et al. References https://doi.org/10.2147/cia.S65921 [31] Harvey J, Solovyova N, Irving A. Leptin and its role in hippocampal synaptic plasticity. Prog Lipid Res. 2006;45(5):369-78. https://doi.org/10.1016/j.plipres.2006.03.001 [32] Oomura Y, Hori N, Shiraishi T, et al. Leptin facilitates learning and memory performance and enhances hippocampal CA1 long-term potentiation and CaMK II phosphorylation in rats. Peptides. 2006;27(11):2738- 49. https://doi.org/10.1016/j.peptides.2006.07.001 [33] van Dam R, Van Ancum JM, Verlaan S, Scheerman K, Meskers CGM, Maier AB. Lower Cognitive Function in Older Patients with Lower Muscle Strength and Muscle Mass. Dement Geriatr Cogn Disord. 2018;45(3-4):243-50. https://doi.org/10.1159/000486711 [34] Levine ME, Crimmins EM. Sarcopenic obesity and cognitive functioning: the mediating roles of insulin resistance and inflammation? Curr Gerontol Geriatr Res. 2012;2012:826398. https://doi.org/10.1155/2012/826398 [34] Levine ME, Crimmins EM. Sarcopenic obesity and cognitive functioning: the mediating roles of insulin resistance and inflammation? Curr Gerontol Geriatr Res. 2012;2012:826398. https://doi.org/10.1155/2012/826398 [35] Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing. 2010;39(4):412-23. https://doi.org/10.1093/ageing/afq034 [35] Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing. 2010;39(4):412-23. https://doi.org/10.1093/ageing/afq034 References Association between late-life body mass index, waist circumference, and dementia: Kahrizak Elderly Study. J Am Geriatr Soc. 2012;60(1):173-4. https://doi.org/10.1111/j.1532-5415.2011.03751.x [22] Fitzpatrick AL, Kuller LH, Lopez OL, et al. Midlife and late-life obesity and the risk of dementia: cardiovascular health study. Arch Neurol. 2009;66(3):336-42. https://doi.org/10.1001/archneurol.2008.582 [23] Suemoto CK, Gilsanz P, Mayeda ER, Glymour MM. Body mass index and cognitive function: the potential for reverse causation. Int J Obes (Lond). 2015;39(9):1383-9. https://doi.org/10.1038/ijo.2015.83 [24] Dahl AK, Löppönen M, Isoaho R, Berg S, Kivelä SL. Overweight and obesity in old age are not associated with greater dementia risk. J Am Geriatr Soc. 2008;56(12):2261-6. https://doi.org/10.1111/j.1532- 5415.2008.01958.x [25] Kim G, Choi S, Lyu J. Body mass index and trajectories of cognitive decline among older Korean adults. Aging Ment Health. 2020;24(5):758-64. https://doi.org/10.1080/13607863.2018.1550628 [26] Kiesswetter E, Schrader E, Diekmann R, Sieber CC, Volkert D. Varying Associations Between Body Mass Index and Physical and Cognitive Function in Three Samples of Older Adults Living in Different Settings. J Gerontol A Biol Sci Med Sci. 2015;70(10):1255-61. https://doi.org/10.1093/gerona/glv048 [27] Kim S, Kim Y, Park SM. Body Mass Index and Decline of Cognitive Function. PLoS One. 2016;11(2):e0148908. https://doi.org/10.1371/journal.pone.0148908 [28] Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients: a prospective study. Jama. 1999;281(21):2013-9. https://doi.org/10.1001/jama.281.21.2013 [28] Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients: a prospective study. Jama. 1999;281(21):2013-9. https://doi.org/10.1001/jama.281.21.2013 [29] Zhang J, Lyu YB, Yin ZX, Luo JS, Shi WH, Shi XM. [Follow-up study of body mass index and risk of cognitive impairment among elderly adults aged ≥65 years old from longevity areas of China]. Zhonghua Yu Fang Yi Xue Za Zhi. 2017;51(11):1019-23. https://doi.org/10.3760/cma.j.issn.0253-9624.2017.11.012 [29] Zhang J, Lyu YB, Yin ZX, Luo JS, Shi WH, Shi XM. [Follow-up study of body mass index and risk of cognitive impairment among elderly adults aged ≥65 years old from longevity areas of China]. Zhonghua Yu Fang Yi Xue Za Zhi. 2017;51(11):1019-23. https://doi.org/10.3760/cma.j.issn.0253-9624.2017.11.012 Page 12/21 Page 12/21 [30] Cheng D, Liang B, Hao Y, Zhou W. Estrogen receptor α gene polymorphisms and risk of Alzheimer's disease: evidence from a meta-analysis. Clin Interv Aging. 2014;9:1031-8. https://doi.org/10.2147/cia.S65921 [30] Cheng D, Liang B, Hao Y, Zhou W. Estrogen receptor α gene polymorphisms and risk of Alzheimer's disease: evidence from a meta-analysis. Clin Interv Aging. 2014;9:1031-8. https://doi.org/10.2147/cia.S65921 [30] Cheng D, Liang B, Hao Y, Zhou W. Estrogen receptor α gene polymorphisms and risk of Alzheimer's disease: evidence from a meta-analysis. Clin Interv Aging. 2014;9:1031-8. Tables Table 1. Characteristics of 12027 participants at baseline* Page 13/21 Variables Total Underweight Normal Overweight Obese P value No. of participants 12027(100.0) 4074(33.87) 6181(51.39) 1370(11.39) 402(3.34) - Age, years, mean (SD) 81.23±10.72 84.57±10.41 80.34±10.49 76.66±9.77 76.59±10.35 <0.001 Age group <0.001 65-79 years 5401(44.91) 1274(31.27) 2983(48.26) 880(64.23) 264(65.67) ≥ 80 years 6626(55.09) 2800(68.73) 3198(51.74) 490(35.77) 138(34.33) Sex <0.001 Male 5722(47.58) 1624(39.86) 3190(51.61) 720(52.55) 188(46.77) Female 6305(52.42) 2450(60.14) 2991(48.39) 650(47.45) 214(53.23) Type of residence <0.001 Urban 2458(20.44) 579(14.21) 1299(21.02) 443(32.34) 137(34.08) Rural 9569(79.56) 3495(85.79) 4882(78.98) 927(67.66) 265(65.92) Marital status <0.001 Married 5404(44.93) 1369(33.60) 2986(48.31) 817(59.64) 232(57.71) Not married 6623(55.07) 2705(66.40) 3195(51.69) 553(40.36) 170(42.29) Education <0.001 Illiterate 6606(54.93) 2673(65.61) 3200(51.77) 547(39.93) 186(46.27) Literate 5421(45.07) 1401(34.39) 2981(48.23) 823(60.07) 216(53.73) Living arrangement <0.001 With family member 9982(83.00) 3286(80.66) 5167(83.59) 1180(86.13) 349(86.82) Alone or in nursing home 2045(17.00) 788(19.34) 1014(16.41) 190(13.87) 53(13.18) Smoking status <0.001 Never smoking 7584(63.06) 2730(67.01) 3752(60.07) 846(61.75) 256(63.68) Former smoking 1716(14.27) 511(12.54) 923(14.93) 220(16.06) 62(15.42) Drinking status <0.001 Never drinking 7988(66.42) 2832(69.51) 4021(65.05) 872(63.65) 263(65.42) Current drinking 2753(22.89) 829(20.35) 1491(24.12) 340(24.82) 93(23.13) Former drinking 1286(10.69) 413(10.14) 669(10.82) 158(11.53) 46(11.44) Regular exercise <0.001 Yes 4342(36.10) 1182(29.01) 2326(37.63) 647(47.23) 187(46.52) No 7685(63.90) 2892(70.99) 3855(62.37) 723(52.77) 215(53.48) Vegetables intake <0.001 Always 7295(60.68) 2304(56.58) 3807(61.61) 913(66.64) 271(67.58) Not always 4727(39.32) 1768(43.42) 2372(38.39) 457(33.36) 130(32.42) Fruits intake <0.001 Always 1374(11.43) 294(7.22) 739(11.96) 252(18.39) 89(22.19) Not always 10651(88.57) 3780(92.78) 5441(88.04) 1118(81.61) 312(77.81) Duration of follow-up, months, median (25th, 75th percentile) 70(37-109) 54(36-94) 71(37-110) 74(38-111) 71(37-111) <0.001 Note: All participants were classified as underweight ((body mass index (BMI) <18.5kg/m2), normal (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2). Note: All participants were classified as underweight ((body mass index (BMI) <18.5kg/m2), normal (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2). * Data are expressed as counts (percentages), except for age and duration of follow-up. Table 2. Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval; Baseline BMI was classified as underweight (BMI <18.5kg/m2), normal (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2). Tables Page 16/21 Cognitive impairment Model Categorical Events Participants HR (95%CI) P value Model 1a Large weight loss 966 2912 1.77 (1.61, 1.95) <0.001 Small weight loss 325 1511 1.01 (0.89, 1.15) 0.834 Stable weight status 783 3592 Reference - Small weight gain 266 1186 1.04 (0.90, 1.19) 0.587 Large weight gain 746 2826 1.29 (1.16, 1.42) <0.001 Model 2b Large weight loss 966 2912 1.38 (1.26, 1.52) <0.001 Small weight loss 325 1511 1.00 (0.88, 1.14) 1 Stable weight status 783 3592 Reference - Small weight gain 266 1186 1.09 (0.94, 1.25) 0.249 Large weight gain 966 2912 1.38 (1.26, 1.52) 0.073 Model 3c Large weight loss 966 2912 1.42 (1.29, 1.56) <0.001 Small weight loss 325 1511 1.01 (0.89, 1.15) 0.909 Stable weight status 783 3592 Reference - Small weight gain 266 1186 1.08 (0.94, 1.24) 0.312 Large weight gain 746 2826 1.04 (0.94, 1.16) 0.414 Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval; BMI change was classified as large weight loss (BMI change <-10%), small weight loss (-10%≤ BMI change <-5%), stable weight status (-5%≤ BMI change <5%), small weight gain (5%≤ BMI change <10%) and large weight gain (BMI change>10%). a: Univariable analysis b: Age and sex were adjusted c: Type of residence, marital status, education, living arrangement, smoking status, drinking status, regular exercise, vegetables, fruit intake and baseline BMI group were additionally adjusted. Tables The association between baseline BMI and the incidence of cognitive impairment Page 15/21 Cognitive impairment Model Categorical Events Participants HR (95%CI) P value Model 1a Underweight 1323 4074 1.55 (1.44, 1.67) <0.001 Normal 1464 6181 Reference - Overweight 224 1370 0.63 (0.55, 0.73) <0.001 Obese 75 402 0.77 (0.61, 0.97) 0.025 Model 2b Underweight 1323 4074 1.05 (0.98, 1.14) 0.177 Normal 1464 6181 Reference - Overweight 224 1370 0.84 (0.73, 0.96) 0.013 Obese 75 402 1.01 (0.80, 1.27) 0.937 Model 3c Underweight 1323 4074 1.02 (0.94, 1.10) 0.648 Normal 1464 6181 Reference - Overweight 224 1370 0.86 (0.75, 0.99) 0.036 Obese 75 402 1.01 (0.80, 1.28) 0.906 Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval; Baseline BMI was classified as underweight (BMI <18.5kg/m2), normal (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) an obesity (BMI ≥28 kg/m2). Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval; Baseline BMI was classified as underweight (BMI <18.5kg/m2), normal (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) a obesity (BMI ≥28 kg/m2). a: Univariable analysis b: Age and sex were adjusted c: Type of residence, marital status, education, living arrangement, smoking status, drinking status, regular exercise, vegetables and fruit intake were additionally adjusted. Table 3. The association between BMI change and the incidence of cognitive impairment. Figures Page 17/21 Page 17/21 Figure 1 Figure 1 Flow chart of the study population. Note: CLHLS: Chinese Longitudinal Health Longevity Study. Flow chart of the study population. Note: CLHLS: Chinese Longitudinal Health Longevity Study. Page 18/21 Page 18/21 Figure 2 Restricted cubic spline for the association of baseline BMI and BMI change with cognitive impairment. A. Baseline BMI and cognitive impairment: reference point is the median value of baseline BMI (19.84 kg/m2) , after adjusting for age, sex, type of residence, marital status, education, living arrangement, smoking status, drinking status, regular exercise, vegetables and fruit intake. B. BMI change and cognitive impairment: reference point is the median value of BMI change (0%), with knots placed at 10th, 50th and 90th percentiles, after adjusting for age, sex, type of residence, marital status, education, living arrangement, smoking status, drinking status, regular exercise, vegetables, fruit intake and baseline BMI. Note: Hazard ratios are indicated by solid lines and 95% confidence intervals by dashed lines, with knots placed at 10th, 50th and 90th percentiles. Page 19/21 Page 20/21 Figure 3 Subgroup analysis for the association of baseline BMI and BMI change with cognitive impairment. A. Subgroup analysis by age and gender for baseline BMI; B. Subgroup analysis by age and gender for BMI change. Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval. Adjusted covariates include age, sex, education background, marital status, smoking status, alcohol drinking, fruit intake, vegetables intake, type of residence, regular exercise and living arrangement. Baseline BMI group was additionally adjusted for results of BMI change. Baseline BMI was classified as underweight (BMI <18.5kg/m2), normal Figure 3 Subgroup analysis for the association of baseline BMI and BMI change with cognitive impairment. A. Subgroup analysis by age and gender for baseline BMI; B. Subgroup analysis by age and gender for BMI change. Note: BMI: body mass index; HR: hazard ratio; CI: confidence interval. Adjusted covariates include age, sex, education background, marital status, smoking status, alcohol drinking, fruit intake, vegetables intake, type of residence, regular exercise and living arrangement. Baseline BMI group was additionally adjusted for results of BMI change. Baseline BMI was classified as underweight (BMI <18.5kg/m2), normal Page 20/21 (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2); BMI change was classified as large weight loss (BMI change <-10%), small weight loss (-10%≤ BMI change <-5%), stable weight status (-5%≤ BMI change <5%), small weight gain (5%≤ BMI change <10%) and large weight gain (BMI change>10%). (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2); BMI change was classified as large weight loss (BMI change <-10%), small weight loss (-10%≤ BMI change <-5%), stable weight status (-5%≤ BMI change <5%), small weight gain (5%≤ BMI change <10%) and large weight gain (BMI change>10%). (18.5≤ BMI <24.0 kg/m2), overweight (24.0≤ BMI <28.0 kg/m2) and obesity (BMI ≥28 kg/m2); BMI change was classified as large weight loss (BMI change <-10%), small weight loss (-10%≤ BMI change <-5%), stable weight status (-5%≤ BMI change <5%), small weight gain (5%≤ BMI change <10%) and large weight gain (BMI change>10%). SupplementaryMaterials.docx Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. Page 21/21
W2062570494.txt	https://downloads.hindawi.com/journals/acp/2000/673017.pdf	fr		New Approaches to Molecular Profiling of Tissue Samples	Analytical cellular pathology	2,000	cc-by	3,663	1 Mini review New approaches to molecular profiling of tissue samples Carolyn J.M. Best a , John W. Gillespie a,c , Chad R. Englert a , Jennifer I. Swalwell a , John Pfeifer b , David B. Krizman c,d , Emanuel F. Petricoin e , Lance A. Liotta d and Michael R. Emmert-Buck a,∗ a Pathogenetics Unit, Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA b Center for Information Technology, National Institutes of Health, Bethesda, MD, USA c Cancer Genome Anatomy Project, Advanced Technology Center, National Cancer Institute, Gaithersburg, MD, USA d Office of the Chief, Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA e Division of Cytokine Biology, Center for Biologics and Research, Food and Drug Administration, Bethesda, MD, USA Received 31 March 2000 Accepted 4 April 2000 Figures on http://www.esacp.org/acp/2000/20-1/best.htm. The past decade has witnessed several technological advances that are having a significant impact on the field of molecular pathology. For example, widespread use of the polymerase chain reaction (PCR) now provides for routine amplification of DNA and mRNA from small amounts of template, exponentially less than was previously needed, allowing for more thorough analyses of nucleic acids. In parallel, several * Correspondencing author: Michael R. Emmert-Buck, M.D., Ph.D., Pathogenetics Unit, Laboratory of Pathology, National Cancer Institute, Building 10, Room 2A33, 9000 Rockville Pike, Bethesda, MD 20892, USA. Tel.: +1 301 496 2912; Fax: +1 301 594 7582; E-mail: mbuck@helix.nih.gov. Analytical Cellular Pathology 20 (2000) 1–6 ISSN 0921-8912 / $8.00  2000, IOS Press. All rights reserved high-throughput platforms have been developed that are capable of performing global measurements of cell transcriptomes and proteomes [1,2,7,10,11,13,26,28, 29,37,38]. These methods have the potential to identify individual genes and pathways that are important in both normal cellular physiology and pathology, as well as to determine the patterns of expression that mediate cellular behavior. For studies of tissue samples it is often desirable to selectively procure specific normal or diseased cells from a complex tissue milieu. Studies of this type are being advanced by several new tissue microdissection methods. Taken together, these technical advances have the potential to greatly improve our understanding of the molecular profiles that underlie normal cellular physiology and disease processes in a variety of biological systems. Our laboratory utilizes a tissue microdissectionbased approach to perform molecular profiling studies of human cancer. The importance of tissue microdissection to molecular pathology research is evident by the fact that therapies targeted to specific cellular abnormalities can be developed only when the responsible molecular alterations are precisely identified and understood. To obtain a clear representation of normal physiology and disease, one must be able to specifically isolate and examine aberrant cells. Tissue microdissection provides for the procurement of specific cell types from a specimen. For example, a pure population of tumor cells can be analyzed without any interference from neighboring non-tumor cells. Additionally, investigators can recover select sub-populations of cells such as those of pre-malignant lesions that cannot be studied in bulk tissue specimens. Several methods of tissue microdissection have been described in the literature [4,5,15,25,30]. The most important parameters are speed, precision, and avoidance of contamination, and any method that achieves these to the satisfaction of the dissector is adequate. Our group developed a laser capture microdissection (LCM) system at the 2 C.J.M. Best et al. / New approaches to molecular profiling of tissue samples National Cancer Institute, which greatly increases the speed and efficiency of tissue microdissection [6,12]. The system works by placing an ultrathin, transparent, thermoplastic film on top of a routinely prepared tissue section and activating the film with a pulse from a focused laser beam. The laser pulse is very brief (approximately 5 ms) and the membrane is activated at 90◦ C, resulting in only a brief thermal transient in the tissue. Once the laser has been fired, which can be done at various laser diameters, the activated film adheres tightly to the underlying cell or cells, which are then selectively procured from the tissue section when the film is removed. The laser capture is performed while the investigator observes the tissue through the microscope, thus precise and selective transfer is ensured. In addition, images of tissue and cells can be captured before, during, and after microdissection, which is critical in maintaining an accurate record of each dissection and correlating histopathology with subsequent molecular results. After microdissection is complete, the cells are extracted from the film using an appropriate buffer and the biomolecules recovered for analysis. Laser capture microdissection is applicable to many types of molecular pathology research, and its development is now providing access to populations of cells that were previously difficult to study. For example, LCM has recently been used to differentiate the genetic profiles of small (less than or equal to 6 mm) and large (7–30 mm) foci of hepatocellular carcinoma, enabling the investigators to show that tumor progression correlates with tumor expansion in this tissue [32]. LCM has also been used to differentiate the morphologically distinct but interwoven compartments of B-cell lymphomas. Whereas DNA from whole tissue failed to reveal evidence for biclonality, an LCM-based study revealed multiple independent clonal rearrangements in tumors, suggesting distinct clonal origins [20]. In addition, the power of LCM is well suited for use in conjunction with immunostaining [19]. In identifying p53 point mutations, Tam and colleagues recently demonstrated that there is only a weak concordance between immunostaining and molecular analysis of whole tumors, presumably due to the fact that tumor staining is highly heterogeneous. When LCM was employed to isolate specific p53-stained tumor cells, the concordance of p53 mutation detection between immunostaining and molecular analysis improved fourfold [36]. Importantly, LCM has proven useful in concert with high-throughput molecular analyses, as Erlander and colleagues used LCM to procure individual neuronal subtypes from rodent brains and study gene expression patterns by cDNA microarrays [24]. In our laboratory, tissue microdissection was utilized to assist in the identification and cloning of the gene for multiple endocrine neoplasia type 1 by allowing procurement and allelic deletion analysis of multiple small neuroendocrine tumors from patients in affected kindreds [8,14]. Currently, we are using a tissue microdissection-based approach in an effort to begin elucidating the molecular events that underlie prostate cancer as a part of the Cancer Genome Anatomy Project (CGAP) of the National Cancer Institute [33, 34]. All of the data associated with CGAP are made immediately available to the research community via the Internet (http://www.ncbi.nlm.nih.gov//ncicgap/). Through molecular profiling of microdissected normal prostate tissue, prostatic intraepithelial neoplasia (PIN), and invasive carcinoma of the prostate (see Fig. 1), there exists a unique opportunity to examine the nature and sequence of genetic alterations that occur during tumor progression. Therefore, we constructed representative cDNA libraries from 12 microdissected prostate samples that included a spectrum of normal and neoplastic phenotypes [21]. The libraries were subjected to EST sequencing and the data analyzed by a variety of statistical tests. These data have been used for a number of studies including construction of a prostate epithelial unigene set, identification of prostate-unique genes, and comparison of gene expression profiles that occur during tumor progression [16]. Tissue microdissection also expands the opportunity for disease gene hunting. For example, prostate cancer shows a high rate of allelic loss on chromosome band 8p21. Thus, this region of the genome may contain a tumor suppressor gene (TSG) that is important in the development of prostate neoplasia. Since the use of tissue microdissection and PCR analysis of loss of heterozygosity allows for definitive scoring of allelic loss, we have been able to determine a minimal gene interval on chromosome band 8p21 that is likely to harbor the responsible TSG [35]. Microdissection has also been critical in allowing us to examine allelic loss patterns in PIN, the putative precursor lesion of prostate cancer. PIN exists in the prostate as discrete microscopic foci generally found in association with tumor. Previous studies in our group showed that PIN exhibits a high level of allelic loss on chromosome band 8p21, and, in fact, the minimal gene interval in our recent study was defined by a case of PIN [17,35]. Although the power of tissue microdissection has become quite clear in research involving the molecular analysis of nucleic acids, there is currently less data C.J.M. Best et al. / New approaches to molecular profiling of tissue samples 3 Fig. 1. Molecular profiling of LCM-derived prostate tissue. Following microdissection of normal epithelium, PIN (shown above), and invasive carcinoma of the prostate, RNA was isolated and subjected to RT-PCR. Linkers were attached to the double-stranded cDNA, which was then amplified by PCR to generate cDNA libraries. Libraries were cloned via UDG vectors and the clones subjected to sequencing. The sequence data was filtered and subsequently entered into dbEST. The flow of reagents and information essentially followed that initially designed by the I.M.A.G.E. consortium [23]. This approach allows for identification of transcripts specifically expressed in cells of a distinct origin and tumorigenic stage. available regarding its efficacy in proteomics and the determination of protein profiles from microdissected samples. In one study, Banks and colleagues recently demonstrated that LCM-based analysis was useful in the study of protein patterns in the normal cervix [3]. In addition, our laboratory has been assessing the utility of LCM for studying protein profiles using three analysis methods: two-dimensional gel electrophoresis (2D-PAGE), surface enhanced laser desorption ionization (SELDI) and quantitative LCM [13,27,31]. Results to date indicate that proteomic studies of microdissected cell samples are a powerful method to examine protein alterations that occur during the evolution of a disease [13,15,31]. For example, initial 2DPAGE-based experiments comparing dissected normal prostate epithelium and patient-matched tumors of three cases demonstrated that 98% of the observed proteins were identical between the cell types. However, eight proteins were discovered that showed an identical change (up- or down-regulation) in each tumor. Determination of the identity of these proteins is underway. Surface Enhanced Laser Desorption Ionization (SELDI) is a new technology that utilizes matrixassisted laser desorption and time-of-flight analysis to study proteins [22]. Protein profiles can be generated from lysates of 1500 LCM-procured epithelial cells and proteins as small as 2,000 kD can be reliably detected. Applied to prostate cancer progression, we have found that SELDI protein profiles differ reproducibly among benign and malignant prostate epithelial cells [27]. Quantitative LCM is a newly developed method that permits determination of protein levels from microdissected cell samples using an automated sandwich chemiluminescent immunoassay. Initial studies have demonstrated the utility of this approach by precisely measuring prostate specific antigen (PSA) from microdissected prostate cells [31]. Lastly, our laboratory is currently developing two new strategies for molecular profiling of human tumorigenesis. The first is a novel analytical method that utilizes a layered array of capture membranes to perform high-throughput DNA, mRNA, or protein measurements from biological samples [18]. The method works by transferring cell or tissue samples through a series of individual capture layers, each linked to a separate antibody or DNA sequence. As the samples traverse the membrane set, each targeted protein or mRNA is specifically captured by the layer containing its antibody or complementary DNA sequence. The two-dimensional relationship of the cell populations is maintained during the transfer process, thereby producing a molecular profile of each cell type present. The second new strategy is a three-dimensional approach to the study of prostate cancer [9]. As shown in Fig. 2, this approach allows investigators to query gene expression and proteomic data and view these data in 4 C.J.M. Best et al. / New approaches to molecular profiling of tissue samples Fig. 2. Three-dimensional molecular analysis approach to the study of prostate cancer. Panel A shows an overview of a whole prostate showing the number, extent and anatomic locations of tumors at various levels of the gland. In this patient, the prostate has two tumors present, T1 and T2. The transverse view of block E (panel B) demonstrates there are two areas of high-grade cancer, T1A and T1B. In addition, there are three areas of prostatic intraepithelial neoplasia (PIN), one near the tumor (P1) and two distant (P2 and P3). There is also an area of low-grade cancer with nearby lymphocytes (L1) and abundant histologically normal epithelium (N1–4). The hyperlinked box (P2) links to a representative hematoxylin and eosin stained, high-resolution image of the region of study (panel C). The hyperlinked cube links to a serial sectioning study that has been performed on this microstructure (panel D). Panel E shows an example of a global gene expression study that compared PIN and prostate cancer using sequence analysis of cDNA libraries. Several ribosomal protein genes were expressed at significantly higher levels in cancer than PIN, including RPL11 and RPL39. C.J.M. Best et al. / New approaches to molecular profiling of tissue samples concert with the relevant histopathology. A 3D reconstruction approach allows both the physical and molecular relationship of all entities to be determined, e.g., which dysplastic lesions are developing into tumor and how the gene expression profiles change as this process unfolds. A second advantage of the 3D analytical approach is the opportunity to study the complete spectrum of tumor progression in proper context. Ultimately, the goal of this approach is a better understanding of prostate tumorigenesis at a molecular level that leads to improved diagnosis and treatment for patients, and the ability to generate data from microdissected cell samples is an integral and invaluable part of this endeavor. References [1] A. Abbott, DNA chips intensify the sequence search, Nature 379(6564) (1996), 392. [2] N.L. Anderson and N.G. Anderson, Proteome and proteomics: new technologies, new concepts, and new words, Electrophoresis 19 (1998), 1853–1861. [3] R.E. Banks, M.J. Dunn, M.A. Forbes, A. Stanley, D. Pappin, T. Naven, M. Gough, P. Harnden and P.J. Selby, The potential use of laser capture microdissection to selectively obtain distinct populations of cells for proteomic analysis – preliminary findings, Electrophoresis 20(4–5) (1999), 689–700. [4] I. Becker, K. Becker, M.H. Rohrl and H. Hofler, Laser-assisted preparation of single cells from stained histological slides for gene analysis, Histochem. Cell Biol. 108 (1997), 447–451. [5] M. Bohm, I. Wieland, K. Schutze and H. Rubben, Non-contact laser microdissection of membrane-mounted native tissue, Am. J. Pathol. 151 (1997), 63–67. [6] R.F. Bonner, M.R. Emmert-Buck, K.A. Cole, T. Pohida, R.F. Chuaqui and S.R.L. Goldstein, and L.A. Liotta, Laser capture microdissection: molecular analysis of tissue, Science 278 (1997), 1481–1483. [7] J.E. Celis, M. Ostergaard, N.A. Jensen, I. Gromova, H.H. Rasmussen and P. Gromov, Human and mouse proteomic databases: novel resources in the protein universe, FEBS Lett. 430 (1998), 64–72. [8] S.C. Chandrasekharappa, S.C. Guru, P. Manickam, S.E. Olufemi, F.S. Collins, M.R. Emmert-Buck, L.V. Debelenko, Z. Zhuang, I.A. Lubensky, L.A. Liotta, J.S. Crabtree, Y. Wang, B.A. Roe, J. Weisemann, M.S. Boguski, S.K. Agarwal, M.B. Kester, Y.S.H. Kim, C.Q. Dong, A.M. Spiegel, A.L. Burns and S.J. Marx, Positional cloning of the gene for multiple endocrine neoplasia-type 1, Science 276(5311) (1997), 404–407. [9] K.A. Cole, D.B. Krizman and M.R. Emmert-Buck, The genetics of cancer – a 3D model, Nature Genet. 21(1 Suppl.) (1999), 38–41. [10] J. DeRisi, L. Penland, P.O. Brown, M.L. Bittner, P.S. Meltzer, M. Ray, Y. Chen, Y.A. Su and J.M. Trent, Use of a cDNA microarray to analyse gene expression patterns in human cancer, Nature Genet. 14 (1996), 457–460. 5 [11] A. Dove, Proteomics: translating genomics into products?, Nat. Biotechnol. 17 (1999), 233–236. [12] M.R. Emmert-Buck, R.F. Bonner, P.D. Smith, R.F. Chuaqui, S.R. Goldstein, Z. Zhuang, R.A. Weiss and L.A. Liotta, Laser capture microdissection, Science 274 (1996), 998–1001. [13] M.R. Emmert-Buck, J.W. Gillespie, C.P. Pawletz, D.K. Ornstein, V. Basrur, E. Appella, Q.-H. Wang, J. Huang, P. Taylor and E.F. Petricoin, III, A strategic approach for proteomic analysis of human tumors, Mol. Carcinog. 27 (2000), 1–8. [14] M.R. Emmert-Buck, I.A. Lubensky, Q. Dong, P. Manickam, S.C. Guru, M.B. Kester, S.E. Olufemi, S. Agarwal, A.L. Burns, A.M. Spiegel, F.S. Collins, S.J. Marx, Z. Zhuang, L.A. Liotta, S.C. Chandrasekharappa and L.V. Debelenko, Localization of the multiple endocrine neoplasia type I (MEN1) gene based on tumor loss of heterozygosity analysis, Cancer Res. 57 (1997), 1855–1858. [15] M.R. Emmert-Buck, M.J. Roth, Z. Zhuang, E. Campo, J. Rozhin, S.B. F, L.A. Liotta and W.G. Stetler-Stevenson, Increased gelatinase A (MMP-2) and cathepsin B activity in invasive tumor regions of human colon cancer samples, Am. J. Pathol. 145 (1994), 1285–1290. [16] M.R. Emmert-Buck, R.L. Strausberg, D.B. Krizman, M.F. Bonaldo, R.F. Bonner, D.G. Bostwick, M.R. Brown, K.H. Buetow, R. Chuaqui, K.A. Cole, P.F. Duray, C.R. Englert, J.W. Gillespie, S. Greenhut, G. Grouse, L.W. Hillier, K.S. Katz, R.D. Klausner, V. Kuznetzov, A.E. Lash, G. Lennon, W.M. Linehan, L.A. Liotta, M.A. Marra, P.J. Munson, D.K. Ornstein, V. Prabhu, C. Prange, G.D. Schuler, M.B. Soares, C.M. Tolstoshev, C.D. Vocke and R.H. Waterston, Molecular profiling of clinical tissue specimens: feasibility and applications, Am. J. Pathol. (2000), in press. [17] M.R. Emmert-Buck, C.D. Vocke, R.O. Pozzatti, P.H. Duray, S.B. Jennings, C.D. Florence, Z. Zhuang, D.G. Bostwick, L.A. Liotta and W.M. Linehan, Allelic loss on chromosome 8p12–21 in microdissected prostatic intraepithelial neoplasia (PIN), Cancer Res. 55(14) (1995), 2959–2962. [18] C.R. Englert, G.V. Baibakov and M.R. Emmert-Buck, Layered expression scanning: rapid molecular profiling of tumor samples, Cancer Res. 60 (2000), 1526–1530. [19] F. Fend, M.R. Emmert-Buck, R. Chuaqui, K. Cole, J. Lee, L.A. Liotta and M. Raffeld, Immuno-LCM: laser capture microdissection of immunostained frozen sections for mRNA analysis, Am. J. Pathol. 154(1) (1999), 61–66. [20] F. Fend, L. Quintanilla-Martinez, S. Kumar, M.W. Beaty, L. Blum, L. Sorbara, E.S. Jaffe and M. Raffeld, Composite lowgrade B-cell lymphomas with two immunophenotypically distinct cell populations are true biclonal lymphomas. A molecular analysis using laser capture microdissection, Am. J. Pathol. 154(6) (1999), 1857–1866. [21] D.B. Krizman, R.F. Chuaqui, P.S. Meltzer, J.M. Trent, P.H. Duray, W.M. Linehan, L.A. Liotta and M.R. Emmert-Buck, Construction of a representative cDNA library from prostatic intraepithelial neoplasia (PIN), Cancer Res. 56(23) (1996), 5380–5383. [22] H. Kuwata, T.T. Yip, C.L. Yip, M. Tomita and T.W. Hutchens, Bactericidal domain of lactoferrin: detection, quantitation, and characterization of lactoferricin in serum by SELDI affinity mass spectrometry, Biochem. Biophys. Res. Commun. 245 (1998), 764–773. 6 C.J.M. Best et al. / New approaches to molecular profiling of tissue samples [23] G. Lennon, C. Auffray, M. Polymeropolous and M.B. Soares, The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression, Genomics 33 (1996), 151– 152. [24] L. Luo, R.C. Salunga, H. Guo, A. Bittner, K.C. Joy, J.E. Galindo, H. Xiao, K.E. Rogers, J.S. Wan, M.R. Jackson and M.G. Erlander, Gene expression profiles of laser-captured adjacent neuronal subtypes, Nature Med. 5 (1999), 112–122. [25] C.A. Moskaluk and S.E. Kern, Microdissection and polymerase chain reaction amplification of genomic DNA from histological tissue sections, Am. J. Pathol. 150 (1997), 1547–1552. [26] R. Nowak, Entering the postgenome era, Science 270 (1995), 368–371. [27] C.P. Pawletz, D. Ornstein, J.W. Gillespie, K.A. Cole, N.L. Simone, Y. Yip, T.W. Hutchens, M.R. Emmert-Buck, L.A. Liotta and E.F. Petricoin, Tissue proteomics at the microscale: a novel approach to the study of disease progression, Drug Devel. Res. (2000), in press. [28] G. Ramsay, DNA chips: state of the art, Nat. Biotechnol. 16 (1998), 40–44. [29] M. Schena, R.A. Heller, T.P. Theriault, K. Konrad, E. Lachenmeier and R.W. Davis, Microarrays: biotechnology’s discovery platform for functional genomics, Trends Biotechnol. 16 (1998), 301–306. [30] D. Shibata, D. Hawes, Z.H. Li, A.M. Hernandez, C.H. Spruck and P.W. Nichols, Specific genetic analysis of microscopic tissue after selective ultraviolet radiation fractionation and the polymerase chain reaction, Am. J. Pathol. 141(3) (1992), 539– 543. [31] N.L. Simone, A.T. Remaley, L. Charboneau, E.F. Petricoin, III, J.W. Glickman, M.R. Emmert-Buck, T.A. Fleisher and L.A. Liotta, Sensitive immunoassay of tissue cell proteins procured by laser capture microdissection, Am. J. Pathol. 156 (2000), 445– 452. [32] Y. Sirivatanauksorn, V. Sirivatanauksorn, S. Bhattacharya, B.R. Davidson, A.P. Dhillon, A.K. Kakkar, R.C. Williamson and N.R. Lemione, Evolution of genetic abnormalities in hepatocellular carcinomas demonstrated by DNA fingerprinting, J. Pathol. 189(3) (1999), 344–350. [33] R.L. Strausberg, D.C. A and K.R. D, New opportunities for uncovering the molecular basis of cancer, Nature Genet. 15 (1997), 415–416. [34] R.L. Strausberg, K.H. Buetow, M.R. Emmert-Buck and R.D. Klausner, The Cancer Genome Anatomy Project: Building an annotated gene index, Trends Genet. 16 (2000), 103– 106. [35] J.I. Swalwell, C.D. Vocke, L. Grouse, Y. Yang, T. Thomas Ried, S.H. Myers, J.W. Gillespie, E. Konieczynski, J. Lee, P.H. Duray, D.G. Bostwick, W.M. Linehan and M.R. EmmertBuck, Determination of a minimal tumor suppressor gene interval on chromosome band 8p21 in sporadic prostate cancer, Manuscript submitted. [36] A.S. Tam, J.F. Foley, T.R. Devereux, R.R. Maronpot and T.E. Massey, High frequency and heterogeneous distribution of p53 mutations in aflatoxin B1-induced mouse lung tumors, Cancer Res. 58(15) (1999), 3634–3640. [37] V. Velculescu, L. Zhang, B. Vogelstein and K. Kinzler, Serial analysis of gene expression, Science 270 (1995), 484–487. [38] L. Zhang, W. Zhou, V.E. Velculescu, S.E. Kern, R.H. Hruban, S.R. Hamilton, B. Vogelstein and K.W. Kinzler, Gene expression profiles in normal and cancer cells, Science 276(5316) (1997), 1268–1272. MEDIATORS of INFLAMMATION The Scientific World Journal Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Gastroenterology Research and Practice Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Diabetes Research Disease Markers Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 International Journal of Journal of Immunology Research Endocrinology Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 Submit your manuscripts at http://www.hindawi.com BioMed Research International PPAR Research Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 Journal of Obesity Ophthalmology Stem Cells International Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Journal of Volume 2014 Evidence-Based Complementary and Alternative Medicine Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Oncology Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Parkinson’s Disease Computational and Mathematical Methods in Medicine Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 AIDS Behavioural Neurology Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Research and Treatment Oxidative Medicine and Cellular Longevity Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014
https://openalex.org/W2063468623	https://journal-bcs.springeropen.com/track/pdf/10.1007/s13173-011-0050-6	English	null	Branch and bound algorithms for the maximum clique problem under a unified framework	Journal of the Brazilian Computer Society	2,011	cc-by	12,333	J Braz Comput Soc (2012) 18:137–151 DOI 10.1007/s13173-011-0050-6 J Braz Comput Soc (2012) 18:137–151 DOI 10.1007/s13173-011-0050-6 SI: GRAPHCLIQUES Branch and bound algorithms for the maximum clique problem under a uniﬁed framework Renato Carmo · Alexandre Züge Received: 14 November 2011 / Accepted: 23 November 2011 / Published online: 24 December 2011 © The Brazilian Computer Society 2011 bound-based schemes stand out in the literature as one of the best approaches in practice. Abstract In this paper we review branch and bound-based algorithms proposed for the exact solution of the maximum clique problem and describe them under a unifying concep- tual framework. As a proof of concept, we actually imple- mented eight of these algorithms as parametrized versions of one single general branch and bound algorithm. More often than not, these algorithms are published from an experimental standpoint, where running times for several testing benchmarks are given and commented upon, but lit- tle or no analytic results are given in support of the veriﬁed performance. On the other hand the currently available re- sults on the asymptotic behavior of algorithms for MC seem to leave a considerable gap between the worst case perfor- mance and the one actually reported by experimental results. The purpose of the present work is double folded. In the one hand, the implementation of several different algo- rithms under the same computational environment allows for a more precise assessment of their comparative perfor- mance at the experimental level. On the other hand we see the unifying conceptual framework provided by such de- scription as a valuable step toward a more ﬁne grained anal- ysis of these algorithms. In this paper we review eight of these algorithms and de- scribe them under a unifying conceptual framework. Besides surveying some of the best performing algorithms published to date, we aim to contribute with some perspective on the subject of branch and bound algorithms for MC from both, conceptual and experimental standpoints. Keywords Maximum clique · Exact solution · Branch and bound The uniﬁed framework introduced in the following sec- tions invites to an implementation in which each of the al- gorithms discussed becomes a particular variation of a gen- eral branch and bound algorithm for MC. We actually imple- mented each of them in this way and present experimental results on their performance under the same computational environment, something which is not available in the litera- ture to the best of our knowledge. R. Carmo () · A. Züge Departamento de Informática da UFPR Centro Politécnico da Universidade Federal do Paraná, Curitiba, PR, Brasil 81531-990, P.O. Box 19081 e-mail: renato.carmo.rc@gmail.com A. Züge e-mail: alexandrezuge@gmail.com 1.1 Deﬁnitions and notation A graph on n vertices can have as much as 3n/3 differ- ent maximal cliques [11]. Therefore, any algorithm which enumerates all maximal cliques of a graph on n vertices must have worst case running time of Ω(3n/3). An algo- rithm matching this lower bound with worst case running time of O(3n/3) was introduced in [19]. Given a set S and an integer k we denote by S k the set of subsets of S of size k. A graph G is a pair (V (G),E(G)) where V (G) is a ﬁnite set and E(G) ⊆ V (G) 2 . The elements of V (G) and E(G) are called vertices and edges of G, respectively. Two vertices u and v are said to be neighbors in G if {u,v} is an edge in G. The neighborhood of a vertex v in G is the set of its neighbors in G and is denoted ΓG(v). The degree of a vertex v in G is the size of its neighborhood in G. Given a set S ⊆V (G) the common neighborhood of S in G is the set of vertices in G which are neighbors to all vertices in S and is denoted Γ ∩ G(S). It will be convenient to adopt the convention that Γ ∩ G(∅) = V (G). On the other hand, ﬁnding the maximum clique of a graph does not require to actually examine all of its max- imal cliques. Along the search among the maximal cliques of the graph, some non-maximal cliques can be discarded as soon as they are identiﬁed as not contained in a clique larger than another already known. That the number of discarded cliques in such a strategy can be signiﬁcant is shown in [15] which introduces an algorithm for MC with worst case run- ning time of O(2n/3). This bound was later improved to O(20.304n) [8] and further to O(20.276n) [13]. Presently the value of this bound is set to O(20.249n) [14]. A graph G is said to be complete if E(G) = V (G) 2 . I S ⊆ V (G), the subgraph of G induced by a set S ⊆V (G) is the graph G[S] = S, S 2 ∩E(G) and G−S denotes the graph G[V (G) −S]. A clique in G is a set of vertices of G that induces a complete graph. 1These instances are publicly available from http://dimacs.rutgers.edu/ Challenges. 1.1 Deﬁnitions and notation The size of a maximum clique in G is denoted ω(G). Reports from the “experimental front”, however, sug- gest that worst case estimates do not tell the whole story. Indeed, several authors who implemented BK-based algo- rithms for MC report running times which may be surpris- ing when confronted to the best known worst case estimates (besides [5, 6, 9, 16–18] which are discussed below, see also [10, 12]). Given an integer k, a k-coloring of G is a surjective func- tion γ :V (G) →{1,...,k} satisfying γ (u) ̸= γ (v) for every {u,v} ∈E(G). The value of γ (v) is called the color of v and the integer k is called the number of colors in γ . A coloring of G is a k-coloring of G for some integer k. We note that ω(G) ≤k for any graph G and any k-coloring of G. Experimental results for MC in the literature are usually obtained using two main classes of instances, namely, A list of vertices of G is a sequence L = (v1,...,vn) of distinct vertices of G with n = \|V (G)\|. A coloring γ of G is greedy with respect to the list L if the color of each vertex vi:1 ≤i ≤n is the minimum not in {γ (vj):1 ≤j < i}. random graphs: sets of graphs generated according to the Gn,p model [2], for different values of the parameters n (number of vertices) and p (edge probability). random graphs: sets of graphs generated according to the Gn,p model [2], for different values of the parameters n (number of vertices) and p (edge probability). In the use of the notation above, we omit subscripts and superscripts whenever this can be done without ambiguity. DIMACS graphs: a set of 66 graphs from the DIMACS Sec- ond Implementation Challenge.1 Many of these algorithms are published with the focus on the benchmarking, while little (if any at all) concern is given to the analysis of the proposed algorithm. Explaining the gap between the disheartening worst case estimates and what has actually already been achieved in practice seems to be an interesting challenge. With this long term goal in mind, in the next section we reframe the BK-based algorithms in a unifying form. 1 Introduction The Maximum Clique problem (MC) is the problem of ﬁnd- ing a clique of maximum size on a given graph. There are a number of proposed algorithms for the ex- act solution of MC which are reported to effectively solve instances of practical interest (some of them of considerable size) in several domains [3, 9, 17]. Among these, branch and The text is organized as follows. In Sect. 2 we review some theoretical and experimental results on the solution of MC. In Sect. 3 we focus on a certain class of these al- gorithms, namely, branch and bound approaches based on the enumerating algorithm of Bron–Kerbosch [4]. We state Bron–Kerbosch’s algorithm in a form that highlights its main idea and then convert it into a general branch and bound algorithm from which many of the previously pro- posed algorithms can be easily derived. In Sect. 4 we focus on eight particular algorithms and discuss their implemen- tation as particular instances of the general algorithm stated A. Züge e-mail: alexandrezuge@gmail.com J Braz Comput Soc (2012) 18:137–151 138 in Sect. 3, and we present comparative experimental results from our implementation, all obtained under the same com- putational environment. In Sect. 5 we discuss some imple- mentation details and make our concluding remarks. In the Appendix, we include the unabridged version of the experi- mental results presented in Sect. 4. at general (as opposed to particular classes of) instances, branch and bound-based approaches stand out with respect to the veriﬁed performance, besides being relatively easy to implement. In a sense that will be made more precise in Sect. 3, the majority of branch and bound algorithms for MC can be seen as based on the algorithm of Bron–Kerbosch [4] for enumerating all maximal cliques of a graph. We will refer to such algorithms as BK-based algorithms. 3 A general branch and bound algorithm MAXCLIQUEBB(G) (C,S ) ←pre-process(G) 1 While S ̸= ∅ 2 (Q,K) ←pre-process-state(G, pop(S ),C) 3 While K ̸= ∅and \|C\| < \|Q\| + bound(G,Q,K) 4 v ←remove(K) 5 S ←push(G,Q,K) 6 (Q,K) ←pre-process-state(G,Q ∪{v}, 7 K ∩Γ (v),C) If \|C\| < \|Q\| 8 (C,S ) ←update(G,C,S ,Q) 9 Return post-process(G,C) 10 BK(G,Q,N) If Γ ∩ G(Q) = ∅ 1 Return {Q} 2 If Γ ∩ G(Q) −N ⊆ΓG(v) for some v ∈N 3 Return ∅ 4 v ←a vertex from Γ ∩ G(Q) −N 5 Return BK(G,Q ∪{v},N) ∪BK(G,Q,N ∪{v}) 6 BK(G,Q,N) If Γ ∩ G(Q) = ∅ 1 Return {Q} 2 If Γ ∩ G(Q) −N ⊆ΓG(v) for some v ∈N 3 Return ∅ 4 v ←a vertex from Γ ∩ G(Q) −N 5 Return BK(G,Q ∪{v},N) ∪BK(G,Q,N ∪{v}) 6 Algorithm BK makes explicit the enumeration scheme proposed in [4]. The idea of the algorithm is clear once we notice that if G is a graph, Q is a non-empty clique in G and N ⊆V (G) −Q, then BK(G,Q,N) is the set of all maximal cliques of G containing Q which do not intersect N. In Algorithm MAXCLIQUEBB we have the following. 1. pre-process(G) returns the initial values of C and S . 1. pre-process(G) returns the initial values of C and S . Algorithm BK can be converted into algorithm MAXCLIQUE below in a straightforward way as follows. 2. pre-process-state(G,Q,K,C) returns a state (Q′,K′) where Q′ −Q ⊆K −K′. Note that this includes the case (Q′,K′) = (Q,K). MAXCLIQUE(G) C ←∅ 1 S ←{(∅,V (G))} 2 While S ̸= ∅ 3 (Q,K) ←pop(S ) 4 While K ̸= ∅ 5 v ←remove(K) 6 S ←push(Q,K) 7 (Q,K) ←(Q ∪{v},K ∩Γ (v)) 8 If \|C\| < \|Q\| 9 C ←Q 10 Return C 11 MAXCLIQUE(G) C ←∅ 1 S ←{(∅,V (G))} 2 While S ̸= ∅ 3 (Q,K) ←pop(S ) 4 While K ̸= ∅ 5 v ←remove(K) 6 S ←push(Q,K) 7 (Q,K) ←(Q ∪{v},K ∩Γ (v)) 8 If \|C\| < \|Q\| 9 C ←Q 10 Return C 11 3. bound(G,Q,K) returns an integer b ≥ω(G[K]). 4. remove(K) is as in MAXCLIQUE. 5. update(G,C,S ,Q) returns the clique Q and updates stack S . 5. update(G,C,S ,Q) returns the clique Q and updates stack S . 6. post-process(G,C) returns C. In the sequel, we refer collectively to these six routines as the custom routines of Algorithm MAXCLIQUEBB. 2 Exact solution of the maximum clique problem The Maximum Clique problem (MC) is the problem of ﬁnd- ing a clique of maximum size on a given graph. More pre- cisely, an instance to MC is a graph G and a solution to in- stance G is a clique of maximum size in G. The problem is NP-hard [7] and cannot even be approximated in polyno- mial time up to a factor of \|V (G)\|1/3 [1]. Several approaches have been proposed to the exact so- lution of MC. A nice survey can be found in [3]. Concerning the actual implementation of exact solutions for MC targeted J Braz Comput Soc (2012) 18:137–151 139 3 A general branch and bound algorithm 2. The set S is a stack (“last in ﬁrst out”) data structure, implementing the recursion in Algorithm BK. 2. The set S is a stack (“last in ﬁrst out”) data structure, implementing the recursion in Algorithm BK. In this section we focus on algorithms for MC based on the enumerating algorithm of Bron–Kerbosch [4], which we call BK-based algorithms. We start by stating Bron–Kerbosch’s algorithm in a form that highlights its main idea. The Bron– Kerbosch’s algorithm is not an algorithm for MC. Rather, it solves the related problem of enumerating all maximal cliques of a given graph. We proceed to a straightforward conversion of our statement of Bron–Kerbosch algorithm into a non-recursive algorithm for MC and thence to a gen- eral branch and bound algorithm from which BK-based al- gorithms can be easily derived. 3. The statement v ←remove(K) means that some vertex is removed from set K and left in variable v. 4. Each pair (Q,K) corresponds to the pair of sets (Q,Γ ∩ G(Q) −N) in Algorithm BK. 4. Each pair (Q,K) corresponds to the pair of sets (Q,Γ ∩ G(Q) −N) in Algorithm BK. In order to convert Algorithm MAXCLIQUE into a branch and bound algorithm for MC, we add a bounding scheme which allows us to discard a pair (Q,K) from the stack S if we detect that this pair cannot possibly lead to a clique larger than C. We also add some pre and post-processing routines which will be discussed in the sequel. Let G be a graph and consider the Algorithm BK(G,Q,N) below, where Q is a clique in G and N ⊆ V (G) −Q. 4 Eight branch and bound algorithms for MC In this section we select eight BK-based algorithms and discuss them as particular variations of Algorithm MAXCLIQUEBB. The chosen algorithms, to which we will refer collectively as “MCBB algorithms”, are the following. Down to implementation level, leaving the pivoting strat- egy unspeciﬁed (or underspeciﬁed) amounts to having a piv- oting strategy where the order of the vertices is induced by the data structure representing the graph. It is worth noting that the data structure representing the graph, for its turn, is sensitive to the way the input data are organized and parsed. cp: the algorithm MAXCLIQ as described in [5]. chi: the algorithm χ as described in [6]. df: the algorithm DF as described in [6]. chi + df: the algorithm χ + DF as described in [6]. mcq: the algorithm MCQ as described in [17]. mcr: the algorithm MCR as described in [16]. mcs: the algorithm MCS as described in [18]. dyn: the algorithm MAXCLIQUEDYN as described in [9]. cp: the algorithm MAXCLIQ as described in [5]. chi: the algorithm χ as described in [6]. df: the algorithm DF as described in [6]. chi + df: the algorithm χ + DF as described in [6]. mcq: the algorithm MCQ as described in [17]. mcr: the algorithm MCR as described in [16]. mcs: the algorithm MCS as described in [18]. dyn: the algorithm MAXCLIQUEDYN as described in [9]. The tables in the following sections show the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of the other MCBB algorithms for 21 of the DIMACS instances. These instances were selected because they were the ones for which each of the MCBB algorithms ran to completion within the time limit of three hours (10800 seconds). Com- plete tables, showing these values for all of the 66 DIMACS instances, are presented in the Appendix. The diagram in Fig. 1 displays the publishing timeline for the MCBB algorithms. An arrow in the diagram means that the author of the latter work explicitly builds upon the work of the former. We actually implemented Algorithm MAXCLIQUEBB and the variations corresponding to each of the MCBB algo- rithms in order to effect a comparative experimental analysis of their performance under the same computational environ- ment. 3 A general branch and bound algorithm pre-process(G): returns the pair (G,{(∅,V (G))}), pre-process-state(G,Q,K,C): returns the pair (Q,K), bound(G,Q,K): returns the value of \|K\|, remove(K): removes (and returns) a vertex from K, update(G,C,S ,Q): returns the pair (Q,S ), post-process(G,C): returns the set C. This is the most basic variation of Algorithm MAXCLIQUEBB in the sense that each of the operations above performs trivial processing in time Θ(1). We refer to this particular variation of Algorithm MAXCLIQUEBB as the “basic algorithm”, and include it into the collective “MCBB algorithms”. 4.2 Non-trivial branching: cp Among the MCBB algorithms, cp is the ﬁrst to explore a non-trivial pivoting strategy, which is to privilege low de- gree vertices as pivots. The intuitive idea behind this strat- egy is that low degree vertices are “unlikely” to be part of a maximum clique and so should be examined and discarded as soon as possible. 4.1 The basic algorithm 4 Eight branch and bound algorithms for MC For details on this implementation as well as on the computational environment in which the experimental data presented were gathered, we refer the reader to Sect. 5. 3 A general branch and bound algorithm As the name suggests, branch and bound schemes of op- timization own their performance in great measure to the choice of the branching and the bounding steps. In the reference frame provided by Algorithm MAXCLIQUEBB, the branching step corresponds to remove(K) and the bounding step corresponds to bound(G,Q,K). The rou- tines pre-process(G), pre-process-state(G,Q,K,C) and update(G,C,S ,Q) serve the purpose of creating and up- dating data structures to aid the branching and bounding steps. We refer to these as the branching and bounding strategies, respectively. The vertex v returned by remove(K) In Algorithm MAXCLIQUE we have the following. 1. At any given point of the execution, the set C stores the maximum clique in G found up to that point of the exe- cution. 1. At any given point of the execution, the set C stores the maximum clique in G found up to that point of the exe- cution. 140 J Braz Comput Soc (2012) 18:137–151 Fig. 1 Publishing timeline of the MCBB algorithms is often called the pivot of the branching step and, accord- ingly, branching strategies are often called pivoting strate- gies in this context, emphasizing the fact that the branching strategy reduces to the strategy for choosing the pivot. It is usual to picture branch and bound schemes as the process of transversing of a tree, often referred to as the search tree. In Algorithm MAXCLIQUEBB the search tree can be viewed as one in which the nodes are the pairs (Q,K). Each branching step corresponds to choosing a pivot v ∈K and then adding the nodes (Q,K −{v}) and pre-process-state(G,Q ∪{v},K ∩Γ (v),C) as children of node (Q,K). We will call these the right and left child of the node (Q,K), respectively. Fig. 1 Publishing timeline of the MCBB algorithms Each round of the main loop of Algorithm MAXCLIQUEBB can then be described as follows. Visit the leftmost unvisited leaf of the search tree and examine the pair (Q,K) which constitutes this leaf and either attach new children to this leaf (branch) or mark it as visited (bound). The overall running time of the algorithm is then the time spent while visiting each node of the tree summed over all nodes of the tree. 4.1 The basic algorithm Each of these coloring algo- rithms work by choosing at each step the next vertex to be colored. Different colorings are obtained by changing the policy for choosing the next vertex to be colored. As a vari- ation of Algorithm MAXCLIQUEBB, Algorithm chi is the one obtained when the routine bound(G,Q,K) performs this computation and all other custom routines are as in the basic algorithm. v3 is the vertex of minimum degree in G −{v1,v2}, and so on, and remove(K) returns the lowest indexed vertex in K according to L. Moreover, if G is a “dense” graph, then pre-process-state(G,Q,K,C) recomputes this list re- stricted to G[K]. A precise deﬁnition of “dense” is not given by the authors of [5]. In our implementation of cp we do not treat differently graphs according to their density leav- ing pre-process-state(G,Q,K,C) the same as in the basic algorithm. Algorithm df explores what the author of [6] calls do- main ﬁlters. The idea is to “clean-up the graph” before each branching step. More precisely, when at node (Q,K) in the search tree, every vertex of degree less than \|C\| −\|Q\| in G[K] is removed from K, since such vertices cannot be part of a clique larger than C in G. Then, every vertex of degree \|K\|−1 in (the resulting) G[K] is moved to Q. This amounts to collapsing several branching steps in one, because when a vertex of degree \|K\| −1 is chosen as the pivot at node (Q,K), both children of this node will be equal or, equiv- alently, this node will have only one child. Algorithm df is, thus, the variation of Algorithm MAXCLIQUEBB obtained when the routine pre-process-state(G,Q,K,C) performs the above described processing, and all other custom rou- tines are as in basic. In Table 1 we show the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of cp on the selected DI- MACS instances. These values show clearly that the branch- ing strategy in cp substantially reduces the number of search tree nodes with respect to the basic algorithm. Indeed, ex- cepting the extreme cases on both ends of the sample, we see that the number of search tree nodes for basic is 1.65 larger than the number of search tree nodes for cp (with a standard deviation of 0.43). 4.1 The basic algorithm Viewed as a variation of Algorithm MAXCLIQUEBB, cp is obtained when pre-process(G) orders V (G) into a list L = (v1,...,vn) where v1 is the vertex of minimum de- gree in G, v2 is the vertex of minimum degree in G −{v1}, We begin by pointing out that Algorithm MAXCLIQUE cor- responds to the variation of Algorithm MAXCLIQUEBB where J Braz Comput Soc (2012) 18:137–151 141 Table 1 Number of search tree nodes and execution time for basic and cp Instance # Search tree nodes Time (s) basic cp basic cp brock200_2 655863 454569 8.06 5.48 brock200_3 9025785 3610663 108.50 45.76 brock200_4 29071199 15103689 375.90 191.39 c-fat200-1 847 515 0.03 0.03 c-fat200-2 2373 2405 0.06 0.06 c-fat500-1 2419 1187 0.11 0.19 c-fat500-2 9775 4227 0.23 0.22 c-fat500-5 812265 32215829 22.61 772.37 hamming6-2 62239 61479 1.10 1.05 hamming6-4 1949 1657 0.01 0.02 hamming8-4 50630977 34762759 668.19 459.77 johnson8-2-4 721 571 0.01 0.00 johnson8-4-4 160555 83555 1.95 1.02 johnson16-2-4 29897217 23319913 277.13 217.54 keller4 13347251 8588277 148.24 96.35 MANN_a9 5194193 1247265 54.00 13.41 p_hat300-1 127783 96531 1.62 1.15 p_hat300-2 218320753 10212863 3333.80 164.79 p_hat500-1 1192121 776357 15.19 9.94 p_hat700-1 4405187 3067767 58.24 38.60 sanr200_0.7 122475879 54201495 1597.00 714.08 A word seems to be due with respect to instance c-fat500-5, whose data seem to be so discrepant with re- spect to the other c-fat* instances. The reason for this is the fact that the maximum clique found by the algorithm for this instance is formed by vertices of maximum degree in the graph and, as explained above, such vertices are the last ones to be considered by the algorithm. The reader will ob- serve the same phenomenon for other instances of the “c-fat* family” in the Appendix. In all three of them, pre-process(G) orders V (G) into a list L = (v1,...,vn). No details are given by the author about this ordering, as its only purpose is to keep the ver- tices easily indexable so that remove(K) returns the lowest indexed vertex in K according to L. In our implementation we use the trivial pre-process(G) of the basic algorithm. In Algorithm chi the bounding step is performed by com- puting four colorings of G[K] and returning the number of colors of the one which uses the least number of col- ors. These four colorings are obtained using two different greedy coloring algorithms. 4.1 The basic algorithm J Braz Comput Soc (2012) 18:137–151 142 Table 2 Number of search tree nodes for basic, chi, df and chi + df Instance basic df chi chi + df brock200_2 655863 55853 8451 10515 brock200_3 9025785 606435 82721 61131 brock200_4 29071199 2475297 127653 160001 c-fat200-1 847 19 33 19 c-fat200-2 2373 65 145 11 c-fat500-1 2419 67 97 27 c-fat500-2 9775 163 169 27 c-fat500-5 812265 269 397 27 hamming6-2 62239 10577 151 131 hamming6-4 1949 343 237 227 hamming8-4 50630977 3167847 20413 21049 johnson8-2-4 721 137 59 77 johnson8-4-4 160555 14125 253 219 johnson16-2-4 29897217 5228329 1379359 1614925 keller4 13347251 1187295 63271 65453 MANN_a9 5194193 772459 1347 1323 p_hat300-1 127783 10025 4879 4397 p_hat300-2 218320753 9604633 78333 74219 p_hat500-1 1192121 85221 31901 30143 p_hat700-1 4405187 285863 90371 93259 sanr200_0.7 122475879 9945037 397241 386685 Table 3 Running time (s) for basic, chi, df and chi + df Instance basic df chi chi + df brock200_2 8.06 13.40 170.98 189.73 brock200_3 108.50 152.46 1262.22 1065.25 brock200_4 375.90 701.33 2803.73 3302.96 c-fat200-1 0.03 0.04 0.62 0.96 c-fat200-2 0.06 0.08 1.59 0.67 c-fat500-1 0.11 0.29 5.96 6.21 c-fat500-2 0.23 0.44 9.70 8.92 c-fat500-5 22.61 1.43 24.68 18.85 hamming6-2 1.10 6.26 0.72 0.99 hamming6-4 0.01 0.06 0.81 0.72 hamming8-4 668.19 973.37 1172.32 1244.42 johnson8-2-4 0.01 0.01 0.11 0.14 johnson8-4-4 1.95 2.93 2.87 2.99 johnson16-2-4 277.13 363.29 4068.67 4460.17 keller4 148.24 192.10 958.69 984.13 MANN_a9 54.00 66.18 9.19 9.35 p_hat300-1 1.62 2.58 103.97 103.22 p_hat300-2 3333.80 5108.47 4018.11 3976.74 p_hat500-1 15.19 26.16 1088.94 1022.10 p_hat700-1 58.24 95.80 3922.90 3934.76 sanr200_0.7 1597.00 3003.31 8052.30 8458.74 Table 2 Number of search tree nodes for basic, chi, df and chi + df Instance basic df chi chi + df brock200_2 655863 55853 8451 10515 brock200_3 9025785 606435 82721 61131 brock200_4 29071199 2475297 127653 160001 c-fat200-1 847 19 33 19 c-fat200-2 2373 65 145 11 c-fat500-1 2419 67 97 27 c-fat500-2 9775 163 169 27 c-fat500-5 812265 269 397 27 hamming6-2 62239 10577 151 131 hamming6-4 1949 343 237 227 hamming8-4 50630977 3167847 20413 21049 johnson8-2-4 721 137 59 77 johnson8-4-4 160555 14125 253 219 johnson16-2-4 29897217 5228329 1379359 1614925 keller4 13347251 1187295 63271 65453 MANN_a9 5194193 772459 1347 1323 p_hat300-1 127783 10025 4879 4397 p_hat300-2 218320753 9604633 78333 74219 p_hat500-1 1192121 85221 31901 30143 p_hat700-1 4405187 285863 90371 93259 sanr200_0.7 122475879 9945037 397241 386685 As we shall see in the sequel, there are ways to beneﬁt from the idea of coloring-based bounding at a less demand- ing cost in execution time. 4.1 The basic algorithm Algorithm chi + df is simply the union of the bounding strategy of Algorithm chi and the branching strategy of Al- gorithm df into a single algorithm. Table 2 shows side by side the number of search tree nodes in the execution of the same instances as in Table 1. 4.4 Color-based branching and bounding: mcq, mcr, mcs and dyn As was the case in the discussion of algorithm cp, the val- ues for the “c-fat* family” are remarkable when compared to the others. Indeed the author of [6] himself observes that these instances “are quite easy to solve using domain ﬁlter- ing”. The algorithms discussed above (among others) showed ex- perimental evidence that the use of non-trivial branching and bounding strategies were worth the processing time over- head per node of the search tree. Table 3 shows the running times corresponding to the val- ues in Table 2. Here the improvement, when there is im- provement, is much less pronounced than what one sees when comparing the number of search tree nodes. The con- clusion is that, differently from what we observed about cp, the overhead incurred in the more elaborate strategies of branching and bounding is not negligible and can be such as to actually increase the overall running time when compared to the basic algorithm. This is not completely surprising if we keep in mind the complexity of the processing which takes place at each branching and bounding step. Besides, as discussed in Sect. 5, it may be the case that some of the im- plementation details contribute to make this overhead even more pronounced. The algorithm mcq goes one step further, using the idea of coloring the vertices of the graph not only as a bounding strategy but also as a branching strategy. The idea is that the coloring of the graph will not only provide a bound on the size of the maximum clique, but also serve as an ordering of the vertices guiding the choice of the pivot at each branching step. This is done as follows. First, the vertices of the graph are initially ordered into a list L = (v1,...,vn) where the ver- tices are in non-increasing degree order. This corresponds to the routine pre-process(G) in Algorithm MAXCLIQUEBB. At each branching step, the routine pre-process-state(G, Q,K,C) computes a greedy coloring of G[K] with respect to the list L and then returns (Q,K). 4.1 The basic algorithm The running times, for their turn, show clearly that the overhead imposed by the pre- processing does not compromise the overall performance of the algorithm. 4.1 The basic algorithm Ties are broken in such a way that if vi−1 and vi have the same degree, then the sum of the degrees of the neigh- bors of vi in G−{vi+1,...,vn} is less or equal than the sum of the degrees of the neighbors of vi−1 in G −{vi,...,vn}. Everything else proceeds as in mcq. Algorithm mcs further improves pre-process(G) by modifying the adjacency matrix representing the graph so that the order of the neighbors of each vertex is compatible with the order of the vertices in the initial list L. In Algorithm dyn the branching step keeps track of the sizes of the set Q′ in each node (Q′,K′) visited in the search tree. More precisely, the routine pre-process-state(G,Q, K,C) computes the number of nodes (Q′,K′) of the search tree visited so far satisfying \|Q′\| ≤\|Q\|. Whenever this num- ber is less than 2.5% of the number of search tree nodes vis- ited so far, the vertices in K are ordered into a list as the one computed in mcq and a coloring of G[K] with the same properties as the one in the pre-processing of mcq is recom- puted. This coloring, however, has the additional property of keeping the relative order of all vertices of color less or equal \|C\| −\|Q\|. The routine remove(K) returns a vertex of maximum color from this coloring and the routine bound(G,Q,K) just returns the number of colors in this coloring. All other custom routines are as in the basic algorithm. There are two noteworthy differences in the use of color- ing in mcq with respect to algorithms chi and chi + df. First, only one coloring is computed at each branching step, in- stead of the four in algorithms chi and chi + df. Second, the way the pivot is chosen at the branching step is such that the coloring does not need to be recomputed for the right child of each node. As the pivot v is chosen so that it is col- ored with the maximum color in K and has a neighbor of each color less than its own color, this coloring restricted to K −{v} preserves these properties. Table 5 shows the number of search tree nodes in the ex- ecution of basic, mcr, mcs and dyn. These values show that each of these algorithms effectively reduces the size of the search tree with respect to mcq. 4.1 The basic algorithm This coloring, however, has the additional property of keeping the relative order of all vertices of color less or equal \|C\| −\|Q\|. Table 5 shows the number of search tree nodes in the ex- ecution of basic, mcr, mcs and dyn. These values show that each of these algorithms effectively reduces the size of the search tree with respect to mcq. On the other hand, differently from the observed with respect to algorithms df, chi and chi + df, the processing overhead incurred because of the non-trivial branching and bounding computation needed at each step does not cancel out the gain obtained by reducing the size of the search tree, as can be seen in Table 6, which shows the running times corresponding to the entries in Table 5. Moreover, the size of the search tree for mcs is further reduced with respect to mcr for most instances. When this is not the case and both search trees have the same size, then both consume about the same processing time. We note that it is to be expected that the number of nodes Table 4 Number of search tree nodes and execution time for basic and mcq Instance # Search tree nodes Time (s) basic mcq basic mcq brock200_2 655863 8649 8.06 0.94 brock200_3 9025785 33359 108.50 5.21 brock200_4 29071199 127691 375.90 19.26 c-fat200-1 847 437 0.03 0.03 c-fat200-2 2373 487 0.06 0.06 c-fat500-1 2419 1045 0.11 0.12 c-fat500-2 9775 1093 0.23 0.19 c-fat500-5 812265 1245 22.61 0.79 hamming6-2 62239 127 1.10 0.03 hamming6-4 1949 257 0.01 0.01 hamming8-4 50630977 29529 668.19 7.85 johnson8-2-4 721 103 0.01 0.00 johnson8-4-4 160555 433 1.95 0.06 johnson16-2-4 29897217 707187 277.13 26.75 keller4 13347251 31597 148.24 3.64 MANN_a9 5194193 191 54.00 0.02 p_hat300-1 127783 4301 1.62 0.32 p_hat300-2 218320753 19535 3333.80 6.23 p_hat500-1 1192121 22457 15.19 2.62 p_hat700-1 4405187 68671 58.24 8.83 sanr200_0.7 122475879 365773 1597.00 60.72 The only difference between mcq and mcr is in the pre- processing of the graph, before the actual branch and bound is executed. In Algorithm mcr the set V (G) is ordered into a list L = (v1,...,vn) where vn is the vertex of minimum de- gree in G, vn−1 is the vertex of minimum degree in G−{vn}, vn−2 is the vertex of minimum degree in G−{vn−1,vn}, and so on. 4.1 The basic algorithm This is done as follows. First, the vertices of the graph are initially ordered into a list L = (v1,...,vn) where the ver- tices are in non-increasing degree order. This corresponds to the routine pre-process(G) in Algorithm MAXCLIQUEBB. At each branching step, the routine pre-process-state(G, Q,K,C) computes a greedy coloring of G[K] with respect to the list L and then returns (Q,K). J Braz Comput Soc (2012) 18:137–151 143 Table 4 Number of search tree nodes and execution time for basic and mcq Instance # Search tree nodes Time (s) basic mcq basic mcq brock200_2 655863 8649 8.06 0.94 brock200_3 9025785 33359 108.50 5.21 brock200_4 29071199 127691 375.90 19.26 c-fat200-1 847 437 0.03 0.03 c-fat200-2 2373 487 0.06 0.06 c-fat500-1 2419 1045 0.11 0.12 c-fat500-2 9775 1093 0.23 0.19 c-fat500-5 812265 1245 22.61 0.79 hamming6-2 62239 127 1.10 0.03 hamming6-4 1949 257 0.01 0.01 hamming8-4 50630977 29529 668.19 7.85 johnson8-2-4 721 103 0.01 0.00 johnson8-4-4 160555 433 1.95 0.06 johnson16-2-4 29897217 707187 277.13 26.75 keller4 13347251 31597 148.24 3.64 MANN_a9 5194193 191 54.00 0.02 p_hat300-1 127783 4301 1.62 0.32 p_hat300-2 218320753 19535 3333.80 6.23 p_hat500-1 1192121 22457 15.19 2.62 p_hat700-1 4405187 68671 58.24 8.83 sanr200_0.7 122475879 365773 1597.00 60.72 The routine remove(K) returns a vertex of maximum color from this coloring and the routine bound(G,Q,K) just returns the number of colors in this coloring. All other custom routines are as in the basic algorithm. There are two noteworthy differences in the use of color- ing in mcq with respect to algorithms chi and chi + df. First, only one coloring is computed at each branching step, in- stead of the four in algorithms chi and chi + df. Second, the way the pivot is chosen at the branching step is such that the coloring does not need to be recomputed for the right child of each node. As the pivot v is chosen so that it is col- ored with the maximum color in K and has a neighbor of each color less than its own color, this coloring restricted to K −{v} preserves these properties. Table 4 shows the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of mcq. 4.1 The basic algorithm Differently from what happens with algorithms chi, df and chi + df, the reduc- ing of the number of search tree nodes reﬂects directly in the running time, and the differences are even more pronounced. The algorithms mcr, mcs and dyn are improvements on mcq. The ﬁrst two are proposed by some of the same authors The only difference between mcq and mcr is in the pre- processing of the graph, before the actual branch and bound is executed. In Algorithm mcr the set V (G) is ordered into a list L = (v1,...,vn) where vn is the vertex of minimum de- gree in G, vn−1 is the vertex of minimum degree in G−{vn}, vn−2 is the vertex of minimum degree in G−{vn−1,vn}, and so on. Ties are broken in such a way that if vi−1 and vi have the same degree, then the sum of the degrees of the neigh- bors of vi in G−{vi+1,...,vn} is less or equal than the sum of the degrees of the neighbors of vi−1 in G −{vi,...,vn}. Everything else proceeds as in mcq. Algorithm mcs further improves pre-process(G) by modifying the adjacency matrix representing the graph so that the order of the neighbors of each vertex is compatible with the order of the vertices in the initial list L. Besides that, pre-process-state(G,Q,K,C) computes a coloring γ of G[K] similar to the one computed by mcq, but with the following difference. If a vertex v has neighbors colored with all the lowest \|C\|−\|Q\| colors, then a color with only one neighbor u is searched for and the algorithm tries to recolor v and u so that both vertices stay on the lowest \|C\| −\|Q\| colors. At the end, post-process(G,C) undoes the modiﬁcation in the adjacency matrix made in pre-process(G). In Algorithm dyn the branching step keeps track of the sizes of the set Q′ in each node (Q′,K′) visited in the search tree. More precisely, the routine pre-process-state(G,Q, K,C) computes the number of nodes (Q′,K′) of the search tree visited so far satisfying \|Q′\| ≤\|Q\|. Whenever this num- ber is less than 2.5% of the number of search tree nodes vis- ited so far, the vertices in K are ordered into a list as the one computed in mcq and a coloring of G[K] with the same properties as the one in the pre-processing of mcq is recom- puted. 4.1 The basic algorithm As a consequence our presentation and discussion of experi- mental results does not aim at more than giving the reader a “sense of proportion” between the different approaches to the problem and their practical impact as well as a modest historical perspective on the evolution of ideas toward the solution of MC. Our experimental data are in agreement with Table 5 Number of search tree nodes for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_2 655863 7825 4937 7183 brock200_3 9025785 30619 15927 25821 brock200_4 29071199 142077 64799 94351 c-fat200-1 847 377 377 437 c-fat200-2 2373 353 353 487 c-fat500-1 2419 973 973 1045 c-fat500-2 9775 949 949 1093 c-fat500-5 812265 873 873 1245 hamming6-2 62239 135 129 127 hamming6-4 1949 153 153 257 hamming8-4 50630977 16899 9707 25765 johnson8-2-4 721 59 59 103 johnson8-4-4 160555 247 171 393 johnson16-2-4 29897217 533629 474647 1264845 keller4 13347251 23995 11749 17625 MANN_a9 5194193 75 57 191 p_hat300-1 127783 4209 3091 4267 p_hat300-2 218320753 15677 6957 15039 p_hat500-1 1192121 21635 16079 21821 p_hat700-1 4405187 65849 43751 56061 sanr200_0.7 122475879 325415 132903 211053 stances. This is because mcs can be viewed as an improve- ment over mcr which “may or may not be triggered” depend- ing on the instance. The same is true for dyn with respect to mcq Table 6 Running time (s) for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_2 8.06 0.90 0.74 1.35 brock200_3 108.50 4.88 3.23 5.67 brock200_4 375.90 20.35 12.47 19.47 c-fat200-1 0.03 0.05 0.05 0.04 c-fat200-2 0.06 0.07 0.07 0.06 c-fat500-1 0.11 0.25 0.26 0.14 c-fat500-2 0.23 0.33 0.34 0.24 c-fat500-5 22.61 0.80 0.77 0.95 hamming6-2 1.10 0.02 0.03 0.03 hamming6-4 0.01 0.00 0.01 0.01 hamming8-4 668.19 4.64 3.03 9.47 johnson8-2-4 0.01 0.01 0.00 0.01 johnson8-4-4 1.95 0.03 0.03 0.06 johnson16-2-4 277.13 19.84 18.28 64.38 keller4 148.24 2.82 1.73 3.36 MANN_a9 54.00 0.00 0.01 0.02 p_hat300-1 1.62 0.37 0.34 0.39 p_hat300-2 3333.80 4.92 2.76 5.63 p_hat500-1 15.19 2.73 2.21 2.99 p_hat700-1 58.24 8.86 6.97 13.81 sanr200_0.7 1597.00 53.26 29.67 50.90 language. This is to note that the running times in the ex- perimental data presented are to be taken mainly as a qual- itative assessment. Indeed, implementations geared toward stances. This is because mcs can be viewed as an improve- ment over mcr which “may or may not be triggered” depend- ing on the instance. 4.1 The basic algorithm The same is true for dyn with respect to mcq. language. This is to note that the running times in the ex- perimental data presented are to be taken mainly as a qual- itative assessment. Indeed, implementations geared toward maximum efﬁciency are reported to run the same algorithms discussed here in time orders of magnitude smaller, even in computational environments less powerful than the one available here. For the purpose of this work, however, the language is very suitable in the ﬂexibility it offers to the pro- grammer. The conclusion is that the use of coloring as an aid for both, the branching and the bounding strategies yields al- gorithms that perform better than the ones discussed in the previous sections. This is because, (i) the coloring algorithm is simple; (ii) the choice of the pivot based on the coloring is a good pivoting strategy and (iii) colorings can be inherited and reused along some branches of the search tree. The primary goal of this work is the introduction of a uniﬁed conceptual framework which may serve as a starting point toward more ﬁne grained analysis of such algorithms, and thus contributing to closing the gap between known the- oretical bounds and observed experimental performance. As a consequence our presentation and discussion of experi- mental results does not aim at more than giving the reader a “sense of proportion” between the different approaches to the problem and their practical impact as well as a modest historical perspective on the evolution of ideas toward the solution of MC. Our experimental data are in agreement with the experimental data available from the papers where each of the MCBB algorithms was originally proposed. While the values are of course different, their qualitative relationship is the same. The reader interested in a more focused discus- 4.1 The basic algorithm On the other hand, differently from the observed with respect to algorithms df, chi and chi + df, the processing overhead incurred because of the non-trivial branching and bounding computation needed at each step does not cancel out the gain obtained by reducing the size of the search tree, as can be seen in Table 6, which shows the running times corresponding to the entries in Table 5. Table 4 shows the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of mcq. Differently from what happens with algorithms chi, df and chi + df, the reduc- ing of the number of search tree nodes reﬂects directly in the running time, and the differences are even more pronounced. Moreover, the size of the search tree for mcs is further reduced with respect to mcr for most instances. When this is not the case and both search trees have the same size, then both consume about the same processing time. The algorithms mcr, mcs and dyn are improvements on mcq. The ﬁrst two are proposed by some of the same authors of mcq. We note that it is to be expected that the number of nodes in the search tree of mcr and mcs is the same for some in- J Braz Comput Soc (2012) 18:137–151 144 J C p S ( 0 ) 8 3 5 Table 5 Number of search tree nodes for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_2 655863 7825 4937 7183 brock200_3 9025785 30619 15927 25821 brock200_4 29071199 142077 64799 94351 c-fat200-1 847 377 377 437 c-fat200-2 2373 353 353 487 c-fat500-1 2419 973 973 1045 c-fat500-2 9775 949 949 1093 c-fat500-5 812265 873 873 1245 hamming6-2 62239 135 129 127 hamming6-4 1949 153 153 257 hamming8-4 50630977 16899 9707 25765 johnson8-2-4 721 59 59 103 johnson8-4-4 160555 247 171 393 johnson16-2-4 29897217 533629 474647 1264845 keller4 13347251 23995 11749 17625 MANN_a9 5194193 75 57 191 p_hat300-1 127783 4209 3091 4267 p_hat300-2 218320753 15677 6957 15039 p_hat500-1 1192121 21635 16079 21821 p_hat700-1 4405187 65849 43751 56061 sanr200_0.7 122475879 325415 132903 211053 stances. This is because mcs can be viewed as an improve- ment over mcr which “may or may not be triggered” depend- ing on the instance. The same is true for dyn with respect to mcq. 4.1 The basic algorithm The conclusion is that the use of coloring as an aid for both, the branching and the bounding strategies yields al- gorithms that perform better than the ones discussed in the previous sections. This is because, (i) the coloring algorithm is simple; (ii) the choice of the pivot based on the coloring is a good pivoting strategy and (iii) colorings can be inherited and reused along some branches of the search tree. 5 Implementation details and concluding remarks The implementation of Algorithm MAXCLIQUEBB to which our experimental results refer was made in the Python language using the framework provided by the module Net- workX. The running times were taken in a GNU/Linux sys- tem running on a 2.4 GHz, 32-core machine with 128 GB of memory. Each process was allowed to run for a maxi- Table 6 Running time (s) for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_2 8.06 0.90 0.74 1.35 brock200_3 108.50 4.88 3.23 5.67 brock200_4 375.90 20.35 12.47 19.47 c-fat200-1 0.03 0.05 0.05 0.04 c-fat200-2 0.06 0.07 0.07 0.06 c-fat500-1 0.11 0.25 0.26 0.14 c-fat500-2 0.23 0.33 0.34 0.24 c-fat500-5 22.61 0.80 0.77 0.95 hamming6-2 1.10 0.02 0.03 0.03 hamming6-4 0.01 0.00 0.01 0.01 hamming8-4 668.19 4.64 3.03 9.47 johnson8-2-4 0.01 0.01 0.00 0.01 johnson8-4-4 1.95 0.03 0.03 0.06 johnson16-2-4 277.13 19.84 18.28 64.38 keller4 148.24 2.82 1.73 3.36 MANN_a9 54.00 0.00 0.01 0.02 p_hat300-1 1.62 0.37 0.34 0.39 p_hat300-2 3333.80 4.92 2.76 5.63 p_hat500-1 15.19 2.73 2.21 2.99 p_hat700-1 58.24 8.86 6.97 13.81 sanr200_0.7 1597.00 53.26 29.67 50.90 language. This is to note that the running times in the ex- perimental data presented are to be taken mainly as a qual- itative assessment. Indeed, implementations geared toward maximum efﬁciency are reported to run the same algorithms discussed here in time orders of magnitude smaller, even in computational environments less powerful than the one available here. For the purpose of this work, however, the language is very suitable in the ﬂexibility it offers to the pro- grammer. The primary goal of this work is the introduction of a uniﬁed conceptual framework which may serve as a starting point toward more ﬁne grained analysis of such algorithms, and thus contributing to closing the gap between known the- oretical bounds and observed experimental performance. 5 Implementation details and concluding remarks In h tree mined nd the onted is the in the idged of df, of Ta- in the is the on of ersion s, the cond Table 7 Instances from the DIMACS second implementation chal- lenge G \|V (G)\| \|E(G)\| ω(G) brock200_1 200 14834 21 brock200_2 200 9876 12 brock200_3 200 12048 15 brock200_4 200 13089 17 brock400_1 400 59723 27 brock400_2 400 59786 29 brock400_3 400 59681 31 brock400_4 400 59765 33 brock800_1 800 207505 23 brock800_2 800 208166 24 brock800_3 800 207333 25 brock800_4 800 207643 26 c-fat200-1 200 1534 12 c-fat200-2 200 3235 24 c-fat200-5 200 8473 58 c-fat500-1 500 4459 14 c-fat500-2 500 9139 26 c-fat500-5 500 23191 64 c-fat500-10 500 46627 126 hamming6-2 64 1824 32 hamming6-4 64 704 4 hamming8-2 256 31616 128 hamming8-4 256 20864 16 hamming10-2 1024 518656 512 hamming10-4 1024 434176 ≥32 johnson8-2-4 28 210 4 johnson8-4-4 70 1855 14 johnson16-2-4 120 5460 8 johnson32-2-4 496 107880 ≥16 keller4 171 9435 11 keller5 776 225990 27 keller6 3361 4619898 ≥59 MANN_a9 45 918 16 MANN_a27 378 70551 126 MANN_a45 1035 533115 345 MANN_a81 3321 5506380 ≥1100 p_hat300-1 300 10933 8 p_hat300-2 300 21928 25 p_hat300-3 300 33390 36 p_hat500-1 500 31569 9 p_hat500-2 500 62946 36 p_hat500-3 500 93800 50 p_hat700-1 700 60999 11 p_hat700-2 700 121728 44 p_hat700-3 700 183010 62 p hat1000-1 1000 122253 10 Table 7 Instances from the DIMACS second implementation chal- lenge sion of the experimental results is encouraged to refer to the respective references. Finally, even under the limitations above pointed, the ex- perimental data presented here seem to be enough to select mcs and dyn as the best algorithms for MC among the MCBB algorithms. Indeed, mcs is the one algorithm which shows more consistently the lowest values for the number of nodes in the search tree and processing time. By a different count, however, dyn is the one algorithm which solved the largest number of instances in the prescribed time, as shown in the Appendix. Acknowledgements R. Carmo was supported by CNPq Proc. 308692/2008-0. A. Züge was partially supported by CAPES. 5 Implementation details and concluding remarks The implementation of Algorithm MAXCLIQUEBB to which our experimental results refer was made in the Python language using the framework provided by the module Net- workX. The running times were taken in a GNU/Linux sys- tem running on a 2.4 GHz, 32-core machine with 128 GB of memory. Each process was allowed to run for a maxi- mum processing time of three hours (10800 seconds). The machine was not dedicated to these experiments. Since Python is an interpreted language, its programs will run substantially slower than the equivalent in a compiled J Braz Comput Soc (2012) 18:137–151 145 ) 18:137–151 145 tal results is encouraged to refer to the r the limitations above pointed, the ex- nted here seem to be enough to select st algorithms for MC among the MCBB mcs is the one algorithm which shows lowest values for the number of nodes processing time. By a different count, ne algorithm which solved the largest n the prescribed time, as shown in the Carmo was supported by CNPq Proc. was partially supported by CAPES. ged experimental data resent the experimental data obtained B algorithms on each of the DIMACS e names, the number of vertices, the the size of a maximum clique for each ere the size of the maximum clique is indicates that the exact value is only . number of search tree nodes and the e for the basic algorithm confronted for executions of cp. In other words, ged version of Table 1. owing tables, an entry marked “–” in eans that the corresponding execution e hours of processing (“time-out”). In presented as the number of search tree ood as the number of nodes examined processing. number of search tree nodes and the e for the basic algorithm confronted for executions of mcq. Table 9 is the Table 4. e number of search tree nodes in the nd chi + df. Table 10 is the unabridged e running times in the execution of df, e 11 is the unabridged version of Ta- e number of search tree nodes in the mcr, mcs and dyn. Table 12 is the Table 5. he running times in the execution of yn. Table 13 is the unabridged version or each of the MCBB algorithms, the arch tree nodes examined per second. 5 Implementation details and concluding remarks Table 7 Instances from the DIMACS second implementation chal- lenge G \|V (G)\| \|E(G)\| ω(G) brock200_1 200 14834 21 brock200_2 200 9876 12 brock200_3 200 12048 15 brock200_4 200 13089 17 brock400_1 400 59723 27 brock400_2 400 59786 29 brock400_3 400 59681 31 brock400_4 400 59765 33 brock800_1 800 207505 23 brock800_2 800 208166 24 brock800_3 800 207333 25 brock800_4 800 207643 26 c-fat200-1 200 1534 12 c-fat200-2 200 3235 24 c-fat200-5 200 8473 58 c-fat500-1 500 4459 14 c-fat500-2 500 9139 26 c-fat500-5 500 23191 64 c-fat500-10 500 46627 126 hamming6-2 64 1824 32 hamming6-4 64 704 4 hamming8-2 256 31616 128 hamming8-4 256 20864 16 hamming10-2 1024 518656 512 hamming10-4 1024 434176 ≥32 johnson8-2-4 28 210 4 johnson8-4-4 70 1855 14 johnson16-2-4 120 5460 8 johnson32-2-4 496 107880 ≥16 keller4 171 9435 11 keller5 776 225990 27 keller6 3361 4619898 ≥59 MANN_a9 45 918 16 MANN_a27 378 70551 126 MANN_a45 1035 533115 345 MANN_a81 3321 5506380 ≥1100 p_hat300-1 300 10933 8 p_hat300-2 300 21928 25 p_hat300-3 300 33390 36 p_hat500-1 500 31569 9 p_hat500-2 500 62946 36 p_hat500-3 500 93800 50 p_hat700-1 700 60999 11 p_hat700-2 700 121728 44 p_hat700-3 700 183010 62 p_hat1000-1 1000 122253 10 to the he ex- select MCBB shows nodes count, argest in the Proc. tained MACS s, the r each que is s only nd the onted words, “–” in cution t”). Appendix: Unabridged experimental data In this appendix we present the experimental data obtained for each of the MCBB algorithms on each of the DIMACS instances. Table 7 shows the names, the number of vertices, the number of edges and the size of a maximum clique for each of the instances. Where the size of the maximum clique is shown as “≥k”, this indicates that the exact value is only known to be at least k. Table 8 shows the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of cp. In other words, Table 8 is the unabridged version of Table 1. In this and the following tables, an entry marked “–” in the column “time” means that the corresponding execution was aborted after three hours of processing (“time-out”). In such cases, the value presented as the number of search tree nodes is to be understood as the number of nodes examined up to that point in the processing. Table 9 shows the number of search tree nodes and the overall execution time for the basic algorithm confronted with the same values for executions of mcq. Table 9 is the unabridged version of Table 4. Table 10 shows the number of search tree nodes in the execution of df, chi and chi + df. Table 10 is the unabridged version of Table 2. Table 11 shows the running times in the execution of df, chi and chi + df. Table 11 is the unabridged version of Ta- ble 3. Table 12 shows the number of search tree nodes in the execution of basic, mcr, mcs and dyn. Table 12 is the unabridged version of Table 5. Table 14 shows, for each of the MCBB algorithms, the average number of search tree nodes examined per second. Appendix: Unabridged experimental data J Braz Comput Soc (2012) 18:137–151 146 146 J Braz Comput Soc (2012) 18:137–151 Table 7 (Continued) G \|V (G)\| \|E(G)\| ω(G) p_hat1000-2 1000 244799 46 p_hat1000-3 1000 371746 68 p_hat1500-1 1500 284923 12 p_hat1500-2 1500 568960 65 p_hat1500-3 1500 847244 ≥56 san200_0.7_1 200 13930 30 san200_0.7_2 200 13930 18 san200_0.9_1 200 17910 70 san200_0.9_2 200 17910 60 san200_0.9_3 200 17910 44 san400_0.5_1 400 39900 13 san400_0.7_1 400 55860 40 san400_0.7_2 400 55860 30 san400_0.7_3 400 55860 22 san400_0.9_1 400 71820 100 san1000 1000 250500 15 sanr200_0.7 200 13868 18 sanr200_0.9 200 17863 42 sanr400_0.5 400 39984 13 sanr400_0.7 400 55869 21 Table 8 Number of search tree nodes and execution time for basic and cp Instance # Search tree nodes Time (s) basic cp basic cp brock200_1 760463351 251690269 10268.47 3502.69 brock200_2 655863 454569 8.06 5.48 brock200_3 9025785 3610663 108.50 45.76 brock200_4 29071199 15103689 375.90 191.39 brock400_1 781878631 781941457 – – brock400_2 779146349 755850137 – – brock400_3 744501333 715649169 – – brock400_4 750841656 732504582 – – brock800_1 773878067 776717005 – – brock800_2 755264235 768571657 – – brock800_3 745732597 763313958 – – brock800_4 782408235 760254294 – – c-fat200-1 847 515 0.03 0.03 c-fat200-2 2373 2405 0.06 0.06 c-fat200-5 197957 458462948 5.28 – c-fat500-1 2419 1187 0.11 0.19 c-fat500-2 9775 4227 0.23 0.22 c-fat500-5 812265 32215829 22.61 772.37 c-fat500-10 222003341 298372854 – – hamming6-2 62239 61479 1.10 1.05 hamming6-4 1949 1657 0.01 0.02 Table 8 (Continued) Instance # Search tree nodes Time (s) basic cp basic cp hamming8-2 479589100 352543447 – – hamming8-4 50630977 34762759 668.19 459.77 hamming10-2 477939325 402027810 – – hamming10-4 799627648 778850521 – – johnson8-2-4 721 571 0.01 0.00 johnson8-4-4 160555 83555 1.95 1.02 johnson16-2-4 29897217 23319913 277.13 217.54 johnson32-2-4 1124521577 1104958353 – – keller4 13347251 8588277 148.24 96.35 keller5 788436558 834071065 – – keller6 814607071 779038755 – – MANN_a9 5194193 1247265 54.00 13.41 MANN_a27 1039572148 470689919 – – MANN_a45 896111360 293734044 – – MANN_A81 649082104 166411314 – – p_hat300-1 127783 96531 1.62 1.15 p_hat300-2 218320753 10212863 3333.80 164.79 p_hat300-3 694764638 632946226 – – p_hat500-1 1192121 776357 15.19 9.94 p_hat500-2 666419883 601945642 – – p_hat500-3 627863675 603266182 – – p_hat700-1 4405187 3067767 58.24 38.60 p_hat700-2 674310972 570261174 – – p_hat700-3 575688852 578871234 – – p_hat1000-1 25084283 16177539 327.15 204.89 p_hat1000-2 651274462 589428559 – – p_hat1000-3 609819818 546143432 – – p_hat1500-1 232906339 139624485 3473.86 2151.07 p_hat1500-2 577460315 556716701 – – p_hat1500-3 609271041 542561068 – – san200_0.7_1 1199510748 1285897159 – – san200_0.7_2 1162303528 1040102008 – – san200_0.9_1 712616501 857295178 – – san200_0.9_2 580762880 708916119 – – san200_0.9_3 788583024 702721963 – – san400_0.5_1 1057751310 1072592630 – – san400_0.7_1 1159107699 1093205907 – – san400_0.7_2 1091551627 1289478422 – – san400_0.7_3 1011697883 1037906517 – – san400_0.9_1 868327833 818891530 – – san1000 894222972 912859488 – – sanr200_0.7 122475879 54201495 1597.00 714.08 sanr200_0.9 708848293 626288059 – – sanr400_0.5 53900541 36182977 680.22 457.73 sanr400_0.7 791421504 749744630 – – Table 7 (Continued) Table 8 (Continued) J Braz Comput Soc (2012) 18:137–151 147 Table 9 Number of search tree nodes and execution time for basic and mcq Instance # Search tree nodes Time (s) basic mcq basic mcq brock200_1 760463351 934071 10268.47 194.72 brock200_2 655863 8649 8.06 0.94 brock200_3 9025785 33359 108.50 5.21 brock200_4 29071199 127691 375.90 19.26 brock400_1 781878631 46078511 – – brock400_2 779146349 48936847 – – brock400_3 744501333 40736005 – – brock400_4 750841656 36592912 – – brock800_1 773878067 49759405 – – brock800_2 755264235 50923139 – – brock800_3 745732597 49157697 – – brock800_4 782408235 50647778 – – c-fat200-1 847 437 0.03 0.03 c-fat200-2 2373 487 0.06 0.06 c-fat200-5 197957 621 5.28 0.31 c-fat500-1 2419 1045 0.11 0.12 c-fat500-2 9775 1093 0.23 0.19 c-fat500-5 812265 1245 22.61 0.79 c-fat500-10 222003341 1493 – 19.16 hamming6-2 62239 127 1.10 0.03 hamming6-4 1949 257 0.01 0.01 hamming8-2 479589100 519 – 1.79 hamming8-4 50630977 29529 668.19 7.85 hamming10-2 477939325 2433 – 178.36 hamming10-4 799627648 38911002 – – johnson8-2-4 721 103 0.01 0.00 johnson8-4-4 160555 433 1.95 0.06 johnson16-2-4 29897217 707187 277.13 26.75 johnson32-2-4 1124521577 286689139 – – keller4 13347251 31597 148.24 3.64 keller5 788436558 21418630 – – keller6 814607071 18188831 – – MANN_a9 5194193 191 54.00 0.02 MANN_a27 1039572148 86157 – 919.57 MANN_a45 896111360 129263 – – MANN_A81 649082104 13107 – – p_hat300-1 127783 4301 1.62 0.32 p_hat300-2 218320753 19535 3333.80 6.23 p_hat300-3 694764638 5122669 – 2508.94 p_hat500-1 1192121 22457 15.19 2.62 p_hat500-2 666419883 1084401 – 566.91 p_hat500-3 627863675 16142732 – – p_hat700-1 4405187 68671 58.24 8.83 p_hat700-2 674310972 8798091 – 7296.62 p_hat700-3 575688852 11194735 – – Table 9 (Continued) Instance # Search tree nodes Time (s) basic mcq basic mcq p_hat1000-1 25084283 406251 327.15 50.43 p_hat1000-2 651274462 13673923 – – p_hat1000-3 609819818 16712232 – – p_hat1500-1 232906339 2567285 3473.86 442.68 p_hat1500-2 577460315 13699588 – – p_hat1500-3 609271041 14540018 – – san200_0.7_1 1199510748 3551 – 1.53 san200_0.7_2 1162303528 3531 – 0.72 san200_0.9_1 712616501 453207 – 214.31 san200_0.9_2 580762880 2166287 – 1315.95 san200_0.9_3 788583024 1854567 – 1450.32 san400_0.5_1 1057751310 5601 – 3.00 san400_0.7_1 1159107699 174471 – 136.18 san400_0.7_2 1091551627 134609 – 119.02 san400_0.7_3 1011697883 895391 – 383.13 san400_0.9_1 868327833 83553153 – – san1000 894222972 511507 – 1093.09 sanr200_0.7 122475879 365773 1597.00 60.72 sanr200_0.9 708848293 19847341 – – sanr400_0.5 53900541 610909 680.22 75.05 sanr400_0.7 791421504 48382248 – – Table 10 Number of search tree nodes for basic, chi, df and chi + df Instance basic df chi chi + df brock200_1 760463351 32933229 625739 485819 brock200_2 655863 55853 8451 10515 brock200_3 9025785 606435 82721 61131 brock200_4 29071199 2475297 127653 160001 brock400_1 781878631 35364477 419494 426556 brock400_2 779146349 33170085 461625 408386 brock400_3 744501333 34509290 404194 424708 brock400_4 750841656 32556199 421270 376092 brock800_1 773878067 37311293 295164 286255 brock800_2 755264235 36583861 296180 285318 brock800_3 745732597 32631672 291028 276027 brock800_4 782408235 39054988 310928 306408 c-fat200-1 847 19 33 19 c-fat200-2 2373 65 145 11 c-fat200-5 197957 127 229 127 c-fat500-1 2419 67 97 27 c-fat500-2 9775 163 169 27 c-fat500-5 812265 269 397 27 c-fat500-10 222003341 299 503 11 hamming6-2 62239 10577 151 131 Table 9 (Continued) Table 9 (Continued) J Braz Comput Soc (2012) 18:137–151 148 148 J Braz Comput Soc (2012) 18:137 151 Table 10 (Continued) Instance basic df chi chi + df hamming6-4 1949 343 237 227 hamming8-2 479589100 7705015 3973 1841 hamming8-4 50630977 3167847 20413 21049 hamming10-2 477939325 24989216 15841 15880 hamming10-4 799627648 34293269 147962 100629 johnson8-2-4 721 137 59 77 johnson8-4-4 160555 14125 253 219 johnson16-2-4 29897217 5228329 1379359 1614925 johnson32-2-4 1124521577 163763590 2158150 2374215 keller4 13347251 1187295 63271 65453 keller5 788436558 36800640 152609 158865 keller6 814607071 31790821 933 790 MANN_a9 5194193 772459 1347 1323 MANN_a27 1039572148 276077527 45063 73981 MANN_a45 896111360 247828444 2585 2332 MANN_A81 649082104 134484013 18 15 p_hat300-1 127783 10025 4879 4397 p_hat300-2 218320753 9604633 78333 74219 p_hat300-3 694764638 25445287 257474 267335 p_hat500-1 1192121 85221 31901 30143 p_hat500-2 666419883 19861603 168191 156518 p_hat500-3 627863675 26495825 148979 197406 p_hat700-1 4405187 285863 90371 93259 p_hat700-2 674310972 19704073 150286 124285 p_hat700-3 575688852 21204840 82150 115173 p_hat1000-1 25084283 1688731 273593 287859 p_hat1000-2 651274462 22031073 132626 142135 p_hat1000-3 609819818 20670410 90971 100185 p_hat1500-1 232906339 13776857 179966 186962 p_hat1500-2 577460315 21490388 60226 68721 p_hat1500-3 609271041 21103903 69038 44671 san200_0.7_1 1199510748 230612837 35959 999 san200_0.7_2 1162303528 139202470 20719 20585 san200_0.9_1 712616501 95545497 160730 189882 san200_0.9_2 580762880 22944996 58131 184741 san200_0.9_3 788583024 30003374 413304 371226 san400_0.5_1 1057751310 83927533 4029 3081 san400_0.7_1 1159107699 293738048 108630 173528 san400_0.7_2 1091551627 191488689 240064 200279 san400_0.7_3 1011697883 173777224 438690 243408 san400_0.9_1 868327833 106086764 104439 398631 san1000 894222972 31153745 20004 20194 sanr200_0.7 122475879 9945037 397241 386685 sanr200_0.9 708848293 33205279 360044 339146 sanr400_0.5 53900541 4214565 558533 544368 sanr400_0.7 791421504 36721656 432554 454927 Table 11 Running time (s) for basic, chi, df and chi + df Instance basic df chi chi + df brock200_1 10268.47 – – – brock200_2 8.06 13.40 170.98 189.73 brock200_3 108.50 152.46 1262.22 1065.25 brock200_4 375.90 701.33 2803.73 3302.96 brock400_1 – – – – brock400_2 – – – – brock400_3 – – – – brock400_4 – – – – brock800_1 – – – – brock800_2 – – – – brock800_3 – – – – brock800_4 – – – – c-fat200-1 0.03 0.04 0.62 0.96 c-fat200-2 0.06 0.08 1.59 0.67 c-fat200-5 5.28 0.43 22.05 21.26 c-fat500-1 0.11 0.29 5.96 6.21 c-fat500-2 0.23 0.44 9.70 8.92 c-fat500-5 22.61 1.43 24.68 18.85 c-fat500-10 – 11.91 94.35 10.00 hamming6-2 1.10 6.26 0.72 0.99 hamming6-4 0.01 0.06 0.81 0.72 hamming8-2 – – 541.34 531.47 hamming8-4 668.19 973.37 1172.32 1244.42 hamming10-2 – – – – hamming10-4 – – – – johnson8-2-4 0.01 0.01 0.11 0.14 johnson8-4-4 1.95 2.93 2.87 2.99 johnson16-2-4 277.13 363.29 4068.67 4460.17 johnson32-2-4 – – – – keller4 148.24 192.10 958.69 984.13 keller5 – – – – keller6 – – – – MANN_a9 54 66.18 9.19 9.35 MANN_a27 – – – – MANN_a45 – – – – MANN_A81 – – – – p_hat300-1 1.62 2.58 103.97 103.22 p_hat300-2 3333.80 5108.47 4018.11 3976.74 p_hat300-3 – – – – p_hat500-1 15.19 26.16 1088.94 1022.10 p_hat500-2 – – – – p_hat500-3 – – – – p_hat700-1 58.24 95.80 3922.90 3934.76 p_hat700-2 – – – – p_hat700-3 – – – – p_hat1000-1 327.15 539.14 – – J Braz Comput Soc (2012) 18:137–151 149 J Braz Comput Soc (2012) 18:137–151 149 Table 11 (Continued) Instance basic df chi chi + df p_hat1000-2 – – – – p_hat1000-3 – – – – p_hat1500-1 3473.86 5638.03 – – p_hat1500-2 – – – – p_hat1500-3 – – – – san200_0.7_1 – – 968.54 41.62 san200_0.7_2 – – 686.48 1005.41 san200_0.9_1 – – – – san200_0.9_2 – – 6522.12 – san200_0.9_3 – – – – san400_0.5_1 – – 276.68 198.35 san400_0.7_1 – – – – san400_0.7_2 – – – – san400_0.7_3 – – – – san400_0.9_1 – – – – san1000 – – – – sanr200_0.7 1597 3003.31 8052.30 8458.74 sanr200_0.9 – – – – sanr400_0.5 680.22 1050.91 – – sanr400_0.7 – – – – Table 12 Number of search tree nodes for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_1 760463351 813301 284689 466715 brock200_2 655863 7825 4937 7183 brock200_3 9025785 30619 15927 25821 brock200_4 29071199 142077 64799 94351 brock400_1 781878631 48146206 35069235 33124132 brock400_2 779146349 48813539 33430506 32451585 brock400_3 744501333 41982519 33296604 32832879 brock400_4 750841656 35973707 29794439 28115613 brock800_1 773878067 50773718 43825898 38468378 brock800_2 755264235 53435742 44837753 36531613 brock800_3 745732597 49566028 42160113 37270885 brock800_4 782408235 51112500 44532590 33801284 c-fat200-1 847 377 377 437 c-fat200-2 2373 353 353 487 c-fat200-5 197957 285 285 621 c-fat500-1 2419 973 973 1045 c-fat500-2 9775 949 949 1093 c-fat500-5 812265 873 873 1245 c-fat500-10 222003341 1 1 1493 hamming6-2 62239 135 129 127 hamming6-4 1949 153 153 257 Table 12 (Continued) Instance basic mcr mcs dyn hamming8-2 479589100 1207 1057 537 hamming8-4 50630977 16899 9707 25765 hamming10-2 477939325 23141 2497 2167 hamming10-4 799627648 37220492 15618125 19888977 johnson8-2-4 721 59 59 103 johnson8-4-4 160555 247 171 393 johnson16-2-4 29897217 533629 474647 1264845 johnson32-2-4 1124521577 288086577 282675611 236697572 keller4 13347251 23995 11749 17625 keller5 788436558 16911848 17475952 13452717 keller6 814607071 13800894 16179301 6555099 MANN_a9 5194193 75 57 191 MANN_a27 1039572148 29763 7615 86157 MANN_a45 896111360 129059 164086 94849 MANN_A81 649082104 12942 12942 9555 p_hat300-1 127783 4209 3091 4267 p_hat300-2 218320753 15677 6957 15039 p_hat300-3 694764638 3103461 523957 1257961 p_hat500-1 1192121 21635 16079 21821 p_hat500-2 666419883 627809 137045 379719 p_hat500-3 627863675 15527986 7860621 7118821 p_hat700-1 4405187 65849 43751 56061 p_hat700-2 674310972 5583275 700379 2195677 p_hat700-3 575688852 10991858 6126902 5211837 p_hat1000-1 25084283 398281 234723 339555 p_hat1000-2 651274462 12990532 8868752 7810188 p_hat1000-3 609819818 16924970 7386341 6864282 p_hat1500-1 232906339 2481911 1564723 2277359 p_hat1500-2 577460315 12680238 6257860 6025488 p_hat1500-3 609271041 14830728 4880291 5044208 san200_0.7_1 1199510748 7779 1739 2103 san200_0.7_2 1162303528 3145 1537 4603 san200_0.9_1 712616501 456721 47993 77603 san200_0.9_2 580762880 717197 26629 586527 san200_0.9_3 788583024 126665 7039 585453 san400_0.5_1 1057751310 4719 3023 4811 san400_0.7_1 1159107699 217471 48269 90853 san400_0.7_2 1091551627 42627 26811 18985 san400_0.7_3 1011697883 810445 248085 763393 san400_0.9_1 868327833 16184155 9529297 766201 san1000 894222972 426757 168419 251773 sanr200_0.7 122475879 325415 132903 211053 sanr200_0.9 708848293 19698991 6130061 12709733 sanr400_0.5 53900541 586913 329631 509701 sanr400_0.7 791421504 48490681 38355044 34153618 Table 11 (Continued) Instance basic df chi chi + df p_hat1000-2 – – – – p_hat1000-3 – – – – p_hat1500-1 3473.86 5638.03 – – p_hat1500-2 – – – – p_hat1500-3 – – – – san200_0.7_1 – – 968.54 41.62 san200_0.7_2 – – 686.48 1005.41 san200_0.9_1 – – – – san200_0.9_2 – – 6522.12 – san200_0.9_3 – – – – san400_0.5_1 – – 276.68 198.35 san400_0.7_1 – – – – san400_0.7_2 – – – – san400_0.7_3 – – – – san400_0.9_1 – – – – san1000 – – – – sanr200_0.7 1597 3003.31 8052.30 8458.74 sanr200_0.9 – – – – sanr400_0.5 680.22 1050.91 – – sanr400_0.7 – – – – J Braz Comput Soc (2012) 18:137–151 150 150 J Braz Comput Soc (2012) 18:137–151 Table 13 Running time (s) for basic, mcr, mcs and dyn Instance basic mcr mcs dyn brock200_1 10268.47 168.36 79.72 140.60 brock200_2 8.06 0.90 0.74 1.35 brock200_3 108.50 4.88 3.23 5.67 brock200_4 375.90 20.35 12.47 19.47 brock400_1 – – – – brock400_2 – – – – brock400_3 – – – – brock400_4 – – – – brock800_1 – – – – brock800_2 – – – – brock800_3 – – – – brock800_4 – – – – c-fat200-1 0.03 0.05 0.05 0.04 c-fat200-2 0.06 0.07 0.07 0.06 c-fat200-5 5.28 0.21 0.21 0.36 c-fat500-1 0.11 0.25 0.26 0.14 c-fat500-2 0.23 0.33 0.34 0.24 c-fat500-5 22.61 0.80 0.77 0.95 c-fat500-10 – 1.57 0.68 10.29 hamming6-2 1.10 0.02 0.03 0.03 hamming6-4 0.01 0.00 0.01 0.01 hamming8-2 – 4.24 4.67 2.46 hamming8-4 668.19 4.64 3.03 9.47 hamming10-2 – 3269.93 190.30 195.95 hamming10-4 – – – – johnson8-2-4 0.01 0.01 0.00 0.01 johnson8-4-4 1.95 0.03 0.03 0.06 johnson16-2-4 277.13 19.84 18.28 64.38 johnson32-2-4 – – – – keller4 148.24 2.82 1.73 3.36 keller5 – – – – keller6 – – – – MANN_a9 54.00 0.00 0.01 0.02 MANN_a27 – 302.00 63.32 1280.38 MANN_a45 – – – – MANN_A81 – – – – p_hat300-1 1.62 0.37 0.34 0.39 p_hat300-2 3333.80 4.92 2.76 5.63 p_hat300-3 – 1586.09 366.70 951.82 p_hat500-1 15.19 2.73 2.21 2.99 p_hat500-2 – 349.39 94.19 281.62 p_hat500-3 – – – – p_hat700-1 58.24 8.86 6.97 13.81 p_hat700-2 – 4743.96 763.28 2637.62 p_hat700-3 – – – – p_hat1000-1 327.15 49.21 34.61 66.25 Table 13 (Continued) Instance basic mcr mcs dyn p_hat1000-2 – – – – p_hat1000-3 – – – – p_hat1500-1 3473.86 428.89 274.09 529.73 p_hat1500-2 – – – – p_hat1500-3 – – – – san200_0.7_1 – 2.51 0.82 1.24 san200_0.7_2 – 0.63 0.41 1.46 san200_0.9_1 – 206.17 35.23 68.44 san200_0.9_2 – 391.74 23.37 581.02 san200_0.9_3 – 43.70 4.72 691.17 san400_0.5_1 – 2.79 1.79 2.28 san400_0.7_1 – 212.86 50.85 75.95 san400_0.7_2 – 30.13 15.67 26.69 san400_0.7_3 – 356.42 110.19 293.37 san400_0.9_1 – – – 4906.83 san1000 – 910.12 298.10 147.11 sanr200_0.7 1597.00 53.26 29.67 50.90 sanr200_0.9 – – 5223.79 10692.55 sanr400_0.5 680.22 72.19 51.47 83.51 sanr400_0.7 – – – – Table 14 Average search tree node per second and number of solved instances for each algorithm Algorithm (# Search tree nodes)/s Solved instances basic 72066.01 26 cp 66490.03 25 df 6648.27 26 chi 28.87 28 chi + df 29.58 27 mcq 3778.20 43 dyn 2579.93 45 mcr 3523.45 43 mcs 3014.30 44 This is the number of search tree nodes summed over all the instances, divided by the total time summed over all the in- stances. 1. Bellare M, Goldreich O, Sudan M (1995) Free bits, pcps and non- approximability-towards tight results. In: Proceedings, 36th an- nual symposium on foundations of computer science, 1995. IEEE Comput Soc, Los Alamitos, pp 422–431 Appendix: Unabridged experimental data The last column shows how many of the instances were solved to completion within the prescribed time of three hours. References 1. Bellare M, Goldreich O, Sudan M (1995) Free bits, pcps and non- approximability-towards tight results. In: Proceedings, 36th an- nual symposium on foundations of computer science, 1995. IEEE Comput Soc, Los Alamitos, pp 422–431 This is the number of search tree nodes summed over all the instances, divided by the total time summed over all the in- stances. The last column shows how many of the instances were solved to completion within the prescribed time of three hours. References 151 J Braz Comput Soc (2012) 18:137–151 11. Moon J, Moser L (1965) On cliques in graphs. Isr J Math 3(1):23– 28. doi:10.1007/BF02760024 2. Bollobás B (2001) Random graphs. Cambridge University Press, Cambridge. http://books.google.com/books?hl=en&lr=&id= o9WecWgilzYC&oi=fnd&pg=PR10&dq= bollobas%2Brandom%2Bgraphs&ots=YyFTnSQpVh&sig= 7GrvDOb_MJLesgbjLvQj0TeNG8U#PPP1,M1 12. Östergård PR (2002) A fast algorithm for the maximum clique problem. Discrete Appl Math 120(1–3):197–207. doi:10.1016/ S0166-218X(01)00290-6 g j Qj 3. Bomze IM, Budinich M, Pardalos PM, Pelillo M (1999) The maximum clique problem. In: Handbook of combinatorial op- timization, vol 4, pp 1–74. http://citeseerx.ist.psu.edu/viewdoc/ summary?doi=10.1.1.56.6221 13. Robson J (1986) Algorithms for maximum independent sets. J Al- gorithms 7(3):425–440. doi:10.1016/0196-6774(86)90032-5 14. Robson J (2001) Finding a maximum independent set in time o(2(n/4)). http://www.labri.fr/perso/robson/mis/techrep.html 4. Bron C, Kerbosch J (1973) Algorithm 457: ﬁnding all cliques of an undirected graph. Commun ACM 16(9):575–577. doi:10.1145/ 362342.362367 15. Tarjan RE, Trojanowski AE (1976) Finding a maximum indepen- dent set. Tech. rep., Computer Science Department, School of Hu- manities and Sciences, Stanford University, Stanford, CA, USA. http://portal.acm.org/citation.cfm?id=892099 5. Carraghan R, Pardalos PM (1990) An exact algorithm for the maximum clique problem. Oper Res Lett 9(6). doi:10.1016/0167- 6377(90)90057-C 16. Tomita E, Kameda T (2007) An efﬁcient branch-and-bound al- gorithm for ﬁnding a maximum clique with computational ex- periments. J Glob Optim 37(1):95–111. doi:10.1007/s10898-006- 9039-7 ( ) 6. Fahle T (2002) Simple and fast: Improving a branch-and-bound algorithm for maximum clique. In: Lecture notes in computer sci- ence. Springer, Berlin, pp 47–86. doi:10.1007/3-540-45749-6_44 17. Tomita E, Seki T (2003) An efﬁcient branch-and-bound algo- rithm for ﬁnding a maximum clique. Springer, Berlin. http://www. springerlink.com/content/7jbjyglyqc8ca5n9 7. Garey M, Johnson D (1979) Computers and intractability. Free- man, San Francisco 8. Jian T (1986) An o(20.304n) algorithm for solving maximum inde- pendent set problem. IEEE Trans Comput 35(9):847–851. doi:10. 1109/TC.1986.1676847 18. Tomita E, Sutani Y, Higashi T, Takahashi S, Wakatsuki M (2010) A simple and faster branch-and-bound algorithm for ﬁnding a maximum clique. In: Rahman M, Fujita S (eds) WALCOM: Algo- rithms and computation, vol 5942. Springer, Berlin, pp 191–203. doi:10.1007/978-3-642-11440-3_18. Chap. 18 9. Konc J, Janezic D (2007) An improved branch and bound algo- rithm for the maximum clique problem. MATCH Commun Math Comput Chem. http://www.sicmm.org/~konc/%C4%8CLANKI/ MATCH58(3)569-590.pdf 19. Tomita E, Tanaka A, Takahashi H (2006) The worst-case time complexity for generating all maximal cliques and computa- tional experiments. Theor Comput Sci 363(1):28–42. doi:10.1016/ j.tcs.2006.06.015 10. Li CM, Quan Z (2010) An efﬁcient branch-and-bound algorithm based on maxsat for the maximum clique problem. References In: Twenty- fourth AAAI conference on artiﬁcial intelligence. http://www. aaai.org/ocs/index.php/AAAI/AAAI10/paper/view/1611
https://openalex.org/W4293086888	https://link.springer.com/content/pdf/10.1007/s10579-022-09588-2.pdf	English	null	Predicting lexical complexity in English texts: the Complex 2.0 dataset	Language resources and evaluation	2,022	cc-by	22,405	* Matthew Shardlow m.shardlow@mmu.ac.uk Language Resources and Evaluation (2022) 56:1153–1194 https://doi.org/10.1007/s10579-022-09588-2 Language Resources and Evaluation (2022) 56:1153–1194 https://doi.org/10.1007/s10579-022-09588-2 Language Resources and Evaluation (2022) 56:1153–1194 https://doi.org/10.1007/s10579-022-09588-2 ORIGINAL PAPER Predicting lexical complexity in English texts: the Complex 2.0 dataset Matthew Shardlow1 · Richard Evans2 · Marcos Zampieri3 Accepted: 7 March 2022 / Published online: 23 March 2022 © The Author(s) 2022 Keywords Complex word identification · Lexical complexity · Text simplification Abstract Identifying words which may cause difficulty for a reader is an essential step in most lexical text simplification systems prior to lexical substitution and can also be used for assessing the readability of a text. This task is commonly referred to as complex word identification (CWI) and is often modelled as a supervised classification prob- lem. For training such systems, annotated datasets in which words and sometimes multi-word expressions are labelled regarding complexity are required. In this paper we analyze previous work carried out in this task and investigate the properties of CWI datasets for English. We develop a protocol for the annotation of lexical com- plexity and use this to annotate a new dataset, CompLex 2.0. We present experi- ments using both new and old datasets to investigate the nature of lexical complex- ity. We found that a Likert-scale annotation protocol provides an objective setting that is superior for identifying the complexity of words compared to a binary anno- tation protocol. We release a new dataset using our new protocol to promote the task of Lexical Complexity Prediction. Keywords Complex word identification · Lexical complexity · Text simplification 1 Introduction Predicting lexical complexity can enable systems to better guide a user to an appropriate text, or tailor it to their needs. The task of automatically identifying which words are likely to be considered complex by a given target population is known as Complex Word Identification (CWI) and it constitutes an important step in most lexical simplification pipelines (Paetzold & Specia, 2017).i p pi p p p The topic has gained significant attention in the last few years, particularly for English—which is also the focus of our study. A number of studies have been pub- lished on predicting complexity of both single words and multi-word expressions (MWEs) including two recent competitions organized on the topic, CWI 2016 and CWI 2018, discussed in detail in Sect. 2. The first shared task on CWI was organ- ized at SemEval in 2016 (Paetzold & Specia, 2016a) providing participants with an English dataset in which words in context were annotated as non-complex (0) or complex (1) by a pool of human annotators. The goal was to predict this binary value for the target words in the test set. A post-competition analysis of the CWI 2016 results (Zampieri et al., 2017) examined the performance of the participating systems and evidenced how challenging CWI 2016 was with respect to the distri- bution (more testing than training instances) and annotation type. The second edition of the CWI shared task was organized in 2018 at the BEA workshop (Yimam et al., 2018). CWI 2018 featured multilingual (English, Span- ish, German, and French) and multi-domain datasets (Yimam et al., 2017). Unlike in CWI 2016, predictions were evaluated not only in a binary classification set- ting but also in terms of probabilistic classification in which systems were asked to assign the probability of the given target word in its particular context being complex. Although CWI 2018 provided an element of regression, the continuous complexity value of each word was calculated as the proportion of annotators that found a word complex. For example, if 5 out of 10 annotators labeled a word as complex then the word was given a score of 0.5. This measure relies on an aggre- gation of absolute binary judgments of complexity to give a continuous value. Instead of using binary judgments, the CompLex dataset uses Likert Scale judg- ments (Shardlow et al., 2020), for which the specification is discussed in depth in Sect. 4. 23456789) 1 3 1154 M. Shardlow et al. 1 3 1 Introduction The main contributions of this paper are: – A concise yet comprehensive survey of the two editions of the CWI shared tasks organized in 2016 and 2018; – A concise yet comprehensive survey of the two editions of the CWI shared tasks organized in 2016 and 2018; A i i i i h f f h l i h l i l l i – An investigation into the types of features that correlate with lexical complexity – A qualitative analysis of the CWI–2016 (Paetzold and Specia 2016a), CWI–2018 (Yimam et al., 2018) and Maddela–2018 (Maddela and Xu 2018) datasets, high- lighting issues with the annotation protocols that were used;i – The specification of a new annotation protocol for the CWI task;i i – An implementation of our specification, describing the annotation of a new data- set for CWI (CompLex 1.0 and 2.0);f – Experiments comparing the features affecting lexical complexity in our dataset, as compared to others;f – Experiments using our dataset, demonstrating the effects of genre on CWI. The remainder of this paper is organized as follows. Section 2 provides an overview of the previous CWI shared tasks. Section 3 provides a preliminary investigation into the types of features that correlate with complexity labels in previous CWI data- sets. Section 4 firstly discusses the datasets that have previously been used for CWI, highlighting issues in their annotation protocols in Sect. 4.1, and then proposes a new protocol for constructing CWI datasets in Sect. 4.5. Section 5 reports on the construction of a new dataset following the specification previously laid out. Sec- tion 6 compares the annotations in our new dataset to those of previous datasets by developing a categorical annotation scheme. Section 7 shows further experiments demonstrating how our new corpus can be used to investigate the nature of lexical complexity. Finally, a discussion of our main thesis and conclusions of our work are presented in Sects. 8 and 9 respectively. We have previously published the CompLex 1.0 data as a workshop paper (Shardlow et al., 2020). The CompLex 2.0 data was also described in the SemEval task description paper (Shardlow et al., 2021). In this paper, we seek to build upon these prior works to give an in depth and rounded treatment to the lexical complex- ity problem. 1 Introduction CompLex is a multi-domain English dataset annotated with a 5-point Likert scale (1-5) corresponding to the annotators comprehension and familiarity with the words in which 1 represents very easy and 5 represents very difficult. The CompLex dataset was used as the official dataset of SemEval-2021 Task 1: Lexical Complexity Prediction (LCP) (Shardlow et al., 2021). The goal of LCP 2021 is to predict this complexity score for each target word in context in the test set. In this paper, we investigate properties of multiple annotated English lexical com- plexity datasets such as the aformentioned CWI datasets and others from the litera- ture (Maddela & Xu, 2018). We investigate the types of features that make words complex. We analyse the shortcomings of the previous CWI datasets and use this to motivate the specification of a new type of CWI dataset, focusing not on complex- word identification (CWI), but instead on lexical complexity prediction (LCP), that is CWI in a continuous-label setting. We further develop a dataset based on adding 1 3 1 3 Predicting lexical complexity in English texts: the Complex… 1155 additional annotations to the existing CompLex 1.0 to create our new dataset, Com- pLex 2.0, and use this to provide experiments into the nature of lexical complexity. additional annotations to the existing CompLex 1.0 to create our new dataset, Com- pLex 2.0, and use this to provide experiments into the nature of lexical complexity. 2 Related work There have been various studies which have both created datasets and explored computational models for CWI, particularly focusing on English texts (Shardlow, 2013b, a; Gooding and Kochmar, 2019; Finnimore et al., 2019). These studies have addressed CWI as a stand-alone task or as part of lexical simplification pipelines.i i Given the direct application of CWI to lexical simplification systems, where the goal is to decide whether or not a word needs to be substituted for a simpler one, the clear majority of studies have addressed CWI as a binary classification task. 1 3 3 1156 M. Shardlow et al. That said, there have been multiple studies analyzing the shortcomings of approach- ing CWI as a binary classification task. Some studies have studied the relationship between classification performance and dataset annotation in an attempt to esti- mate the theoretical upper boundary of binary CWI systems (Zampieri et al., 2017) while others have investigated alternative ways to model the task. One study posed that comparative judgments are more consistent than binary classification for CWI (Gooding et al., 2019).i CWI is of direct interest to those working in lexical simplification as it forms the first part of the lexical simplification pipeline (Devlin & Tait, 1998). Before a word can be simplified, a decision must be made as to whether or not that word requires simplification. Simplification systems (Biran et al., 2011; Bott et al., 2012), then generate potential candidates for simplification and use a similar process to CWI to select the most simple candidate (Paetzold et al., 2017). Comparative complexity is a related but distinct task to Lexical Complexity Prediction. In this task, two words are taken and a judgment is given to determine which is the most complex. A recent study found that annotations for comparative complexity were more consistent than binary classification (Gooding et al., 2019). Nonetheless, we have not focussed on comparative complexity in this work, but rather on continuous complexity. We are most interested in the complexity of a word in it’s original context, rather than in relation to another word. The increased interest from the research community in CWI was the primary motivation for the organisation of the two editions of the aforementioned CWI shared task in 2016 and 2018. These shared tasks have made important benchmark datasets available to the community that are widely used beyond these competitions. 2 Related work In the next sub-sections we provide an overview of these two editions: CWI–2016 organized at SemEval 2016 (Paetzold & Specia, 2016a) and CWI–2018 organized at the BEA workshop in 2018 (Yimam et al., 2018). We describe the task setup, present the datasets, and briefly discuss the approaches submitted by participants in the two editions of the competition. We also present the approaches and the features used by each system. Finally, we analyze the results obtained by the participants and the main challenges of each edition of the CWI Shared Task. 1 http://alt.qcri.org/semeval2016/task11/. 2.1 CWI–2016 The first shared task on CWI was organized as Task 11 at the International Workshop on Semantic Evaluation (SemEval) in 2016.1 CWI–2016 provided participants with a manually annotated dataset in which words in context were labeled as complex or non-complex, where complexity is interpreted as whether a word was understood or not by a pool of 400 non-native speakers of English. CWI–2016 was therefore modelled as a binary text classification task at the word level. Participants were required to build systems to predict lexical complexity in sentences of the unlabeled test set and assign label 0 to non-complex words and 1 to complex ones. Two examples from the CWI–2016 dataset are shown below: 1 3 3 Predicting lexical complexity in English texts: the Complex… 1157 (1) A frenulum is a small fold of tissue that secures or restricts the motion of mobile organ in the body. (1) A frenulum is a small fold of tissue that secures or restricts the motion of a mobile organ in the body. (2) The name ‘kangaroo mouse’ refers to the species’ extraordinary jumping abil- ity, as well as its habit of bipedal locomotion. (2) The name ‘kangaroo mouse’ refers to the species’ extraordinary jumping abil- ity, as well as its habit of bipedal locomotion. The words in bold: frenulum, restricts, and motion in Example 1, and extraor- dinary, bipedal, and locomotion in Example 2 were annotated by at least one of the annotators as complex and thus they were labeled as such in the training set. Adjacent words like bipedal locomotion do not represent multi-word expressions (MWEs) as they were annotated in isolation because the task set-up of CWI–2016 only considered single word annotations. Whilst MWEs were not considered in CWI–2016, they were studied in CWI–2018 (see Sect. 2.2). The dataset provided by the organizers of CWI–2016 contained a training set of 2237 target words in 200 sentences. The training set was annotated by 20 anno- tators and a word was considered complex in the training set if at least one of the 20 annotators assigned it as so. The test set included 88,221 target words in 9,000 sentences and each word was annotated by only one annotator. Therefore, the ground truth label for each word in the test was attributed based on a single complexity judgement. 2.1 CWI–2016 According to the organisers of CWI–2016, this setup was devised to imitate a realistic scenario where the goal was to predict the individual needs of a speaker based on the needs of the target group (Paetzold & Specia, 2016a). Finally, the data included in the CWI–2016 dataset comes from various sources such as the CW Corpus (Shardlow, 2013a), the LexMTurk Corpus (Horn et al., 2014), and Simple Wikipedia (Kauchak, 2013). p p CWI–2016 attracted a large number of participants. A total of 21 teams sub- mitted 42 systems to the competition. A wide range of features such as word embeddings, word and character n-grams, word frequency, Zipfian frequency- based features, word length, morphological, syntactic, semantic, and psycholin- guistic features were used by participants. A number of different approaches to classification were tested, ranging from traditional machine learning classifiers such as support vector machines (SVM), decision trees, random forest, and maxi- mum entropy classifiers to deep learning classifiers, such as recurrent neural net- works. In Table 1, we list the approaches submitted to CWI–2016 by the 19 teams who wrote system description papers presented at SemEval. In terms of performance the top-3 systems were team PLUJAGH (Wróbel, 2016), LTG (Malmasi et al., 2016), and MAZA (Malmasi & Zampieri, 2016) which obtained 0.353, 0.312, and 0.308 F1-score respectively. The three teams used rather simple probabilistic models trained on features such as n-grams, word frequency, word length, and the presence of words in vocabulary lists extracted from Simple Wikipedia, introduced by PLUJAGH. The relatively low perfor- mance obtained by all teams, including the top-3 systems, evidences how chal- lenging the CWI–2016 shared task was. Both the data annotation protocol and the training/test split, where 40 times more testing data than training data is avail- able, contributed to making CWI–2016 a difficult task. fi A post-competition analysis was carried out using the output of all 42 systems submitted to CWI–2016 (Zampieri et al., 2017). Each system output to each test 1 3 3 3 1158 M. Shardlow et al. able 1 Systems submitted to the CWI–2016 in alphabetical order. We include team names and a brief description of each system including features and classifiers used. 2.1 CWI–2016 A reference to each system description paper is provided for more information Team Classifiers Features References AI-KU SVM Word embeddings of the target and surrounding words Kuru (2016) Amrita-CEN SVM Word embeddings and various semantic and morphological features Sanjay and Soman (2016) BHASHA SVM, Decision Tree Lexical and morphological features Choubey and Pateria (2016) ClacEDLK Random Forests Semantic, morphological, and psycholinguistic features Davoodi and Kosseim (2016) CoastalCPH Neural Network, Logistic Regression Word frequencies and word embeddings Bingel et al. (2016) HMC Decision Tree Lexical, semantic, syntactic and psycholinguistic features Quijada and Medero (2016) IIT Nearest Centroid Semantic and morphological features Palakurthi and Mamidi (2016) UNLP Random Forest, Naive Bayes Semantic, lexicon-based, morphological and syntactic features Mukherjee et al. (2016) LTG Decision Tree n-grams and word length Malmasi et al. (2016) MACSAAR Random Forest, SVM Zipfian frequency distribution, word length Zampieri et al. (2016) MAZA Meta-classifier n-grams, word probability, word length (Malmasi and Zampieri 2016) Melbourne Weighted Random Forests Lexical and semantic features Brooke et al. (2016) LUJAGH Threshold-based methods Features extracted from Simple Wikipedia Wróbel (2016) omona Threshold-based methods Word frequencies Kauchak (2016) ensible Ensemble Recurrent Neural Networks Word embeddings Gillin (2016) V000gg System voting with threshold Morphological, lexical, and semantic features Paetzold and Specia (2016b) TALN Random Forest Lexical, morphological, semantic, and syntactic features Ronzano et al. (2016) USAAR Bayesian Ridge classifiers Hand-crafted word sense entropy metric and language model perplexity Martínez Martínez and Tan (2016) UWB Maximum Entropy Word occurrence counts on Wikipedia documents Konkol (2016) 3 Table 1 Systems submitted to the CWI–2016 in alphabetical order. We include team names and a brief description of each system including features and classifiers used. A reference to each system description paper is provided for more information Team Classifiers Features References AI-KU SVM Word embeddings of the target and surrounding words Kuru (2016) Amrita-CEN SVM Word embeddings and various semantic and morphological features Sanjay and Soman (2016) BHASHA SVM, Decision Tree Lexical and morphological features Choubey and Pateria (2016) ClacEDLK Random Forests Semantic, morphological, and psycholinguistic features Davoodi and Kosseim (2016) CoastalCPH Neural Network, Logistic Regression Word frequencies and word embeddings Bingel et al. (2016) HMC Decision Tree Lexical, semantic, syntactic and psycholinguistic features Quijada and Medero (2016) IIIT Nearest Centroid Semantic and morphological features Palakurthi and Mamidi (2016) JUNLP Random Forest, Naive Bayes Semantic, lexicon-based, morphological and syntactic features Mukherjee et al. 2.1 CWI–2016 (2016) LTG Decision Tree n-grams and word length Malmasi et al. (2016) MACSAAR Random Forest, SVM Zipfian frequency distribution, word length Zampieri et al. (2016) MAZA Meta-classifier n-grams, word probability, word length (Malmasi and Zampieri 2016) Melbourne Weighted Random Forests Lexical and semantic features Brooke et al. (2016) PLUJAGH Threshold-based methods Features extracted from Simple Wikipedia Wróbel (2016) Pomona Threshold-based methods Word frequencies Kauchak (2016) Sensible Ensemble Recurrent Neural Networks Word embeddings Gillin (2016) SV000gg System voting with threshold Morphological, lexical, and semantic features Paetzold and Specia (2016b) TALN Random Forest Lexical, morphological, semantic, and syntactic features Ronzano et al. (2016) USAAR Bayesian Ridge classifiers Hand-crafted word sense entropy metric and language model perplexity Martínez Martínez and Tan (2016) UWB Maximum Entropy Word occurrence counts on Wikipedia documents Konkol (2016) 1 Predicting lexical complexity in English texts: the Complex… 1159 instance was used as a vote to build two ensemble models. The ensemble models were built using plurality voting which assigns the highest number of votes as the label of a given instance, and exploits an oracle which assigns the correct label for an instance if at least one of the systems predicted the ground truth label for that instance. The plurality vote serves to better understand the performance of the sys- tems using the same dataset while the oracle is used to quantify the theoretical upper limit performance on the dataset (Kuncheva et al., 2001). The study showed that the potential upper limit for the CWI–2016 dataset considering the output of the partici- pating systems is 0.60 F1 score for the complex word class. The outcome confirms that the low performance of the systems is related to the way the data has been anno- tated. Finally, this study also confirmed the relationship between word length and lexical complexity annotation in this dataset, a feature used by many of the teams participating in CWI–2016 as well as in our present work. 2 https://www.sites.google.com/view/cwisharedtask2018/. 2.2 CWI–2018 Following the success of CWI–2016, the second edition, CWI–2018, was organized at the Workshop on the Innovative Use of NLP for Building Educational Applica- tions (BEA) in 2018.2 Unlike CWI–2016 which focused only on English, CWI–2018 featured English, French, German, and Spanish datasets opening new perspectives in research in this area. A total of four tracks were available at CWI–2018: English, German, and Span- ish, in which training and testing data was available for each language, and French. The organizers released a French test set with no corresponding training set with the goal of deriving models for French CWI from the English, Spanish, and German datasets. CWI–2018 featured two sub-tasks: (i) a binary classification task similar to CWI–2016 where participants were asked to label the given target word in a par- ticular context as complex or simple; (ii) a probabilistic classification task where participants were asked to give a probability of the given target word in a particular context being complex. In terms of data, CWI–2018 used the CWIG3G2 dataset (Yimam et al., 2017) in English, German, and Spanish. The English dataset contains texts from three domains, News, WikiNews, and Wikipedia articles and the evaluation was carried out per domain. To allow cross-lingual learning, a dataset for French was collected using the same methodology as the one used for the CWIG3G2 corpus. Another important difference between CWI–2016 and CWI–2018 is that the CWIG3G2 featured anno- tation of both single words and MWEs while the dataset used in CWI–2016 only considered single words.f In terms of participation, CWI–2018 attracted 12 teams in different task/track combinations. In Table 2, we list the approaches submitted to the English binary classification single word track by the 10 teams who wrote system description papers presented at BEA. Most teams tried multiple approaches and here we describe the teams’ best-performing ones according to their system description papers. 1 3 1160 M. Shardlow et al. k. We include team names and a brief description of each system including fea- information Table 2 Systems submitted to the CWI–2018 English binary classification single word track. We include team names and a brief description of each system including fea- tures and classifiers used. A reference to each system description paper is provided for more information Team Classifiers Features Paper CAMB Adaboost N-grams, WordNet features, POS tags, dependency parsing relations, psycholin- guistic features. 2.2 CWI–2018 Gooding and Kochmar (2018) CFILT_IITB Voting ensemble Word length, syllable counts, vowel counts, WordNet-based features. Wani et al. (2018) hu-berlin Naive Bayes Character n-grams Popović (2018) ITEC LSTM Word length, word and character embeddings, frequency count, psycholinguistics features. De Hertog and Tack (2018) LaSTUS/TALN SVM, Random Forest Word length, word embeddings, semantic and contextual features. AbuRa’ed and Saggion (2018) NILC XGBoost N-grams, word length, number of syllables, WordNet-based features. Hartmann and dos Santos (2018) NLP-CIC Tree ensembles and CNNs Word frequency, syntactic and lexical features, psycholinguistic features, and word embeddings. Aroyehun et al.et al. (2018) SB@GU Extra trees Word length, number of syllables, n-grams, frequency distribution. Alfter and Pilán (2018) TMU Random Forest Word length, word frequency, probability features derived from corpora. Kajiwara and Komachi (2018) UnibucKernel Kernel-based learning with SVMs. Character n-grams, semantic features, and word embeddings. Butnaru and Ionescu (2018) 1 3 Predicting lexical complexity in English texts: the Complex… 1161 For the English binary classification single word track, the organizers reported the performance by all teams per domain. Team CAMB obtained the best perfor- mance for the three domains: 0.8736 F1-score on News, 0.8400 F1-score on Wiki- News, and 0.8115 F1-score on Wikipedia. We observed that for all teams the perfor- mance on the News domain was generally substantially higher than the performance obtained in the two other domains. Several teams used the opportunity to compare multiple approaches for this task and many of them reported that traditional machine learning classifiers were more accurate than deep neural networks (Hartmann & dos Santos, 2018; Alfter &Pilán, 2018). 3 https://en.wikipedia.org/wiki/Help:Infobox. Last accessed 16th September 2021. 3 Analysis of features of complex words Upon analysing the datasets and system features used in CWI–2016 and CWI–2018, we noticed several intuitive explanations as to why a word may be judged as com- plex, or not: – The word is archaic. – The word is archaic. – The word is a borrowing from another language or refers to a concept that is atypical in the culture of the reader. – The word is a borrowing from another language or refers to a concept that is atypical in the culture of the reader. – The word is uncommon and many people are not generally exposed to it. – The word refers to a very specialised concept. – Although the word is common, it is being used with an uncommon meaning in the given context. These possible characteristics motivated us to represent input words as sets of indic- ative linguistic features for the purpose of CWI. We used 378 features to represent words in our data set. These include psycholinguistic features derived from the MRC database (Wilson, 1988), word embeddings, and several other features with the potential to capture our intuitions about lexical complexity. Values of the psycholinguistic features of words were obtained using the API to the MRC database. Many of the resources included in the database were built before 1998. These were derived through rigorous psycholinguistic testing, and as a result are of restricted size (offering relatively poor coverage of current English vocabu- lary). For this reason, in addition to specifying the values of these features directly from the database, we included binary features to indicate whether or not the word occurs in the MRC database. We used information about whether or not the Wikipedia entry for the word includes an infobox element to indicate its degree of specialisation. Wiki-pedia3 describes infoboxes as: […] a fixed-format table usually added to the top right-hand corner of arti- […] a fixed-format table usually added to the top right-hand corner of arti- cles to consistently present a summary of some unifying aspect that the arti- cles share and sometimes to improve navigation to other interrelated articles. 3 Analysis of features of complex words Many infoboxes also emit structured metadata which is sourced by DBpedia [ ]i y p g cles to consistently present a summary of some unifying aspect that the arti- l h d i i i i h i l d i l y p y y g p cles share and sometimes to improve navigation to other interrelated articles. Many infoboxes also emit structured metadata which is sourced by DBpedia 1 3 1162 M. Shardlow et al. and other third party re-users. The generalized infobox feature grew out of the original taxoboxes (taxonomy infoboxes) that editors developed to visually express the scientific classification of organisms. and other third party re-users. The generalized infobox feature grew out of the original taxoboxes (taxonomy infoboxes) that editors developed to visually express the scientific classification of organisms. We observed that entries for specialised vocabulary (e.g. Gharial) frequently con- tain infobox elements of various types (e.g. biota). We extracted features encoding information about the occurrence and type of infobox element as an indicator of the level of specialisation of the word. We view this as a type of coarse-grained seman- tic information which is available for a relatively large proportion of words: more than 76% of those occurring in the CWI-2016 and CWI-2018 datasets. The full feature set is displayed in Tables 3 and 4. Given that it encodes well- motivated psycholinguistic information and includes features which capture our intuitions about lexical complexity, we consider this feature set to be suitable for use in the derivation of models for CWI. We processed the human-annotated CWI–2016 and CWI–2018 datasets to represent words as feature vectors using the features in these tables. Features P and S (Table 4) can be categorised as high coverage (holding for more than two thirds of the tokens in the annotated corpora); features G, E, J, H, Q, and I (Tables 3 and 4) as medium coverage (holding for more than one third but less than two thirds of the tokens in the corpora); and features F, R, N, O, K, B, D, M, L, and O (Tables 3 and 4) as low coverage (holding for less than one third of the tokens in the corpora).4 Considered individually, the great majority of features/feature sets listed in Table 3 have no linear relationship with the averaged human judgement of word complexity in the CWI 2016 and CWI 2018 datasets. 4 These features are listed in decreasing order of coverage provided. 3 Analysis of features of complex words The only exceptions are word length (feature group C) and the word’s frequency count in the London- Lund corpus (feature group H). As the distributions of these two features are non-normal, we measured correlation with the averaged complexity ratings of words using Spearman’s rho. We found that normalised word length has a low positive correlation ( 𝜌(28 677) = 0.435, p < 0.001 ) while the frequency of the word in the Brown corpus has a low negative correlation with word complexity ( 𝜌(28 677) = −0.354, p < 0.001 ). It is worth noting that MWEs in the CWI-2018 data are always complex and this may have influenced the results for word-length as MWEs are typically longer than single words. There is no linear relationship between the values of features/feature sets listed in rows K–S of Table 4 and the averaged values of word complexity assigned by the anno- tators. In our experiments, we did not investigate the strength of correlations between individual word embedding features and average complexity ratings. Given that the distributions of our features are non-normal, we used Levene’s test (Levene, 1960) to assess the homogeneity of variance between word feature values and complexity scores. In all cases, the Levene test statistic exceeded critical values and obtained p < 0.01 , indicating no equality of variance between complexity scores and feature values. 1 3 Predicting lexical complexity in English texts: the Complex… 1163 able 3 Features (A–J) used to represent words Available at https://archive.org/stream/dictionaryofarch028421mbp/dictionaryofarch~028421mbp_djvu.txt. Last accessed 26th February 2019 Longest word in English being 45 characters (pneumonoultramicroscopicsilicovolcanoconiosis) D Feature Type Definition A Frequent Binary One of the 10 000 most frequent words listed in Wiktionary B Archaic Binary Listed in an archaic word list.† C Length (normalised) Numerical Length of the word divided by 50.‡ D Plurality Binary 5 features indicating whether the word is plural, has no plural form, is a singular form, is both singular and plural form, or is plural but acts singular. E Familiarity Numerical (100-700) Familiarity score, derived by merging three sets of norms: Paivio (unpublished; these are an expansion of the norms of Paivio et al. (1968)), Toglia and Battig (1978), and Gilhooly and Logie (1980)). 3 Analysis of features of complex words See Wilson (1988) for more details on these metrics F Concreteness Numerical (100-700) Concreteness score, listed in the MRC Database G Imageability Numerical (100-700) Imageability score of the word, listed in the MRC Database H Brown Numerical Frequency count of the word in the London-Lund Corpus of English Conversation (Svartvik and Quirk 1980) KFFREQ Numerical Frequency count of the word in the Kučera and Francis (1967) frequency list, derived from the Brown corpus. TLFREQ Numerical Frequency listed in Thorndike and Lorge (1944) L count, which combines the counts of morphological variants of the word in a reference corpus Table 3 Features (A–J) used to represent words † Available at https://archive.org/stream/dictionaryofarch028421mbp/dictionaryofarch~028421mbp_djvu.txt. Last accessed 26th February 2019 ‡ Longest word in English being 45 characters (pneumonoultramicroscopicsilicovolcanoconiosis) ID Feature Type Definition A Frequent Binary One of the 10 000 most frequent words listed in Wiktionary B Archaic Binary Listed in an archaic word list.† C Length (normalised) Numerical Length of the word divided by 50.‡ D Plurality Binary 5 features indicating whether the word is plural, has no plural form, is a singular form, is both singular and plural form, or is plural but acts singular. E Familiarity Numerical (100-700) Familiarity score, derived by merging three sets of norms: Paivio (unpublished; these are an expansion of the norms of Paivio et al. (1968)), Toglia and Battig (1978), and Gilhooly and Logie (1980)). See Wilson (1988) for more details on these metrics F Concreteness Numerical (100-700) Concreteness score, listed in the MRC Database G Imageability Numerical (100-700) Imageability score of the word, listed in the MRC Database H Brown Numerical Frequency count of the word in the London-Lund Corpus of English Conversation (Svartvik and Quirk 1980) I KFFREQ Numerical Frequency count of the word in the Kučera and Francis (1967) frequency list, derived from the Brown corpus. J TLFREQ Numerical Frequency listed in Thorndike and Lorge (1944) L count, which combines the counts of morphological variants of the word in a reference corpus 3 1164 M. Shardlow et al. 3 Analysis of features of complex words Table 4 Features (K–T) used to represent words ID Feature Type Definition K MEANC Numerical (100-700) Meaningfulness rating of the word as provided by the Colorado norms of Toglia and Battig (1978) L MEANP Numerical (100-700) Meaningfulness rating of the word as provided by the norms of Paivio (unpublished) M AOA Numerical (100-700) Age of acquisition, as provided by the norms of Gilhooly and Logie (1980) N TQ2Q Binary Morphological variant of another word in the dictionary O TQ22 Binary Ends in the letter R and this R is not pronounced except when the next word begins with a vowel P WTYPE Binary 9 features indicating the word type (adverb, conjunction, interjection, adjective, noun, past participle, pronoun, verb, or other) as listed in the Shorter Oxford English Dictionary or Webster’s New Interna- tional Dictionary Q STATUS Binary 7 features indicating the word status (archaic, alien, obsolete, colloquial, rare, and standard) as listed in the Dolby database (Dolby et al., 1963) R STRESS Binary 14 features indicating the stress pattern of the word when pronounced. Where 2 is a strongly stressed syllable, 1 is medium stressed, and 0 is an unstressed syllable, the 14 stress patterns are: 0, 01020, 010200, 02, 020, 0200, 10020, 102, 1020, 10200, 20, 200, 2000, and 22 S INFOBOX Binary 13 features indicating the type of infobox present in the English Wikipedia page for the word. Infobox types are: AMBIGUOUS, BIOGRAPHY_VCARD, BIOTA, BORDERED, COLLAPS- IBLE_ AUTOCOLLAPSE, DEFAULT, GEOGRAPHY_VCARD, HPRODUCT, NONE, VCARD, VCARD_PLAINLIST, VEVENT, and VEVENT_HAUDIO T Word Embeddings Numerical 300 features are the vector representation of the word derived using GloVe (Pennington et al., 2014) 1 Predicting lexical complexity in English texts: the Complex… 1165 Clearly, this is a surprising result. Research in psycholinguistics indicates, for exam- ple, that the frequency of a given word (feature groups A, H, I, and J) affects its percep- tion (Segui et al., 1982; Dupoux & Mehler, 1990; Marslen-Wilson, 1990), that word familiarity (feature group E) and frequency affect visual and auditory word recogni- tion (Connine et al., 1990), and that word imageability (feature group G) significantly impacts word reading accuracy and rate of word learning among first and second grad- ers at risk for reading disabilities (Steacy & Compton, 2019). 3 Analysis of features of complex words Further, the word “con- creteness effect” (feature group F) is a well-established concept in psycholinguistics with the tendency of words with tangible physical referents being learned earlier, rec- ognised faster, and recalled with less effort than words with abstract referents (Paivio, 1991; Schwanenflugel, 1991). Schwanenflugel et al. (1988) proposed that abstract words are more difficult to recognise because their interpretation is more reliant on context than is the case for concrete words. Word meaningfulness (feature groups K and L) has been observed to have a positive effect on word recognition (Leeds, 1976) and words with great meaningfulness have been found to be easier to recall than words with less meaningfulness (Kinoshita, 1989). Finally, the age of acquisition of words (feature group M) has been reported to be a predictor of the speed of reading words aloud and lexical decision tasks (in which participants are asked to judge whether par- ticular sequences of characters are real words), with words acquired early in life being responded to more quickly than words acquired later in life (Morrison & Ellis, 2000). We would therefore expect to see more of our features correlating with complexity. This is likely to be a factor of the annotation protocols used in the datasets we analysed and motivates our wider argument in this work that there is a need for new CWI data- sets. The two features that we did identify as showing correlation with word complexity (length and frequency) are both features that are used in almost all of the systems for the shared tasks at CWI–2016 and CWI–2018 as shown in Tables 1 and 2 respectively. This indicates that these features are useful for complexity both in our correlation anal- ysis and in the empirical results of the systems that have submitted using these features. We include this here to show the lack of correlation between sensible features and those datasets. In our next section, we will discuss the deficiencies of these datasets, as well as proposed our specification for an improved CWI dataset. 4.1 Building on previous datasets The previous datasets for CWI have interesting characteristics that make them useful for the CWI task. A quick overview of these datasets is presented in Table 5, where they are compared according to some of their basic features.i The first dataset we have considered is the CWI–2016 dataset, which provides binary annotations on words in context. 9,200 sentences were selected and the anno- tation was performed as described below in Paetzold and Specia (2016a): Volunteers were instructed to annotate all words that they could not under- stand…A subset of 200 sentences was split into 20 sub-sets of 10 sentences, and each subset was annotated by a total of 20 volunteers. The remaining 9,000 sentences were split into 300 subsets of 30 sentences, each of which was annotated by a single volunteer. The annotators were asked to identify any words for which they did not know the meaning. Each annotator had a different proficiency level and therefore will find dif- ferent words more or less complex - giving rise to a varied dataset with different portions of the data reflecting differing complexity levels. Further, each instance in the test set was annotated by 20 annotators, whereas each instance in the training set was annotated by a single annotator. For the test set, any word which was anno- tated as complex by at least one annotator was marked complex (even if the other 19 annotators disagreed). This is problematic as the training data is not representa- tive of the testing data, making it hard for supervised systems to do well on this task. Binary annotation of complexity requires an annotator to impose a subjective threshold on the level at which they transition from considering a word complex as opposed to simple. An annotator’s background, education, etc. may affect where this threshold between complex and simple terms should be set. Further, it is likely that one annotator may find words difficult that another finds simple and vice-versa. Fac- tors such as the annotator’s native language, educational background, dialect, etc. all affect the type of words they are familiar with. In the case of the training data where 20 annotators have all annotated the same instance and any instance with at least one annotation is considered complex, it may be taken that the annotations rep- resent some form of maximum complexity - i.e., that any word is above the lowest possible threshold of complexity. 4 Specification for CWI data protocol In the previous Section we analysed differing features of complexity. In this sec- tion, we first highlight some of the design decisions that were taken in the creation of prior CWI datasets. We continue by proposing a specification, based on our prior analysis, for a new CWI dataset that improves on prior work. Our specification is designed to enable CWI research in areas that have not previously been explored. As well as providing a specification, we also provide a list of features for future datasets to implement in Table 6. 1 3 M. Shardlow et al. 1166 Table 5 CWI Datasets compared according to their features ‘Binary’, ‘Probabilistic’ and ‘Continuous’ refer to the nature of the annotated labels. ‘Context’ refers to the presence of sentential context at annotation time and ‘Multi-Genre’ refers to the dataset drawing from sources across many genres Dataset Binary Probabilistic Continuous Context Multi-Genre CWI–2016 × × × CWI–2018 × × × × Maddela–2018 (Mad- dela and Xu 2018) × Table 5 CWI Datasets compared according to their features ‘Binary’, ‘Probabilistic’ and ‘Continuous’ refer to the nature of the annotated labels. ‘Context’ refers to the presence of sentential context at annotation time and ‘Multi-Genre’ refers to the dataset drawing from sources across many genres Dataset Binary Probabilistic Continuous Context Multi-Genre CWI–2016 × × × CWI–2018 × × × × Maddela–2018 (Mad- dela and Xu 2018) × ‘Binary’, ‘Probabilistic’ and ‘Continuous’ refer to the nature of the annotated labels. ‘Context’ refers to the presence of sentential context at annotation time and ‘Multi-Genre’ refers to the dataset drawing from sources across many genres 4.1 Building on previous datasets However, in the case of the test set where each 1 3 Predicting lexical complexity in English texts: the Complex… 1167 word is annotated by a single annotator, the annotations are harder to interpret. Each instance is personal, reflecting only a single annotator’s judgment. word is annotated by a single annotator, the annotations are harder to interpret. Each instance is personal, reflecting only a single annotator’s judgment. l Moving on from the CWI–2016 dataset, the CWI–2018 dataset also provides binary annotations, which were aggregated to give a ‘probabilistic’ measure of com- plexity. CWI–2018 invited participants to submit results on both the binary com- plexity annotation setting and the probabilistic annotation setting. To collect their data, the organisers of CWI–2018 followed a similar principle as in CWI–2016. Sentences were presented to annotators and the annotators were asked to select any words or phrases that they found to be complex. As in CWI–2016, the annotation task in CWI–2018 was subjective, with potentially low agreement between annota- tors. In the probabilistic setting, at least 20 annotations were collected from native and non-native speakers and each word was given a score indicating the proportion of annotators that found that word to be complex. (i.e., if 10 out of 20 annotators marked the word, then it would be given a score of 0.5). A useful property of this style of annotation is that words are seen on a probablistic scale of complexity. How- ever, the aggregation of binary annotations to give continuous annotations does not necessarily tell us about the complexity of the word itself. Instead it tells us about the annotators, and how many of them will consider a word complex. So, for exam- ple a score of 0.5 does not indicate a median level of complexity (or some sort of neutrality between simple and complex), but instead should be interpreted as indi- cating that 50% of the annotator pool will consider this word complex.i The final dataset we have covered was published in 2018 by Maddela and Xu (2018). We refer to this as Maddela–2018 for brevity. In this dataset, 11 annota- tors who spoke English as a second language were employed to annotate a por- tion of 15,000 words on a 6-point Likert scale with 5–7 annotations being col- lected for each vocabulary item. 4.1 Building on previous datasets Words were presented without context, with the annotators guessing or making assumptions about the sense of the word at annotation. Different annotators may have considered the word to have a different sense or to have been used in a different context. Almost all words are polyse- mous and the different senses of the words are likely to have different levels of complexity - particularly in a coarse grained sense setting (e.g., mean average vs. a mean person). The main effect here is that the varied complexities of the mul- tiple senses and usages of a word are conflated into a single annotation. There is no information as to which word sense the annotators were giving the annotations for, and as such the annotations may be unreliable in cases where a word is used in an uncommon sense. In the Likert-scale type annotation, it is less of an issue that annotators’ opinions will vary than in the binary setting used in CWI–2016 and CWI–2018, as each annotator’s judgment is aggregated on a common con- tinuous scale. This means that the final averaged annotation is reflective of the average complexity that a word might have in a general setting. This is making an assumption that the annotations are normally distributed and that a mean average is valid in this case. A normality test could be used to quantify whether instances are likely to have normal distributions, however with only 5–7 annotations per instance, this may not be reliable. 3 1168 M. Shardlow et al. So far, we have mainly considered complex words. However, the complexity of multi-word expressions is a valuable addition to the CWI literature. MWEs can be considered as compositional or non-compositional. Compositional MWEs (e.g., christmas tree, notice board, golf cart, etc.) take their meaning from the constituent words in the MWE, whereas non-compositional MWEs do not (e.g., hot dog, red herring, reverse ferret,etc.). It is reasonable to assume that complexity will follow a similar pattern to semantics and that compositional MWEs will be dependent on the constituent words to give the complexity of the expression, whereas the complexity of non-compositional MWEs will be independent of the constituent words. In the previous datasets, only the CWI–2018 dataset asked annotators to highlight phrases as well as single words, giving a limit of 50 characters to prevent overreaching. 4.1 Building on previous datasets Par- ticipants in the task were asked to also give complexity annotations for the high- lighted phrases. The system with the highest overall score reported that they found it easier to always consider MWEs as complex in the binary setting (Gooding & Koch- mar, 2018). The work of Maddela and Xu (2018) also considers MWEs. Although they do not annotate for these, instead using average pooling to combine the embed- dings of each token in a phrase into a single embedding, which is then processed in the same way as for single words. As described previously, this assumes composi- tionality, which will not always be the case.f Little treatment has been given to the variations in complexity between differ- ent parts of speech. None of the previous datasets annotate specifically for part of speech except for the CWI–2016 shared task data, which explicitly asks annotators to only highlight content words in the target sentences. Again, this is an important consideration as the roles of nouns, verbs, adjectives and adverbs are different in a sentence and considering them as different entities during annotation will help to better structure corpora. Developers of the existing corpora that span POS tags all suggest the use of POS as a feature for classification—demonstrating its importance in CWI. All of the corpora recognise the importance of a diversity of reader backgrounds in their corpus construction. Native speakers of English might not realise that cer- tain words they know well (depending on their socio-cultural biases) are not com- monly known or may falsely assume that they “find all words easy”. All three of the corpora that we have studied include annotations by non-native speakers. The CWI–2016 dataset used crowdsourcing to get annotations from 400 non-native speakers, the CWI–2018 dataset used native and non-native speakers (collecting at least 10 annotations from each for every instance). The Maddela–2018 data used 11 non-native speakers. The use of non-native speakers for CWI annotation may lead to models trained using these datasets being useful for identifying words which are complex to non–native speakers, but may not be applicable to other groups. All the datasets are heavily biased towards text which has not been professionally edited. The CWI–2016 dataset compiles a number of sources taken from Wikipedia and Simple Wikipedia, the CWI–2018 dataset takes Wikipedia, WikiNews and one formal set of news text sources. 1 3 4.2 Specification In the remainder of this section we will describe some of the qualities of an ideal dataset for CWI. Our recommendations are summarised at the end of this Section in Table 6. This specification is intended to give general purpose recommendations for anyone seeking to develop a new CWI dataset. The key issue with the shared task datasets was the subjectivity that arose dur- ing the annotation process due to their treatment of complexity as a binary notion. When multiple annotators are asked to “mark any complex word” they will each draw on their subjective definition of complexity, and each will choose a different subset of words to be annotated as complex. The annotations that result from this are probabilistic in nature and tell us more about the annotators than the words them- selves. Future datasets should consider providing measures which attempt to give more objectivity and move towards consensus between annotators. Of course, any complexity annotation involving human participants will always rely on the partici- pants subjective knowledge and hence will be dependent on the participants. More objective measures of continuous complexity could be given by asking annotators to mark words on a Likert scale as by Maddela-2018, or by looking at external meas- urements of the ability of people to read the words, such as lexical access time, eye tracking, etc. There are two factors to be considered here when measuring word complexity. One is the perceived complexity of a word (how difficult an annotator estimates a word to be) and the other is the actual complexity of a word (how much difficulty that word presents to the reader) (Leroy et al., 2013). Clearly these are both impor- tant factors in estimating a word’s complexity and although we may expect them to be correlated there is no guarantee they will be aligned. Whereas perceived com- plexity affects how a user may prejudge a text, actual complexity determines the degree with which a reader is likely to struggle. Of course, any measure of complexity which is derived by asking humans to give a subjective judgment of how difficult they find a word is bound to give a meas- ure of perceived rather than actual complexity. 4.1 Building on previous datasets The Maddela-2018 dataset uses the Google Web1T (Brants and Franz 2006) (taken from a large web-crawl) to identify the most fre- quent 15,000 words in English and re-annotates each for complexity. Except for the news texts in the 2018 data, all of these sources are written for informal purposes 1 3 Predicting lexical complexity in English texts: the Complex… 1169 and will contain spelling mistakes, idioms, etc. There has been little prior work exploring cross-genre learning for CWI, however it is unlikely that models trained on such informal text will be appropriate for identifying complexity in formal texts. and will contain spelling mistakes, idioms, etc. There has been little prior work exploring cross-genre learning for CWI, however it is unlikely that models trained on such informal text will be appropriate for identifying complexity in formal texts. 4.2 Specification In fact, measuring actual complexity would only be possible if the human was taken out of the loop altogether (even a set- ting where the reader doesn’t know they are being assessed would rely on a partici- pant’s innately subjective assessment of each word). Any annotation scheme which focusses on continuous complexity judgments is still inviting perceived complexity assessment. By giving more levels to the assessment of complexity (i.e., through a Likert Scale assessment) the annotators have more ability to better record their per- ception of the complexity of the words that are being assessed. The only previous dataset to present continuous annotations (Maddela-2018) did so in the absence of context. Context is key to determining the usage and meaning of a word and the same word used in different contexts can vary greatly in both 1 3 3 M. Shardlow et al. 1170 Table 6 A list of recommended features for future CWI dataset development ID Feature Description 1 Continuous annotations Complexity labels should be on a continuous scale ranging from least to most difficult 2 Context Tokens should be presented in their original contexts of usage 3 Multiple token instances Each token should be included several times in a dataset 4 Multiple token annotations Each token should receive many annotations from dif- ferent annotators 5 Diverse annotators The fluency and background of annotators should be as diverse as possible 6 Multiple genres The text sources used to select contexts should cover diverse genres 7 Multi-word expressions These should be considered alongside single word tokens as part of an annotation scheme Table 6 A list of recommended features for future CWI dataset development ID Feature Description 7 semantics and complexity. Indeed, a familiar word in an unfamiliar context may be just as jarring as a rare word for a reader, who is forced to quickly update their mental lexicon with the new sense of the word they have encountered (e.g. words like base, boss, and fanning in the domains of chemistry, architecture, and geology, respectively). Datasets should include context for any words that annotations are provided for. This will help systems to identify how contextual factors affect the complexity of a given instance. 1 3 4.2 Specification When presenting context, researchers may wish to either ask annotators to mark every word in a sentence according to some complex- ity judgment (dense annotation) or they may wish to pick a target word in a context and ask only for a judgment of the complexity of this word (sparse annotation). In the dense annotation setting, it is likely to be possible to get a much higher through- put of complexity annotations, as the reader will need to only read a sentence once to give multiple annotations, however they are likely to be deeply influenced by the meaning of the sentence, and may struggle to disassociate this from their annotation of complex words themselves. In the sparse annotation setting more contexts are required to give a comparative number of instances compared to the dense annota- tion setting, however the annotation given is more likely to be a direct result of the token itself, rather than the sentence. Any such sparse annotation task should be set up to ensure that an annotator gives judgments based on the word in its context (i.e., that they read and understand the context), rather than just giving a judgment based on the word, as if no context were presented. Given that we are recommending that the data is presented in context, there is a strong argument for presenting multiple instances of each word. If only one instance of a word were presented in context, then it may be the case that this word had a spe- cific usage that was not representative of its general usage. Words are polysemous (Fellbaum, 2010) and this is true both at the coarse grained (tennis bat vs. fruit bat) and narrow grained levels (I love you vs. I love London). The coarse grained level represents different meanings or etymologies, whereas the fine-grained level may 1 3 Predicting lexical complexity in English texts: the Complex… 1171 represent a similar meaning but a different intensity (as in our example). The provi- sion of multiple instances of a word allows both of these factors to be taken into account. This consideration should be held in balance with the need to have a diver- sity of tokens. If a dataset has N instances, constituting P occurrences of R words, then we suggest that R ≫P . 4.2 Specification I.e., the number of total words should be much larger than the number of instances of each word. There is more to be gained in a dataset by having a diversity of tokens than by having many annotations on each token. An interesting separate task would be to annotate many instances of one word form for complexity and analyse how the context affects this. However, this is a secondary task to the one we are presenting here of assessing a word’s complexity. Each instance in a new CWI dataset should be viewed and annotated by multiple people, ideally from a spectrum of ability levels. Multiple annotations have been a common theme of the previous CWI datasets we have discussed, with datasets using as many as 20 annotators per instance. All subsets (train, dev, test) of a dataset should be annotated by the same number of annotators, or at the very least by anno- tators drawn from the same distribution. This ensures that all subsets of the data are comparable. More annotations allows us to capture a wider array of viewpoints from annotators of varying ability levels. If the annotators are carefully selected to ensure they represent a mixture of ability levels then this will lead to annotations that are representative. Consider the case where all annotators are of low ability, or of high ability. The resulting annotations may lead to all words being assigned to the most or least complex categories respectively. This may be desirable in user- or genre-spe- cific settings, but is not desirable for general-purpose LCP. There are two potential approaches to selecting a pool of annotators and distributing annotations between them. Firstly, a researcher may choose to use a fixed number of annotators, such that each annotator views every data instance once. In this setting, each data instance receives N annotations, where N is the number of annotators chosen. Secondly, the annotations may be distributed across a wider pool of annotators, where given N annotators each sees a randomised subset of the data. In this setting, a researcher may choose to control how many instances each annotator sees, ensuring an even distribution of annotators across the data instances. The second approach is more appropriate in a crowd-sourcing setting, where a researcher has diminished ability to control who takes on which job. 4.2 Specification Previous CWI datasets for English have given a strong focus on non-native speak- ers as discussed above. Non-native speakers have learnt English as a foreign lan- guage and the assumption in using them for CWI research is that they will have only learnt a simple subset of English that allows them to get by in daily tasks. However, a non-native speaker may range from a new immigrant who has recently arrived in an English speaking country to someone who has lived there for decades. Fur- ther, both native and non-native speakers may simultaneously be specialists in some domains and novices in other domains. Non-natives may be specialists in domains where natives are not, and vice versa, influencing their complexity judgments. We would suggest, that whilst non-native speakers should not be excluded from the CWI annotation process, they should not be relied upon either. Instead the pool of annotators should be selected for their general ability in English, not for their mother tongue. Indeed, when selecting non-native speakers it may be worth considering 1 3 3 1172 M. Shardlow et al. selecting a variety of mother-tongues, as it is the case that different languages, or language families will have cognates and near-cognates with English, making it eas- ier for non-native speakers of certain backgrounds to understand words in English with roots in their mother tongue. selecting a variety of mother-tongues, as it is the case that different languages, or language families will have cognates and near-cognates with English, making it eas- ier for non-native speakers of certain backgrounds to understand words in English with roots in their mother tongue. Allowing for multiple genres gives more diversity in the type of text studied and allows systems that are trained on it to generalise better to unseen texts. This pre- vents overfitting to one text-type, leading to results being more reliable and hence more interpretable, and ultimately leads to the creation of useful models that can be applied across genres. CWI resources should name the source genres that their texts are taken from and comply with licences placed on those genres. Whilst informal, or amateur text is in abundance (e.g., Twitter or Wikipedia), formal texts should also be considered for CWI such as professionally written news, scientific articles, par- liament proceedings, legal texts or any other such texts that are written for a pro- fessional audience. 4.2 Specification These texts provide well structured language, which is typically targeted at a specific audience and is of a difficult quality for those outside that audi- ence. These texts contain a higher density of complex words and as such are useful examples of the types of text that might need interventions to improve their readabil- ity for a lay reader. As discussed previously, MWEs are an important element in complexity as pre- vious studies have shown that MWEs are generally considered more complex by a user than individual words (Gooding & Kochmar, 2018). Any new CWI dataset should consider incorporating MWEs as they will certainly be useful for future CWI research. When we consider that MWEs can range from simple collocations (White House), to verbal phrases (pick up) and may span 2 or more words, across parts of speech—including phrasal MWEs (it’s raining cats and dogs)—it is clear that the number of potential MWEs to consider is much wider than the number of single tokens. How do we select appropriate MWEs to cover? There is no particular advan- tage to CWI in selecting one category of MWE over another, but we suggest that any dataset covering MWEs explicitly names the types of MWE that it has covered. By incorporating MWEs, a dataset may be used to investigate both the nature of complexity in those MWEs and in the constituent tokens. Strategies for identifying MWEs, as well as the different types of MWEs are beyond the scope of this work and we would direct the reader to the MWE literature (Sag et al., 2002; Schneider et al., 2014) for a more comprehensive treatment of this problem. 5 CompLex 2.0 In this Section we describe a new CWI dataset that we have collected. Our new dataset, dubbed ‘CompLex 2.0’ builds on prior work (CompLex 1.0 Shardlow et al., 2020), in which we collected and annotated tokens in context for complexity. We have described the data collection process for CompLex 1.0 as below and then the annotation process that we undertook to extend this data to CompLex 2.0. CompLex 2.0 covers more instances than CompLex 1.0 and crucially, has more annotations per instance than CompLex 1.0, making it more reliable. We present statistics on 1 3 Predicting lexical complexity in English texts: the Complex… 1173 our new dataset and describe how it fits the recommendations we have made in our specification for new CWI datasets above. CompLex 2.0 was used as the dataset for the SemEval Shared Task on Lexical Complexity Prediction in 2021. our new dataset and describe how it fits the recommendations we have made in our specification for new CWI datasets above. CompLex 2.0 was used as the dataset for the SemEval Shared Task on Lexical Complexity Prediction in 2021. 5.1 Data collection The first challenge in dataset creation is the collection of appropriate source texts. We have followed our specification above and selected three sources that give a suf- ficient level of complexity. We aimed to select sources that were sufficiently differ- ent from one another to prevent trained models generalising to any one source text. The sources that we used are described below. – Bible: We selected the World English Bible translation (Christodouloupoulos & Steedman, 2015). This is a modern translation, so does not contain archaic words (thee, thou, etc.), but still contains religious language that may be complex. The inclusion of this text gives language that combines narrative and poetic text that uses language typically familiar for a reader, yet interspersed with unfamiliar named entities and terms with specific religious meanings (propitiation, atone- ment, etc.). – Europarl: We used the English portion of the European Pariliament proceedings selected from europarl (Koehn 2005). This is a very varied corpus concerning a wide range of issues related to European policy. As this is speech transcription, it is often dialogical in nature in contrast to our other two corpora. Again, the style of text is generally familiar as it is transcriptions of debates. However technical terminology relating to the topics of discussion is present, raising the difficulty level of this text for a reader. – Biomedical: We selected articles from the CRAFT corpus (Bada et al., 2012), which are all in the biomedical domain. These present a very specialised type of language that will be unfamiliar to non-domain experts. Academic articles pre- sent a classic challenge in understanding for a reader and are typically written for a very narrow audience. We expect these texts to be particularly dense with complex words. In addition to single words, we also selected targets containing two tokens. We used syntactic patterns to identify these MWEs, selecting for adjective-noun or noun- noun patterns. We discounted any syntactic pattern that was followed by a further noun to avoid splitting complex noun phrases (e.g., noun-noun-noun, or adjective- noun-noun). We used the StanfordCoreNLP tagger (Manning et al., 2014) to get part-of-speech tags for each sentence and then applied our syntactic patterns to iden- tify candidate MWEs. In addition to single words, we also selected targets containing two tokens. We used syntactic patterns to identify these MWEs, selecting for adjective-noun or noun- noun patterns. 5.1 Data collection We discounted any syntactic pattern that was followed by a further noun to avoid splitting complex noun phrases (e.g., noun-noun-noun, or adjective- noun-noun). We used the StanfordCoreNLP tagger (Manning et al., 2014) to get part-of-speech tags for each sentence and then applied our syntactic patterns to iden- tify candidate MWEs. Clearly this approach does not capture the full variation of MWEs. It limits the length of each to 2 tokens and only identifies compound or described nouns. Some examples of the types of MWE that we identify with this scheme are given in Table 7. Whilst this inhibits the scope of MWEs that are present in our corpus, this does allow us to make a focused investigation on these types of MWEs. Notably, the 1 3 3 3 1174 M. Shardlow et al. Table 7 The varied types of MWEs that can be captured by our syntactic pattern matching. NN indicates a Noun-Noun pattern, whereas JN indicates an Adjective-Noun pattern Pattern MWE Type NN storage box Compound Noun JN ready meal Described Noun JN electric vehicle Compositional NN hot dog Non-compositional JN European Union Named Entity types of MWE that we have identified are those that are the most common (com- pound nouns, described nouns, compositional, non-compositional and named enti- ties). The investigation of other types of MWEs may be addressed by other, more targeted studies following our recommendations for CWI annotation. g g For each corpus we selected words using frequency bands, ensuring that words in our corpus were distributed across the range of low to high frequency. We selected the following eight frequency bands according to the SUBTLEX frequencies in order of least to most frequent (i.e., most to least complex): 2–4, 5–10, 11–50, 51–250, 251–500, 501–1400, 1401–3100, 3101–10000. We excluded the rarest words (those with a frequency of only 1) as well as the most frequent (those above 10,000) in order to ensure that our instances were well-attested content words. As frequency is correlated to complexity (Brysbaert et al., 2011), this ensures that our final corpus will have a range of high and low complexity targets. We chose to select 3000 single words and 600 MWEs from each corpus to give a total of 10,800 instances in our corpus. We selected a representative number of instances from each frequency band to give the desired total number of instances in each corpus. 5.1 Data collection We automatically anno- tated each sentence with POS tags and only selected nouns as our targets, in-keeping with our MWE selection strategy. We allowed a maximum of 5 instances of a token to be selected in each genre (ensuring that contexts were different). This maximises the total number of examples of each instance, whilst still allowing some variation in the selection of tokens. There is a theoretical minimum of 600 instances of sin- gle words and 120 MWEs that could occur in our corpus (each with 5 occurrences in each of the three genres. Table 11 shows that the number of repeated instances is much lower. This is a factor of the stochastic selection procedure that we have employed. We have included examples of the contexts and target words in Table 8. A word which was neither difficult nor easy. fi Words for which an annotator was unclear of the meaning, but may have been able to infer the meaning from the sentence. Words that an annotator had never seen before, or were very unclear. We used the following key to transform the numerical labels to a 0-1 range when aggregating the annotations: 1 →0 , 2 →0.25 , 3 →0.5 , 4 →0.75 , 5 →1 . This allowed us to ensure that our complexity labels were normalised in the range 0–1. We initially employed crowd workers through the Figure Eight platform (for- merly CrowdFlower), requesting 20 annotations per data instance and paying $0.03 per annotation. We selected annotators from English speaking countries (UK, USA and Australia). In addition, we used the annotation platform’s in-built quality control metrics to filter out annotators who failed pre-set test questions, or who answered a set of questions too quickly. After we had collected these results, we further analysed the data to detect instances where annotators had not fully participated in the task. We specifically analysed instances where an annotator had given the exact same annotation for all instances (usually these were all ’Neutral’) and discarded these from our data. We retained any data instance that had at least 4 valid annotations in our final dataset. i This led to the version of the dataset we described as CompLex 1.0. Whilst this dataset evidenced the trends we expected to see, the conclusions we were able to draw from it were weaker than we hoped (Shardlow et al., 2020). The median num- ber of annotators was 7 per instance, and we identified this as an area for improve- ment. The involvement of more annotators would allow more opinions to be expressed, leading to better average judgments. For the second round of annotations we used the Amazon Mechanical Turk plat- form. We used exactly the same data as in the original annotation of CompLex 1.0 and requested new annotations for each instance. We gave the same instructions to annotators regarding the Likert-scale points. As there is no in-built quality control in Mechanical Turk, we opted to release the data in batches (1200 instances at a time). We asked for a further 10 annotations per instance and paid at a rate of $0.03 per annotation. A word which was neither difficult nor easy. We reviewed the annotators work in between batches, rejecting accounts which submitted annotations too quickly, or without correlation to the other annota- tor’s judgments. We also measured the correlation with lexical frequency to ensure that the annotations we were receiving were in the range we expected. This allowed us to gather a further 108,000 annotations on the CompLex data. These new judgments were aggregated with those from CompLex 1.0 to give a new dataset—CompLex 2.0. We used this data to run a shared task on Lexical Complex- ity Prediction at SemEval 2021 (Shardlow et al., 2021). 5.2 Data labelling As has been previously mentioned, prior datasets have focused on either (a) binary complexity or (b) probabilistic complexity. Neither of which give a true representa- tion of the complexity of a word. In our annotation we chose to annotate each word on a 5-point Likert scale, where each point was given the following descriptor: Words which were very familiar to an annotator. Words for which an annotator was aware of the meaning. 1. Very Easy: 2. Easy: 1 3 Predicting lexical complexity in English texts: the Complex… 1175 3. Neutral: A word which was neither difficult nor easy. 4. Difficult: Words for which an annotator was unclear of the meaning, but may have been able to infer the meaning from the sentence. 5. Very Difficult: Words that an annotator had never seen before, or were very unclear. A word which was neither difficult nor easy. 5.3 Corpus statistics The first round of annotations led to an initial version of the Corpus (CompLex 1.0), for which we have shown the statistics originally reported in Table 9. Due to 3 M. Shardlow et al. 1176 Table 8 Examples from our corpus, the target word is highlighted in bold text The field Complexity refers to perceived complexity Corpus Context Complexity Bible This was the length of Sarah’s life Low Biomed […] cell growth rates were reported to be 50% lower […] Low Europarl Could you tell me under which rule they were enabled to extend this item to have four rather than three debates? Low Europarl These agencies have gradually become very important in the financial world, for a variety of reasons Medium Biomed […] leads to the hallmark loss of striatal neurons […] Medium Bible The idols of Egypt will tremble at his presence […] Medium Bible This is the law of the trespass offering High Europarl They do hold elections, but candidates have to be endorsed by the conservative clergy, so dissenters are by definition excluded High Biomed [..] due to a reduction in adipose tissue High The field Complexity refers to perceived complexity the quality control that we employed for this round of annotation, we discarded a large portion of our original judgments and only kept instances with four or more annotations. This is evident in the fact that only 9476 instances out of our original 10,800 are present in this iteration of the corpus. Additionally, the median num- ber of annotators was 7 across our corpus (with the range being from 4 to 20). Retaining only the annotations in which we could be certain of the quality was a difficult choice, as it reduced the amount of data available. However, the mean complexities of the sub-corpora were in line with our expectations. With Biomed- ical text being on average more complex than the other two genres. This led us to undertake our second round of annotation in order to develop CompLex 2.0 ready for the SemEval shared task. We have included statistics on the annotations aggregated from both rounds in Table 10. 513 separate annotators viewed our data, with each annotator seeing on average 542 instances across all rounds of annotation (around 5% of our corpus). We gathered a total of 278,093 annotations, paying $0.03 per annotation. 1 3 5.3 Corpus statistics The average time spent per annotation was 21.61 seconds, which means that we paid our workers at an average rate of 5 US Dollars per hour. The task received reviews indicating that annotators found it to be well paid in comparison to other tasks on the platform. We gathered an average of 25.75 annotations per instance, this is an increase over CompLex 1.0, which only had on average 7 annotations per instance. We expect that by having more annotations per instance, we will have more reliable average estimates of the complexity of each word. We report detailed statistics on our new dataset, CompLex 2.0, in Table 11. We can see that in total 5,617 unique tokens covering single words and multi-word expressions are distributed across 10,800 contexts. Whilst the contexts are split evenly between each genre (3,600 each) the number of repeated words is higher in the Biomed and Bible corpora, with more distinct words occurring in the Europarl corpus. The complexity annotations are low at 0.321 for the entire corpus, indicating 1 3 1 3 Predicting lexical complexity in English texts: the Complex… Predicting lexical complexity in English texts: the Complex… 1177 Table 9 The statistics for CompLex 1.0. We report on the entire corpus and also present a breakdown of statistics by Genre We include statistics on the number of Contexts, the number of Unique Words and the mean Complexity in each partition Genre Contexts Unique words Complexity All 9476 5166 0.394 Europarl 3496 2194 0.390 Biomed 2960 1670 0.407 Bible 3020 1705 0.385 We include statistics on the number of Contexts, the number of Unique Words and the mean Complexity in each partition that the average complexity of words is somewhere between points 2 (0.25—a word which that the annotator was aware of the meaning) and 3 (0.5—A word which was neither difficult nor easy) on our Likert scale. This indicates that annotators gen- erally understood the words in our dataset. The annotations did use the full range of our Likert scale and the dataset contains words of all complexities. We can see from the data that the Biomedical genre was on average more difficult to understand (0.353) than the other genres (0.303 for Europarl and 0.307 for Bible respectively). Multiword expressions are markedly more complex (0.419) than single words (0.302), with the same genre distinctions as in the full data. 5.4 Inter‑annotator agreement Achieving strict adherence to annotation guidelines is difficult in the crowd-sourcing setting as there is little time to train, test or survey annotators. As a result, inter- annotator agreement tends to be lower in this context. We provided some controls as outlined above to ensure that annotators were fully participating in the task and that their annotations aligned with those of other annotators. In our setting, we do not necessarily expect annotators to agree in every case as one may legitimately con- sider a word to be complex, whilst another considers it to be simple. A reasonable expectation is that annotators will provide similar annotations to each other, and that the annotations will mostly fall into one category. We expect the distribution of annotations for one instance to be normally distributed. We have already made this assumption, as we take the mean to give the average complexity. To test this, we used a Shapiro-Wilk test (Shapiro & Wilk, 1965), which gives a number in the range of 0-1 indicating how likely a given distribution is to fol- low the normal distribution. For each of our instances, we perform the test on the annotations for that instance. A higher number indicates that the instance has anno- tations which are more likely to be normally distributed, whereas a low number on this test indicates a non-Gaussian distribution, such as a multi-modal distribution. A histogram of this data is displayed in Fig. 1. This shows that the majority of our data obtains a score between 0.7 and 0.9 according to the Shapiro-Wilk test, with a peak around 0.85. This indicates that our data is generally normally distributed, and hence that annotators generally gave annotations that centered around a mean value. In Table 12 we have shown a number of examples from our corpus that do not follow the distribution that we may have expected. These were infrequent in 1 3 3 1178 M. Shardlow et al. Table 10 Statistics on the round of evaluation undertaken with Mechanical Turk Number of annotators 513 Number of instances 10,800 Number of annotations 278,093 Annotations per Instance 25.75 Instances per annotator 542.09 Time per annotation 21.61 (s) Table 11 The statistics for CompLex 2.0 We report on the entire corpus and also present a breakdown of sta- tistics by Genre and by Subset. 5.4 Inter‑annotator agreement We include statistics on the number of Contexts, the number of Unique Words and the mean Complex- ity in each partition Subset Genre Contexts Unique words Complexity All Total 10,800 5617 0.321 Europarl 3600 2227 0.303 Biomed 3600 1904 0.353 Bible 3600 1934 0.307 Single Total 9000 4129 0.302 Europarl 3000 1725 0.286 Biomed 3000 1388 0.325 Bible 3000 1462 0.293 MWE Total 1800 1488 0.419 Europarl 600 502 0.388 Biomed 600 516 0.491 Bible 600 472 0.377 We report on the entire corpus and also present a breakdown of sta- tistics by Genre and by Subset. We include statistics on the number of Contexts, the number of Unique Words and the mean Complex- ity in each partition We report on the entire corpus and also present a breakdown of sta- tistics by Genre and by Subset. We include statistics on the number of Contexts, the number of Unique Words and the mean Complex- ity in each partition our corpus, but are displayed here to help the reader understand where annotators may have disagreed. In example 1 the simple word ‘heaven’ was given to annota- tors, most of whom assigned it to the Very Easy category. However, 3 annota- tors disagreed with this, assigning it to the Neutral category. Possibly, the annota- tors found the word easy, but the metaphorical usage harder to grasp. Example 2 shows a similar disagreement, albeit around a more difficult word. ‘Election’ is a word that most people living in a democracy will have encountered, yet 5 people felt it was neither easy, nor difficult—placing it in the Neutral category. Our third example, taken from the Biomedical genre, demonstrates a word (Gran- ules) which is considered Easy by 14 annotators, yet is considered Difficult by 4 annotators. Whilst ‘Granules’ is not a particularly rare word, it may be considered complex by some in this instance due to its contextual usage in the biomedical lit- erature. Example 4 shows a word which is specific to biblical language (‘Cubit’). Although the annotations gave a reasonably Gaussian set of annotations (0.848 according to the Shapiro-Wilk statistic), they were split over all 5 potential cate- gories. This is an example of annotators’ previous familiarity with the text. 5.5 CompLex 2.0 features We have presented a corpus that was developed according to the recommendations that we have set out earlier in this work (see Sect. 4.1). Whilst we have made every effort to follow these, practical concerns have led us to pragmatic design decisions that made the development of our corpus feasible. In the following list, we itemise the design decisions that were made during the construction of our corpus and show how these link to the recommendations from Table 6. 1. Continuous annotations: We have implemented this using a Likert Scale as described above. Unlike Maddela-2018 who used a 4-point Likert scale, we chose a 5-point Likert scale to allow annotators to give a neutral judgment. To give final complexity values we took the mean average of these annotations, transforming the complexity labels in the range 0–1. 2. Context: We presented annotations in context to the annotators and explicitly asked annotators to judge a word based on its contextual usage (but not on the context itself). Following (Peirce 1906), we distinguish between word types (the distinct words used in a text, which comprise its vocabulary) and word tokens (the different occurrences or instances of those words throughout the text). There are clear variations in the complexity of different tokens sharing the same word type. For example, the word ‘table’ receives a higher complexity rating in the less common sense of ‘table a motion’ than in the more frequent sense of something being ‘on the table’. 3. Multiple tokens: We presented a maximum of 5 tokens per word type, per genre. This led to 5,617 word types across 10,800 tokens and contexts giving an average density of 1.92 contexts per word type. Although some word types do appear in multiple contexts 3,423 words appear with only a single context. 671 word types feature 5 or more tokens (and contexts). This is a compromise between our desire to include a wide variety of word types in the dataset and to include multiple tokens of each type. A dataset featuring a more rigorous treatment of contexts may reveal the role of context in complexity estimation in a way that our data is not able to. 4. Multiple token annotations: We have described our process of gathering an aver- age of 25.75 annotations per token. 5.4 Inter‑annotator agreement Those who know a cubit is an ancient measure of length will score it on the easier side our corpus, but are displayed here to help the reader understand where annotators may have disagreed. In example 1 the simple word ‘heaven’ was given to annota- tors, most of whom assigned it to the Very Easy category. However, 3 annota- tors disagreed with this, assigning it to the Neutral category. Possibly, the annota- tors found the word easy, but the metaphorical usage harder to grasp. Example 2 shows a similar disagreement, albeit around a more difficult word. ‘Election’ is a word that most people living in a democracy will have encountered, yet 5 people felt it was neither easy, nor difficult—placing it in the Neutral category. Our third example, taken from the Biomedical genre, demonstrates a word (Gran- ules) which is considered Easy by 14 annotators, yet is considered Difficult by 4 annotators. Whilst ‘Granules’ is not a particularly rare word, it may be considered complex by some in this instance due to its contextual usage in the biomedical lit- erature. Example 4 shows a word which is specific to biblical language (‘Cubit’). Although the annotations gave a reasonably Gaussian set of annotations (0.848 according to the Shapiro-Wilk statistic), they were split over all 5 potential cate- gories. This is an example of annotators’ previous familiarity with the text. Those who know a cubit is an ancient measure of length will score it on the easier side 1 3 Predicting lexical complexity in English texts: the Complex… 1179 of the Likert scale, whereas those who have not seen the word before will score it as more difficult. The remaining three examples (5–7) all score similarly highly on the Shapiro-Wilk test, however they have a wide spread of annotations. Again, this is likely due to the familiarity of the annotators with each word. 5.5 CompLex 2.0 features We could have chosen to do fewer annotations in favour of annotating more tokens, however we prioritised having a large num- ber of judgments per token to give a more consistent and representative averaged annotation. 5. Diverse annotators: We did not place many restrictions, or record demographic information regarding our annotators. Doing so may have helped to better under- stand the makeup of our annotations and identify potential biases. We did not 5. Diverse annotators: We did not place many restrictions, or record demographic information regarding our annotators. Doing so may have helped to better under- stand the makeup of our annotations and identify potential biases. We did not 1 3 3 1180 M. Shardlow et al. Fig. 1 A histogram of Shapiro-Wilk’s test statistics, demonstrating the likelihood for each instance that the annotations are normally distributed Fig. 1 A histogram of Shapiro-Wilk’s test statistics, demonstrating the likelihood for each instance t Fig. 1 A histogram of Shapiro-Wilk’s test statistics, demonstrating the likelihood for each instance that the annotations are normally distributed record this information due to the crowd-sourcing setting that we used. This is something for future LCP annotation efforts to consider. f 6. Multiple genres: We have selected three diverse genres with a potential for com- plex language. We deliberately avoided the use of Wiki text as this has been studied widely already in CWI. f 6. Multiple genres: We have selected three diverse genres with a potential for com- plex language. We deliberately avoided the use of Wiki text as this has been studied widely already in CWI. 7. Multi-word expressions: We have included these in a limited form as part of our corpus. The MWEs make up 16.66% of our corpus. We have included these as an interesting area to study and we hope that their inclusion will shed light on the complexity of MWEs. Further studies could focus on specific types of MWE, extending our research. The CompLex 2.0 corpus is designed according to the recommendations we have set out. In particular, we do not record demographic information on our participants and as such cannot make reasonable claims as to the diversity of our annotators. Our corpus is intended as a starting point for future LCP researchers to build on. 6 Predicting categorical complexity We represented words and multiword units in the CWI–2016 (Paetzold & Specia, 2016a), CWI–2018 (Yimam et al., 2018), and the new CompLex 2.0 single word and multiword datasets using features which, on the basis of previous work in lexical simplification (Paetzold, 2016), text readability (Yaneva et al., 2017; Deutsch et al., 2020), psycholinguistics/neuroscience (Yonelinas et al., 2005), and our inspection of the annotated data, we consider likely to be predictive of their complexity (Sect. 3). We used the trees.RandomForest method distributed with Weka (Hall et al., 2009) to build baseline lexical complexity prediction models exploiting the features presented in Sect. 3. In the experiments described in the current Section, we framed the prediction as a classification task with continuous complexity scores mapped to a 5-point scale. The points on these scales denote the proportions of anno- tators who consider the word complex (c): few ( 0 ≤c < 0.2 ), some ( 0.2 ≤c < 0.4 ), half ( 0.4 ≤c < 0.6 ), most ( 0.6 ≤c < 0.8 ), and all ( 0.8 ≤c ≤1). Table 13 displays weighted average F-scores and mean absolute error (MAE) scores obtained by the baseline models in the ten-fold cross validation setting. This table includes statistics on the number of instances to be classified in each dataset. i Table 14 displays the results of an ablation study performed in order to assess the contribution of various groups of features to the word complexity prediction task applied in the four datasets: CWI–2016, CWI–2018, CompLex (single words), and CompLex (multi-words). The feature sets refer to those studied previously in this work in Sect. 3. In the table, negative values of ΔMAE indicate that the features are helpful, reducing the mean absolute error of the classifier. The reverse is true of positive values. Our results indicate that for prediction of lexical complexity in the CWI–2016 dataset, five of the ablated feature groups are useful. Features encoding information about word length and the regularity of the singular/plural forms of nouns, the typi- cal age of acquisition of the words, and the broad syntactic categories of the words improve the accuracy of the classifier, as do word embeddings. i For words in the CWI–2018 dataset, no feature group was found to be particularly useful for prediction of lexical complexity, though a simple model based only on word length information outperformed the default baseline exploiting all features. 5.5 CompLex 2.0 features Using the methodology described in this section, further datasets encoding information about complex words can be annotated, focusing on the remaining open research questions in lexical complexity prediction. 3 3 1 Predicting lexical complexity in English texts: the Complex… 1181 Table 12 Examples of annotations with interesting distributions indicating disagreement among annotators respectively The target word in the context is highlighted in bold. S-W stands for Shapiro-Wilk. Levels of Annotations are Very Easy (VE), Easy (E), Neutral (N), Difficult (D), and V Diffi lt (D) ID Corpus Context Annotations VE E N D VD S-W 1 Bible You will have treasure in heaven 24 1 3 0.423 2 Europarl Election of Vice-Presidents (first, second and third ballots) 19 1 5 0.544 3 Biomed Annexin A7 was isolated as the agent that mediated aggregation of chromaffin granules and fusion of… 14 2 4 0.612 4 Bible Ehud made him a sword which had two edges, a cubit in length; and he wore it under his clothing on his right thigh 2 3 4 12 8 0.848 5 Europarl I therefore wanted to tell you that I inadvertently voted “yes’ in the vote on the Cornelissen report on the first part of recital 0, when I intended to vote “no’ 1 11 3 9 2 0.848 6 Europarl The Rospuda valley is the last peat bog system of its kind in Europe 2 11 3 4 4 0.848 7 Biomed Amyloid burden worsens significantly with age, and by 9 mo, the hippocampus and cortex of untreated mice are largely filled with aggregated peptide 1 8 7 9 2 0.901 M. Shardlow et al. 1182 6 Predicting categorical complexity Again, this may be due to the typically longer MWEs present in the CWI–2018 dataset, which are exclusively labelled as complex. When predicting the lexical complexity of individual words in the CompLex 2.0 data, features encoding information about whether or not the word was archaic, about the regularity of the singular/plural forms of nouns, and about the stress pat- terns of the words were all found to be useful. When considering multiword units (bigrams), a far larger proportion of the feature groups was observed to be useful for lexical complexity prediction. In our ablation study of bigrams, we assigned the bigram the average value of each feature (all of the features were represented numerically, including binary and one hot representations, and none of the features were symbolic). We found that features encoding information about word frequency, whether or not the words were archaic, word length, regularity of singular/plural 3 Predicting lexical complexity in English texts: the Complex… Predicting lexical complexity in English texts: the Complex… 1183 Table 13 Evaluation results of the baseline trees. RandomForest classifier Dataset F-score (weighted average) MAE Instances CWI 2016 0.915 0.04 2237 CWI 2018 0.843 0.0681 11, 949 CompLex 2.0 (single) 0.607 0.1782 7233 CompLex 2.0 (MWE) 0.568 0.2137 1465 forms, standard age of acquisition, broad syntactic category, the word’s status as either archaic, alien, obsolete, colloquial, rare, or standard, the stress pattern of the word, and the occurrence of an INFOBOX element in the Wikipedia entry for the word were all useful predictors of lexical complexity. Averaged word embeddings also improved the accuracy of predictions made by the trees.RandomForest classifier in the CompLex (multi) dataset. forms, standard age of acquisition, broad syntactic category, the word’s status as either archaic, alien, obsolete, colloquial, rare, or standard, the stress pattern of the word, and the occurrence of an INFOBOX element in the Wikipedia entry for the word were all useful predictors of lexical complexity. Averaged word embeddings also improved the accuracy of predictions made by the trees.RandomForest classifier in the CompLex (multi) dataset. i In the CWI–2016 and CWI–2018 datasets, we applied Weka’s attribute (fea- ture) ranking method with the unsupervised Principal Components Attrib- ute Transformer evaluator to the 378 numerical features described previously in Tables 3 and 4 (Sect. 3). Table 15 displays the ten top-ranked groups of features for the four datasets. 6 Predicting categorical complexity The main observations to be drawn from the feature selection study is the usefulness of information related to word familiarity, concreteness, and imageability in all datasets and information from the vector representations of words derived using GloVe (Pennington et al., 2014). These features occur in the systems that participated in the CWI Shared Tasks as shown in Tables 1 and 2. This corrobo- rates our findings in line with previous work. i Interestingly, whereas the results presented previously using a correlation analy- sis did not find psycholinguistic features (Groups E,F,G,K) to be correlative with complexity, the principal component analysis indicates that these features are in fact likely to be useful for prediction in these datasets. These results demonstrate that by using our new data from CompLex 2.0, the features that we expect to correlate well with complexity judgments are more likely to be effective features for classification than when annotations are done in a binary setting as in the CWI–2016 and CWI–2018 datasets. 7 Predicting continuous complexity In our final section, we use the data we have collected to discuss the nature of com- plex words from a different perspective than in Sect. 6. Whereas in the previous Section we converted all labels into a categorical format to allow comparison, in this Section we use the labels assigned to CompLex 2.0 to discuss factors affecting the nature of lexical complexity, and its prediction. We first look at the effects of genre on CWI. We then continue in our exploration to study the distribution of annota- tions, to determine how and when annotators agree on the complexity of a word. 1 3 M. Shardlow et al. 1184 Table 14 Results of feature ablation Table 14 Results of feature ablation Positive numbers represent a higher MAE after the named feature group was removed (hence the feature was helpful), whereas negative numbers represent the opposite. Most deltas are small, indicating minimal effect from many features. Values above 0.001 or below -0.001 are highlighted in bold Ablated CompLex CompLex feature group CWI–2016 CWI–2018 (single) (multi) ΔMAE A +1E-04 0 +0.0002 −0.0002 B +0.0002 +0.0002 −1E-04 −0.0006 C −0.0001 +0.0004 +1E-04 −0.0004 D −0.0001 +0.0001 −1E-04 −0.0002 E, F, G 0 +0.0001 +0.0003 +0.0006 H, I, J +0.0002 +0.0004 +0.0007 +0.0012 M −0.0001 +0.0001 0 −0.001 N 0 0 +1E-04 +0.0005 P −0.0002 +0.0001 +0.0002 −0.0007 Q 0 +0.0001 +1E-04 −0.0002 R +1E−04 0 −1E−04 −0.0009 S +1E−04 +0.0001 0 −1E-04 Linguistic features (A-S) 0 +0.0009 +0.0018 +0.0027 T − 0.0029 +0.002 +0.001 − 0.0065 All but C +0.0093 − 0.0681 +0.0469 +0.0396 Positive numbers represent a higher MAE after the named feature group was removed (hence the feature was helpful), whereas negative numbers represent the opposite. Most deltas are small, indicating minimal effect from many features. Values above 0.001 or below -0.001 are highlighted in bold 1 3 7.1 Prediction of complexity across genres To better understand the effect of text genre on the LCP task we designed the experi- ments described in this Section. For these, we employed a simple linear regression with the features described previously in Sect. 3. We use the single words in the corpus and split the data into training and test portions, with 90% of the data in the training portion and 10% of the data in the test portion. We first created our linear regression using all the available training data and evaluated this using Pearson’s Correlation. We used the labels given to the data during the annotation round we undertook to create CompLex 2.0. The prediction model based on linear regression achieved a score of 0.771, indicating a reasonably high level of correlation between its predictions and the labels of the test set. This result is recorded in Table 16, where we also show the results for each genre. In each case, we have selected only data from a given genre and followed the same procedure as above, splitting into train and test and evaluating using Pearson’s cor- relation. The linear regression model is less closely correlated when making lexical complexity predictions in the Europarl (0.724) and in the Bible data (0.735). This is expected, given the reduction in size of the training data. It is surprising to see that the linear regression model worked better for the Biomedical data than for any other subset (0.784). This may indicate that simple and complex words are more distinct in this corpus and that this distinction can be learnt from a more focused training set. 7.1 Prediction of complexity across genres 1 3 Predicting lexical complexity in English texts: the Complex… Predicting lexical complexity in English texts: the Complex… 1185 Table 15 Results of feature selection (PrincipalComponents) Rank CWI–2016 CWI–2018 CompLex CompLex (single) (multi) 1 E, F, G, K E, F, G E, F, G T (subset) 2 T (subset) E, F, G T (subset) T (subset) 3 H, I, J , T (subset) T (subset) E, F, G 4 T (subset) T (subset) T (subset) T (subset) 5 T (subset) T (subset) D, N, A T (subset) 6 T (subset) D, T (subset) T (subset) T (subset) 7 T (subset) T (subset) T (subset) T (subset) 8 T (subset) T (subset) T (subset) T (subset) 9 T (subset) T (subset) T (subset) T (subset) 10 T (subset) P, T (subset) T (subset) T (subset) To further determine the effects of genre on lexical complexity prediction, we constructed a new linear regression model that was trained and tested using specific genres selected from our corpus. We trained on single genres and tested on each of the other 2 genres, as well as training on a combined subset of 2 genres and testing on the remaining genre. The results for this experiment are shown in Table 17. We were able to build a reliable predictive model for cross-genre complexity prediction in each case. Our results show that there is a drop in performance when training on out-of- domain data, compared to training on in-domain data. This is true across all genres, where a reduction of between 0.119 and 0.297 can be observed in Pearson’s correla- tion. In each genre, the scores improve when training on the other two genres, rather than just on one. This may be due to the effect of multiple genres helping the linear regression to generalise to global complexity effects, rather than overfitting to spe- cific complexity features in one genre. If we were to test our results on an additional genre/domain, we may hope to see that training on three genres (as are present in our corpus) would yield even more generalised results. 7.2 Subjectivity We previously used a Shapiro-Wilk test to demonstrate that our annotations are gen- erally normally distributed. We obtained the mean of each annotation distribution to give a complexity score for each instance in our dataset. An interesting question to ask is how representative these means are of the true complexity of a word. One word may be considered easy by one annotator, yet difficult by another. Factors such as age, education and background may well affect which words a reader is familiar with. We can use the normally distributed annotations to understand this phenom- enon by investigating the standard deviations of the annotations for each instance. We have provided examples from our corpus in Table 18 with both the mean com- plexity and the standard deviation ( 휎 ) of the annotations. The top three rows show 1 3 3 1186 M. Shardlow et al. Table 16 Results of training a linear regression on all the data, and on each genre Subset Correlation All 0.771 Europarl 0.724 Biomed 0.784 Bible 0.735 Table 17 Results of training a linear regression on one genre, or pair of genres and testing on a different genre Train Test Correlation Biomed Europarl 0.542 Bible Europarl 0.484 Biomed + Bible Europarl 0.651 Bible Biomed 0.487 Europarl Biomed 0.630 Bible + Europarl Biomed 0.723 Biomed Bible 0.605 Europarl Bible 0.616 Biomed + Europarl Bible 0.692 examples of high standard deviation, whereas the bottom three rows show examples of low standard deviations. It is clear from the table that annotators generally agree more about words which are less complex, with disagreements tending to happen around the more difficult words. An analysis of the mean complexity and standard deviation of the complexity yields a Pearson’s correlation of 0.621, indicating that these are moderately correlated (disagreement is linearly related to complexity). Table 17 Results of training a linear regression on one genre, or pair of genres and testing on a different genre 1 3 8 Discussion Our work has sought to introduce a new definition of lexical complexity to the research community. Whereas previous treatments of lexical complexity have con- sidered it a binary phenomenon in the Complex Word Identification (CWI) setting, we have extended this definition to lexical complexity prediction (LCP), consider- ing complexity as a continuous value associated with a word. This new definition asks the question of ‘how complex is a word’ rather than ‘is this word complex or not?’. This question allows us to give each token a complexity rating on a continuous scale, rather than a binary judgment. If binary judgments were required, it would be easy to create them using our dataset by imposing a threshold at some point in the data. By imposing thresholds at different points, binary labels can be obtained to suit different subjective definitions of complexity. Further, by implementing multiple thresholds, multiple categorical labels can be recovered from the data. In Sect. 3 we showed that the types of features we would typically expect to cor- relate with word complexity did not show any correlation with the CWI–2016 and 1 3 Predicting lexical complexity in English texts: the Complex… 1187 CWI–2018 datasets. This motivated our analysis of the protocol underlying the annotation of these datasets and our development of a new protocol for CWI anno- tation. In Sect. 6, we were able to show through the use of feature ablation experi- ments that more of the feature sets that we used were relevant to the classification of CompLex 2.0, than were relevant to the annotation of CWI–2016 or CWI–2018. This implies that the annotations in our new dataset are more reflective of traditional measures of complexity. We discussed the existing CWI datasets at length (Sect. 3), culminating in our new specification for LCP datasets in Sect. 4. Whilst we have gone on to develop our own dataset (CompLex 2.0), we also hope to see future work developing new CWI datasets following the principles that we have laid out. Future datasets could focus on multilinguality, multi-word expressions, further genres, or simply extending our analysis to further tokens and contexts. Certainly, we do not see the production of CompLex 2.0 as an end point in LCP research, but rather a starting point for other researchers to build from. 5 https://github.com/MMU-TDMLab/CompLex. 8 Discussion This is why we have included our protocol in detail—in order to ensure the replicability of our work in future research. In moving from binary annotations to Likert-scale annotations, we have pro- vided a new dataset, which gives continuous annotations based on a more objective measure of complexity. The binary setting could also be improved if more objective guidelines were provided to the annotators (e.g., instructions such as “identify words that are appropriate for an adult”, or “identify words that are specific to a domain”, as opposed to “identify words that you find difficult). In our comparison, we are comparing a subjective binary dataset to a (more) objective continuous dataset (of course, our dataset still relies on some degree of annotator interpretation of the Lik- ert scale labels). We do not have the ability to compare an objective binary dataset to our data, as it does not exist to the best of the author’s knowledge, however doing so would likely yield further interesting insights into the differences between continu- ous and binary lexical complexity.i We implemented our specification for a new LCP dataset, following the rec- ommendations established in Sect. 3. This led to the creation of CompLex 2.0. In Sect. 5.5 we have explicitly compared our dataset to the recommendations we made in Table 6, and we would encourage the creators of future LCP datasets to do the same. This will ensure that datasets can be easily evaluated and compared at a fea- ture level. The CompLex 2.0 dataset is available via GitHub5. We have made this data available under a CC-BY licence, facilitating its reuse and reproducibility out- side of our work.i Our new LCP dataset is the first to provide continuous complexity annotations for words in context. The role of context in lexical complexity has not been widely studied and we hope that this dataset will go some way towards allowing research- ers to work on this topic. Indeed, the evidence from our annotations shows that for a single token in multiple contexts, the complexity annotation of that token does vary. Further work is needed to prove that the variation is an effect of the contextual 1 3 1188 M. Shardlow et al. 8 Discussion Table 18 Examples of instances with subjective (wide standard deviation) and certain (narrow standard deviation) annotations Corpus Context Com- plexity 휎 Biomed The first step requires generating a floxed allele in ES cells that will serve as the substrate for subsequent exchanges (RMCE-ready ES cell, Fig. 1) 0.556 0.433 Bible The second came, saying, ’Your mina, Lord, has made five minas 0.433 0.423 Europarl ’Budget support’ refers to the transfer of financial resources from a funding agency outside the partner country’s treasury, under the proviso that the country abide by the agreed conditions governing payments 0.567 0.382 Biomed Similarly, changes in synaptic plasticity due to Ca2+-permeable AMPARs [51,52,60], e.g., in piriform cortex, might alter odor memorization processes 0.975 0.077 Bible Or were you baptized into the name of Paul? 0.000 0.000 Europarl Therefore, I would like to ask, in accordance with the Rules of Pro- cedure, for the matter to be referred to the competent body 0.175 0.118 occurrence, or difference in sense and not due to the stochastic nature of annotations resulting from crowdsourcing. occurrence, or difference in sense and not due to the stochastic nature of annotations resulting from crowdsourcing. Although we gave annotators in our task a 5-point scale ranging from Very Easy to Very Difficult, we chose to aggregate the annotations to give a mean-average for each instance. This makes a fundamental assumption that the distance in continu- ous complexity space between each point on the Likert scale is constant. Obviously, there is no guarantee that such an assumption is true. The danger of this is that anno- tations may be falsely biased towards one end of the scale. For instance, if the dis- tance between Very Easy and Easy is shorter than the distance between Easy and Neutral, then considering these as the same distance will falsely inflate complexity ratings. Another strategy could have been to take the median or mode of the com- plexity annotations to give a final value. The disadvantage of that approach would be that every instance would have an ordinal categorical label instead of a continuous label as we have advocated for. This would be a different problem to the one we have explored, and is left to future research. We used categorical complexity to provide a feature analysis of our dataset and the prior CWI datasets in Sect. 6. 8 Discussion We observed that a number of features were iden- tified as useful for the prediction task, indicating that complexity is a matter of many factors and no single factor can be used to determine a word’s complexity. Inter- estingly, in Table 13, we showed that in the categorical setting the CWI-2018 and CWI-2016 datasets both outperformed the CompLex 2.0 dataset. We are not trying to use the dataset here to demonstrate some superior performance, but rather dem- onstrate a comparative analysis of features that are useful for complexity prediction. This may indicate that systems wishing to return a categorical label (as those used in Sect. 6), could use probabilistic or categorical data for training and get better results than when using our data. Our continuous labels allow us to perform further inter- esting analyses into the nature of complex words as presented in Sect. 7. 1 3 Predicting lexical complexity in English texts: the Complex… 1189 We were able to use our data to show that complexity can be predicted across genres. This is encouraging as our dataset contains three diverse genres, and we can expect that the complexity annotations we have identified will generalise well to other genres. A model trained on all three genres will learn features of complexity that are common to all genres, rather than to any one specific genre. We also dem- onstrated that our instances vary in subjectivity of complexity, with those rated as more complex typically being more subjective. Identifying the factors that make a word subjectively complex would be an interesting line of study, but is left for future research. The ambiguity of a word is likely to play a role in its complexity. Words which are often mistaken for others are more likely to be confused and hence are likely to be rated as more difficult to understand by a reader. Conversely however, there is a well documented direct correlation between polysemy and frequency (i.e., infre- quent words are typically monosemes, whereas frequent words have many senses. See the WordNet entries for ‘run’, ‘bat’, ‘cat’, etc.). It may be hypothesised that ambiguity and frequency need to be jointly taken into account when investigating lexical complexity, with a likely ordering from least to most complex being: (high- frequency, monosemous), (high-frequency, polysemous), (low-frequency, monose- mous), (low-frequency, polysemous). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com- mons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/. 8 Discussion Prior efforts have been undertaken to create sense annotated complexity datasets (Strohmaier et al., 2020), and building upon our research with sense annotations, using the specification given in Sect. 4.2 will lead to fruitful research outcomes. References AbuRa’ed, A., Saggion, H. (2018). LaSTUS/TALN at Complex Word Identification (CWI) 2018 Shared Task. In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Computational Linguistics, New Orleans, United States.i Alfter, D., Pilán, I. (2018). SB@GU at the Complex Word Identification 2018 Shared Task. In Proceed- ings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Asso- ciation for Computational Linguistics, New Orleans, United States.i omputational Linguistics, New Orleans, United States Aroyehun, S. T., Angel, J., Alvarez, D. A. P., Gelbukh, A. (2018). Complex word identification: convolu- tional neural network vs. feature engineering . In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Computational Linguistics, New Orleans, United States. Bada, M., Eckert, M., Evans, D., Garcia, K., Shipley, K., Sitnikov, D., Baumgartner, W. A., Cohen, K. B., Verspoor, K., Blake, J. A., et al. (2012). Concept annotation in the craft corpus. BMC Bioinformat- ics, 13(1), 161. Bingel, J., Schluter, N., Martínez Alonso, H. (2016). CoastalCPH at SemEval-2016 Task 11: The impor- tance of designing your Neural Networks right. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 1028–1033. Association for Computational Linguis- tics, San Diego, California. g Biran, O., Brody, S., Elhadad, N. (2011). Putting it simply: A context-aware approach to lexical simpli- fication. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguis- tics: shortpapers (ACL-2011), pp. 496–501. Portland, Oregon. Bott, S., Rello, L., Drndarevic, B., Saggion, H. (2012). Can Spanish be simpler? LexSiS: Lexical Sim- plification for Spanish. In Proceedings of the 12th International Conference on Intelligent Text Pro- cessing and Computational Linguistics, Lecture Notes in Computer Science (pp. 8–15). Springer, Samos, Greece. Brants, T., Franz, A. (2006). The google web 1t 5-gram corpus version 1.1. LDC2006T13. Brooke, J., Uitdenbogerd, A., Baldwin, T. (2016). Melbourne at SemEval 2016 Task 11: Classifying Type-level Word Complexity using Random Forests with Corpus and Word List Features. In Pro- ceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 975–981. Association for Computational Linguistics, San Diego, California. Brysbaert, M., Buchmeier, M., Conrad, M., Jacobs, A.M., Bölte, J., Böhl, A. (2011). The word frequency effect. Experimental Psychology. f Butnaru, A., Ionescu, R. T. (2018). UnibucKernel: A kernel-based learning method for complex word identification . 9 Conclusion We have demonstrated that previous datasets are insufficient for the task of Complex Word Identification. In fact, the very definition of the task—identifying complex words in a subjective binary setting rather than on an objective continuous scale is at fault. We have advocated for a generalisation of this task to Lexical Complex- ity Prediction and we have provided recommendations for datasets approaching this task. Further to this we have provided a new dataset, CompLex 2.0, which is the first publicly available dataset to provide continuous complexity annotations for words in context. We release the data in full to allow future researchers to join us in this excit- ing task of Lexical Complexity Prediction. 1 3 M. Shardlow et al. 1190 References References In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educa- tional Applications. Association for Computational Linguistics, New Orleans, United States.i Choubey, P., Pateria, S. (2016). Garuda & Bhasha at SemEval-2016 Task 11: Complex Word Identifica- tion Using Aggregated Learning Models. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 1006–1010. Association for Computational Linguistics, San Diego, California. Christodouloupoulos, C., & Steedman, M. (2015). A massively parallel corpus: The bible in 100 languages. Language Resources and Evaluation, 49(2), 375–395. https://doi.org/10.1007/ s10579-014-9287-yf Connine, C., Mullennix, J., Shernoff, E., & Yelen, J. (1990). Word familiarity and frequency in visual and auditory word recognition. Journal of Experimental Psychology: Learning, Memory, and Cogni- tion, 16(6), 1084–1096. Davoodi, E., Kosseim, L. (2016). CLaC at SemEval-2016 Task 11: Exploring linguistic and psycho-lin- guistic Features for Complex Word Identification. In Proceedings of the 10th International Work- shop on Semantic Evaluation (SemEval-2016), pp. 982–985. Association for Computational Lin- guistics, San Diego, California.i De Hertog, D., Tack, A. (2018). Deep Learning Architecture for Complex Word Identification . In Pro- ceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Computational Linguistics, New Orleans, United States. Deutsch, T., Jasbi, M., Shieber, S. (2020). Linguistic features for readability assessment. In Proceed- ings of the Fifteenth Workshop on Innovative Use of NLP for Building Educational Applications, pp. 1–17. Association for Computational Linguistics, Seattle, WA, USA Online. https://doi.org/10. 18653/v1/2020.bea-1.1 1 3 Predicting lexical complexity in English texts: the Complex… 1191 Devlin, S., Tait, J. (1998). The use of a psycholinguistic database in the simplification of text for aphasic readers. Linguistic Databases pp. 161–173.f Devlin, S., Tait, J. (1998). The use of a psycholinguistic database in the simplification of text for aphasic readers. Linguistic Databases pp. 161–173.f Dolby, J. L., Resnikoff, H. L., MacMurray, F.L. (1963). A tape dictionary for linguistic experiments. In Proceedings of the American Federation of information processing societies: Fall Joint Computer Conference, pp. 419 – 423. Spartan Books, Baltimore, MD. Dupoux, E., & Mehler, J. (1990). Monitoring the lexicon with normal and compressed speech: Frequency effects and the prelexical code. Journal of Memory & Language, 29, 316–335. f Fellbaum, C. (2010). Wordnet. In R. Poli, M. Healy, & A. Kameas (Eds.), Theory and applications ontology: Computer applications (pp. 231–243). Amsterdam: Springer. Finnimore, P., Fritzsch, E., King, D., Sneyd, A., Ur Rehman, A., Alva-Manchego, F., Vlachos, A. (2019). References Strong baselines for complex word identification across multiple languages. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp. 970–977. Association for Computational Linguistics, Minneapolis, Minnesota. Strong baselines for complex word identification across multiple languages. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp. 970–977. Association for Computational Linguistics Minneapolis Minnesota Computational Linguistics, Minneapolis, Minnesota. lhooly, K., & Logie, R. (1980). Age-of-acquisition, imagery, concreteness, familiarity, and ambiguity measures for 1,944 words. Behavior Research Methods, 12(4), 396–427. Gillin, N. (2016). Sensible at SemEval-2016 Task 11: Neural Nonsense mangled in ensemble mess. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 963– 968. Association for Computational Linguistics, San Diego, California.i Gooding, S., & Kochmar, E. (2018). CAMB at CWI Shared Task 2018: Complex Word Identification with Ensemble-Based Voting . In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Computational Linguistics, New Orleans, United States. Gooding, S., Kochmar, E. (2019). Complex word identification as a sequence labelling task. In Proceed- ings of the 57th Annual Meeting of the Association for Computational Linguistics, pp. 1148–1153. Association for Computational Linguistics, Florence, Italy. ation for Computational Linguistics, Florence, Italy. Gooding, S., Kochmar, E., Sarkar, A., Blackwell, A. (2019). Comparative judgments are more consistent than binary classification for labelling word complexity. In Proceedings of the 13th Linguistic Anno- tation Workshop, pp. 208–214. Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann, P., & Witten, I. H. (2009). The weka data mining software: An update. SIGKDD Explorations Newsletter, 11, 10–18. https://doi.org/10.1145/ 1656274.1656278 Hartmann, N., & dos Santos, L. B. (2018). NILC at CWI 2018: Exploring Feature Engineering and Fea- ture Learning . In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educa- tional Applications. Association for Computational Linguistics, New Orleans, United States.i Horn, C., Manduca, C., & Kauchak, D. (2014). Learning a lexical simplifier using Wikipedia. In Pro- ceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers, pp. 458–463). Association for Computational Linguistics, Baltimore, Maryland.i g pi g p ceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers, pp. 458–463). References Olkin (Ed.), Contributions to probability and statistics: Essays in Honor of Harold Hotelling. Palo Alto, CA: Stanford University Press. Maddela, M., & Xu, W. (2018). A word-complexity lexicon and a neural readability ranking model for lexical simplification. In Proceedings of the 2018 Conference on Empirical Methods in Natural Lan- guage Processing, pp. 3749–3760. Association for Computational Linguistics, Brussels, Belgium. https://doi.org/10.18653/v1/D18-1410i Malmasi, S., Dras, M., & Zampieri, M. (2016). LTG at SemEval-2016 Task 11: Complex Word Identifi- cation with Classifier Ensembles. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 996–1000. Association for Computational Linguistics, San Diego, California. Malmasi, S., & Zampieri, M. (2016). MAZA at SemEval-2016 Task 11: Detecting Lexical Complex- ity Using a Decision Stump Meta-Classifier. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 991–995. Association for Computational Linguistics, San Diego, California. Manning, C. D., Surdeanu, M., Bauer, J., Finkel, J. R., Bethard, S., & McClosky, D. (2014). The stanford corenlp natural language processing toolkit. In Proceedings of 52nd Annual Meeting of the Associa- tion for Computational Linguistics: System Demonstrations, pp. 55–60. f p g y pp Marslen-Wilson, W. (1990). Activation, competition, and frequency in lexical acess. Cambridge, MA: MIT Press. Martínez Martínez, J. M., & Tan, L. (2016). USAAR at SemEval-2016 Task 11: Complex Word Iden- tification with Sense Entropy and Sentence Perplexity. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 958–962. Association for Computational Linguistics, San Diego, California.f g , g , Morrison, C., & Ellis, A. W. (2000). Real age of acquisition effects in word naming and lexical decisio British Journal of Psychology, 91 Pt. 2(2), 167-180 180. Mukherjee, N., Patra, B. G., Das, D., & Bandyopadhyay, S. (2016). JU_NLP at SemEval-2016 Task 11: Identifying Complex Words in a Sentence. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 986–990. Association for Computational Linguistics, San Diego, California.i g Paetzold, G., & Specia, L. (2016). SemEval 2016 Task 11: Complex Word Identification. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 560–569. Asso- ciation for Computational Linguistics, San Diego, California. Paetzold, G., & Specia, L. (2016). SV000gg at SemEval-2016 Task 11: Heavy Gauge Complex Word Identification with System Voting. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 969–974. Association for Computational Linguistics, San Diego, California.i Paetzold, G. H. (2016). Lexical simplification for non-native english speakers. References Association for Computational Linguistics, Baltimore, Maryland.i p pp p g y Kajiwara, T., & Komachi, M. (2018). Complex word identification based on frequency in a learner corpus . In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applica- tions. Association for Computational Linguistics, New Orleans, United States. Kauchak, D. (2013). Improving text simplification language modeling using unsimplified text data. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers, pp. 1537–1546). Association for Computational Linguistics, Sofia, Bulgaria. i Kauchak, D. (2016). Pomona at SemEval-2016 Task 11: Predicting Word Complexity Based on Cor- pus Frequency. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemE- val-2016), pp. 1047–1051. Association for Computational Linguistics, San Diego, California. Kinoshita, S. (1989). Generation enhances semantic processing? The role of distinctiveness in the gener tion effect. Memory & Cognition, 17, 563–571. f Koehn, P. (2005). Europarl: A parallel corpus for statistical machine translation. In Proceedings of M Summit.i Konkol, M. (2016). UWB at SemEval-2016 Task 11: Exploring Features for Complex Word Identifica- tion. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 1038–1041. Association for Computational Linguistics, San Diego, California.i Kuncheva, L. I., Bezdek, J. C., & Duin, R. P. (2001). Decision templates for multiple classifier fusion: A experimental comparison. Pattern Recognition, 34(2), 299–314. Kuru, O. (2016). AI-KU at SemEval-2016 Task 11: Word Embeddings and Substring Features f Complex Word Identification. In Proceedings of the 10th International Workshop on Semant 1 3 3 1192 M. Shardlow et al. Evaluation (SemEval-2016, pp. 1042–1046). Association for Computational Linguistics, San Diego, California. Kučera, H., & Francis, W. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press.l Leeds, B.G. (1976). Kindergarten children and the influence of letter shapes and meaningfulness of vocabulary as factors influencing word recognition. Technical Report ED136250, Department of Health, Education & Welfare, National Institute of Education.fi Leroy, G., Kauchak, D., & Mouradi, O. (2013). A user-study measuring the effects of lexical simplifica- tion and coherence enhancement on perceived and actual text difficulty. International Journal of Medical Informatics, 82(8), 717–730. Levene, H. (1960). Robust tests for equality of variances. In I. Olkin (Ed.), Contributions to probability and statistics: Essays in Honor of Harold Hotelling. Palo Alto, CA: Stanford University Press. Levene, H. (1960). Robust tests for equality of variances. In I. References Phd thesis, University of Sheffield. fi Paetzold, G. H., Alva-Manchego, F., Specia, L. (2017). MASSAlign: Alignment and Annotation of Com- parable Documents. In The Companion Volume of the IJCNLP 2017 Proceedings: System Demon- strations, pp. 1–4.ii Paetzold, G. H., & Specia, L. (2017). A survey on lexical simplification. Journal of Artificial Intelligence Research, 60, 549–593. Paivio, A. (1991). Dual coding theory: Retrospect and current status. Canadian Journal of Psychology/ Revue canadienne de psychologie, 45(3), 255–287. Paivio, A., Yuille, J. C., & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychology, 76(1, Pt.2), 1–25. 1 3 Predicting lexical complexity in English texts: the Complex… 1193 Palakurthi, A., & Mamidi, R. (2016). IIIT at SemEval-2016 Task 11: Complex Word Identification using Nearest Centroid Classification. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 1017–1021. Association for Computational Linguistics, San Diego, California. Peirce, C. S. S. (1906). Prolegomena to an apology for pragmaticism. The Monist, 6, 492–546. Pennington, J., Socher, R., & Manning, C. D. (2014). Glove: Global vectors for word representation. Proceedings of EMNLP.i Popović, M. (2018). Complex Word Identification using Character n-grams. In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Com- putational Linguistics, New Orleans, United States. p g Quijada, M., & Medero, J. (2016). HMC at SemEval-2016 Task 11: Identifying Complex Words Using Depth-limited Decision Trees. In Proceedings of the 10th International Workshop on Seman- tic Evaluation (SemEval-2016), pp. 1034–1037. Association for Computational Linguistics, San Diego, California. Ronzano, F., Abura’ed, A., Espinosa Anke, L., & Saggion, H. (2016). TALN at SemEval-2016 Task 11: Modelling Complex Words by Contextual, Lexical and Semantic Features. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016, pp. 1011–1016). Association for Computational Linguistics, San Diego, California. Sag, I. A., Baldwin, T., Bond, F., Copestake, A. A., & Flickinger, D. (2002). Multiword expressions: A pain in the neck for nlp. In Proceedings of the Third International Conference on Computational Linguistics and Intelligent Text Processing, CICLing ’02, p. 1-15. Springer, Berlin, Heidelberg.f Schneider, N., Onuffer, S., Kazour, N., Danchik, E., Mordowanec, M. T., Conrad, H., & Smith, N. A. (2014). Comprehensive annotation of multiword expressions in a social web corpus. In Proceed- ings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pp. 455–461. References European Language Resources Association (ELRA), Reykjavik, Iceland.ll ( ) p p p ings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pp. 455–461. European Language Resources Association (ELRA), Reykjavik, Iceland.ll Schwanenflugel, P. J. (1991). Why are abstract concepts hard to understand? In P. J. Schwanenflugel (Ed.), The psychology of word meaning (pp. 223–250). Mahwah, NJ: Erlbaum.l Schwanenflugel, P. J., Harnishfeger, K. K., & Stowe, R. W. (1988). Context availability and lexical dec sions for abstract and concrete words. Journal of Memory and Language, 27(5), 499–520.f Segui, J., Mehler, J., Frauenfelder, U., & Morton, J. (1982). The word frequency effect and lexical acces Neuropsychologia, 20(6), 615–627. Shapiro, S. S., & Wilk, M. B. (1965). An analysis of variance test for normality (complete samples)† . Biometrika, 52(3–4), 591–611. https://doi.org/10.1093/biomet/52.3-4.591 Shardlow, M. (2013). A comparison of techniques to automatically identify complex words. In 51st Annual Meeting of the Association for Computational Linguistics Proceedings of the Student Research Workshop, pp. 103–109. Association for Computational Linguistics, Sofia, Bulgaria.i i Shardlow, M. (2013). The CW corpus: A new resource for evaluating the identification of complex words. In The Second Workshop on Predicting and Improving Text Readability for Target Reader Popula- tions (PITR 2013). Association for Computational Linguistics, Sofia, Bulgaria. i In The Second Workshop on Predicting and Improving Text Readability for Target Reader Popula- tions (PITR 2013). Association for Computational Linguistics, Sofia, Bulgaria. i Shardlow, M., Cooper, M., & Zampieri, M. (2020). CompLex—A new corpus for lexical complexity prediction from Likert Scale data. In Proceedings of the 1st Workshop on Tools and Resources to Empower People with REAding DIfficulties (READI), pp. 57–62. European Language Resources Association, Marseille, France. Shardlow, M., Evans, R., Paetzold, G., & Zampieri, M. (2021). Semeval-2021 task 1: Lexical com- plexity prediction. In Proceedings of the 14th International Workshop on Semantic Evaluation (SemEval-2021).i Sanjay, S. P., & Soman, K. P. (2016). AmritaCEN at SemEval-2016 Task 11: Complex Word Identifica- tion using Word Embedding. In Proceedings of the 10th International Workshop on Semantic Eval- uation (SemEval-2016), pp. 1022–1027. Association for Computational Linguistics, San Diego, California. Steacy, L. M., & Compton, D. L. (2019). Examining the role of imageability and regularity in word read- ing accuracy and learning efficiency among first and second graders at risk for reading disabilities. Journal of Experimental Child Psychology, 178, 226–250. Strohmaier, D., Gooding, S., Taslimipoor, S., & Kochmar, E. (2020). SeCoDa: Sense complexity dataset. References In Proceedings of the 12th Language Resources and Evaluation Conference, pp. 5962–5967. Euro- pean Language Resources Association, Marseille, France. https://aclanthology.org/2020.lrec-1.730 p g g , , p gy g Svartvik, J., & Quirk, R. (Eds.). (1980). Handbook of semantic word norms. Lund: Liver/Gleerups. Svartvik, J., & Quirk, R. (Eds.). (1980). Handbook of semantic word norms. Lund: Liver/Gleerups. 1 3 1194 M. Shardlow et al. Thorndike, E. L., & Lorge, I. (1944). The teacher’s word book of 30,000 words. Teacher’s Colleg Columbia University, New York, NY, USA. Toglia, M. P., & Battig, W. F. (1978). Handbook of semantic word norms. Hillsdale, NJ, USA: Erlbaum. Wani, N., Mathias, S., Gajjam, J. A., & Bhattacharyya, P. (2018). The Whole is Greater than the Sum of its Parts: Towards the Effectiveness of Voting Ensemble Classifiers for Complex Word Identi- fication In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Wani, N., Mathias, S., Gajjam, J. A., & Bhattacharyya, P. (2018). The Whole is Greater than the Sum of its Parts: Towards the Effectiveness of Voting Ensemble Classifiers for Complex Word Identi- fication. In Proceedings of the 13th Workshop on Innovative Use of NLP for Building Educational Applications. Association for Computational Linguistics, New Orleans, United States. Wilson, M. (1988). MRC psycholinguistic database: Machine-usable dictionary, version 2.00. Behavi Research Methods, Instruments, & Computers, 20, 6–10.i Wróbel, K. (2016). PLUJAGH at SemEval-2016 Task 11: Simple System for Complex Word Identifica- tion. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), Wróbel, K. (2016). PLUJAGH at SemEval-2016 Task 11: Simple System for Complex Word Identifica- tion. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 953–957. Association for Computational Linguistics, San Diego, California. tion. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016), pp. 953–957. Association for Computational Linguistics, San Diego, California. 7. Association for Computational Linguistics, San Di Yaneva, V., Orăsan, C., Evans, R., & Rohanian, O. (2017). Combining multiple corpora for readability assessment for people with cognitive disabilities. In Proceedings of the 12th Workshop on Innova- tive Use of NLP for Building Educational Applications, pp. 121–132. Association for Computa- tional Linguistics, Copenhagen, Denmark. https://doi.org/10.18653/v1/W17-5013 Š Yimam, S.M., Biemann, C., Malmasi, S., Paetzold, G., Specia, L., Štajner, S., Tack, A., & Zampieri, M. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References (2018) A report on the complex word identification shared task 2018 In Proceedings of BEA Yimam, S.M., Biemann, C., Malmasi, S., Paetzold, G., Specia, L., Štajner, S., Tack, A., & Zampieri, M. (2018). A report on the complex word identification shared task 2018. In Proceedings of BEA. Š i Yimam, S.M., Štajner, S., Riedl, M., & Biemann, C. (2017). Multilingual and Cross-Lingual Complex Word Identification. In Proceedings of the International Conference Recent Advances in Natural i Yimam, S.M., Štajner, S., Riedl, M., & Biemann, C. (2017). Multilingual and Cross-Lingual Complex Word Identification. In Proceedings of the International Conference Recent Advances in Natural Language Processing, RANLP 2017, pp. 813–822. INCOMA Ltd., Varna, Bulgaria. Yonelinas, A. P., Otten, L. J., Shaw, K. N., & Rugg, M. D. (2005). Separating the brain regions involve in recollection and familiarity in recognition memory. Journal of Neuroscience, 25(11), 3002–300 Zampieri, M., Malmasi, S., Paetzold, G., & Specia, L. (2017). Complex Word Identification: Challenges in Data Annotation and System Performance. In Proceedings of the 4th Workshop on Natural Lan- guage Processing Techniques for Educational Applications (NLPTEA 2017, pp. 59–63). Asian Fed- eration of Natural Language Processing, Taipei, Taiwan.i Zampieri, M., Tan, L., & van Genabith, J. (2016). MacSaar at SemEval-2016 Task 11: Zipfian and Char- acter Features for ComplexWord Identification. In Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016, pp. 1001–1005). Association for Computational Linguis- tics, San Diego, California. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3 3 1 3
https://openalex.org/W4213079781	https://pure.au.dk/ws/files/257333281/Peng_Li_2022_Sound_Decisions_The_Combined_Role_of_Ambient_Noise_and_Cognitive_Regulation_on_the_Neurophysiology_of_Food_Cravings.pdf	English	null	“Sound” Decisions: The Combined Role of Ambient Noise and Cognitive Regulation on the Neurophysiology of Food Cravings	Frontiers in neuroscience	2,022	cc-by	13,918	ORIGINAL RESEARCH published: 16 February 2022 doi: 10.3389/fnins.2022.827021 Keywords: EEG, EDA, cognitive load, emotions, self-regulation, restaurant noise, decision-making, consumer behavior “Sound” Decisions: The Combined Role of Ambient Noise and Cognitive Regulation on the Neurophysiology of Food Cravings Danni Peng-Li1,2,3, Patricia Alves Da Mota1,4, Camile Maria Costa Correa1, Raymond C. K. Chan3,5, Derek Victor Byrne1,2 and Qian Janice Wang1,2 1 Food Quality Perception and Society Team, iSENSE Lab, Department of Food Science, Aarhus University, Aarhus, Denmark, 2 Sino-Danish College (SDC), University of Chinese Academy of Sciences, Beijing, China, 3 Neuropsychology and Applied Cognitive Neuroscience Laboratory, CAS Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China, 4 Department of Clinical Medicine, Center for Music in the Brain, Aarhus University, Aarhus, Denmark, 5 Department of Psychology, University of Chinese Academy of Sciences, Beijing, China Our ability to evaluate long-term goals over immediate rewards is manifested in the brain’s decision circuit. Simplistically, it can be divided into a fast, impulsive, reward “system 1” and a slow, deliberate, control “system 2.” In a noisy eating environment, our cognitive resources may get depleted, potentially leading to cognitive overload, emotional arousal, and consequently more rash decisions, such as unhealthy food choices. Here, we investigated the combined impact of cognitive regulation and ambient noise on food cravings through neurophysiological activity. Thirty-seven participants were recruited for an adapted version of the Regulation of Craving (ROC) task. All participants underwent two sessions of the ROC task; once with soft ambient restaurant noise (∼50 dB) and once with loud ambient restaurant noise (∼70 dB), while data from electroencephalography (EEG), electrodermal activity (EDA), and self-reported craving were collected for all palatable food images presented in the task. The results indicated that thinking about future (“later”) consequences vs. immediate (“now”) sensations associated with the food decreased cravings, which were mediated by frontal EEG alpha power. Likewise, “later” trials also increased frontal alpha asymmetry (FAA) —an index for emotional motivation. Furthermore, loud (vs. soft) noise increased alpha, beta, and theta activity, but for theta activity, this was solely occurring during “later” trials. Similarly, EDA signal peak probability was also higher during loud noise. Collectively, our ﬁndings suggest that the presence of loud ambient noise in conjunction with prospective thinking can lead to the highest emotional arousal and cognitive load as measured by EDA and EEG, respectively, both of which are important in regulating cravings and decisions. Thus, exploring the combined effects of interoceptive regulation and exteroceptive cues on food-related decision-making could be methodologically advantageous in consumer neuroscience and entail theoretical, commercial, and managerial implications. Edited by: Wuke Zhang, Ningbo University, China Reviewed by: Yijie Lai, Shanghai Jiao Tong University, China Mark Selikowitz, Sydney Developmental Clinic, Australia Debo Dong, University of Electronic Science and Technology of China, China Edited by: Wuke Zhang, Ningbo University, China Correspondence: Danni Peng-Li dannipengli@outlook.com Specialty section: This article was submitted to Decision Neuroscience, a section of the journal Frontiers in Neuroscience Received: 01 December 2021 Accepted: 17 January 2022 Published: 16 February 2022 Specialty section: This article was submitted to Decision Neuroscience, a section of the journal Frontiers in Neuroscience Received: 01 December 2021 Accepted: 17 January 2022 Published: 16 February 2022 Citation: Citation: Peng-Li D, Alves Da Mota P, Correa CMC, Chan RCK, Byrne DV and Wang QJ (2022) “Sound” Decisions: The Combined Role of Ambient Noise and Cognitive Regulation on the Neurophysiology of Food Cravings. Front. Neurosci. 16:827021. doi: 10.3389/fnins.2022.827021 February 2022 \| Volume 16 \| Article 827021 1 Frontiers in Neuroscience \| www.frontiersin.org Neurophysiology of Food Cravings Peng-Li et al. Top-Down Cognitive Regulation Top Down Cognitive Regulation In fact, several cognitive strategies have been proposed to facilitate top-down self-regulatory eating behaviors, such as mental imagery (Petit et al., 2017; Zorjan et al., 2020) or episodic future thinking (Dassen et al., 2016; Sun and Kober, 2020). These self-managerial strategies are important components in cognitive-behavioral treatments for treating obesity, eating disorders and addictions (Grilo et al., 2011; Gearhardt et al., 2012) and have been instrumentalized in experimental paradigms (Sun and Kober, 2020). The Regulation of Craving (ROC) task, originally developed by Kober et al. (2010a) attempts to measure the speciﬁc causal eﬀect of regulation strategies on craving for cigarette, alcohol, and/or foods (Kober et al., 2010b; Boswell et al., 2018; Suzuki et al., 2020). The ROC task enables quantiﬁcation and casual inferences of the underlying neural mechanisms of cue-induced cravings from an immediate “now” perspective (anticipatory reward) and a future “later” decision perspective (delayed gratiﬁcation). For instance, using functional Magnetic Resonance Imaging (fMRI), Kober et al. (2010b) demonstrated that cravings for both cigarettes and food decreased when thinking about long-term consequences vs. immediate sensations. These subjective ratings were reﬂected in the blood-oxygen-level- dependent (BOLD) signal which showed that later (vs. now) -trials increased activation in the dorsomedial prefrontal INTRODUCTION cortex (dmPFC), dorsolateral prefrontal cortex (dlPFC), and ventrolateral prefrontal cortex (vlPFC)—all a part of the reﬂective system 2—whereas they decreased activity in brain regions associated with emotion and reward valuation (system 1), i.e., ventral striatum and amygdala. Frontiers in Neuroscience \| www.frontiersin.org Value-Based Decision-Making Our ability to evaluate long-term goals over immediate rewards is encoded in an array of complex computational processes in the brain (Rangel et al., 2008; Levin et al., 2012). These include resisting the impulse of consuming palatable foods, foreseeing the future potential health consequences associated, and at the same time being able to delay one’s gratiﬁcation by valuing the “rational” alternative despite temporal discounting (Volkow and Baler, 2015; Cai et al., 2019). Similarly, an electroencephalogram (EEG) study focusing on event-related potentials (ERPs), showed that a later (vs. now) mindset reduced cravings for high-caloric foods as well as evoked larger late positive potential (LPP) compared to remaining conditions, suggesting that a cognitive focus on negative long- term consequences increases arousal (Meule et al., 2013). Indeed, our choices and decisions ought to fulﬁll both immediate needs and those that are better served for future gains (Motoki et al., 2019). To evolutionarily optimize such balanced utilitarian behaviors, the neural circuitry of human decision- making can simplistically be divided into two neuroanatomically and -functionally distinctive systems—an automatic, emotional, impulsive system (bottom-up) and a deliberate, reﬂective, control system (top-down)—popularly referred to as a fast “system 1” and a slow “system 2” (Evans, 2007; Chen et al., 2018). While the emotional and motivational behaviors of system 1 are manifested in deeper striatal brain structures, the prefrontal cortices govern the cognitive and prospective system 2 functions (Peng-Li et al., 2020c). Bottom-Up Auditory Manipulation Bottom-Up Auditory Manipulation In commercial contexts, consumer researchers and behavioral economists have explored more bottom-up avenues for alleviating the “obesogenic” environment. Such sensory marketing strategies entail changing the so-called choice architecture by nudging consumers toward healthier behaviors through multisensory cues in the environment (Krishna, 2012; Bucher et al., 2016; Seo, 2020). Particularly, auditory contributions to this ﬁeld have in the past decade emerged with numerous studies highlighting the often underestimated power of sound and noise on food choice (Huang and Labroo, 2019), liking (Alamir and Hansen, 2021), attention (Peng-Li et al., 2020b), and perception (Woods et al., 2011). Without cognitive inhibition of system 2, the mere presence of appetitive and salient food cues reinforces anticipatory reward (“wanting”) responses through sensitized neural ﬁring of dopamine, potentially leading to excess food consumption, weight gain, and even addictive behaviors (Burger and Stice, 2012; Schulte et al., 2016; Coccurello and Maccarrone, 2018; Nguyen et al., 2021). p p Louder (vs. softer) ambient noise has consistently shown adverse eﬀects on psychophysiological mechanisms, including increased arousal states (Alvarsson et al., 2010) and cognitive load (Mehta et al., 2012), potentially resulting in poorer decisions and unhealthier food choices (Biswas et al., 2019; Volz et al., 2021; Peng-Li et al., 2022). These phenomena can be explained through the lenses of attentional processes and sensory overload (Doucé and Adams, 2020), whereby “louder noise may diminish the ability to attend to speciﬁc elements of the experience” (Bravo-Moncayo et al., 2020). In fact, attentional distractions have been associated with decreased functional brain connectivity between the inferior frontal gyrus (part of system 2) and the putamen (part of system 1) during goal-directed eﬀort for food rewards (Duif et al., 2020). A complementary mechanism can be reasoned through evidence of sensation transference (Spence and Gallace, 2011), aﬀective priming (Tay and Ng, 2019), or embodied cognition (Zhu and Meyers-Levy, 2005), all in which the ambient sounds physiologically change consumers’ interoceptive, reward, and emotional responses (Salimpoor et al., 2011; Liu et al., 2018; Kantono et al., 2019). Conceptual Framework The evidence highlighted thus far conveys that our food cravings are driven by how we internally are able to regulate our valuation and decisions processes (system 1 or system 2), but at the same time, sensory distractions, such as ambient noise, are also inﬂuencing our cognitive resources and emotional states necessary for controlling and managing these behaviors. This implies that the underlying mechanisms of food-related decision- making are based on an integration of exteroceptive sensory inputs and interoceptive bodily states (Petit et al., 2016; Papies et al., 2020), that translate our somatic signals into feelings of February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 2 Neurophysiology of Food Cravings Peng-Li et al. anticipation, desires, or cravings (Bechara et al., 2005; He et al., 2019). anticipation, desires, or cravings (Bechara et al., 2005; He et al., 2019). quantiﬁcation of cognitive load and related emotional processes, which is not restrained by introspection, verbalization, or any other subjective and self-report limitations. To understand these diﬀerent, yet possibly interacting factors, on a behavioral as well as neural level, the employment of implicit psychophysiological measures can be advantageous. One approach to assess this is through EEG. In addition to the measurement of electrophysiological activity response to a speciﬁc single time-locked stimulus or event as in ERP research (Shang et al., 2018), longer-lasting and continuous functional indices of neural activity are also possible via EEG (Fernandez Rojas et al., 2020; Firestone et al., 2020; Diao et al., 2021). Here, the EEG signal can be decomposed into various frequency spectra representing the oscillatory dynamics in the brain and correlated with certain mental processes (Barlaam et al., 2011; Diao et al., 2017; Aoh et al., 2019). In fact, the power spectral density (PSD) in speciﬁc frequency bands, e.g., theta (4–8), alpha (8–12 Hz), and beta (12–25 Hz), have been associated with various distinct cognitive and emotional states during food viewing (Tashiro et al., 2019; Biehl et al., 2020) and music/noise listening (Gleiss and Kayser, 2014; Chabin et al., 2020). Similarly, measurements based on the sympathetic activity in the peripheral nervous system, including electrodermal activity (EDA), also referred to as galvanic skin response (GSR) can generate complementary biometric information of these aﬀective processes. That is, EDA amplitude ampliﬁcation, thereby higher EDA peak probability has been used to capture increased emotional arousal states. Conceptual Framework With increased sympathetic activity due to interoceptive or exteroceptive triggers, sweat production is elevated, leading to heightened/lowered skin conductance/resistance as an indication of elevated arousal (Kytö et al., 2019; Verastegui-Tena et al., 2019; Pedersen et al., 2021), as determined by the circumplex model of aﬀect (Russell, 1980). In light of the empirical framework, we here investigated the inﬂuences of self-regulatory decision strategy and ambient noise level on cue-induced food cravings by means of neurophysiological activity. We adapted an EEG-based ROC task (Kober et al., 2010b; Meule et al., 2013) in which participants should either focus on the long-term consequences or the immediate rewards of eating high-caloric palatable foods while listening to either soft or loud levels of restaurant noise. We hypothesized that both noise level and decision perspective would aﬀect subjective food cravings as well as objective measures, including EDA and EEG, as measures of emotional arousal/motivation and cognitive load (Figure 1). Speciﬁcally, we expected, as a result of increased emotional arousal and motivation as well as cognitive load, that loud noise would potentially diminish the cognitive resources requisite for more top-down processing, important for especially thinking about future consequences associated with the food. To test this, we examined the PSD in the theta, alpha, and beta frequency bands in the fronto-cortical areas, FAA, as well as EDA during cognitive regulation in the presence of ambient noise and visual food presentation. In the decision and cognitive science literature, both theta and alpha activity in frontal and parietal regions are commonly linked to measures of cognitive load, i.e., the used amount of working memory recourses (Stipacek et al., 2003; Antonenko et al., 2010; Brouwer et al., 2012), including focused attention and sensory processing (Cabañero et al., 2019). Particularly, spectral theta power has been found to increase with sustained concentration and task diﬃculty (Gevins and Smith, 2003), while alpha oscillatory activity has been associated with alertness (Kamzanova et al., 2014) and cognitive fatigue (Borghini et al., 2012). Likewise, a large body of evidence suggests that augmented PSD in the beta frequency band is related to active and analytical thinking (Zhang et al., 2008) as well as short-term memory (Palva et al., 2011) and mental workload (Coelli et al., 2015). Conceptual Framework Of course, delta and gamma band power have also been explored in the context of human behavior (Posada-Quintero et al., 2019), yet they are less related to cognitive and mental workload in decision research (Fernandez Rojas et al., 2020). MATERIALS AND METHODS j Instead, frontal lateralization, commonly referred to as frontal asymmetry (FA; Ramsøy et al., 2018), especially in the alpha frequency range, FA has been employed as an index of mental engagement, reward anticipation, and incentive salience and shown to converge with BOLD activity in frontal cortices (Gorka et al., 2015). In particular, greater right (vs. left) frontal hemispheric alpha power is indexed by a positive frontal alpha asymmetry (FAA) score, denoting emotional motivation and approach, whereas a negative FAA score is linked to avoidance and withdrawal behavior (van Bochove et al., 2016; Fischer et al., 2018). Preliminary evidence even suggests that FAA functions as a potential biomarker for aﬀective neuromodulation (Sun et al., 2017). FAA might therefore be a useful measure for studying aﬀective states and cognitive processes in response to multisensory stimuli. 1https://aucobe.sona-systems.com/default.aspx?logout=Y Frontiers in Neuroscience \| www.frontiersin.org Participants Thirty-seven healthy Danish university students aged 18–35 years were recruited through the Sona recruitment system at the Cognition and Behavior (COBE) Lab, Aarhus University, Denmark.1 The choice of sample size was based on previous EEG literature employing similar designs (n = 25; Meule et al., 2013; n = 28; Biehl et al., 2020; n = 19; Tashiro et al., 2019). As this is the ﬁrst study implementing these conditions/manipulations, we computed a hypothetical power calculation in G∗power (Faul et al., 2009). This yielded a required sample size of at least 28 participants at a power of 0.95, eﬀect size of 0.1, and α of 0.05. All participants fulﬁlled the screening criteria and reported having a normal or corrected-to-normal hearing, normal or corrected-to-normal vision without color blindness, no food allergies, no dietary restraints, and no cardiovascular In short, EEG frequency patterns can be an excellent tool and for measuring the underlying brain dynamics of food- related and managerial decision-making processes. Through spectral analyses, it oﬀers an implicit, objective, and nuanced 1https://aucobe.sona-systems.com/default.aspx?logout=Y February 2022 \| Volume 16 \| Article 827021 3 Peng-Li et al. Neurophysiology of Food Cravings FIGURE 1 \| Conceptual framework. Exploring the effects of top-down cognitive strategy (now vs. later decision perspective) and bottom-up nudging strategy (soft vs. loud ambient restaurant noise) on food cravings by means of cognitive load (EEG), emotional motivation (EEG), and emotional arousal (EDA). GURE 1 \| Conceptual framework. Exploring the effects of top-down cognitive strategy (now vs. later decision perspective) and bot . loud ambient restaurant noise) on food cravings by means of cognitive load (EEG), emotional motivation (EEG), and emotional aro ork. Exploring the effects of top-down cognitive strategy (now vs. later decision perspective) and bottom-up nudging strategy (soft e) on food cravings by means of cognitive load (EEG), emotional motivation (EEG), and emotional arousal (EDA). trial, a jittered 2,000–2,400 ms ﬁxation cross was inserted. We implemented 60 diﬀerent trials (30 now-trials and 30 later-trials) per experimental block, which was repeated for each of the two sound conditions (soft noise vs. loud noise), resulting in a total of 120 trials in the experiment. Trials were presented in a randomized order and blocks were counterbalanced across participants. The adapted ROC task was programmed in the iMotions software (Copenhagen, Denmark)2. or neurological diseases. Participants One participant was omitted from the analysis due to unacceptable data quality, resulting in a valid sample size of 36 (mean age ± SD = 24.22 ± 3.59 years; mean BMI ± SD = 23.52 ± 3.90 kg/m2; 50% females) all of whom provided written informed consent. The study was approved by the Aarhus University Ethics Committee (approval number: 2020-0184772) and conducted in accordance with the ethical standards laid out in the Declaration of Helsinki. All participants were compensated monetarily for their participation (250 DKK). 2https://imotions.com Regulation of Craving Task g g The ROC task experimentally measures the speciﬁc causal eﬀect of regulation strategies and self-management on craving, as well as allows to study its underlying neural mechanisms. The original ROC used images of cigarettes and unhealthy foods to induce cravings among cigarette smokers (Kober et al., 2010a). In our adapted version, we exclusively focused on high-calorie food items as craving cues. During each trial of the adapted ROC task (Figure 2), participants were exposed to one of these cues, preceded by the instruction to follow one of two decision perspectives: “now”—focus on the immediate sensations and feelings associated with consuming the food (e.g., it will taste good and satisfy my cravings), or “later”—focus on the long-term negative consequences associated with repeated consumption (e.g., it will increase my risk for weight gain and heart disease). Participants were then asked to rate their craving for the speciﬁc food they just saw (“how much do you crave this food?”), using a 1 (not at all) to 5 (very much) visual analog scale (VAS). The now or later instructions were presented for 3,000 ms and the subsequent food image for 5,000 ms. Between each The 5-item Self-Regulation of Eating Behavior Questionnaire (SREBQ) is a measure of eating self-regulatory capacity (Kliemann et al., 2016). The SREBQ assesses people’s capacity to control and manage their eating behavior in order to achieve and/or maintain their eating intentions. We adapted the original SREBQ into a Danish version using back-translation. The total score cut-oﬀpoints include < 2.8 = low self-regulation, 2.8– 3.6 = medium self-regulation > 3.6 = high self-regulation. Frontiers in Neuroscience \| www.frontiersin.org Auditory Stimuli Two versions of a restaurant noisescape (chattering and tableware noises) retrieved from Freesound4 were used for the study. The volume level of the noisescape was manipulated based on the Loudness Unit Full Scale (LUFS) by the European Broadcast Union (EBU) standards (European Broadcast Union, 2016). To attain a soft volume version, the noisescape was decreased to approximately –30 LUFS, while the loud version was increased to approximately –4 LUFS via Logic Pro Version FIGURE 2 \| The adapted ROC task. Before each trial, a jittered inter-trial interval (ﬁxation cross) is presented for 2–2.4 s. Then either a now or later cue (nu or senere in Danish) is shown for 3 s, followed by 5 s exposure of a high-calorie food item. Finally, participants rate how much they want the presented food on a VAS from “not at all” to “very much.” Either soft or loud noise is played in the background throughout the entire block. 10.6.1 (Apple Inc.). This was done to ensure the sound intensity (dB) matched 50–55 dB (soft) and 70–75 dB (loud) after audio calibration. The volume levels were chosen based on prior research, which has indicated sound at 80 dB leads to negative aﬀect and even loss of hearing, and sound below 50 dB is often not detected (Witt, 2008). Furthermore, previous food- sound studies have used sound/noise levels in similar ranges (Woods et al., 2011; Biswas et al., 2019). The two noisescapes were ﬁrst validated in a separate online test (N = 91) in which participants listened to each version and rated them in terms of relaxation/arousal on a VAS from 1 to 9. Soft restaurant noise (mean rating ± SD = 4.27 ± 2.25) was expectably perceived as being more relaxing (vs. arousing) compared to loud restaurant noise (mean rating ± SD = 7.49 ± 1.04). The ﬁnal noisescapes used for the study can be heard at: https://soundcloud.com/ danni-peng-li/sets/eeg-roc-t-sound-study. daylight lamps (E27/55W) were used to create optimal lighting conditions. The lens angle was approximately 45◦, the distance from center plate to center tripod was 39.5 cm, and the height of the center of the camera on the tripod was 38 cm to resemble the viewing of a plate of food on a table during mealtime. Each food was presented on a white plate with a diameter of 17.0 cm. A light gray background was chosen to ensure suﬃcient contrast between plate and background. To standardize the background, MeVisLab (MeVis Medical Solutions AG, Bremen, Germany) and the open-source registration software Elastix3 were used (Klein et al., 2010). Each plate was segmented, registered on a standardized background from one image, and smoothened on the plate edges. The complete photographing protocol is described in Charbonnier et al. (2016). 3http://elastix.isi.uu.nl/ 4https://freesound.org Visual Stimuli Thirty high-resolution standardized high-caloric food images from the Full4Health Image Collection (Charbonnier et al., 2016) were selected for the current study (meancalorie = 384 kcal/100 g; meanfat = 21 g/100 g). The images were balanced in terms of taste, such that 15 images were categorized as sweet food items and 15 as savory food items (Table 1). The images were taken in a closed 60 × 60 × 60 cm cubic photo tent. Two February 2022 \| Volume 16 \| Article 827021 4 Peng-Li et al. Neurophysiology of Food Cravings FIGURE 2 \| The adapted ROC task. Before each trial, a jittered inter-trial interval (ﬁxation cross) is presented for 2–2.4 s. Then either a now or later cue (nu or senere in Danish) is shown for 3 s, followed by 5 s exposure of a high-calorie food item. Finally, participants rate how much they want the presented food on a VAS from “not at all” to “very much.” Either soft or loud noise is played in the background throughout the entire block. Frontal Alpha Asymmetry (FAA) = ln(αF4 αF3 ) After ensuring that participants understood the procedure, they initiated the two counterbalanced experimental blocks (conditions) of the adapted ROC-task—one block with soft ambient restaurant noise and one block with loud ambient restaurant noise—with a 5 min break between blocks and an optional break within each block. The adapted ROC task was followed by a manipulation check, i.e., arousal, valence, and distraction ratings of the noisescapes on a 9-point Signal Processing g g EEG data were collected from 32 Ag/AgCl electrodes (Fp1, Fz, F3, F4, FT9, FC5, FC7, C3, T7, TP9, CP5, CP1, Pz, P3, P7, O1, Oz, O2, P4, P9, TP10, CP6, CP2, Cz, C4, T8, FT10, FC6, FC2, F4, F8, Fp2) placed according to the 10–20 system using actiCap (Brain Products GmbH, Gilching, Germany) with a sampling rate of 500 Hz. Raw EEG data were ﬁltered (Butterworth) with a zero phase- lag band-pass ﬁlter [0.5–100 Hz] and a zero phase-lag notch ﬁlter (50 Hz), re-referenced to the mastoid reference electrode placed at TP9. Artifacts were then rejected using an artifact threshold [120 µV] based on the absolute signal value. Power spectra analysis was computed using Fast Fourier Transform (FFT; Welch method; Welch, 1967), by splitting pre-processed data into 1-s time windows with an overlap of 50% and submitted to the FFT, resulting in one power spectrum per 0.5 s. Finally, theta, alpha, and beta activities were calculated by averaging the power spectral density within the standard power bands: theta [4–8 Hz], alpha [8–12 Hz], and beta [12–25 Hz] (Figure 3B). We focused on a hypothesis-based region of interest (ROI) by clustering the frontal electrodes (Fp1, Fz, F3, F4, FT9, FC5, FC7, FT10, FC6, FC2, F4, F8, Fp2; Figure 3A). This electrode clustering was chosen based on previous literature showing various cognitive processes related to the multiple frontal regions as described in the “Introduction” section as well as to avoid loss in statistical power (Moazami-Goudarzi et al., 2008). Furthermore, FAA scores were computed using two frontal electrodes (F3 and F4) on each hemisphere using the formula according to Allen et al. (2004): Frontal Alpha Asymmetry (FAA) = ln(αF4 αF3 ) Frontal Alpha Asymmetry (FAA) = ln(αF4 αF3 ) EDA data was collected from two analog electrode channels placed on the tip of the ﬁngers using a Shimmer3 GSR+ (Shimmer Sensing, Dublin, Ireland). The phasic signal was extracted using a median ﬁlter over a time window of 8,000 ms, and a low-pass Butterworth ﬁlter with a cutoﬀfrequency of 5 Hz was applied to the phasic signal. Peak onset thresholds [0.01 µS] and oﬀset thresholds [0 µS] were then detected on the phasic signal. EDA peak amplitude threshold was set at 0.005 µS with a minimum peak duration of 500 ms. All physiological measures were enclosed to a time window of 5 s, i.e., during food presentation in order to capture audiovisual stimulations of food and noise. Signal processing steps for EEG and EDA were carried out in iMotions through an integrated R algorithm. while EEG and EDA electrodes were applied while checking signal quality in the iMotions software. No natural light entered the room (i.e., only artiﬁcial LED light). To reduce movement artifacts participants rested their heads on a chinrest attached to the table. During the paradigm introduction, participants were instructed to minimize head movements throughout the recordings. They also rated how hungry they were on a 9- point VAS. They then completed 4 practice trials to familiarize themselves with the task. After ensuring that participants understood the procedure, they initiated the two counterbalanced experimental blocks (conditions) of the adapted ROC-task—one block with soft ambient restaurant noise and one block with loud ambient restaurant noise—with a 5 min break between blocks and an optional break within each block. The adapted ROC task was followed by a manipulation check, i.e., arousal, valence, and distraction ratings of the noisescapes on a 9-point while EEG and EDA electrodes were applied while checking signal quality in the iMotions software. No natural light entered the room (i.e., only artiﬁcial LED light). To reduce movement artifacts participants rested their heads on a chinrest attached to the table. During the paradigm introduction, participants were instructed to minimize head movements throughout the recordings. They also rated how hungry they were on a 9- point VAS. They then completed 4 practice trials to familiarize themselves with the task. Frontiers in Neuroscience \| www.frontiersin.org Design and Procedure To control for possible hunger eﬀects, participants were asked to fast for 2 h (i.e., no food intake but water intake was allowed) and refrain from consuming alcoholic drinks for 24 h prior to the study (Frank et al., 2010; Hume et al., 2015; Zhang and Seo, 2015). On testing days (between 9 am to 5 pm), participants arrived at the laboratory for a 1.5 h session where they were informed about the study procedure and provided written informed consent. Participants were seated 70 cm from the HP EliteDisplay E243i, 24,” 16:10 monitor (screen resolution of 1,920 × 1,080 pixels), February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 5 Peng-Li et al. Neurophysiology of Food Cravings TABLE 1 \| Calorie and fat content per 100 g of the 30 food images included in the study. Food item Taste category Calorie (kcal/100 g) Fat (g/100 g) Image no. Potato crisps (natural) Savory 541 33.5 1 Spring rolls Savory 181 8.2 10 Chicken nuggets Savory 272 17.1 12 French fries Savory 306 14.3 16 Nacho-cheese tortilla chips Savory 487 22.3 24 Pepper potato crisps Savory 544 33.0 122 Croissants Savory 424 23.0 130 Wotsits cheesey (chips) Savory 547 33.0 185 Pizza Bolognese Savory 234 9.5 245 Paprika chips Savory 544 33.0 316 Cheese burgers Savory 246 12.0 317 Pita with doner Savory 218 14.0 318 Turkish pizza with doner Savory 233 10.0 319 Pizza margarita Savory 251 12.3 321 French fries with ketchup Savory 268 11.9 322 Donuts with icing Sweet 416 27.8 25 Chocolate chip cookies Sweet 500 25.0 26 Milk chocolate Sweet 546 32.5 32 Chocolate nuts Sweet 584 42.1 36 Brownies Sweet 401 20.0 43 Whipped cream pie Sweet 350 25.0 44 Mini donuts Sweet 358 21.1 100 Pancakes Sweet 196 4.9 101 Syrup wafﬂes Sweet 473 19.3 109 Cake with chocolate Sweet 450 25.0 112 Strawberry pie Sweet 205 11.0 117 Cake Sweet 424 23.9 118 Round pastry/danish Sweet 315 9.0 289 Knoppers Sweet 548 33.4 302 Prince biscuits Sweet 469 17.0 304 Average 384 21 Image no. refers to the Full4Health Image Collection numbering (Charbonnier et al., 2016). TABLE 1 \| Calorie and fat content per 100 g of the 30 food images included in the study. VAS, as well as completion of the SREBQ. Finally, demographic information was collected. VAS, as well as completion of the SREBQ. Design and Procedure Finally, demographic information was collected. VAS, as well as completion of the SREBQ. Finally, demographic information was collected. Data Analysis All physiological and behavioral data were imported and analyzed in R version 4.0.2 for Mac OS. A manipulation check was performed using a pairwise t-test based on VAS ratings to ensure that the two soundscapes were in fact perceived diﬀerently in terms of arousal (1 = very relaxing; 9 = very arousing), valence (1 = very pleasant; 9 = very unpleasant), and distraction (1 = not distracting at all; 9 = very distracting). To investigate the eﬀects of ambient noise and cognitive regulation strategy on EEG, EDA, and self-reported cravings, February 2022 \| Volume 16 \| Article 827021 6 Peng-Li et al. Neurophysiology of Food Cravings FIGURE 3 \| Illustration of (A) channel locations of the 32 electrodes with the frontal ROI highlighted, including Fp1, Fz, F3, F4, FT9, FC5, FC7, FT10, FC6, FC2, F4, F8, Fp2, and (B) example topographical maps of theta (4–8 Hz), alpha (8–12 Hz), and beta (12–25 Hz) power band activity across conditions. FIGURE 3 \| Ill t ti f (A) h l l ti f th 32 l t d ith th f t l ROI hi hli ht d i l di FIGURE 3 \| Illustration of (A) channel locations of the 32 electrodes with the frontal ROI highlighted, including Fp1, Fz, F3, F4, FT9, FC5, FC7, FT10, FC6, FC2, F4, F8, Fp2, and (B) example topographical maps of theta (4–8 Hz), alpha (8–12 Hz), and beta (12–25 Hz) power band activity across conditions. we carried out generalized linear mixed models (GLMMs) via the glmer()-function of the lme4 package. The GLMMs account for the hierarchical structure, non-independence of observations from individual participants in the repeated measure design, and to satisfy the normality assumptions without transformation. EEG and craving data were ﬁtted using a Gaussian distribution with the restricted maximum likelihood (REML) method (Heller et al., 2016), while EDA peak detection was ﬁtted using Poisson distribution (Bolker et al., 2009). In all models, the independent variables were noise level (soft vs. loud) and decision perspective (now vs. later), which were coded as ﬁxed eﬀects. Participant ID entered the model as a random eﬀect. Furthermore, we controlled for possible confounds by adding BMI, hunger status, and SREBQ scores as covariates to the models. Data Analysis The dependent variables of interest included frontal theta power, frontal alpha power, frontal beta power, FAA, EDA peaks, and food craving. Omnibus tests were carried out to test the main Finally, to theorize our conceptual model, we computed four conjoint multiple mediation analyses using the lavaan structural equation modeling package (Rosseel, 2012). Noise level and decision perspective, respectively, entered the models as the binary independent/exogenous variables, craving as the dependent/endogenous variable, and measures of cognitive load (frontal theta power, frontal alpha power, and frontal beta power) as well as emotional arousal (EDA) and emotional motivation (FAA) as the mediators. The multiple mediation analyses were carried out using bootstrapping procedure with the DWLS estimator for 1,000 bootstrapped samples. Data Analysis However, none of the covariates contributed signiﬁcantly to any of the models, and as we did not have any a priori hypotheses regarding these factors, they were therefore removed from the analyses [BMItheta: F(1, 34) = 0.39; p = 0.538; BMIalpha: F(1, 34) = 0.26; p = 0.614; BMIbeta: F(1, 34) = 0.21; p = 0.653; Hungertheta: F(1, 34) = 1.45; p = 0.237; Hungeralpha: F(1, 34) = 0.33; p = 0.567; Hungerbeta: F(1, 34) = 3.35; p = 0.076; SREBQtheta: F(1, 34) = 0.11; p = 0.738; SREBQalpha: F(1, 34) = 0.53; p = 0.470; SREBQbeta: F(1, 34) = 0.46; p = 0.502]. The dependent variables of interest included frontal theta power, frontal alpha power, frontal beta power, FAA, EDA peaks, and food craving. Omnibus tests were carried out to test the main eﬀects and interactions between the ﬁxed independent variables. If a signiﬁcant interaction was indicated by the GLMM, Tukey’s HSD post hoc tests were performed to explore the corrected pairwise comparisons. we carried out generalized linear mixed models (GLMMs) via the glmer()-function of the lme4 package. The GLMMs account for the hierarchical structure, non-independence of observations from individual participants in the repeated measure design, and to satisfy the normality assumptions without transformation. EEG and craving data were ﬁtted using a Gaussian distribution with the restricted maximum likelihood (REML) method (Heller et al., 2016), while EDA peak detection was ﬁtted using Poisson distribution (Bolker et al., 2009). In all models, the independent variables were noise level (soft vs. loud) and decision perspective (now vs. later), which were coded as ﬁxed eﬀects. Participant ID entered the model as a random eﬀect. Furthermore, we controlled for possible confounds by adding BMI, hunger status, and SREBQ scores as covariates to the models. However, none of the covariates contributed signiﬁcantly to any of the models, and as we did not have any a priori hypotheses regarding these factors, they were therefore removed from the analyses [BMItheta: F(1, 34) = 0.39; p = 0.538; BMIalpha: F(1, 34) = 0.26; p = 0.614; BMIbeta: F(1, 34) = 0.21; p = 0.653; Hungertheta: F(1, 34) = 1.45; p = 0.237; Hungeralpha: F(1, 34) = 0.33; p = 0.567; Hungerbeta: F(1, 34) = 3.35; p = 0.076; SREBQtheta: F(1, 34) = 0.11; p = 0.738; SREBQalpha: F(1, 34) = 0.53; p = 0.470; SREBQbeta: F(1, 34) = 0.46; p = 0.502]. Manipulation Check In terms of arousal, the loud noise (mean rating ± SD = 7.22 ± 1.37) compared to the soft noise (mean rating ± SD = 3.99 ± 1.71) was perceived as being signiﬁcantly more arousing [vs. relaxing; t(35) = 10.98; p < 0.001]. For valence, the soft noise (mean rating ± SD = 4.03 ± 1.60) compared to February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 7 Peng-Li et al. Neurophysiology of Food Cravings Figure 5C]. Finally, for FAA, the GLMM did not detect any signiﬁcant interaction, but a main eﬀect of decision perspective was observed with FAA being higher in the later decision perspective condition [F(1, 4222) = 6.08; p = 0.014; Table 3 and Figure 5D]. the loud noise (mean rating ± SD = 6.88 ± 1.64) was likewise perceived as being signiﬁcantly more pleasant [vs. unpleasant; t(35) = 8.53; p < 0.001]. Finally, with regard to distraction, the loud noise (mean rating ± SD = 7.57 ± 1.34) compared to the soft noise (mean rating ± SD = 3.73 ± 1.81) was perceived as being signiﬁcantly more distracting [t(35) = 12.86; p < 0.001]. Behavioral Analysis The EDA-based GLMM did not detect any signiﬁcant interaction, but a main eﬀect of noise level was observed with a higher probability of EDA peak threshold during loud (vs. soft) noise [z(4122) = 3.27; p = 0.001; Table 4 and Figure 6]. The GLMM did not detect any signiﬁcant interaction, but a main eﬀect of decision perspective was observed with food cravings being reportedly signiﬁcantly stronger in now (vs. later) -trials [F(1, 4222) = 1,032.92; p < 0.001; Table 2 and Figure 4]. Electroencephalography Power Spectral Analysis Figure 7 illustrates all of the regression coeﬃcients between independent variables and the mediators as well as the pathways from the mediators onto the dependent variable. With noise level (NL) as the independent variable, the mediation analysis indicated that the standardized indirect eﬀects of neither cognitive load measures (frontal theta power, frontal alpha power, and frontal beta power) nor emotional measures (EDA and FAA) were signiﬁcant, although frontal alpha power denoted a trend (aNL2∗b2; β = 0.01; z = 1.76; p = 0.079). Similarly, the direct eﬀect of noise level on cravings was insigniﬁcant (cNL’; β = 0.02; z = –1.34; p = 0.179). With decision perspective (DP) as the independent variable, the mediation analysis signiﬁed that the standardized indirect eﬀects of frontal alpha power were signiﬁcant (aDP2∗b2; β = 0.01; z = 1.95; p = 0.050), while the remaining mediators were not. Once this mediator was accounted for, there was still a signiﬁcant direct eﬀect of decision perspective on cravings (cDP’; β = –0.41; z = 30.69; p < 0.001), suggesting y For frontal theta power, the GLMM indicated a signiﬁcant interaction eﬀect between noise level and decision perspective [F(1, 4222) = 5.49; p = 0.019; Table 3 and Figure 5A]. Post hoc analyses showed that only in the loud noise condition, the theta band power was stronger for later (vs. now) decisions [z(1609) = 2.72; p = 0.033]. The GLMM for frontal alpha power did not detect any signiﬁcant interaction but, a main eﬀect of both noise level and decision perspective was observed with alpha band power being stronger during the loud noise [F(1, 4222) = 10.59; p = 0.001] and later decision perspective [F(1, 4222) = 16.49; p < 0.001] conditions (Table 3 and Figure 5B). Similarly, the GLMM for frontal beta power did not detect any signiﬁcant interaction, but a main eﬀect of noise level was observed with beta band power being stronger in the loud noise condition [F(1, 4222) = 12.86; p < 0.001; Table 3 and TABLE 2 \| Overview of the GLMM omnibus tests for self-reported cravings. Craving Fixed effects F df p Noise level 2.16 1, 4230 0.141 Decision perspective 1,032.92 1, 4222 < 0.001* Noise level × Decision perspective 0.24 1, 4222 0.623 p < 0.001. FIGURE 4 \| Interaction plot of self-reported cravings between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. Electroencephalography Power Spectral Analysis TABLE 4 \| Overview of the GLMM omnibus tests for EDA peaks. demonstrated in Kober et al. (2010a), but simultaneously our results suggest that the underlying causal mechanisms of these self-regulated cravings may at least partially be explained through frontal brain oscillations. That is, the multiple mediation analysis signiﬁed a partial mediation eﬀect of decision perspective on self-reported cravings through frontal alpha power. This denotes that in particular augmented activity in the alpha frequency range is associated with increased cravings of high-calorie foods and potentially unhealthy eating behavior. Additionally, irrespectively of behavioral ratings, we found that during delayed (vs. immediate) gratiﬁcation of food rewards, i.e., in later-trials, the PSD in both the theta and alpha frequency spectra as well as FAA were increased. p < 0.01. a partial mediation eﬀect of the frontal EEG alpha power on self-reported food cravings. Frontiers in Neuroscience \| www.frontiersin.org Electroencephalography Power Spectral Analysis TABLE 2 \| Overview of the GLMM omnibus tests for self-reported cravings. TABLE 3 \| Overview of the GLMM omnibus tests for frontal theta power, frontal alpha power, frontal beta power, and frontal alpha asymmetry. EEG frontal theta power Fixed effects F df p Noise level 0.31 1, 4222 0.576 Decision perspective 2.18 1, 4222 0.140 Noise level × Decision perspective 5.49 1, 4222 0.019 EEG frontal alpha power Fixed effects F df p Noise level 10.59 1, 4222 0.001 Decision perspective 16.49 1, 4222 <0.001* Noise level × Decision perspective 2.05 1, 4222 0.152 EEG frontal beta power Fixed effects F df p Noise level 12.86 1, 4222 <0.001*** Decision perspective 1.35 1, 4222 0.245 Noise level × Decision perspective 1.29 1, 4222 0.257 EEG frontal alpha asymmetry Fixed effects F df p Noise level 1.92 1, 4222 0.166 Decision perspective 6.08 1, 4222 0.014* Noise level × Decision perspective 0.55 1, 4222 0.457 p < 0.05; p < 0.01; p < 0.001. TABLE 3 \| Overview of the GLMM omnibus tests for frontal theta power, frontal alpha power, frontal beta power, and frontal alpha asymmetry. FIGURE 4 \| Interaction plot of self-reported cravings between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. FIGURE 4 \| Interaction plot of self-reported cravings between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 8 Peng-Li et al. Neurophysiology of Food Cravings FIGURE 5 \| Interaction plots of (A) frontal theta power, (B) frontal alpha power, (C) frontal beta power frontal, and (D) alpha asymmetry between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. FIGURE 5 \| Interaction plots of (A) frontal theta power, (B) frontal alpha power, (C) frontal beta power frontal, and (D) alpha asymmetry between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. TABLE 4 \| Overview of the GLMM omnibus tests for EDA peaks. EDA peaks Fixed effects z df p Noise level 3.27 4122 0.001 Decision perspective 0.15 4122 0.874 Noise level × Decision perspective 0.87 4122 0.384 *p < 0.01. TABLE 4 \| Overview of the GLMM omnibus tests for EDA peaks. DISCUSSION This is in line with previous neuroimaging research using the ROC, which has shown increased BOLD activation in frontal regions associated with cognitive control, including the dmPFC, dlPFC, and vlPFC (Kober et al., 2010b). A hyperactivation of these regions might therefore denote cognitive overload (Matsuo et al., 2007). In fact, structural MRI studies have consistently reported reduced gray matter volume in these frontal regions (Horstmann et al., 2011; He et al., 2013) as well as lower structural connectivity between frontal and limbic structures associated with decision-making, reward, and interoceptive awareness (Gupta et al., 2015; Peng-Li et al., 2020c) in individuals with elevated impulsivity and poorer self-regulation abilities. An EEG study by Meule et al. (2013) also found larger LPP amplitude (350–550 ms after onset)—an ERP component commonly linked While a body of psychiatric and neuroscientiﬁc research has investigated the impact of top-down cognitive strategies, self-regulation, and managerial decision-making on the neurophysiological underpinnings of food cravings, empirical ﬁndings in sensory and consumer science have shown that bottom-up auditory nudging strategies can also inﬂuence eating motivation and food valuation. In the current study, we explored both avenues in a single experimental paradigm employing an adapted version of the ROC task. Our ﬁndings do not only provide direct support for our hypothesis that prospectively thinking about long- term consequences can eﬀectively reduce food cravings as February 2022 \| Volume 16 \| Article 827021 9 Peng-Li et al. Neurophysiology of Food Cravings FIGURE 6 \| Interaction plot of EDA peaks between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. salience and valuation of health beneﬁts. That is, participants may have considered the future rewards of controlling their consumption of unhealthy foods in the presence. Nevertheless, this evidence, across diﬀerent neuroimaging modalities and metrics, suggests increased cognitive demand and emotional engagement, especially when actively deliberating on long- term consequences (system 2) rather than simply evaluating immediate rewards in the present (system 1). Importantly, these psychophysiological processes may be even more intensiﬁed during exteroceptive sensory inputs and distractions including ambient noise, as the increased theta activity in the later-trials was only occurring in the presence of loud (vs. soft) ambient noise. Correspondingly, alpha activity was also augmented during the loud noise condition, yet serving as a main (and not interaction) eﬀect. DISCUSSION As theta and alpha waves are arguably the power spectra mostly associated with cognitive/work load and attention (Klimesch, 1996; O’Keefe and Burgess, 1999; Stipacek et al., 2003; Antonenko et al., 2010; Brouwer et al., 2012; Wang et al., 2019), a combination of reﬂective system 2 thinking during prospective thinking and environmental auditory disturbances requires the most cognitive resources. FIGURE 6 \| Interaction plot of EDA peaks between noise level (soft vs. loud) and decision perspective (now vs. later). Error bars represent standard error. to attention capture (Zorjan et al., 2020) and emotion regulation (Hajcak et al., 2010). Likewise, empirical ﬁndings in consumer neuroscience, popularly referred to as neuromarketing, have utilized FA and FAA to objectively quantify consumer behaviors (Bazzani et al., 2020), such as willingness to pay (Ramsøy et al., 2018), hedonic food valuation (van Bochove et al., 2016), and attention biases (McGeown and Davis, 2018). This suggests that FAA cannot only be used as a measure of cognitive engagement but also as an emotional valence marker denoting aﬀective and reward processes, including anticipatory pleasure and incentive salience (“wanting”). Although, one might expect that the FAA ought to be greater during now-trials due to closer reward proximity and delayed discounting, the manifestation of the opposite pattern can be reasoned through higher incentive However, the power of the cerebral oscillations in the higher beta frequency spectra was not aﬀected by decision perspective but solely augmented in the loud noise condition. Salisbury et al. (2002) similarly observed that background noise increased the latency of the P300 component, even while performance was unaﬀected. In an EEG review, Blume et al. (2019) have highlighted the elevated resting-state beta activity in fronto- central regions in individuals with obesity and binge-eating disorder. The authors argued that this increased beta activity may be the manifestation of the hyper-awareness of food cues and FIGURE 7 \| Diagram of the multiple mediation analyses based on our conceptual framework in Figure 1 with noise level (NL) and decision perspective (DP) as the independent variables, cravings as the dependent variable, and frontal theta power, frontal alpha power, frontal beta power and EDA as the mediators. Paths are shown with standardized regression coefﬁcients and p-values (p < 0.05; p <0 0.01; p < 0.001). DISCUSSION FIGURE 7 \| Diagram of the multiple mediation analyses based on our conceptual framework in Figure 1 with noise level (NL) and decision perspective (DP) as the independent variables, cravings as the dependent variable, and frontal theta power, frontal alpha power, frontal beta power and EDA as the mediators. Paths are shown with standardized regression coefﬁcients and p-values (p < 0.05; p <0 0.01; *p < 0.001). February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 10 Peng-Li et al. Neurophysiology of Food Cravings clinical contexts to help individuals who exhibit maladaptive eating behaviors. maladaptive eating behavior. Through deductive reasoning and in light of these collateral ﬁndings in combination with the results from the present study, it can be inferred that excessively loud noise indeed has neurophysiological impacts. This is measured by means of augmented beta activity, which in turn may provoke adverse eﬀects on food-seeking behavior, even though we did not establish that link between beta activity (only alpha) and behavior (cravings) in the mediation analysis. Secondly, the identiﬁed underlying neurophysiological mechanisms by which top-down self-regulation alleviates cravings, are essential for understanding people’s subconscious and at times suboptimal eating behaviors. In addition, by applying exteroceptive auditory manipulations that analogously aﬀect these fronto-cortical brain oscillations, we emphasize the importance of the power of a well-engineered acoustic environment. Hence, managers and other practitioners, who are at least partly responsible for the consumer, could try to establish eating atmospheres that reinforce healthier eating behavior by reducing stress, arousal and mental load (Doucé and Adams, 2020). Especially, in the times of COVID-19, in which several patients have suﬀered from anosmia (i.e., loss of smell) and/or ageusia (i.e., loss of taste), focusing on other attributes of the food, such as the texture, could help regaining the hedonic eating experience (Høier et al., 2021). Broaden out, one could also imagine that auditory cues, both intrinsic (i.e., the inherent sound of the food) and extrinsic/contextual (e.g., background music), might sensorily compensate for the loss of olfactory and gustatory perception. In addition, we found that the probability for EDA peak detection was also higher during the exposure to loud noise, indicating elevated arousal state (Salimpoor et al., 2011; Kantono et al., 2019). Louder noise may lead to a more stressful mindset that in turn diminishes the cognitive resources requisite for processing and making more rational and healthy decisions (Caviola et al., 2021). DISCUSSION In contrast, when consumers are not interrupted by loud restaurant noises, they are in a more relaxed psychological state, which places them in a better position of restraining and managing their irrational and unhealthy food choices (Peng-Li et al., 2021). In fact, fast tempo and high volume of sound, both of which elevate physiological arousal (Liu et al., 2018; Biswas et al., 2019), have been reported to reduce one’s cognitive abilities, such as decision accuracy (Day et al., 2009), task performance (Nagar and Pandey, 1987), and creative thinking (Mehta et al., 2012). Finally, with the current study being a cross-over between sensory and consumer science and cognitive neuroscience, the framework of the experiment in itself advocates the relevance of robust multidisciplinary research in decision sciences. Particularly, there has been increasing employment of neuroimaging procedures and biometric measurements in (food) market research (Knutson et al., 2007; Plassmann et al., 2008; Clement et al., 2013; Motoki and Suzuki, 2020), and neuromarketing and neuroeconomics have received considerable attention in both the scientiﬁc community and the media (Platt and Huettel, 2008; Ariely and Berns, 2010; Plassmann et al., 2012; Zhang et al., 2019). Thus, with the implementation of both EEG and EDA measurements, the study is of commercial and managerial interest. These tools can oﬀer objective quantitative insights beyond traditional subjective and explicit methods that may be constrained by introspection and verbalization. From an industrial management perspective, consistent utilization of such multimodal methods might enable valid forecasting about consumers’ intentions, behaviors, and ultimately purchases. At the same time, it would at least to some degree increase reproducibility and circumvent the consequences of the replication crisis (Chives, 2019). Yet, to optimally exploit this attention and potential, while preventing it from becoming a mere marketing gimmick, academics in the respective ﬁelds should exploit their experience and ask relevant questions that can in fact provide useful inputs to marketers and managers in addition to conventional marketing research. Altogether, the ﬁndings are partly in line with our hypothesis that both noise level and decision perspective would inﬂuence subjective food cravings and objective measures, including EDA and EEG. However, we did not observe that the manipulations of both noise level and decision perspective had an impact on all measures. Indeed, alpha activity was aﬀected by both loud noise and prospective thinking and could even predict food cravings based on the mediation analysis. Managerial Implications Due to the interdisciplinary nature and methodological novelty of our study, the results have several translational implications both clinically and commercially. First, we have demonstrated that food cravings could be restrained eﬀectively merely via a single cognitive strategy involving deliberately devaluing the immediate rewards and delaying one’s gratiﬁcation for future and long-term health beneﬁts. Thus, we build on the previous literature that has incorporated cognitive strategies to highlight the use of interoceptive regulation and managerial decision-making in food (Kober et al., 2010b; Meule et al., 2013; Boswell et al., 2018) and other substance (Kober et al., 2010a; Naqvi et al., 2015; Suzuki et al., 2020) cravings, which collectively reinforces the theoretical foundation for practically implementing these measures in DISCUSSION Theta activity was inﬂuenced by the interaction of these, i.e., only loud noise and later decision perspective. Yet, beta activity and EDA peak probability were solely determined by noise level, while FAA and food cravings were inﬂuenced by decision perspective only. Hence, it can be inferred that louder noise and prospective thinking strategy can at least to some degree elevate neurophysiological constructs of emotional arousal and motivation as well as cognitive load, but will not necessarily help consumers regulate and manage their ultimate subjective food cravings. Limitations Despite these abovementioned implications, our study involves several limitations. First, it should be noted that the physiological signal analyses were based on rather conservative pre-processing procedures due to the employment of the integrated R algorithm of iMotions. This implicates inﬂexible parameter adjustments during data pre-processing of EEG and EDA. The EEG signal February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org Frontiers in Neuroscience \| www.frontiersin.org 11 Neurophysiology of Food Cravings Peng-Li et al. was referenced to a single mastoid instead of e.g., two mastoids or an average reference, but lateralized metrics, such as FAA can be prone to confounds (Lei and Liao, 2017). Analogously, we could not carry out scrutinized eye-blink detection, manual removal of single trials or events, nor independent component analysis (ICA), but only rely on the simple automated algorithm. Notably, according to a methodological review by Allen et al. (2004), for some spectral computations (e.g., FAA), artifact thresholding alone might be as adequate as using other manual accessorial procedures, such as electrooculography (EOG) and electromyography (EMG). diﬀerent physiological measurements to more holistically, objectively, and optimally study food-related decision-making that can provoke an actual societal and managerial impact. This is not solely conﬁned to the ﬁeld of sensory and consumer neuroscience, but for any decision sciences, this seems applicable and highly pertinent. AUTHOR CONTRIBUTIONS DP-L: conceptualization, methodology, formal analysis, investigation, resources, data curation, project administration, writing—original draft, writing—review and editing, and visualization. PA: conceptualization and writing—review and editing. CC: investigation and writing—review and editing. RC: writing—review and editing and supervision. DB and QW: conceptualization, writing—review and editing, and supervision. All authors contributed to the article and approved the submitted version. Thirdly, we did not incorporate any neutral/silent condition, which could have strengthened the comparability within the study, as done in some previous food, sound, and decision research (Alamir et al., 2020; Peng-Li et al., 2020a). However, the longer design could have been time-consuming and fatiguing for the participants. Besides, one could argue that the soft noise condition would serve as a control condition since complete silence is highly unlikely in a normal eating situation. Finally, we simply conﬁned our EEG analyses to the frontal part of the brain through theta, alpha, and beta activity based on our conceptual framework. While, several studies have investigated the oscillatory power in other or smaller ROIs (Tashiro et al., 2019; Biehl et al., 2020) as well as other frequency bands (i.e., delta and gamma; Colrain et al., 2009; Dimigen et al., 2009) during mental operations, we chose not to, as the analyses would be unreasonably extensive and outside the scope of our framework. DATA AVAILABILITY STATEMENT The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Secondly, due to the nature of our controlled experimental setup, our ﬁndings cannot necessarily be directly generalized to naturalistic food choice settings (Andrade, 2018) in which multiple other external factors (including price, labeling, and social factors) may aﬀect the consumers’ emotional states, cognitive processing, and behaviors (Sørensen et al., 2013; Spence et al., 2014; Petit et al., 2015). Besides, albeit food cravings are strong predictors of eating behavior and food choice (Boswell et al., 2018; Chen et al., 2018; Sun and Kober, 2020), we cannot assure that these independent results encompass ecological validity and are applicable in a real-life managerial decision context. CONCLUSION Data was generated though accessing research infrastructure at AU, including FOODHAY (Food and Health Open Innovation Laboratory, Danish Roadmap for Research Infrastructure). We would furthermore like to thank Kiara Heide and Tue Hvass from the iMotions Client Success Team for facilitating the EEG procedure and Camilla Andersen for helping with the data collection. To conclude, the present study has underlined the combined eﬀects of cognitive regulation and ambient restaurant noise on food cravings through EDA peak probability as well as fronto- cortical brain oscillations as quantitative measures of emotional arousal, motivation, and cognitive load. More broadly, we have highlighted the prospect of and need for considering both interoceptive states and exteroceptive cues, while employing Alamir, M. A., and Hansen, K. (2021). The eﬀect of type and level of background noise on food liking: a laboratory non-focused listening test. Appl. Acoust. 172:107600. doi: 10.1016/j.apacoust.2020.107600 Allen, J. J. B., Coan, J. A., and Nazarian, M. (2004). Issues and assumptions on the road from raw signals to metrics of frontal EEG asymmetry in emotion. Biol. Psychol. 67, 183–218. doi: 10.1016/j.biopsycho.2004.03.007 ETHICS STATEMENT The studies involving human participants were reviewed and approved by the Aarhus University Ethics Committee (approval number: 2020-0184772). The patients/participants provided their written informed consent to participate in this study. FUNDING The research was supported by the Graduate School of Science and Technology, Aarhus University and the Sino-Danish College, University of Chinese Academy of Sciences (DP-L) Project/funding number: 30367. REFERENCES Alamir, M. A., AlHares, A., Hansen, K. L., and Elamer, A. (2020). The eﬀect of age, gender and noise sensitivity on the liking of food in the presence of background noise. Food Qual. Prefer. 84:103950. doi: 10.1016/j.foodqual.2020.10 3950 Alamir, M. A., AlHares, A., Hansen, K. L., and Elamer, A. (2020). The eﬀect of age, gender and noise sensitivity on the liking of food in the presence of background noise. Food Qual. Prefer. 84:103950. doi: 10.1016/j.foodqual.2020.10 3950 February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 12 Peng-Li et al. Neurophysiology of Food Cravings Alvarsson, J. J., Wiens, S., and Nilsson, M. E. (2010). Stress recovery during exposure to nature sound and environmental noise. Int. J. Environ. Res. Public Health 7, 1036–1046. doi: 10.3390/ijerph7031036 Caviola, S., Visentin, C., Borella, E., Mammarella, I., and Prodi, N. (2021). Out of the noise: eﬀects of sound environment on maths performance in middle-school students. J. Environ. Psychol. 73:101552. doi: 10.1016/j.jenvp. 2021.101552 Andrade, C. (2018). Internal, external, and ecological validity in research design, conduct, and evaluation. Indian J. Psychol. Med. 40, 498–499. doi: 10.4103/ IJPSYM.IJPSYM_334_18 Chabin, T., Gabriel, D., Chansophonkul, T., Michelant, L., Joucla, C., Haﬀen, E., et al. (2020). Cortical patterns of pleasurable musical chills revealed by high-density EEG. Front. Neurosci. 14:565815. doi: 10.3389/fnins.2020.565815 Antonenko, P., Paas, F., Grabner, R., and van Gog, T. (2010). Using electroencephalography to measure cognitive load. Educ. Psychol. Rev. 22, 425–438. doi: 10.1007/s10648-010-9130-y g y Charbonnier, L., van Meer, F., van der Laan, L. N., Viergever, M. A., and Smeets, P A M (2016) St d di d f d i h t hi t l d i Charbonnier, L., van Meer, F., van der Laan, L. N., Viergever, M. A., and Smeets, P. A. M. (2016). Standardized food images: a photographing protocol and image database. Appetite 96, 166–173. doi: 10.1016/j.appet.2015.08.041 P. A. M. (2016). Standardized food images: a photographing protocol and image database. Appetite 96, 166–173. doi: 10.1016/j.appet.2015.08.041 Aoh, Y., Hsiao, H.-J., Lu, M.-K., Macerollo, A., Huang, H.-C., Hamada, M., et al. (2019). Event-related desynchronization/synchronization in spinocerebellar ataxia type 3. Front. Neurol. 10:822. doi: 10.3389/fneur.2019. 00822 Chen, R., Peng-Li, D., Turel, O., Sørensen, T. A., Bechara, A., Li, Y., et al. (2018). Decision making deﬁcits in relation to food cues inﬂuence obesity: a triadic neural model of problematic eating. Front. Psychiatry 9:264. doi: 10.3389/fpsyt. 2018.00264 Ariely, D., and Berns, G. S. (2010). REFERENCES Neuromarketing: the hope and hype of neuroimaging in business. Nat. Rev. Neurosci. 11, 284–292. doi: 10.1038/ nrn2795 Chives, T. (2019). A Theroy in Crisis. Nature. Available online at: https://media.nature.com/original/magazine-assets/d41586-019-03755-2/ d41586-019-03755-2.pdf (accessed December 1, 2021). Barlaam, F., Descoins, M., Bertrand, O., Hasbroucq, T., Vidal, F., Assaiante, C., et al. (2011). Time-frequency and ERP analyses of EEG to characterize anticipatory postural adjustments in a bimanual load-lifting task. Front. Hum. Neurosci. 5:163. doi: 10.3389/fnhum.2011.00163 d41586-019-03755-2.pdf (accessed December 1, 2021). Clement, J., Kristensen, T., and Grønhaug, K. (2013). Understanding consumers’ in-store visual perception: the inﬂuence of package design features on visual attention. J. Retail. Consum. Serv. 20, 234–239. doi: 10.1016/j.jretconser.2013. 01.003 Bazzani, A., Ravaioli, S., Trieste, L., Faraguna, U., and Turchetti, G. (2020). Is EEG suitable for marketing research? A systematic review. Front. Neurosci. 14:594566. doi: 10.3389/fnins.2020.594566 Coccurello, R., and Maccarrone, M. (2018). Hedonic eating and the “delicious circle”: from lipid-derived mediators to brain dopamine and back. Front. Neurosci. 12:271. doi: 10.3389/fnins.2018.00271 Bechara, A., Damasio, H., Tranel, D., and Damasio, A. R. (2005). The Iowa Gambling Task and the somatic marker hypothesis: some questions and answers. Trends Cogn. Sci. 9, 159–162. doi: 10.1016/j.tics.2005.02.002 Coelli, S., Sclocco, R., Barbieri, R., Reni, G., Zucca, C., and Bianchi, A. M. (2015). “EEG-based index for engagement level monitoring during sustained attention,” in Proceedings of the 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Milan, 1512–1515. doi: 10.1109/EMBC.2015.7318658 Biehl, S. C., Keil, J., Naumann, E., and Svaldi, J. (2020). ERP and oscillatory diﬀerences in overweight/obese and normal-weight adolescents in response to food stimuli. J. Eat. Disord. 8, 1–11. doi: 10.1186/s40337-020-00290-8 Colrain, I. M., Turlington, S., and Baker, F. C. (2009). Impact of alcoholism on sleep architecture and EEG power spectra in men and women. Sleep 32, 1341–1352. doi: 10.1093/sleep/32.10.1341 Biswas, D., Lund, K., and Szocs, C. (2019). Sounds like a healthy retail atmospheric strategy: eﬀects of ambient music and background noise on food sales. J. Acad. Market. Sci. 47, 37–55. doi: 10.1007/s11747-018-0583-8 Dassen, F. C. M., Jansen, A., Nederkoorn, C., and Houben, K. (2016). Focus on the future: episodic future thinking reduces discount rate and snacking. Appetite 96, 327–332. doi: 10.1016/j.appet.2015.09.032 Blume, M., Schmidt, R., and Hilbert, A. (2019). Abnormalities in the EEG power spectrum in bulimia nervosa, binge-eating disorder, and obesity: a systematic review. Eur. Eat. Disord. Rev. 27, 124–136. doi: 10.1002/erv.2654 Day, R.-F., Lin, C.-H., Huang, W.-H., and Chuang, S.-H. REFERENCES (2009). Eﬀects of music tempo and task diﬃculty on multi-attribute decision-making: an eye-tracking approach. Comput. Hum. Behav. 25, 130–143. doi: 10.1016/j.chb.2008.08.001 Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen, J. R., Stevens, M. H. H., et al. (2009). Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24, 127–135. doi: 10.1016/j.tree.2008. 10.008 Diao, L., Li, W., Zhang, W., Ma, Q., and Jin, J. (2021). Electroencephalographic theta-band oscillatory dynamics represent attentional bias to subjective preferences in value-based decisions. Psychol. Res. Behav. Manag. 14, 149–158. doi: 10.2147/PRBM.S292172 Borghini, G., Vecchiato, G., Toppi, J., Astolﬁ, L., Maglione, A., Isabella, R., et al. (2012). “Assessment of mental fatigue during car driving by using high resolution EEG activity and neurophysiologic indices,” in Proceedings of the 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, San Diego, CA, 6442–6445. doi: 10.1109/EMBC.2012.6347469 Diao, L., Qi, S., Xu, M., Fan, L., and Yang, D. (2017). Electroencephalographic theta oscillatory dynamics reveal attentional bias to angry faces. Neurosci. Lett. 656, 31–36. doi: 10.1016/j.neulet.2017.06.047 Boswell, R. G., Sun, W., Suzuki, S., and Kober, H. (2018). Training in cognitive strategies reduces eating and improves food choice. Proc. Natl. Acad. Sci. U.S.A. 115, E11238–E11247. doi: 10.1073/pnas.1717092115 Dimigen, O., Valsecchi, M., Sommer, W., and Kliegl, R. (2009). Human microsaccade-related visual brain responses. J. Neurosci. 29, 12321–12331. doi: 10.1523/JNEUROSCI.0911-09.2009 Bravo-Moncayo, L., Reinoso-Carvalho, F., and Velasco, C. (2020). The eﬀects of noise control in coﬀee tasting experiences. Food Qual. Prefer. 86:104020. doi: 10.1016/j.foodqual.2020.104020 Doucé, L., and Adams, C. (2020). Sensory overload in a shopping environment: not every sensory modality leads to too much stimulation. J. Retail. Consum. Serv. 57:102154. doi: 10.1016/j.jretconser.2020.102154 Brouwer, A.-M., Hogervorst, M. A., van Erp, J. B. F., Heﬀelaar, T., Zimmerman, P. H., and Oostenveld, R. (2012). Estimating workload using EEG spectral power and ERPs in the n-back task. J. Neural Eng. 9:045008. doi: 10.1088/1741- 2560/9/4/045008 Duif, I., Wegman, J., de Graaf, K., Smeets, P. A. M., and Aarts, E. (2020). Distraction decreases rIFG-putamen connectivity during goal-directed eﬀort for food rewards. Sci. Rep. 10, 1–15. doi: 10.1038/s41598-020-76060-y European Broadcast Union (2016). Guidelines for Production of Programmes in Accordance With EBU R 128. Available online at: https://tech.ebu.ch/docs/tech/ tech3343.pdf (accessed December 1, 2021). Evans, J. S. B. T. (2007). On the resolution of conﬂict in dual process theories of reasoning. Think. Reason. 13, 321–339. doi: 10.1080/13546780601008825 European Broadcast Union (2016). REFERENCES Development and validation of the self-regulation of eating behaviour questionnaire for adults. Int. J. Behav. Nutr. Phys. Act. 13, 1–11. doi: 10.1186/s12966-016-0414-6 Fischer, N. L., Peres, R., and Fiorani, M. (2018). Frontal alpha asymmetry and theta oscillations associated with information sharing intention. Front. Behav. Neurosci. 12:166. doi: 10.3389/fnbeh.2018.00166 Klimesch, W. (1996). Memory processes, brain oscillations and EEG synchronization. Int. J. Psychophysiol. 24, 61–100. doi: 10.1016/S0167- 8760(96)00057-8 Frank, S., Laharnar, N., Kullmann, S., Veit, R., Canova, C., Hegner, Y. L., et al. (2010). Processing of food pictures: inﬂuence of hunger, gender and calorie content. Brain Res. 1350, 159–166. doi: 10.1016/j.brainres.2010.04.030 Knutson, B., Rick, S., Wimmer, G. E., Prelec, D., and Loewenstein, G. (2007). Neural predictors of purchases. Neuron 53, 147–156. doi: 10.1016/j.neuron. 2006.11.010 Gearhardt, A. N., White, M. A., Masheb, R. M., Morgan, P. T., Crosby, R. D., and Grilo, C. M. (2012). An examination of the food addiction construct in obese patients with binge eating disorder. Int. J. Eat. Disord. 45, 657–663. doi: 10.1002/eat.20957 Kober, H., Kross, E. F., Mischel, W., Hart, C. L., and Ochsner, K. N. (2010a). Regulation of craving by cognitive strategies in cigarette smokers. Drug Alcoh. Depend. 106, 52–55. doi: 10.1016/j.drugalcdep.2009.07.017 Kober, H., Mende-Siedlecki, P., Kross, E. F., Weber, J., Mischel, W., Hart, C. L., et al. (2010b). Prefrontal-striatal pathway underlies cognitive regulation of craving. Proc. Natl. Acad. Sci. U.S.A. 107, 14811–14816. doi: 10.1073/pnas.1007779107 Gevins, A., and Smith, M. E. (2003). Neurophysiological measures of cognitive workload during human-computer interaction. Theor. Issues Ergon. Sci. 4, 113–131. doi: 10.1080/14639220210159717 Gleiss, S., and Kayser, C. (2014). Acoustic noise improves visual perception and modulates occipital oscillatory states. J. Cogn. Neurosci. 26, 699–711. doi: 10. 1162/jocn_a_00524 Krishna, A. (2012). An integrative review of sensory marketing: engaging the senses to aﬀect perception, judgment and behavior. J. Consum. Psychol. 22, 332–351. doi: 10.1016/j.jcps.2011.08.003 Kytö, E., Bult, H., Aarts, E., Wegman, J., Ruijschop, R. M. A. J., and Mustonen, S. (2019). Comparison of explicit vs. implicit measurements in predicting food purchases. Food Qual. Prefer. 78:103733. doi: 10.1016/j.foodqual.2019.103733 Gorka, S. M., Phan, K. L., and Shankman, S. A. (2015). Convergence of EEG and fMRI measures of reward anticipation. Biol. Psychol. 112, 12–19. doi: 10.1016/j. biopsycho.2015.09.007 Lei, X., and Liao, K. (2017). Understanding the inﬂuences of EEG reference: a large- scale brain network perspective. Front. Neurosci. 11:205. doi: 10.3389/fnins. 2017.00205 Grilo, C. M., Masheb, R. M., Wilson, G. T., Gueorguieva, R., and White, M. A. (2011). REFERENCES Cognitive–behavioral therapy, behavioral weight loss, and sequential treatment for obese patients with binge-eating disorder: a randomized controlled trial. J. Consult. Clin. Psychol. 79, 675–685. doi: 10.1037/a0025049 Levin, I. P., Xue, G., Weller, J. A., Reimann, M., Lauriola, M., and Bechara, A. (2012). A neuropsychological approach to understanding risk-taking for potential gains and losses. Front. Neurosci. 6:15. doi: 10.3389/fnins.2012.00015 Gupta, A., Mayer, E. A., Sanmiguel, C. P., Van Horn, J. D., Woodworth, D., Ellingson, B. M., et al. (2015). Patterns of brain structural connectivity diﬀerentiate normal weight from overweight subjects. Neuroimage 7, 506–517. doi: 10.1016/j.nicl.2015.01.005 Liu, Y., Liu, G., Wei, D., Li, Q., Yuan, G., Wu, S., et al. (2018). Eﬀects of musical tempo on musicians’ and non-musicians’ emotional experience when listening to music. Front. Psychol. 9:2118. doi: 10.3389/fpsyg.2018.02118 Hajcak, G., MacNamara, A., and Olvet, D. M. (2010). Event-related potentials, emotion, and emotion regulation: an integrative review. Dev. Neuropsychol. 35, 129–155. doi: 10.1080/87565640903526504 Matsuo, K., Glahn, D. C., Peluso, M. A. M., Hatch, J. P., Monkul, E. S., Najt, P., et al. (2007). Prefrontal hyperactivation during working memory task in untreated individuals with major depressive disorder. Mol. Psychiatry 12, 158–166. doi: 10.1038/sj.mp.4001894 He, Q., Chen, C., Dong, Q., Xue, G., Chen, C., Lu, Z. L., et al. (2013). Gray and white matter structures in the midcingulate cortex region contribute to body mass index in Chinese young adults. Brain Struct. Funct. 220, 319–329. doi: 10.1007/s00429-013-0657-9 McGeown, L., and Davis, R. (2018). Frontal EEG asymmetry moderates the association between attentional bias towards food and body mass index. Biol. Psychol. 136, 151–160. doi: 10.1016/j.biopsycho.2018.06.001 He, Q., Huang, X., Zhang, S., Turel, O., Ma, L., and Bechara, A. (2019). Dynamic causal modeling of insular, striatal, and prefrontal cortex activities during a food-speciﬁc Go/NoGo task. Biol. Psychiatry 4, 1080–1089. doi: 10.1016/j.bpsc. 2018.12.005 Mehta, R., Zhu, R., and Cheema, A. (2012). Is noise always bad? Exploring the eﬀects of ambient noise on creative cognition. J. Consum. Res. 39, 784–799. doi: 10.1086/665048 Heller, G. Z., Manuguerra, M., and Chow, R. (2016). How to analyze the visual analogue scale: myths, truths and clinical relevance. Scand. J. Pain 13, 67–75. doi: 10.1016/j.sjpain.2016.06.012 Meule, A., Kübler, A., and Blechert, J. (2013). Time course of electrocortical food- cue responses during cognitive regulation of craving. Front. Psychol. 4:669. doi: 10.3389/fpsyg.2013.00669 Høier, A. T. Z. B., Chaaban, N., and Andersen, B. V. (2021). Possibilities for maintaining appetite in recovering COVID-19 patients. Foods 10:464. REFERENCES Guidelines for Production of Programmes in Accordance With EBU R 128. Available online at: https://tech.ebu.ch/docs/tech/ tech3343.pdf (accessed December 1, 2021). Bucher, T., Collins, C., Rollo, M. E., McCaﬀrey, T. A., De Vlieger, N., Van Der Bend, D., et al. (2016). Nudging consumers towards healthier choices: a systematic review of positional inﬂuences on food choice. Br. J. Nutr. 115, 2252–2263. doi: 10.1017/S0007114516001653 Evans, J. S. B. T. (2007). On the resolution of conﬂict in dual process theories of reasoning. Think. Reason. 13, 321–339. doi: 10.1080/13546780601008825 Burger, K. S., and Stice, E. (2012). Frequent ice cream consumption is associated with reduced striatal response to receipt of an ice cream–based milkshake. Am. J. Clin. Nutr. 95, 810–817. doi: 10.3945/ajcn.111.027003.2 Faul, F., Erdfelder, E., Buchner, A., and Lang, A.-G. (2009). Statistical power analyses using G∗Power 3.1: tests for correlation and regression analyses. Behav. Faul, F., Erdfelder, E., Buchner, A., and Lang, A.-G. (2009). Statistical power analyses using G∗Power 3.1: tests for correlation and regression analyses. Behav. Res. Methods 41, 1149–1160. doi: 10.3758/BRM.41.4.1149 Cabañero, Hervás, R., González, Fontecha, Mondéjar, T., and Bravo, J. (2019). Analysis of cognitive load using EEG when interacting with mobile devices. Proceedings 31:70. doi: 10.3390/proceedings2019031070 Res. Methods 41, 1149–1160. doi: 10.3758/BRM.41.4.1149 Fernandez Rojas, R., Debie, E., Fidock, J., Barlow, M., Kasmarik, K., Anavatti, S., et al. (2020). Electroencephalographic workload indicators during teleoperation of an unmanned aerial vehicle shepherding a swarm of unmanned ground vehicles in contested environments. Front. Neurosci. 14:40. doi: 10.3389/fnins. 2020.00040 Cai, X. l, Weigl, M., Liu, B. H., Cheung, E. F. C., Ding, J. H., and Chan, R. C. K. (2019). Delay discounting and aﬀective priming in individuals with negative schizotypy. Schizophr. Res. 210, 180–187. doi: 10.1016/j.schres.2018.12.040 February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 13 Peng-Li et al. Neurophysiology of Food Cravings Klein, S., Staring, M., Murphy, K., Viergever, M. A., and Pluim, J. (2010). elastix: a toolbox for intensity-based medical image registration. IEEE Trans. Med. Imag. 29, 196–205. doi: 10.1109/TMI.2009.2035616 Firestone, G. M., McGuire, K., Liang, C., Zhang, N., Blankenship, C. M., Xiang, J., et al. (2020). A preliminary study of the eﬀects of attentive music listening on cochlear implant users’ speech perception, quality of life, and behavioral and objective measures of frequency change detection. Front. Hum. Neurosci. 14:110. doi: 10.3389/fnhum.2020.00110 Kliemann, N., Beeken, R. J., Wardle, J., and Johnson, F. (2016). REFERENCES J. Neurosci. 31, 5013–5025. doi: 10.1523/JNEUROSCI. 5592-10.2011 Salisbury, D. F., Desantis, M. A., Shenton, M. E., and McCarley, R. W. (2002). The eﬀect of background noise on P300 to suprathreshold stimuli. Psychophysiology 39, 111–115. doi: 10.1111/1469-8986.3910111 Schulte, E. M., Grilo, C. M., and Gearhardt, A. N. (2016). Shared and unique mechanisms underlying binge eating disorder and addictive disorders. Clin. Psychol. Rev. 44, 125–139. doi: 10.1016/j.cpr.2016.02.001 Papies, E. K., Barsalou, L. W., and Rusz, D. (2020). Understanding desire for food and drink: a grounded-cognition approach. Curr. Dir. Psychol. Sci. 29, 193–198. doi: 10.1177/0963721420904958 Psychol. Rev. 44, 125–139. doi: 10.1016/j.cpr.2016.02.001 Seo, H.-S. (2020). Sensory nudges: the inﬂuences of environmental contexts on consumers’ sensory perception, emotional responses, and behaviors toward foods and beverages. Foods 9:509. doi: 10.3390/foods9040509 Pedersen, H., Quist, J. S., Jensen, M. M., Clemmensen, K. K. B., Vistisen, D., Jørgensen, M. E., et al. (2021). Investigation of eye tracking, electrodermal activity and facial expressions as biometric signatures of food reward and intake in normal weight adults. Food Qual. Prefer. 93:104248. doi: 10.1016/j.foodqual. 2021.104248 Shang, Q., Pei, G., Jin, J., Zhang, W., Wang, Y., and Wang, X. (2018). ERP evidence for consumer evaluation of copycat brands. PLoS One 13:e0191475. doi: 10.1371/journal.pone.0191475 Peng-Li, D., Andersen, T., Finlayson, G., Byrne, D. V., and Wang, Q. J. (2022). The impact of environmental sounds on food reward. Physiol. Behav. 245:113689. doi: 10.1016/j.physbeh.2021.113689 Sørensen, H. S., Holm, L., Møgelvang-Hansen, P., Barratt, D., Qvistgaard, F., and Smith, V. (2013). Consumer understanding of food labels: toward a generic tool for identifying the average consumer. Int. Rev. Retail Distrib. Consum. Res. 23, 291–304. doi: 10.1080/09593969.2013.764339 Peng-Li, D., Byrne, D. V., Chan, R. C. K., and Wang, Q. J. (2020a). The inﬂuence of taste-congruent soundtracks on visual attention and food choice: a cross- cultural eye-tracking study in Chinese and Danish consumers. Food Qual. Prefer. 85:103962. doi: 10.1016/j.foodqual.2020.103962 Spence, C., and Gallace, A. (2011). Multisensory design: reaching out to touch the consumer. Psychol. Market. 28, 267–308. doi: 10.1002/mar.20392 Spence, C., Puccinelli, N. M., Grewal, D., and Roggeveen, A. L. (2014). Store atmospherics: a multisensory perspective. Psychol. Market. 31, 472–488. doi: 10.1002/mar.20709 Peng-Li, D., Chan, R. C. K., Byrne, D. V., and Wang, Q. J. (2020b). The eﬀects of ethnically congruent music on eye movements and food choice—a cross- cultural comparison between Danish and Chinese consumers. Foods 9:1109. doi: 10.3390/foods9081109 Stipacek, A., Grabner, R. H., Neuper, C., Fink, A., and Neubauer, A. C. REFERENCES doi: 10. 3390/foods10020464 Moazami-Goudarzi, M., Sarnthein, J., Michels, L., Moukhtieva, R., and Jeanmonod, D. (2008). Enhanced frontal low and high frequency power and synchronization in the resting EEG of parkinsonian patients. Neuroimage 41, 985–997. doi: 10.1016/j.neuroimage.2008.03.032 Horstmann, A., Busse, F. P., Mathar, D., Müller, K., Lepsien, J., Schlögl, H., et al. (2011). Obesity-related diﬀerences between women and men in brain structure and goal-directed behavior. Front. Hum. Neurosci. 5:58. doi: 10.3389/fnhum. 2011.00058 Motoki, K., Sugiura, M., and Kawashima, R. (2019). Common neural value representations of hedonic and utilitarian products in the ventral striatum: an fMRI study. Sci. Rep. 9:15630. doi: 10.1038/s41598-019-52159-9 Motoki, K., and Suzuki, S. (2020). Extrinsic factors underlying food valuation in the human brain. Front. Behav. Neurosci. 14:131. doi: 10.3389/fnbeh.2020.00131 Huang, X., and Labroo, A. A. (2019). Cueing morality: the eﬀect of high- pitched music on healthy choice. J. Market. 84:002224291881357. doi: 10.1177/ 0022242918813577 Nagar, D., and Pandey, J. (1987). Aﬀect and performance on cognitive task as a function of crowding and noise1. J. Appl. Soc. Psychol. 17, 147–157. doi: 10.1111/j.1559-1816.1987.tb00306.x Hume, D. J., Howells, F. M., Rauch, H. G. L., Kroﬀ, J., and Lambert, E. V. (2015). Electrophysiological indices of visual food cue-reactivity. Diﬀerences in obese, overweight and normal weight women. Appetite 85, 126–137. doi: 10.1016/j. appet.2014.11.012 Naqvi, N. H., Ochsner, K. N., Kober, H., Kuerbis, A., Feng, T., Wall, M., et al. (2015). Cognitive regulation of craving in alcohol-dependent and social drinkers. Alcoholism 39, 343–349. doi: 10.1111/acer.12637 Kamzanova, A. T., Kustubayeva, A. M., and Matthews, G. (2014). Use of EEG workload indices for diagnostic monitoring of vigilance decrement. Hum. Fact. 56, 1136–1149. doi: 10.1177/0018720814526617 Nguyen, D., Naﬀziger, E. E., and Berridge, K. C. (2021). Positive aﬀect: nature and brain bases of liking and wanting. Curr. Opin. Behav. Sci. 39, 72–78. doi: 10.1016/j.cobeha.2021.02.013 Kantono, K., Hamid, N., Shepherd, D., Lin, Y. H. T., Skiredj, S., and Carr, B. T. (2019). Emotional and electrophysiological measures correlate to ﬂavour perception in the presence of music. Physiol. Behav. 199, 154–164. doi: 10.1016/ j.physbeh.2018.11.012 O’Keefe, J., and Burgess, N. (1999). Theta activity, virtual navigation and the human hippocampus. Trends Cogn. Sci. 3, 403–406. doi: 10.1016/S1364- 6613(99)01396-0 February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 14 Peng-Li et al. Neurophysiology of Food Cravings Palva, S., Kulashekhar, S., Hamalainen, M., and Palva, J. M. (2011). Localization of cortical phase and amplitude dynamics during visual working memory encoding and retention. REFERENCES (2003). Sensitivity of human EEG alpha band desynchronization to diﬀerent working memory components and increasing levels of memory load. Neurosci. Lett. 353, 193–196. doi: 10.1016/j.neulet.2003.09.044 Peng-Li, D., Sørensen, T. A., Li, Y., and He, Q. (2020c). Systematically lower structural brain connectivity in individuals with elevated food addiction symptoms. Appetite 155:104850. doi: 10.1016/j.appet.2020.104850 Sun, L., Peräkylä, J., and Hartikainen, K. M. (2017). Frontal Alpha asymmetry, a potential biomarker for the eﬀect of neuromodulation on brain’s aﬀective circuitry—preliminary evidence from a deep brain stimulation study. Front. Hum. Neurosci. 11:584. doi: 10.3389/fnhum.2017.00584 Peng-Li, D., Mathiesen, S. L., Chan, R. C. K., Byrne, D. V., and Wang, Q. J. (2021). Sounds healthy: modelling sound-evoked consumer food choice through visual attention. Appetite 164:105264. doi: 10.1016/j.appet.2021.105264 Sun, W., and Kober, H. (2020). Regulating food craving: from mechanisms to interventions. Physiol. Behav. 222:112878. doi: 10.1016/j.physbeh.2020.1 12878 Petit, O., Basso, F., Merunka, D., Spence, C., Cheok, A. D., and Oullier, O. (2016). Pleasure and the control of food intake: an embodied cognition approach to consumer self-regulation. Psychol. Market. 33, 608–619. doi: 10.1002/mar. 20903 Suzuki, S., Mell, M. M., O’Malley, S. S., Krystal, J. H., Anticevic, A., and Kober, H. (2020). Regulation of craving and negative emotion in alcohol use disorder. Biol. Psychiatry 5, 239–250. doi: 10.1016/j.bpsc.2019.10.005 Petit, O., Spence, C., Velasco, C., Woods, A. T., and Cheok, A. D. (2017). Changing the inﬂuence of portion size on consumer behavior via imagined consumption. J. Bus. Res. 75, 240–248. doi: 10.1016/j.jbusres.2016.07.021 Tashiro, N., Sugata, H., Ikeda, T., Matsushita, K., Hara, M., Kawakami, K., et al. (2019). Eﬀect of individual food preferences on oscillatory brain activity. Brain Behav. 9, 1–7. doi: 10.1002/brb3.1262 Petit, O., Velasco, C., Cheok, A. D., and Spence, C. (2015). “Consumer sensory neuroscience in the context of food marketing,” in Proceedings of the ACM International Conference Proceeding Series, 3–6. (New York, NY: Association for Computing Machinery). doi: 10.1145/2832932.2856226 Tay, R. Y. L., and Ng, B. C. (2019). Eﬀects of aﬀective priming through music on the use of emotion words. PLoS ONE 14, 1–26. doi: 10.1371/journal.pone.0214482 Plassmann, H., O’Doherty, J., Shiv, B., and Rangel, A. (2008). Marketing actions can modulate neural representations of experienced pleasantness. Proc. Natl. Acad. Sci. U.S.A. 105, 1050–1054. doi: 10.1073/pnas.0706929105 van Bochove, M. E., Ketel, E., Wischnewski, M., Wegman, J., Aarts, E., de Jonge, B., et al. (2016). Posterior resting state EEG asymmetries are associated with hedonic valuation of food. Int. J. Psychophysiol. 110, 40–46. REFERENCES doi: 10.1016/j. ijpsycho.2016.10.006 Plassmann, H., Ramsøy, T. Z., and Milosavljevic, M. (2012). Branding the brain: a critical review and outlook. J. Consum. Psychol. 22, 18–36. doi: 10.1016/j.jcps. 2011.11.010 Verastegui-Tena, L., van Trijp, H., and Piqueras-Fiszman, B. (2019). Heart rate, skin conductance, and explicit responses to juice samples with varying levels of expectation (dis)conﬁrmation. Food Qual. Prefer. 71, 320–331. doi: 10.1016/j. foodqual.2018.08.011 Platt, M. L., and Huettel, S. A. (2008). Risky business: the neuroeconomics of decision making under uncertainty. Nat. Neurosci. 11, 398–403. doi: 10.1038/ nn2062 Volkow, N. D., and Baler, R. D. (2015). NOW vs LATER brain circuits: implications for obesity and addiction. Trends Neurosci. 38, 345–352. doi: 10.1016/j.tins. 2015.04.002 Posada-Quintero, H. F., Reljin, N., Bolkhovsky, J. B., Orjuela-Cañón, A. D., and Chon, K. H. (2019). Brain activity correlates with cognitive performance deterioration during sleep deprivation. Front. Neurosci. 13:1001. doi: 10.3389/ fnins.2019.01001 Volz, S., Ward, A., and Mann, T. (2021). Eating up cognitive resources: does attentional consumption lead to food consumption? Appetite 162:105165. doi: 10.1016/j.appet.2021.105165 Ramsøy, T. Z., Skov, M., Christensen, M. K., and Stahlhut, C. (2018). Frontal brain asymmetry and willingness to pay. Front. Neurosci. 12:138. doi: 10.3389/fnins. 2018.00138 Wang, X., Duan, H., Kan, Y., Wang, B., Qi, S., and Hu, W. (2019). The creative thinking cognitive process inﬂuenced by acute stress in humans: an electroencephalography study. Stress 22, 472–481. doi: 10.1080/10253890.2019. 1604665 Rangel, A., Camerer, C., and Montague, P. R. (2008). A framework for studying the neurobiology of value-based decision making. Nat. Rev. Neurosci. 9, 545–556. doi: 10.1038/nrn2357 Welch, P. (1967). The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modiﬁed periodograms. IEEE Trans. Audio Electroacoust. 15, 70–73. doi: 10.1109/TAU.1967.116 1901 Rosseel, Y. (2012). lavaan: an R package for structural equation modeling. J. Stat. Softw. 48, 1–36. doi: 10.18637/jss.v048.i02 Russell, J. A. (1980). A circumplex model of aﬀect. J. Pers. Soc. Psychol. 39, 1161–1178. doi: 10.1037/h0077714 Witt, C. L. (2008). Turn down the noise. Adv. Neonat. Care 8, 137–138. doi: 10.1097/01.ANC.0000324331.63851.46 Salimpoor, V. N., Benovoy, M., Larcher, K., Dagher, A., and Zatorre, R. J. (2011). Anatomically distinct dopamine release during anticipation and experience of peak emotion to music. Nat. Neurosci. 14, 257–262. doi: 10.1038/n n.2726 Woods, A. T., Poliakoﬀ, E., Lloyd, D. M., Kuenzel, J., Hodson, R., Gonda, H., et al. (2011). Eﬀect of background noise on food perception. Food Qual. Prefer. 22, 42–47. Frontiers in Neuroscience \| www.frontiersin.org February 2022 \| Volume 16 \| Article 827021 REFERENCES doi: 10.1016/j.foodqual.2010.07.003 February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 15 Peng-Li et al. Neurophysiology of Food Cravings Zhang, B., and Seo, H.-S. (2015). Visual attention toward food-item images can vary as a function of background saliency and culture: an eye- tracking study. Food Qual. Prefer. 41, 172–179. doi: 10.1016/j.foodqual.2014. 12.004 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Zhang, W., Jin, J., Wang, A., Ma, Q., and Yu, H. (2019). Consumers’ implicit motivation of purchasing luxury brands: an EEG study. Psychol. Res. Behav. Manag. 12, 913–929. doi: 10.2147/PRBM.S21 5751 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Zhang, Y., Chen, Y., Bressler, S. L., and Ding, M. (2008). Response preparation and inhibition: the role of the cortical sensorimotor beta rhythm. Neuroscience 156, 238–246. doi: 10.1016/j.neuroscience.2008.06.061 Copyright © 2022 Peng-Li, Alves Da Mota, Correa, Chan, Byrne and Wang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Zhu, R., and Meyers-Levy, J. (2005). Distinguishing between the Meanings of Music: when background music aﬀects product perceptions. J. Market. Res. 42, 333–345. doi: 10.1509/jmkr.2005.42.3.333 Zorjan, S., Schwab, D., and Schienle, A. (2020). The eﬀects of imaginary eating on visual food cue reactivity: an event-related potential study. Appetite 153:104743. doi: 10.1016/j.appet.2020.104743 Zorjan, S., Schwab, D., and Schienle, A. (2020). The eﬀects of imaginary eating on visual food cue reactivity: an event-related potential study. Appetite 153:104743. doi: 10.1016/j.appet.2020.104743 February 2022 \| Volume 16 \| Article 827021 Frontiers in Neuroscience \| www.frontiersin.org 16
https://openalex.org/W4319831172	https://zenodo.org/record/7671387/files/MS01v102.pdf	English	null	FEX3-ECG/MS01,v1.0.0: Least Squares Approximation of ECG Signals with Rational Functions	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	7,695	Least squares approximation of ECG signals with rational functions Kobzos, Laszlo h k l i@ il Kobzos, Laszlo zehu.kola.ci@gmail.com 2. Introduction In the 1970s, one of the departments of the Research Institute for Telecommunications (TKI, Távközlési Kutató Intézet, Budapest, Hungary) was also involved in the digitisation, transmission, storage and computer processing of ECG signals. Medical expertise was provided by a department of the National Institute of Cardiology (OKI, Országos Kardiológiai Intézet, Budapest, Hungary). Work phases were regularly reported at the Neumann Colloquia in Szeged. The following description is a late continuation [2, 3, 4, 5] of these old works [1]. A lot has changed since then, including the advent of PCs, the Internet, e-mail, GNU SW, and databases. The TKI has ceased to exist, and the name of the OKI has changed, now: Gottsegen National Cardiovascular Center (Gottsegen György Országos Kardiovaszkuláris Intézet). Although a diagnosis cannot be made on the basis of an ECG alone, its findings provide a strong support for the diagnosis. The analysis of the signal is made difficult by noise due to several causes (such as muscle tremors and respiration), and filtering these out by frequency can distort the useful signal. The original signal, even if digitised (sampling and quantization), is generally not suitable for machine learning procedures. In principle, the aim is to determine those parameters of the signal that contain all the biologically useful information, but with as few parameters as possible. Of course, this is only possible with trade-offs, and unfortunately, the ECG signal does not contain such natural parameters. Since cardiologists are very good at evaluating ECG signals by analysing their shapes on the basis of the expertise they have accumulated over almost a century, we also consider the shapes to be important. A common method is to approximate the signal using a function with parameters that can be set, so that the difference between the signal and the shapes of the function (in the biologically relevant parts) is small. Then the parameters of the function that approximate the signal will be considered to be characteristic of it. For this purpose, we have chosen rational functions. For this purpose, we have chosen rational functions. 1. Abstract Our purpose: feature extraction from one period of leads of ECG signals for AI (artificial intelligence), mainly for screening tests purposes. Method: approximation of the signal by changing the parameters of the rational function written in the form below (product of factors), using the Levenberg-Marquardt method. Results: the first 95 signals of the database were all successfully approximated. The drawback is that our procedure is currently time consuming. The reason is the non-automatic generation of the initial parameter sets of the approximations: “view the drawing – select next step – mark points on the drawing”. However, this method has the advantage of not requiring full medical expertise. t: time, the argument of the function t: time, the argument of the function t: time, the argument of the function d: the leading coefficient c: the root of the linear factor in the numerator. This factor may be missing. Quadratic factors may also be missing in the numerator. If any of the quadratic factors in the numerator contains the "–" value (of the "±" signs), then the form of the complex conjugate root pair will be g±a.i; if it contains the "+" value, then the two real roots will be g±a. In the denominator, in all cases the complex conjugate root pair will be h±b.i. The denominator never has a real root, since b² is always non-negative (in practice, it is always positive). The roots of the denominator and the numerator are called poles and zeros, respectively. For complex conjugate root pairs, the root locations are h and g, while b and g are shape parameters. 4.1. The approximation function For the approximation, the rational function was written in the following form (~shape with a root factor): For the approximation, the rational function was written in the following form (~shape with a root factor): For the approximation, the rational function was written in the following form (~shap where · indicates multiplication. It is known about polynomials that: "any polynomial with real coefficients is a product of linear and/or irreducible quadratic factors with real coefficients." For example here:[12] All parameters are real-valued, and the possibility of multiple roots can be ignored. In this form, the parameters to be determined are the values of d c g a h b and not the coefficients of standard form of a polynomial. Each wave of the ECG signal is approximated separately. Therefore, the approximation of the whole period will be the sum of the approximations of the individual waves (plus the baseline). When the approximation of a wave is complete, it will be subtracted from the signal approximated until that point, and the difference will become the new signal to be approximated. In the final step, the difference will be the sum of the noise and the error of the approximation. 3. Objective Our goal is to extract the feature data (FEX, Feature Extraction) from one period of ECG signals needed for artificial intelligence (AI) and more narrowly for machine learning (ML), and even more narrowly for supervised learning (SL). Given the wide variability of ECG signals, we believe it would be useful to first develop a method of classification into three categories, similar to medical screening tests. These categories are: “healthy”, “needs to be seen by a cardiologist”, and “intermediate”. Of course, even when machines will learn to classify into several categories, it would still not be about a machine replacing the cardiologist. The decision is always the responsibility of the medical doctor! Similarly, the result of a laboratory test is also evaluated by a doctor together with a lot of other information (such as symptoms, other findings and anamnesis). gether with a lot of other information (such as symptoms, other findings and anamnesis) together with a lot of other information (such as symptoms, other findings and anamnes ons were chosen to approximate the waves in a period from one lead of the ECG signal Rational functions were chosen to approximate the waves in a period from one lead of the ECG signal. Narrowed down: the coefficients are real; the fraction is a proper fraction (the degree of the denominator is greater than that of the numerator); and the denominator has no real roots. Narrowed down: the coefficients are real; the fraction is a proper fraction (the degree of the denominator is greater than that of the numerator); and the denominator has no real roots If you look at these functions, you can see that they behave similarly to the waves of the ECG signal (P, QRS, and T): they have negligibly low values before and after the wave. 4. Method 4.1. The approximation function 4.5 Data Our ECG data comes from the PhysioNet database [7]. From this database, 549 anonymised ECG data of 290 individuals can be accessed by selecting PTB [11]. In addition to the signals, the “PTB Diagnostic ECG Database” contains other data, as well (such as: age and diagnosis of the individuals). Signals are sampled at 1 kHz (i.e. 1 sample per millisecond), and a 1mV signal corresponds to a quantisation level of 2000. Although our method is independent of the lead system, we have mainly investigated 3 leads of the Frank system (vx, vy and vz; sometimes only x, y and z). Using Dower or inverse-Dower matrix transformations, the data of the conventional 12-lead system can be calculated from the Frank lead system data, and vice versa [13, 14]. 4.3. Properties of the above function form The real roots of the numerator can be very well determined, but not accurately because of the noise, since at these locations (times) the signal is 0. Their number depends on whether there is a linear factor; if their number is odd, and then the answer is yes. From an even number of real zeros, we combine each pair into a single quadratic factor. The real parts (place, time) of the poles and zeros are within the given wave (P, QRS, T) or in the immediate vicinity of the wave. All parameters - except for the leading coefficient d - are time unit, and are almost of the me magnitude, unlike the coefficients of powers of t in the standard form of polynomial same magnitude, unlike the coefficients of powers of t in the standard form of polynomials. Below we also write about other properties of the rational fractional function written in the above form same magnitude, unlike the coefficients of powers of t in the standard form of polynomials. Below we also write about other properties of the rational fractional function written in the above form. lso write about other properties of the rational fractional function written in the above fo Below we also write about other properties of the rational fractional function written in t 4.4 The norm and tools used for approximation Because of the noise in the ECG signal, it was practical to use the standard l2 norm (square root of the sum of the squares of the deviations) to minimize the deviation. The functions written in the form above do not form an orthogonal system of functions, and the standard linearisation method cannot be applied, so we had to use a different, non-linear tool. For the minimisation, we used the MINPACK library downloadable from the Netlib Repository [8]. MINPACK employs a modified Levenberg-Marquardt method (This directory contains the double-precision versions). Choosing the lmdif.f subroutine from MINPACK was significantly convenient when we had to modify our own approximation program (progr1-123). This meant that we did not have to use derivative functions, but instead the difference quotient, which is computed by the routine itself using the subroutine that calculates the value of the above function. To use the lmdif.f routine from MINPACK, you need a few more routines, they can be downloaded together. The minimum value achieved is local. We ran the approximation program (progr1-123) under Windows 7. Clarity was an important aspect when the program was developed, as it had to be modified frequently due to its experimental nature. For this reason, and since the subroutines of the MINPACK package are written in Fortran, we also used Fortran. Within this, we chose Gfortran (in MinGW environment) [10]. Gfortran was developed by the GNU Fortran project. It is a “free” Fortran compiler, part of the GNU Compiler Collection. 4.6. The coordinate system used The x-axis (t-axis) of a period of a lead was determined with a method usually used in cardiology. That is, in the signal-free part before the P wave of the period, we identify a location where the noise is limited, or filter out the noise by “looking”, and we mark the starting point of the baseline. The other point of the baseline is similarly determined at the next P wave. The line passing through these two points is called the baseline. The time value of the maximum point of the signal energy (proportional to the square of the voltage) was chosen as the origin. This point is always within the QRS-complex. In more detail: In the case of a Frank lead system, since it is orthogonal in space, we used the Pythagorean Theorem to determine the curve of the voltage vector length from the vx vy vz leads. Its maximum indicates the maximum activity of the electrical signal of the heart. Because of the noise on the leads, we instead cut the curve at 95% of the maximum voltage vector and choose the middle point between the two intersections as the time value of the origin. Since this does not usually coincide with the location of the sampling point, we round to the nearest location. If we do not use a Frank lead system, but a 12-lead system, then the location of the origin can be determined in the same way using the (nearly perpendicular) V6, II and -0.5×V2 leads [6]. 4.8. Individual waves Ta wave: Ta wave: The segment between the end of the P wave and the beginning of the QRS complex is isoelectric (here: no signal, 0) or nearly isoelectric. Even if it is the latter, the signal is usually so small that it cannot be detected on a standard ECG recording because it is significantly smaller than a small cube of 1 mm (with a height of 0.1mV), i.e. the difference is comparable to the thickness of a hair. This short segment is part of the Ta wave (i.e. atrial repolarisation). If the Ta wave is seen in this lead, it starts slightly after the P wave (the beginning of the depolarisation) and usually ends after the QRS wave. The end of a bigger Ta wave may also slightly affect the ST segment [9]. Although Ta is probably not of significant diagnostic value in general, it reduces the accuracy of the approximation if it is not taken into account. To approximate the Ta wave, we use only 3 parameters: the leading coefficient d and a single factor in the denominator (i.e., an h parameter and a b parameter). P wave: The accuracy of the approximation can be significantly affected by the signal-to-noise ratio. In the case of small P waves and higher noise levels, it is difficult to distinguish accurately between signal and noise. T wave: The next step in our processing is the T wave. According to signal theory, a signal is generated (when it leaves the baseline), followed by the signal phase, which finally disappears (when the signal returns back to the baseline). It is quite common when the T is not zero at the junction of the T wave and the QRS complex, i.e. the T is generated below the QRS complex (similarly to Ta and P). In this case, the wave to be approximated is missing on this segment. However, we have two pieces of information regarding this segment of the T wave: - the ventricular repolarisation cannot start before the beginning of the ventricular depolarisation, and - what is its shape after the QRS complex ends. During the QRS complex, only the sum of the two signals is known. 4.7. Method of approximations The program (progr1-123) that performs the approximation, using the above-mentioned lmdif.f, in a single step The program (progr1-123) that performs the approximation, using the above-mentioned The program (progr1-123) that performs the approximation, using the above-mentioned lmdif.f, in a single step (one run) of approximating 1 wave of 1 period of 1 lead, gradually modifies the initial parameters of the function to minimise the deviation until the stopping criteria are met. In this program (progr1-123), the number of parameters does not change [FEX3-ECG/Charts02: Charts02.zip\s0021_p005-vy\]. The most precisely known parameters are the real zeros. The least precisely known parameter is the leading coefficient. (We will return to complex conjugate poles and complex conjugate zeros later in the description.) The above parameters will define the 3 sub-steps of the approximation: 1k. At first, only the leading coefficient d will change. At this point, the nature of the approximation curve (i.e. its poles and zeros) will not change, but the deviation can be substantially reduced. 2k. Then the denominator parameters (h and b), and the leading coefficient d will chang 3k. Finally, all parameters will change. hen the denominator parameters (h and b), and the leading coefficient d will change. 2k. Then the denominator parameters (h and b), and the leading coefficient d will change. 3k. Finally, all parameters will change. 2k. Then the denominator parameters (h and b), and the leading coefficient d will change. 3k. Finally, all parameters will change. 4.8. Individual waves Ta wave: This segment is short and the T wave is usually already small here, so in general we can estimate the shape of this segment of the T wave well, based on the above two pieces of information (see section 4.11). We checked this with T waves that we were completely familiar with. As a test, we only approximated a longer segment of the T wave and looked at how the approximation behaves on a segment unknown to the program. Even on this segment, the deviation was relatively small. This also depended on the noise of the signal, i.e. this deviation was larger in case of a noisier signal. In the following section, we will therefore slightly depart from the naming system used by cardiologists, because we will collectively refer to the entire signal according to the signal theory as the T wave, that is, together: the part within the possible QRS; the S-T segment; according to the traditional notation, T; and the possible U wave. Although - regardless of the name - the point is: in the end, the calculated full approximation curve should approximate the entire period well. The QRS complex: The QRS complex: Lastly, we approximate the QRS complex and this usually requires a relatively large number of parameters. Lastly, we approximate the QRS complex and this usually requires a relatively large Lastly, we approximate the QRS complex and this usually requires a relatively large number of parameters. 4.9. Tools for approximating a wave In general, approximations require an initial set of parameters, which will be corrected in the “single step”, we described in the above 4.7 section. In the following sections, multiplying the numerator or denominator of an existing approximate function, or both, by new factor(s) is called expansion. These expansions are not done by the program that performs the approximation (progr1-123). The approximating program (progr1-123) will change all old and new parameters after an expansion, when approximating with the new initial parameter set. The accuracy of the values in the initial parameter set is therefore not very important. 3e. Poles: A general quadratic expansion factor is used as the new factor of the denominator. In this case, we take advantage of the fact that, looking at the ECG signal waves, it can be determined that they start from zero and eventually arrive back there. In practice, the following expansion factor with the form (t-h)2+b2 was appropriate: location h is the middle of the wave to be approximated, while b is 1.25 times the width of the wave to be approximated. At the same time, the leading coefficient d should be multiplied by b2. Thus, in the middle of the wave, i.e. at location h, the value of this expansion factor will be 1, and at the edges of the wave, will be 0.86. Therefore, the expansion does not significantly distort the shape of the previous approximation even at the start of the approximation step, i.e. the deviation only increases slightly. Running the program that performs the approximation (progr1-123) generally improves the approximation compared to the previous one. If this multiplier of 1.25, chosen on the basis of experience, would be increased, the initial bias would be If this multiplier of 1.25, chosen on the basis of experience, would be increased, the initial bias would be smaller, but the approximation would converge more slowly (in several internal steps). 2e. The signal intersects the baseline at two nearby points, or the signal almost reache In the case where the noise makes it unclear whether the signal intersect the baseline at two nearby points, or will only almost reach it, then either case can be chosen for the new pair of zeros. Because they can be automatically transformed into each other (two real zeros <==> complex conjugate zero pair) by the program that performs the approximation (progr1-123). If we choose the “almost reach” case, we have to find the timepoint g when the difference between the signal and the baseline is the minimum, the value of this difference is denoted by pp. Then we have to find the time when this distance doubles, i.e. it will be 2·pp. The difference between these two times is denoted by ttt. Thus, in the expansion factor (t-g)2-(-a2), where a2 = ttt2. If one chooses the case of two real roots, see point 1e. 1e. Real zeros: The case of real zeros is the simplest. If there are an odd number of zeros, then linear expansion will also be a factor. From an even number of real zeros, we combine each pair into a single quadratic expansion factor. Let the two real roots be tt1 and tt2, in that order. Then, in the expansion factor in form of (t-g)2-(+a2), where The case of real zeros is the simplest. If there are an odd number of zeros, then linear expansion will also be a factor. From an even number of real zeros, we combine each pair into a single quadratic expansion factor. p , p factor. From an even number of real zeros, we combine each pair into a single quadratic expansion factor. Let the two real roots be tt1 and tt2, in that order. Then, in the expansion factor in form of (t-g)2-(+a2), where g = (tt2 + tt1)/2 and a = (tt2 - tt1)/2. Let the two real roots be tt1 and tt2, in that order. Then, in the expansion factor in form of (t-g)2-(+a2), where g = (tt2 + tt1)/2 and a = (tt2 - tt1)/2. 4e. Complex conjugate zeros: b = \|tt The expansion factor of the denominator, which is (t-h)2 + b2 where h is also the location of the maximum deviation and b2 is equal to tttt2, (i The expansion factor of the denominator, which is (t h) + b where h is also the location of the maximum deviation and b2 is equal to tttt2, (i.e. b = \|tttt is also the location of the maximum deviation and b2 is equal to tttt2, (i.e. b = \|tttt\|). Further, away from the location of the maximum, the expansion factor approaches 1 ( Further, away from the location of the maximum, the expansion factor approaches 1 (~1/t2). Of course, this is a general relation, so it applies even if it is about a maximum location, but a minimum location. It is also independent of whether the signal and the approximation are either both positive or both negative. During the approximation, the initial locations (g and h) of both the numerator and the denominator, which are assumed to be identical, do not usually remain identical. That is, this expansion requires 4 new parameters. Bidirectional differences [FEX3-ECG/Charts02: Charts02.zip\Unid-Alte\alter For example, suppose that for the existing approximation, the signal is smaller than the approximation curve throughout a short segment and then larger after the intersection (alternating). So, in the first segment, the approximation obtained up to that point should be multiplied by numbers less than 1. The minimum of this multiplication factor should be, for example, 0.7, and this is denoted by u (u<1). The distance between the timepoint of this minimum location and the timepoint of the intersection should be, for example, 5 ms, denoted by tttt (tttt>0). Finally, denote the timepoint of the intersection by q. In this case, the zero and pole will have the same shape parameter (the imaginary part of the roots), a = b, that is, they will be located next to each other on the complex number plane, symmetrically on the vertical line that can be drawn through the q-point. After point q, the signal is larger, so the approximation must be increased, staying with the example, at the maximum of the deviation here, it will be multiplied by a factor of 1/0.7 = ~1.42 (1/u). Two more calculations are required to determine the required parameters. The first is to calculate an auxiliary variable v = tttt · (1-u)/(1+u). With the data of this example: v = 5 · (1- 0.7)/(1+0.7) = 0.88. The first is to calculate an auxiliary variable v = tttt · (1-u)/(1+u). With the data of this example: v = 5 · (1- 0.7)/(1+0.7) = 0.88. The second calculation, the calculation of the a (imaginary part): a = √¯(tttt2 - v2). With the data of this example: a = √¯(52 - 0,882)=4.9. The second calculation, the calculation of the a (imaginary part): a = √¯(tttt2 - v2). With the data of this example: a = √¯(52 - 0,882)=4.9. Using this data, a pretty good expansion factor can be created for the function so far. √ this data, a pretty good expansion factor can be created for the function so far. √ The expansion factor of the numerator is (t-g)2-(-a2), where g = q - v and a = √¯(tttt2 - v2). With the data of this example: (t-q + 0.88)2-(-4.92). √ The expansion factor of the numerator is (t-g)2-(-a2), where g = q - v and a = √¯(tttt2 - v2). With the data of this example: (t-q + 0.88)2-(-4.92). 4e. Complex conjugate zeros: The expansion of complex conjugate zeros is done together with the expansion of complex conjugate poles. This is usually only necessary for short signal segments, often only at the end of wave approximation for minor corrections. The following section describes two types of cases. One is the case where there is a unidirectional difference between the signal and the approximation, and the other is the case where there is a bidirectional difference (i.e. a valley and a hill follow each other). The easiest way to show these extensions is through examples. In both cases, the existing parameters change only slightly, as the expansion factor approaches 1 when moving away from the position of the pole-zero pair [FEX3-ECG/Charts02: Charts02.zip\Unid-Alte\]. Unidirectional differences [FEX3-ECG/Charts02: Charts02.zip\Unid-Alte\unidir_rajz.gif]: For example, let's assume that with the existing approximation, the signal is larger than the approximation curve throughout a short segment. At the maximum of the deviation, the signal is 40% greater than the previous approximation; so at this point, the previous approximation should be multiplied by a multiplication factor of 1.4 (denoted by pppp). Find the point in time where the signal is only half pppp (20%) larger than the approximation, so here the multiplier should be 1.2. We denote the time difference between this latter location (20%) and the location of the maximum deviation (40%) by tttt. Using this data, a pretty good expansion factor can be created for the function so far. In this case, the poles and zeros are at a timepoint, one above the other on the complex number plane. The expansion factor of the numerator, which is (t-g)2 - (-a2) where g is the location of the maximum deviation and a2 is equal to tttt2 · pppp, (i.e. a = \|tttt\|·√¯pppp). 2 2 The expansion factor of the numerator, which is (t-g)2 - (-a2) where g is the location of the maximum deviation and a2 is equal to tttt2 The expansion factor of the denominator, which is (t-h)2 + b2 where h is also the location of the maximum deviation and b2 is equal to tttt2, (i.e. b = \|tttt\|). The expansion factor of the denominator, which is (t-h)2 + b2 where h is also the location of the maximum deviation and b2 is equal to tttt2, (i.e. 4.10. Using these tools • Then we increase the number of poles in a cycle up to a given value 3e, after which we look at the completed drawings in order and select the one with a fairly good approximation, but the number of parameters is not yet too large. If there is no such, we have to find the reason by looking at the drawings. The section below describes in detail the possible causes of the problems and some methods that have been developed in practice to eliminate these causes. • Then we increase the number of poles in a cycle up to a given value 3e, after which we look at the completed drawings in order and select the one with a fairly good approximation, but the number of parameters is not yet too large. If there is no such, we have to find the reason by looking at the drawings. 4.11. Problems and their practical solutions - If the approximation at the two edges of the wave (at the baseline) is not good enough, the number of poles should be increased 3e. - If the approximation at the two edges of the wave (at the baseline) is not good enough, the number of poles should be increased 3e. - If the approximate curve between two peaks is farther from the baseline than the original signal, then complex zero is required 4e. Before expansion, in this case it is worth performing a couple of steps backtrack with the number of poles increased in the cycle, so that there are no more poles than necessary. - If the approximate curve between two peaks is farther from the baseline than the original signal, then complex zero is required 4e. Before expansion, in this case it is worth performing a couple of steps backtrack with the number of poles increased in the cycle, so that there are no more poles than necessary. - If the deviation is a narrow “bulge” or “indentation”, or if it is bidirectional (i.e. a valley and a hill follow each other in succession), then the same applies 4e. In such cases, it is also recommended to return a few steps before expansion. - If the approximation of a wave is only slightly better than the previous one, and that step was a pole expansion 3e, then it may be that the pole moved far away from the wave during the approximation. In this case, newer pole expansion 3e usually helps. The reason is usually that the two parameters of the pole are not enough to obtain a better approximation, but a second expansion can lead to a significant reduction of the four parameters. - If the start of the T wave falls below the QRS, do not choose an approximation of T where one of the roots falls within this common segment. Because here we only know the sum of T and QRS. In this case, we apply the “simplest is the most probable” rule (~Occam’s razor). - If we do not find an approximation corresponding to the above rule for a “dome-like” T (in the case of a Frank lead, this can also be negative), we classify it in the “needs to be seen by a cardiologist” category, because the "dom-like" T belongs there from the start. Bidirectional differences [FEX3-ECG/Charts02: Charts02.zip\Unid-Alte\alter The expansion factor of the denominator is (t-h)2 + b2, where h = q + v and b = √¯(tttt2 - v2). With the data The expansion factor of the denominator is (t-h)2 + b2, where h = q + v and b = √¯(tttt2 - v2). With the data of this example: (t-q - 0.88)2+ (4.92). of this example: (t-q - 0.88)2+ (4.92). away from the location of the intersection, the expansion factor approaches 1 (~1/t). Further, away from the location of the intersection, the expansion factor approaches 1 (~ In our example, the zero is first in time, followed by point q, and finally the pole. If the situation is reversed (i.e. if “the signal greater for a short segment than the approximation curve”), then the pole is first, followed by point q, and finally the zero; in this case, the values of g and h would be swapped. (The expansion factor will be the reciprocal of the former one.) During the approximation, the values of both the numerator and the denominator (a and b), which are assumed to be identical, do not usually remain identical. So this expansion also requires 4 new parameters. 4.10. Using these tools Unfortunately, we have not found an exact, programmable method for when and which expansion device to use. After each run of the approximation program (progr1-123), we plot the signal to be approximated and the approximate signal, then look at these and select the next tool 1e - 4e based on the deviations, and finally we stop the program. Of course, after any step, it is possible to backtrack. • To start with, the P, QRS and T waves are approximated using 2, 3 and 2 pairs of poles, respectively (i.e. the denominator is of 4, 6 and 4 degree, and in the numerator only the leading coefficient d will change). • To start with, the P, QRS and T waves are approximated using 2, 3 and 2 pairs of poles, respectively (i.e. the denominator is of 4, 6 and 4 degree, and in the numerator only the leading coefficient d will change). • This is usually followed by the entry of zeros 1e, 2e. This is also possible later. • To start with, the P, QRS and T waves are approximated using 2, 3 and 2 pairs of poles, respectively (i.e. the denominator is of 4, 6 and 4 degree, and in the numerator only the leading coefficient d will change). • This is usually followed by the entry of zeros 1e, 2e. This is also possible later. , g , y g g • This is usually followed by the entry of zeros 1e, 2e. This is also possible later. • This is usually followed by the entry of zeros 1e, 2e. This is also possible later. • Then we increase the number of poles in a cycle up to a given value 3e, after which we look at the completed drawings in order and select the one with a fairly good approximation, but the number of parameters is not yet too large. If there is no such, we have to find the reason by looking at the drawings. The section below describes in detail the possible causes of the problems and some methods that have been developed in practice to eliminate these causes. 4.12. Preparation We use a preparatory program (prep) to cut out the desired period of the signal to be approximated from the raw ECG signals. First, we - temporarily - remove baseline fluctuations and higher frequency noise with frequency filters. Then, similarly to a method described above, we find the locations of the QRS (the time values of the maximum activity point of the ECG signal). Then, by marking the beginnings of the P waves, we get the initial baseline segments of each period. Finally, after omitting noise filtering - we plot the periods of the signal and their initial baseline segments with MS Excel. Now, the entire baseline can be moved by grabbing a point on the baseline segment using the mouse. It is advisable to first choose the captured point close to the beginning of the baseline segment, then - in response to an Excel question - the beginning of the line will move and the end will remain in place. Then the same will happen at a point close to the end. This way we can accurately set the standard final baseline used in cardiology. In what follows, we use the original signal without the temporary filters. 4.11. Problems and their practical solutions If the imaginary part of a root (conjugate complex pole or zero) is much larger than the others, then the factor containing the root can probably be omitted; of course, this must be checked. - If we want to improve the approximation on a signal segment (e.g. 4e), but the program that performs the approximation (progr1-123) reduces the deviation elsewhere, although the former would be a more important location. For example, if we want the program to make a “nicer” approximation on the ST segment than perhaps on the U wave. In this case, it may be useful to increase the weighting of the deviations in this segment (the default value is 1 everywhere). In practice, after one approximation step (one run), the increased weighting can usually be reset to 1. We already use the approximation tools 1e-4e through a framework program (ppp.m) written in the GNU Octave language. (GNU Octave: "The Octave syntax is largely compatible with Matlab".) This framework, which runs in a cycle, sets the starting parameters of the approximation program (progr1-123) based on answer choices and selections with the mouse on the approximation drawings. It then calls the approximation program (progr1-123). Exit from the cycle is not triggered by the reduction of the numerical error of the approximation to a suitable value, but rather by looking at the segments that are important from a diagnostic point of view, the approximation is correct [FEX3-ECG/Charts01]. Then we looked at approximations of some of the more important shape types (not necessarily just with Frank lead)[FEX3-ECG/Charts02: Charts02.zip\Shapes\]: Except for the following 6, because it was immediately obvious (without any examination) that these signals fall into the “needs to be seen by a cardiologist” category: Except for the following 6, because it was immediately obvious (without any examination) that these signals fall into the “needs to be seen by a cardiologist” category: s0007_p146------- arithmia s0018_p113------- atrial fibrillation, no P (in either lead) s0023_p115------- arithmia s0024_p164------- arithmia s0032_p168------- arithmia s0033_p168------- arithmia s0007_p146------- arithmia s0018_p113------- atrial fibrillation, no P (in either lead) s0023_p115------- arithmia s0024_p164------- arithmia s0032_p168------- arithmia s0033_p168------- arithmia s0007_p146------- arithmia s0018_p113------- atrial fibrillation, no P (in either lead) s0023_p115------- arithmia s0024_p164------- arithmia s0032_p168------- arithmia s0033_p168------- arithmia Since there was no healthy patient among the first 35 signals, we also approximated the ECG signal of a healthy patient: s0306_p104 Summarising the above, we have approximated 95 {5×1 + (35-6+1)×3} ECG curves so far using the tools described above 1e-4e, in our opinion all of them of appropriate quality. Then we looked at approximations of some of the more important shape types (not necessarily just with Frank lead)[FEX3-ECG/Charts02: Charts02.zip\Shapes\]: s0025_p005, V3, 7009 Under several names: dómszerű T, dome-like T, T-en-dôme, ST en dôme, Coved ST segment elevation, domed shape, pierre tombale (fr), tombstone (en), scaphoid Ta P32 T34 QRS34 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0025_p005_V3\] s0047_p015, II, 2758 Under several names: sajkaszerű, spoon-shaped, boat-like, navicular, digitalis effect, digoxin effect, reverse tick, shape of Dali’s moustache P22 T33 QRS34 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0047_p015_II\] s0146_p044, V4, 3725 Under several names: koronária T, coronary T, Coronaria T Ta P22 T22 QRS23 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0146_p044_V4\] s0508_p269, aVR, 5019 Under several names: Blokk, Block, Bundle branch block Ta P12 T12 QRS24 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0508_p269_aVR\] s0542_p283, vz, 2581 Under several names: QRS W alakú, QRS W shape (atrial fibrillation, no P) P-- T22 QRS34 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0542_p283_vz\] s0047_p015, II, 2758 Under several names: sajkaszerű, spoon-shaped, boat-like, navicular, digitalis effect, digoxin effect, reverse tick, shape of Dali’s moustache Ta P22 T22 QRS23 [FEX3-ECG/Charts02: Charts02.zip\Shapes\s0146_p044_V4\] s0508 p269, aVR, 5019 Under several names: Blokk, Block, Bundle branch block Next, we approximated all ECG signals (except for 6) in sequence from s0001 to s0035, all successfully. In detail, complete with patient numbers: Next, we approximated all ECG signals (except for 6) in sequence from s0001 to s0035, all successfully. In detail, complete with patient numbers: s0001_p119\ s0002_p196\ s0003_p110\ s0004_p152\ s0005_p138\ s0006_p125\ s0008_p193\ s0009_p175\ s0010_p001\ s0011_p178\ s0012_p114\ s0013_p108\ s0014_p001\ s0015_p002\ s0016_p001\ s0017_p003\ s0019_p140\ s0020_p004\ s0021_p005\ s0022_p006\ s0025_p005\ s0026_p007\ s0027_p006\ s0028_p008\ s0029_p007\ s0030_p106\ s0031_p005\ s0034_p163\ s0035_p009\ Five example drawings [FEX3-ECG/Charts02: Charts02.zip\From_series\]. s0001_p119\ s0002_p196\ s0003_p110\ s0004_p152\ s0005_p138\ s0006_p125\ s0008_p193\ s0009_p175\ s0010_p001\ s0011_p178\ s0012_p114\ s0013_p108\ s0014_p001\ s0015_p002\ s0016_p001\ s0017_p003\ s0019_p140\ s0020_p004\ s0021_p005\ s0022_p006\ s0025_p005\ s0026_p007\ s0027_p006\ s0028_p008\ s0029_p007\ s0030_p106\ s0031_p005\ s0034_p163\ s0035_p009\ Five example drawings [FEX3-ECG/Charts02: Charts02.zip\From_series\]. 5. Results The signals tested are denoted as follows: s.... = serial number, p... = patient number. For example "s0025_p005". If we want to be more precise about the location within the record, we also indicate the lead and the location of the QRS. For example “s0025_p005, V3, 7009”. In the section below, the two numbers after the wave letter are related to the number of approximation parameters. For example: P35 stands for wave P, 3 is the number of parameters of the numerator (including the leading coefficient d), while 5 is the number of pole pairs (factors) of the denominator. Thus, the number of parameters of the approximation is 3 + 2×5=13. Ta means: there is wave Ta. The 2×6 digit number (for example in the filename: yymmdd hhmmss) means “year month day hour minute second” (time stamp). After our initial trials, we first approximated all 125 periods of the s0177_p050 signal vx vy vz (Frank lead) (location of the first QRS: 1627 ms). The goal was to test the stability of the parameters. The results of the approximations show that the stability of the parameters is adequate. An example of this is the histogram of one of the location parameters (h) for the denominator of the T wave on the vz lead [FEX3-ECG/Charts02: Charts02.zip\Histogram\]. This is the location of a pole near the peak of the T wave. To do this, after omitting two outliers (in QRS Periods 47 and 56), the histogram of the remaining 123 values was plotted, which was similar to a bell curve. The mean was 257 (with all the values between 251 and 263), or 257±6 ms. For a conventional recording paper speed, this is equivalent to ±0.15 mm. The accuracy of the other parameters was similar to this one. The accuracy of the other parameters was similar to this one. The accuracy of the other parameters was similar to this one. The accuracy of the other parameters was similar to this one. Then we looked at approximations of some of the more important shape types (not necessarily just with Frank lead)[FEX3-ECG/Charts02: Charts02.zip\Shapes\]: 6. Conclusions The used MINPACK algorithm had no problems, even with a large number of parameters (e.g. 26). With some other algorithms, with a large number of parameters, the “ill-conditioned matrix” problem occurs often. At the beginning of our article, in the “Objective” section, we suggested that initially it would be appropriate to classify the data into only 3 categories. This suggestion is supported by the relatively large number of parameters, as well. 6. Conclusions The process of approximating a wave (P, QRS, or T) usually consists of applying the above tools 1e-4e in the sequence of our choice. It is relatively rarely necessary to revert to a previous approximation, and so far no approximation has ever required returning twice. It is often difficult to decide what level of accuracy is needed. What part of the signal from the database is noise and what part is the signal generated by the electrical activity of the heart. The number of approximation parameters also depends on this. The biggest problem at the moment is processing time. This is of course related to the desired accuracy of the approximation and the number of parameters. When running the approximation program itself (progr1-123), The biggest problem at the moment is processing time. This is of course related to the desired accuracy of the approximation and the number of parameters. When running the approximation program itself (progr1-123), an approximation step takes only a few seconds. However, the settings during the preparation (prep), the review of the drawings after each extension, and the total time for preparing the initial parameters based on them: for the waves of 1 period of 1 lead, it can take up to half an hour in total, even using the framework program written in the "GNU Octave language" (ppp.m) as well. This can be as much as one and a half hours for three leads! an approximation step takes only a few seconds. However, the settings during the preparation (prep), the review of the drawings after each extension, and the total time for preparing the initial parameters based on them: for the waves of 1 period of 1 lead, it can take up to half an hour in total, even using the framework program written in the "GNU Octave language" (ppp.m) as well. This can be as much as one and a half hours for three leads! A trained cardiologist rarely needs more than 5 minutes to analyse an ECG for screening tests (in simple cases). (Of course, in more serious cases, they may need much more time and may recommend other tests.) The approximation program (progr1-123) worked reliably. The used MINPACK algorithm had no problems, even with a large number of parameters (e.g. 26). With some other algorithms, with a large number of parameters, the “ill-conditioned matrix” problem occurs often. The approximation program (progr1-123) worked reliably. 7. References [1] Kobzos László, TKI. “Az EKG parametrizálásának egy biztató kísérlete”. Számítástechnikai és kibernetikai módszerek... 6. kollokvium, Szeged, 1975, pp. 181-183 [1] Kobzos László, TKI. “Az EKG parametrizálásának egy biztató kísérlete”. Számítástechnikai és kibernetikai módszerek... 6. kollokvium, Szeged, 1975, pp. 181-183 [2] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete II”. Orvosi Informatika 2014., A XXVII. Neumann Kollokvium konferencia-kiadványa, ISBN 978-963-396-040-0, https://doi.org/10.5281/zenodo.3648260 pp. 25-28 [2] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete II”. Orvosi Informatika 2014., A XXVII. Neumann Kollokvium konferencia-kiadványa, ISBN 978-963-396-040-0, https://doi.org/10.5281/zenodo.3648260 pp. 25-28 [3] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete III”. Orvosi Informatika 2016., A XXIX. Neumann Kollokvium konferencia-kiadványa, ISBN 978-963-306-514-3, https://doi.org/10.5281/zenodo.3666134 pp. 103-106 [3] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete III”. Orvosi Informatika 2016., A XXIX. Neumann Kollokvium konferencia-kiadványa, ISBN 978-963-306-514-3, https://doi.org/10.5281/zenodo.3666134 pp. 103-106 [4] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete IV”. Orvosi Informatika 2018., A XXXI. Neumann Kollokvium konferencia-kiadványa, ISBN 978-615-5036-14-9, https://doi.org/10.5281/zenodo.3674617 pp. 13-19 [4] dr. Kobzos László, “Az EKG parametrizálásának egy biztató kísérlete IV”. Orvosi Informatika 2018., A XXXI. Neumann Kollokvium konferencia-kiadványa, ISBN 978-615-5036-14-9, https://doi.org/10.5281/zenodo.3674617 pp. 13-19 [5] dr. Kobzos László, “Az EKG parametrizálásának egy bíztató kísérlete V.”. Orvosi Informatika 2020., A XXXIII. Neumann Kollokvium konferencia, Online presentation only [5] dr. Kobzos László, “Az EKG parametrizálásának egy bíztató kísérlete V.”. Orvosi Informatika 2020., A XXXIII. Neumann Kollokvium konferencia, Online presentation only [6] J.A. Kors, J.L. Talmon, J.H. Van BemmeL Multilead ECG analysis Comput. Biomed. Res., 19 (1) (1986), pp. 28-46 [ ] , , y p , ( ) ( ), pp [7] Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, et al. "PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals" Circulation 101(23):e215-e220 [Circulation Electronic Pages; http://circ.ahajournals.org/cgi/content/full/101/23/e215]; 2000 (June 13). PMID: 10851218; doi: 10.1161/01.CIR.101.23.e215 [7] Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, et al. "PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals" Circulation 101(23):e215-e220 [Circulation Electronic Pages; http://circ.ahajournals.org/cgi/content/full/101/23/e215]; 2000 (June 13). PMID: 10851218; doi: 10.1161/01.CIR.101.23.e215 [ ] p g p [9] http://hqmeded-ecg.blogspot.hu/2011/11/atrial-repolarization-wave-mimicking-st.html [10] http://gcc.gnu.org/fortran/ [11] https://doi.org/10.13026/C28C71 [12] https://math.stackexchange.com/questions/1041908/can-all-real-polynomials-be-factored-into-quadratic-and-linear-factors [13] https://www.sciencedirect.com/science/article/abs/pii/S0022073668800135 [14] https://www.sciencedirect.com/science/article/abs/pii/0022073688901136
https://openalex.org/W4379255801	https://zenodo.org/records/8000878/files/CAJEI%200571.pdf	Russian	null	ПРИМЕНЕНИЕ "LINI SEMINA" В МЕДИЦИНЕ ДЛЯ ЛЕЧЕНИЕ ЗАБОЛЕВАНИЙ ЖЕЛУДОЧНО-КИШЕЧНОГО ТРАКТА	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	3,705	CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz Цвет масла светло-желтый с буроватым оттенком, запах характерный, вкус приятный [1, 11-12]. Масло состоит в основном из двух-, трехкислотных триглицеридов олеиновой, линолевой и линоленовой кислот. Кислотное число не более 5. Содержание масла в семенах зависит от разновидности льна, районов произрастания и почвенно- климатических условий, поэтому колеблется в широких пределах - от 24 до 44%. Еще большее разнообразие наблюдается в составе масла: по мере продвижения посевов с севера на юг в составе триглицеридов уменьшается количество линолевой (с 60 до 25%) и линоленовой (с 45 до 20 %) кислот. В соответствии с этим колеблется и величина йодного числа: масло из северных районов имеет более высокое йодное число [2, 13-14]. Долгое время применение льняного масла в медицинской практике ограничивалось приготовлением линимента от ожогов. Открытие способности полиненасыщенных жирных кислот ускорять распад и выводить липиды из организма сразу и резко подняло лекарственную ценность льняного масла [2, 15-16]. Более того, оказалось, что полиненасыщенные жирные кислоты являются биологически исходными веществами для биосинтеза в человеческом и животном организме простагландинов, проявляющих разнообразное физиологическое действие. Из льняного масла получают препарат линетол, представляющий собой смесь этиловых эфиров линолевой (около 15%), линоленовой (около 57%), олеиновой (около 15%) кислот; остальное количество приходится на предельные кислоты [2, 17-18]. Этерификация жирных кислот улучшает органолептические свойства масла; препарат лучше переносится. Применяется линетол при атеросклерозе и в виде мази наружно при ожогах, лучевых поражениях и других болезнях кожи [2, 19-20]. В состав льняного семени входят три вида ценных полиненасыщенных жирных кислот (Омега-3, Омега-6 и Омега-9), правильный баланс которых необходим для всех процессов жизнедеятельности человеческого организма. По содержанию Омега-3 семена льна превосходят все пищевые растительные масла (этой кислоты в семени льна в 3 раза больше, чем в рыбьем жире) [1, 21-22]. Аминокислотный состав белка льняного семени аналогичен составу растительных протеинов сои, знаменитых своей пищевой ценностью. Семена льна являются отличным источником растительной клетчатки, повышенное содержание которой в продуктах способствует снижению риска онкологических заболеваний и положительно влияет на иммунитет [1, 23-24]. Благодаря высокому содержанию полисахаридов отвар из семян льна оказывает обволакивающее и бактерицидное действие при гастрите и язве желудка [1, 25-26]. Лигнаны («растительные гормоны»), которых в семенах льна в 100 раз больше, чем в других растительных продуктах, хорошо известны как антиоксиданты, препятствующие развитию рака. Они также обладают антибактериальным и антивирусным эффектом [1, 27-28]. Витамины F, A, E, B. Семена льна - отличный внешний источник важного для организма витамина F, активно участвующего в жировом и холестериновом обмене (этот витамин не синтезируется в организме). ПРИМЕНЕНИЕ “LINI SEMINA” В МЕДИЦИНЕ ДЛЯ ЛЕЧЕНИЕ ЗАБОЛЕВАНИЙ ЖЕЛУДОЧНО-КИШЕЧНОГО ТРАКТА Музаффарова Нигора Сафаровна Бухарский государственный медицинский институт имени Абу Али ибн Сино https://doi.org/10.5281/zenodo.8000878 Музаффарова Нигора Сафаровна Бухарский государственный медицинский институт имени Абу Али ибн Сино https://doi.org/10.5281/zenodo.8000878 ABSTRACT Применение в народной медицине препараты на основе лекарственных растений не потеряет свое значение и особо уделяется внимание на применении этих средств. Лекарственные препараты на основе семена льна является обволакивающие средство растительного происхождения, и мы рассматриваем в данном обзоре Льна посевного семена производимым Фармацевтическим компаниям «ФармаЦвет». ABSTRACT Применение в народной медицине препараты на основе лекарственных растений не потеряет свое значение и особо уделяется внимание на применении этих средств. Лекарственные препараты на основе семена льна является обволакивающие средство растительного происхождения, и мы рассматриваем в данном обзоре Льна посевного семена производимым Фармацевтическим компаниям «ФармаЦвет». Qabul qilindi: 25-May 2023 yil Ma’qullandi: 28-May 2023 yil Nashr qilindi: 31-May 2023 yil KEY WORDS Семена льна, лень посевной, Lini semina, применение в медицине, язвенная болезнь, колит, энтероколит, воспаления пищевода, вяжущие средство. Введение. Первооткрывателем лечебных свойств семени льна был еще Гиппократ, поведавший миру рецепт отвара из льняных семян, помогающий при болезнях желудка. А в более поздние времена семена льна нашли применение в народной медицине благодаря их смягчающим, очищающим и бактерицидным свойствам [1, 3-4]. В 80-х годах прошлого века диетологи разных стран мира стали активно и глубоко изучать свойства льняного семени, как натурального и полезного для здоровья продукта. Ну а в XXI веке семя льна стало играть уже значительную и важную роль в диетическом питании человека, принимая, таким образом, участие в профилактике и лечении ряда заболеваний [1, 5-7]. Это подтверждают данные: В Германии в сфере хлебопечения используется более 60000 тонн семени льна ежегодно (в среднем это около 1 кг на 1 человека в год). А в Канаде семя льна рассматривают уже не как пищевую добавку, а как отдельный продукт питания, в связи с чем принята специальная Национальная программа, рекомендующая включение до 12 % семян льна в хлебобулочные изделия [1, 8-10]. Материалы и методы исследования. Материалам данного исследования является лекарственные растения Лен обыкновенный, лекарственным веществом относящего к группе лекарственные растение и сырье, содержащие жирные масла. Химический состав. Семена льна содержат слизь, жирное масло, белковые вещества, витамин А, органические кислоты, гликозид линамарин и другие биологически активные вещества. Результаты исследования. Льняное масло получают путем горячего прессования из измельченных семян. Volume 2, Issue 5, Part 3 May 2023 CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION Витамины А и Е («витамины молодости») оказывают благотворное влияние на кожу - именно благодаря им льняные семена нашли применение во множестве косметических рецептов [1, 29-30]. Page 356 Volume 2, Issue 5, Part 3 May 2023 CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz Кроме того, семена льна являются важным источником селена, который, в свою очередь, препятствует развитию опухолей, очищает организм от тяжелых металлов, помогает улучшить зрение и мозговую деятельность. Богато семя льна также и лецитином, столь полезным для человека [1, 31-32]. При заболеваниях пищеварительной системы отвар из семян льна благодаря своему обволакивающему и смягчающему действию оказывает защитное действие на слизистую оболочку пищевода и желудка, и может применяться для лечения язвы желудка и гастрита. Клетчатка, которой богаты семена льна, активизирует прежде всего деятельность кишечника, помогая человеку справиться с хроническими запорами. Такое «слабительное» действие семян льна связано с тем, что, разбухая в кишечнике, они увеличивают объем содержимого кишечника и стимулируют таким образом его опорожнение. Ежедневный прием 50 граммов семян льна в течении двух недель - эффективное средство народной медицины для лечения заболеваний кишечника даже у пожилых людей со слабым иммунитетом. Регулярное употребление семян льна помогает значительно улучшить функцию печени, значительно тормозит всасывание токсинов, способствуют очистке организма от шлаков [1, 33-34]. При сердечно-сосудистых заболеваниях семена льна содержат в большом количестве полиненасыщенную жирную кислоту Омега-3, присутствие которой в организме способствует снижению уровня холестерина в крови и кровяного давления. Это нашло применение в профилактики и лечении атеросклероза, инсульта, инфаркта, тромбозов и других расстройств сердечно-сосудистой системы [1, 35-36]. При онкологических заболеваниях семена льна содержат в себе два важных компонента, снижающих риск возникновения гормонально зависимых раковых заболеваний - это Омега-3 кислота и «растительные гормоны» лигнаны [1, 37-38]. При сахарном диабете семена льна обладает усиливающим действием инсулина, находят также применение в профилактике и лечении диабета [1, 39-40]. При воспалительных заболеваниях омега-3, содержащаяся в семенах льна, повышает в целом общий иммунитет организма, что препятствует возникновению и развитию воспалительных процессов. Компрессы из льняного семени размягчают фурункулы и нарывы, находят применение в лечении заболеваний суставов, а настойки из семян льна используют в лечении воспалительных процессов в полости рта и горла [1, 41-43]. Для очищения организма от радионуклеидов семена льна считается мощным сорбентом, по своим свойствам не уступающий активированному углю, и в отличие от искусственных сорбентов не влияет на разрушение клеток организма. На этом основано его применение в лечении больных лучевой болезнью [1, 44-46]. В народной медицине семя льна известно, как хорошее отхаркивающее средство. Регулярное употребление в пищу семени льна облегчает приступы астмы и способствует облегчению течения аллергических заболеваний, улучшает функции почек, щитовидной железы, нормализует гормональный баланс в организме женщины [1, 47-49]. Выводы. CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz организма. Настой семян льна обладает обволакивающим, противовоспалительным и мягким слабительным действием. Набухшие в воде семена льна обладают слабительным действием. Перспективы дальнейшего исследования связано с изыскание новых свойств и создавать современные лекарственные формы с дополнительным фармакологическим свойством лекарственного растения. организма. Настой семян льна обладает обволакивающим, противовоспалительным и мягким слабительным действием. Набухшие в воде семена льна обладают слабительным действием. Перспективы дальнейшего исследования связано с изыскание новых свойств и создавать современные лекарственные формы с дополнительным фармакологическим свойством лекарственного растения. CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION Из вышеуказанной информации приводим к выводу, что лекарственный препарат производимым фармацевтическим компаниям «ФармаЦвет» обладает ряд полезных свойств для желудочно-кишечного тракта человеческого Page 357 Volume 2, Issue 5, Part 3 May 2023 References: 1.https://pharmatsvet.ru/products/lna- 1.https://pharmatsvet.ru/products/lna semena/#:~:text=%D0%91%D0%BB%D0%B0%D0%B3%D0%BE%D0%B4%D0%B0%D1% 80%D1%8F%20%D1%81%D0%BE%D0%B4%D0%B5%D1%80%D0%B6%D0%B0%D0%B D%D0%B8%D1%8E%20%D0%BB%D0%B5%D1%87%D0%B5%D0%B1%D0%BD%D0%B E%D0%B9%20%D1%81%D0%BB%D0%B8%D0%B7%D0%B8%2C%20%D1%81%D0%BF %D0%BE%D1%81%D0%BE%D0%B1%D0%BD%D0%BE%D0%B9,%D0%BA%D0%BE%D0 %BB%D0%B8%D1%82%D0%B0%2C%20%D1%8D%D0%BD%D1%82%D0%B5%D1%80 %D0%BE%D0%BA%D0%BE%D0%BB%D0%B8%D1%82%D0%B0%2C%20%D0%B2%D0 %BE%D1%81%D0%BF%D0%B0%D0%BB%D0%B5%D0%BD%D0%B8%D0%B8%20%D0 %BF%D0%B8%D1%89%D0%B5%D0%B2%D0%BE%D0%B4%D0%B0. 2. Мусаева, Д.М., Самадов, Б.Ш., Очилова, Г.С., Лекарственные растения [Текст] : учебник / Д.М.Мусаева, Б.Ш.Самадов, Г.С.Очилова.-Бухара: “Sadriddin Salim Buxoriy” Durdona, 2021.-184 с. 2. Мусаева, Д.М., Самадов, Б.Ш., Очилова, Г.С., Лекарственные растения [Текст] : учебник / Д.М.Мусаева, Б.Ш.Самадов, Г.С.Очилова.-Бухара: “Sadriddin Salim Buxoriy” Durdona, 2021.-184 с. 3. Б.Ш. Самадов, Ф.С. Жалилова, Ф.С. Жалилов, Н.А. Муродова., Фармакологическая свойства и химический состав лекарственного растительного сырья “Momordica Charantia L”. Матеріали ІV Міжнародної науково-практичної конференції. Харків, НФаУ, 2020. С. 426-430. 3. Б.Ш. Самадов, Ф.С. Жалилова, Ф.С. Жалилов, Н.А. Муродова., Фармакологическая свойства и химический состав лекарственного растительного сырья “Momordica Charantia L”. Матеріали ІV Міжнародної науково-практичної конференції. Харків, НФаУ, 2020. С. 426-430. 4. Самадов, Б. Ш., Жалилова, Ф. С., Жалилов, Ф. С., & Муродова, Н. А. (2020). ФАРМАКОЛОГИЧЕСКАЯ СВОЙСТВА И ХИМИЧЕСКИЙ СОСТАВ ЛЕКАРСТВЕННОГО РАСТИТЕЛЬНОГО СЫРЬЯ “MOMOR-DICA CHARANTIA L”. Новый день в медицине. Научно-реферативный, духовно-просветительский журнал, 1, 29. 5. Дубинина, Н. В., Дубініна, Н. В., Самадов, Б. Ш., Тищенко, И. Ю., & Тіщенко, І. Ю. (2020). Перспективы использования лекарственного сырья момордика харанция для создания новых лекарственных средств. 6. Самадов, Б. Ш., & Мусаева, Д. М. (2020). Тенденция развития эпидемического процесса гепатита С в Узбекистане. Матеріали ІV Міжнародної науково-практичної конференції. НФаУ, Харьков. Украина, 430-437. 6. Самадов, Б. Ш., & Мусаева, Д. М. (2020). Тенденция развития эпидемического процесса гепатита С в Узбекистане. Матеріали ІV Міжнародної науково-практичної конференції. НФаУ, Харьков. Украина, 430-437. 7. Samadov, B. S., & Dubinina, N. V. (2016). Characteristics and trends of epidemic of hepatitis C in Uzbekistan and Ukraine. 7. Samadov, B. S., & Dubinina, N. V. (2016). Characteristics and trends of epidemic of hepatitis C in Uzbekistan and Ukraine. 8. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. (2020). ВЫРАЩИВАНИЕ ЛЕКАРСТВЕННОГО РАСТЕНИЯ «MOMORDICA CHARANTIA L» В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ. Вестник науки и образования, (21-1 (99)), 92-98. 9. Дубинина, Н. В., Самадов, Б. Ш., Тищенко, И. Ю., Дубініна, Н. В., & Тіщенко, І. Ю. (2020). Вирусные гепатиты с парентеральным механизмом передачи: современные подходы к 8. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. (2020). ВЫРАЩИВАНИЕ ЛЕКАРСТВЕННОГО РАСТЕНИЯ «MOMORDICA CHARANTIA L» В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ. Вестник науки и образования, (21-1 (99)), 92-98. ОБЛАСТИ. Вестник науки и образования, (21 1 (99)), 92 98. 9. Дубинина, Н. В., Самадов, Б. Ш., Тищенко, И. Ю., Дубініна, Н. В., & Тіщенко, І. Ю. (2020). Вирусные гепатиты с парентеральным механизмом передачи: современные подходы к лечению. 9. Дубинина, Н. CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz pharmaceutical ingredients in combined dosage form. pharmaceutical ingredients in combined dosage form. 11. Швець, І. О., Самадов, Б. Ш., Ільїна, Т. В., & Ильина, Т. В. (2017). Навчальна практика з фармакогнозії–складова частина професійної підготовки провізора. 11. Швець, І. О., Самадов, Б. Ш., Ільїна, Т. В., & Ильина, Т. В. (2017). Навчальна практика з фармакогнозії–складова частина професійної підготовки провізора. 12. Samadov, B., Sych, I. A., Shpychak, T. V., & Kiz, O. V. (2017). Quantitative determination by potentiometric titration method of active pharmaceutical ingredients in complex dosage form. 12. Samadov, B., Sych, I. A., Shpychak, T. V., & Kiz, O. V. (2017). Quantitative determination by potentiometric titration method of active pharmaceutical ingredients in complex dosage form. 13. Самадов, Б. Ш., Жалилов, Ф. С., Жалилова, Ф. С., & Шарипова Э.М. (2021). ХИМИЧЕСКИЙ СОСТАВ ЛЕКАРСТВЕННОГО СЫРЬЯ “MOMORDICA CHARANTIA L”, ВЫРАЩИВАННОГО В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ РЕСПУБЛИКИ УЗБЕКИСТАН. Вестник науки и образования, (15-1), 106-110. 13. Самадов, Б. Ш., Жалилов, Ф. С., Жалилова, Ф. С., & Шарипова Э.М. (2021). ХИМИЧЕСКИЙ СОСТАВ ЛЕКАРСТВЕННОГО СЫРЬЯ “MOMORDICA CHARANTIA L”, ВЫРАЩИВАННОГО В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ РЕСПУБЛИКИ УЗБЕКИСТАН. Вестник науки и образования, (15-1), 106-110. 14. Дубинина, Н. В., Самадов, Б. Ш., & Тищенко, И. Ю. (2021). Создание вакцин для профилактики и лечения ВИЧ. 14. Дубинина, Н. В., Самадов, Б. Ш., & Тищенко, И. Ю. (2021). Создание вакцин для профилактики и лечения ВИЧ. 15. Samadov, B. S. (2022). THE USE OF THE MEDICINAL PLANT MOMORDICA CHARANTIA L IN FOLK MEDICINE. Asian journal of pharmaceutical and biological research, 11(2). 16. Bakhodirjon Sharipovich Samadov. (2022). THE CHEMICAL COMPOSITION OF THE MEDICINAL PLANT MOMORDICA CHARANTIA L USED IN FOLK MEDICINE. Thematics Journal of Chemistry, 6(1). 17. Samadov, B. S. (2022). ANATOMICAL STRUCTURE OF THE MEDICINAL PLANT MOMORDICA CHARANTIA L. Thematics Journal of Botany, 6(1). 18. Самадов, Б. Ш., Болтаев, М. М., Мелибоева, Ш. Ш., & Жалилов, Ф. С. (2022). ГИПОЛИПИМИДЕМИЧЕСКАЯ АКТИВНОСТЬ СЫРЬЯ ПЛОДЫ МОМОРДИКА ХАРАНЦИЯ (MOMORDICA CHARANTIA L). Central Asian Academic Journal of Scientific Research, 2(8), 26- 35. 19. Samadov, B. S., Jalilova, F. S., Ziyaeva, D. A., Sharipova, D. S., Ozodova, N. X., & Norova, H. U. & Kudina, OV (2020). Pharmacological properties and chemical composition “Momordica charantia l. 20. Самадов, Б. Ш. (2020). Жалилов Фазлиддин Содикович, Жалилова Феруза Содиковна. ВЫРАЩИВАНИЕ ЛЕКАРСТВЕННОГО РАСТЕНИЯ «MOMORDICA CHARANTIA L» В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ. Вестник науки и образования, (21-1), 99. 21. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). References: В., Самадов, Б. Ш., Тищенко, И. Ю., Дубініна, Н. В., & Тіщенко, І. Ю. (2020). Вирусные гепатиты с парентеральным механизмом передачи: современные подходы к лечению. Page 358 Volume 2, Issue 5, Part 3 May 2023 CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION CHARANTIA L DORIVOR O’SIMLIGINING KIMYOVIY TARKIBI. Журнал химии товаров и народной медицины, 1(4), 134-161. DOI: https://doi.org/10.55475/jcgtm/vol1.iss4.2022.86 27. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). PROSPECTS FOR OBTAINING DOSAGE FORMS BASED ON MOMORDICA CHARANTIA L. Scientific progress, 3(8), 29-32. 28. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). PROSPECTS FOR OBTAINING DOSAGE FORMS BASED ON LOCALIZED INDIAN POMEGRANATE. Scientific progress, 3(8), 33-41. 29. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). COMPOSITION AND TECHNOLOGY OF COLLECTION OF MOMORDICA CHARANTIA L OBTAINED FROM MEDICINAL PLANT RAW MATERIALS. Scientific progress, 3(8), 42-48. 29. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). COMPOSITION AND TECHNOLOGY OF COLLECTION OF MOMORDICA CHARANTIA L OBTAINED FROM MEDICINAL PLANT RAW MATERIALS. Scientific progress, 3(8), 42-48. 30. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). ANALYSIS OF THE COMPONENTS OF THE COLLECTION OF MEDICINAL PLANT RAW MATERIALS OF MOMORDICA CHARANTIA L. Scientific progress, 3(8), 49-57. 31. Samadov, B. S., Zhalilov, F. S., & Zhalilova, F. S. (2022). HYPOLIPIDEMIC ACTIVITY OF THE MEDICINAL PLANT MOMORDICA HARANTIA. Medical Scientific Bulletin of Central Chernozemye (Naučno-medicinskij vestnik Centralʹnogo Černozemʹâ), (89), 57-69. 31. Samadov, B. S., Zhalilov, F. S., & Zhalilova, F. S. (2022). HYPOLIPIDEMIC ACTIVITY OF THE MEDICINAL PLANT MOMORDICA HARANTIA. Medical Scientific Bulletin of Central Chernozemye (Naučno-medicinskij vestnik Centralʹnogo Černozemʹâ), (89), 57-69. 32. Самадов, Б. Ш., Джалилов, Ф. С., & Джалилова, Ф. С. (2022). MOMORDICA CHARANTIA L DORIVOR O’SIMLIGINING ANATOMIK TUZILISHI. Журнал химии товаров и народной медицины, 1(5), 123-149. https://doi.org/10.55475/jcgtm/vol1.iss5.2022.109 33. Samadov, B. S., Jalilov, F. S., Yuldasheva, D. H., Jalilova, F. S., Boltayev, M. M., & qizi Meliboyeva, S. S. APPLICATION IN FOLK MEDICINE FRUITS OF THE MEDICINAL PLANT MOMORDICA CHARANTIA L. 34. Samadov, B. S., Jalilov, F. S., Yuldasheva, D. H., Boltayev, M. M., & qizi Meliboyeva, S. S. THE CHEMICAL COMPOSITION OF THE MEDICINAL PLANT MOMORDICA CHARANTIA L USED IN TRADITIONAL MEDICINE. 35. Samadov, B. S., & Musaeva, D. M. (2020, March). Trends in the development of the epidemic process of hepatitis C in Uzbekistan. In Proceedings of the 4th International Scientific and Practical Conference “Faces-people. Current problems of pharmacotherapy and recognition of medicinal benefits. Kharkiv (Vol. 1, p. 431). 36. Samadov, B. S., Musaeva, D. M., & Dubinina, N. V. (2020). CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION COMPOSITION AND TECHNOLOGY OF COLLECTION OF INDIAN POMEGRANATE OBTAINED FROM MEDICINAL PLANT RAW MATERIALS. Редакційна колегія, 40. 22. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). ANALYSIS OF THE COMPONENTS OF THE COLLECTION OF MEDICINAL PLANT RAW MATERIALS OF INDIAN POMEGRANATE. Редакційна колегія, 43. 22. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). ANALYSIS OF THE COMPONENTS OF THE COLLECTION OF MEDICINAL PLANT RAW MATERIALS OF INDIAN POMEGRANATE. Редакційна колегія, 43. д ц , 23. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). PROSPECTS FOR OBTAINING DOSAGE FORMS BASED ON MOMORDICA CHARANTIAL. Редакційна колегія, 37. 24. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). PROSPECTS FOR OBTAINING DOSAGE FORMS BASED ON LOCALIZED INDIAN POMEGRANATE. Редакційна колегія, 169. 25. Самадов, Б. Ш., Джалилов, Ф. С., Юлдашева, Д. Х., Джалилова, Ф. С., Болтаев, М. М., & Мелибоева, Ш. Ш. к. (2022). ПРИМЕНЕНИЕ В НАРОДНЫЕ МЕДИЦИНЫ ПЛОДЫ ЛЕКАРСТВЕННОГО РАСТЕНИЯ MOMORDICA CHARANTIA L. Журнал химии товаров и народной медицины, 1(4), 117–133. https://doi.org/10.55475/jcgtm/vol1.iss4.2022.76 26. Самадов, Б. Ш., Джалилов, Ф. С., Юлдашева, Д. Х., Джалилова, Ф. С., Болтаев, М. М., & кизи Мелибоева, Ш. Ш. (2022). XALQ TABOBATIDA ISHLATILADIGAN MOMORDICA 23. Samadov, B. S., Jalilova, F. S., & Jalilov, F. S. (2022). PROSPECTS FOR OBTAINING DOSAGE FORMS BASED ON MOMORDICA CHARANTIAL. Редакційна колегія, 37. Page 359 Volume 2, Issue 5, Part 3 May 2023 CENTRAL ASIAN JOURNAL OF EDUCATION CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION Comparative characteristics and trends in the development of the epidemic process of hepatitis C in Ukraine and Uzbekistan. New Day in Medicine, 1(29), 284-290. trends in the development of the epidemic process of hepatitis C in Ukraine and Uzbekistan. New Day in Medicine, 1(29), 284-290. 37. Samadov, B. S., Jalilov, F. S., & Jalilova, F. S. (2022). DOSAGE FORMS BASED ON THE MEDICINAL PLANT MOMORDICA CHARANTIA L. Medical Scientific Bulletin of Central Chernozemye (Naučno-medicinskij vestnik Centralʹnogo Černozemʹâ), (90), 10-18. 38. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. ГИПОЛИПИДЕМИЧЕСКАЯ АКТИВНОСТЬ ЛЕКАРСТВЕННОГО РАСТЕНИЯ МОМОРДИКА ХАРАНЦИЯ. 39. Samadov B. S. MAGNESIUM DEFICIENCY AND ITS CORRECTION WITH VEGETABLE TINCTURE TINCTURAE MORUS //Scientific progress. – 2023. – Т. 4. – №. 3. – С. 4-12. 40. Samadov B. S. CORRECTION MAGNESIUM DEFICIENCY WITH TINCTURE TINCTURAE MORUS //Scientific progress. – 2023. – Т. 4. – №. 2. – С. 369-377. 41. Самадов, Б. Ш., Жалилов, Ф. С., Жалилова, Ф. С., & Дубинина, Н. В. (2022). Антимикробная активность лекарственного растительного сырья “Momordica charantia L.”. 37. Samadov, B. S., Jalilov, F. S., & Jalilova, F. S. (2022). DOSAGE FORMS BASED ON THE MEDICINAL PLANT MOMORDICA CHARANTIA L. Medical Scientific Bulletin of Central Chernozemye (Naučno-medicinskij vestnik Centralʹnogo Černozemʹâ), (90), 10-18. y ( j g ), ( ), 38. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. ГИПОЛИПИДЕМИЧЕСКАЯ АКТИВНОСТЬ ЛЕКАРСТВЕННОГО РАСТЕНИЯ МОМОРДИКА ХАРАНЦИЯ. 39. Samadov B. S. MAGNESIUM DEFICIENCY AND ITS CORRECTION WITH VEGETABLE TINCTURE TINCTURAE MORUS //Scientific progress. – 2023. – Т. 4. – №. 3. – С. 4-12. 40. Samadov B. S. CORRECTION MAGNESIUM DEFICIENCY WITH TINCTURE TINCTURAE MORUS //Scientific progress. – 2023. – Т. 4. – №. 2. – С. 369-377. 41. Самадов, Б. Ш., Жалилов, Ф. С., Жалилова, Ф. С., & Дубинина, Н. В. (2022). Антимикробная активность лекарственного растительного сырья “Momordica charantia L.”. 38. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. ГИПОЛИПИДЕМИЧЕСКАЯ АКТИВНОСТЬ ЛЕКАРСТВЕННОГО РАСТЕНИЯ МОМОРДИКА ХАРАНЦИЯ. 38. Самадов, Б. Ш., Жалилов, Ф. С., & Жалилова, Ф. С. ГИПОЛИПИДЕМИЧЕСКАЯ АКТИВНОСТЬ ЛЕКАРСТВЕННОГО РАСТЕНИЯ МОМОРДИКА ХАРАНЦИЯ. 39. Samadov B. S. MAGNESIUM DEFICIENCY AND ITS CORRECTION WITH VEGETABLE TINCTURE TINCTURAE MORUS //Scientific progress – 2023 – Т 4 – №3 – С 4-12 // p g 41. Самадов, Б. Ш., Жалилов, Ф. С., Жалилова, Ф. С., & Дубинина, Н. В. (2022). Антимикробная активность лекарственного растительного сырья “Momordica charantia L.”. Page 360 Volume 2, Issue 5, Part 3 May 2023 CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION www.in-academy.uz 42. Самадов, Б. Ш., Джалилов, Ф. С., Мусазода, С. М., & Джалилова, Ф. С. (2023). ЛЕКАРСТВЕННЫЕ ФОРМЫ НА ОСНОВЕ ЛЕКАРСТВЕННОГО РАСТЕНИЯ MOMORDICA CHARANTIA L. Журнал химии товаров и народной медицины, 2(1), 139–162. https://doi.org/10.55475/jcgtm/vol2.iss1.2023.149 43. Самадов, Б. Ш., Джалилов, Ф. С., Мусазода, С. М., & Джалилова, Ф. С. (2023). MOMORDICA CHARANTIA L DORIVOR O’SIMLIGI ASOSIDAGI DORI SHAKLLARI. Журнал химии товаров и народной медицины, 2(1), 139-162. https://doi.org/10.55475/jcgtm/vol2.iss1.2023.149 44. Самадов, Б. Ш., Мусаева, Д. М., & Дубинина, Н. В. (2019). Сравнительная MOMORDICA CHARANTIA L DORIVOR O’SIMLIGI ASOSIDAGI DORI SHAKLLARI. Журнал химии товаров и народной медицины, 2(1), 139-162. https://doi.org/10.55475/jcgtm/vol2.iss1.2023.149 44. Самадов, Б. Ш., Мусаева, Д. М., & Дубинина, Н. В. (2019). Сравнительная характеристика и тенденции развития эпидемического процесса гепатита С в Украине и в Узбекистане. Новый день в медицине, (4), 284-290. 45. Самадов Б. Ш., Жалилова Ф. С., Жалилов Ф. С. ХИМИЧЕСКИЙ СОСТАВ ПЛОДЫ “MOMORDICA CHARANTIA L” ВЫРАЩЕННОГО В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ 44. Самадов, Б. Ш., Мусаева, Д. М., & Дубинина, Н. В. (2019). Сравнительная характеристика и тенденции развития эпидемического процесса гепатита С в Украине и в Узбекистане. Новый день в медицине, (4), 284-290. 44. Самадов, Б. Ш., Мусаева, Д. М., & Дубинина, Н. В. (2019). Сравнительная характеристика и тенденции развития эпидемического процесса гепатита С в Украине и в Узбекистане. Новый день в медицине, (4), 284-290. 45. Самадов Б. Ш., Жалилова Ф. С., Жалилов Ф. С. ХИМИЧЕСКИЙ СОСТАВ ПЛОДЫ “MOMORDICA CHARANTIA L” ВЫРАЩЕННОГО В УСЛОВИЯХ БУХАРСКОЙ ОБЛАСТИ РЕСПУБЛИКИ УЗБЕКИСТАН. Матеріали IX Міжнародної науково-практичної internet- конференціі «Сучасні досягнення фармацевтичної технології». Харків, НФаУ. Редакційна колегія. – 2021. – С. 3-7. 46. Bakhodirovich H. D. MAGNESIUM AND POTASSIUM DEFICIENCY AND ITS CORRECTION WITH VEGETABLE TINCTURE TINCTURAE MORUS //AMALIY VA TIBBIYOT FANLARI ILMIY JURNALI. – 2023. – Т. 2. – №. 4. – С. 139-145. 46. Bakhodirovich H. D. MAGNESIUM AND POTASSIUM DEFICIENCY AND ITS CORRECTION WITH VEGETABLE TINCTURE TINCTURAE MORUS //AMALIY VA TIBBIYOT FANLARI ILMIY JURNALI. – 2023. – Т. 2. – №. 4. – С. 139-145. 47. Khaydarov D. PHARMACOLOGICAL ANALYSIS OF THE DRUG “SIRIMOL” //Eurasian Journal of Medical and Natural Sciences. – 2022. – Т. 2. – №. 13. – С. 274-279. 48. Bahodirovich H. D. Use of “Zingiber Officinale Roscoe L” Root in Various Diseases //Central Asian Journal of Medical and Natural Science. – 2023. – Т. 4. – №. 3. – С. 415-422. 49. Haydarov D. TURLI KASALLIKLAR DAVOLSHDA “ZINGIBER OFFICINALE ROSCOE L” DAN 48. Bahodirovich H. D. CENTRAL ASIAN JOURNAL OF EDUCATION AND INNOVATION Use of “Zingiber Officinale Roscoe L” Root in Various Diseases //Central Asian Journal of Medical and Natural Science. – 2023. – Т. 4. – №. 3. – С. 415-422. 49. Haydarov D. TURLI KASALLIKLAR DAVOLSHDA “ZINGIBER OFFICINALE ROSCOE L” DAN FOYDALANISH //Центральноазиатский журнал образования и инноваций. – 2023. – Т. 2. – №. 5 Part 3. – С. 220-226. Page 361 Page 361 Volume 2, Issue 5, Part 3 May 2023
https://openalex.org/W3149246347	http://pure-oai.bham.ac.uk/ws/files/57180601/Acharya_etal_Anisotropic_Flow_Of_Identified_Par_JHEP.pdf	English	null	Anisotropic flow of identified particles in Pb–Pb collisions at √S<sub>NN</sub> = 5.02 TeV	EPJ web of conferences	2,018	cc-by	31,717	Link to publication on Research at Birmingham portal Publisher Rights Statement: checked on 18/1/19 General rights U l li Anisotropic flow of identified particles in Pb-Pb collisions at √sNN=5.02 TeV ALICE Collaboration DOI: 10.1007/JHEP09(2018)006 License: Creative Commons: Attribution (CC BY) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): ALICE Collaboration 2018, 'Anisotropic flow of identified particles in Pb-Pb collisions at √sNN=5.02 TeV', Journal of High Energy Physics, vol. 2018, no. 9. https://doi.org/10.1007/JHEP09(2018)006 Link to publication on Research at Birmingham portal Download date: 24. Oct. 2024 Citation for published version (Harvard): ALICE Collaboration 2018, 'Anisotropic flow of identified particles in Pb-Pb collisions at √sNN=5.02 TeV', Journal of High Energy Physics, vol. 2018, no. 9. https://doi.org/10.1007/JHEP09(2018)006 General rights l li General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. •Users may freely distribute the URL that is used to identify this publication. •Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. y p oad and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private mercial research y y y p sers may download and/or print one copy of the publication from the University of Birmingham research portal for the pu udy or non-commercial research. y •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and P •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of th When citing, please reference the published version. Anisotropic ﬂow of identiﬁed particles in Pb-Pb collisions at √sNN = 5.02 TeV JHEP09(2018)006 Published for SISSA by Springer Received: May 23, 2018 Accepted: August 28, 2018 Published: September 3, 2018 Received: May 23, 2018 Accepted: August 28, 2018 Published: September 3, 2018 Received: May 23, 2018 Accepted: August 28, 2018 Published: September 3, 2018 Open Access, Copyright CERN, for the beneﬁt of the ALICE Collaboration. Article funded by SCOAP3. The ALICE collaboration E-mail: ALICE-publications@cern.ch y g for the beneﬁt of the ALICE Collaboration. Take down policy y ty of Birmingham exercises care and attention in making items available there are rare occasions when an item has bee or has been deemed to be commercially or otherwise sensitive. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions wh uploaded in error or has been deemed to be commercially or otherwise sensitive. this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we will remove access ely and investigate. If you believe that this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we the work immediately and investigate. Download date: 24. Oct. 2024 E-mail: ALICE-publications@cern.ch Contents 1 Introduction 1 2 Experimental setup and data analysis 3 2.1 Event and track selection 4 2.2 Identiﬁcation of π±, K± and p+p 4 2.3 Reconstruction of K0 S and Λ + Λ 5 2.4 Reconstruction of φ-mesons 6 2.5 Flow analysis techniques 7 3 Systematic uncertainties 8 4 Results and discussion 13 4.1 Centrality and pT dependence of ﬂow coeﬃcients 13 4.2 Scaling properties 18 4.3 Comparison with model calculations 20 4.4 Shape evolution of vn(pT) as function of centrality 21 4.5 Comparison with vn of identiﬁed particles at √sNN = 2.76 TeV 27 5 Summary 28 The ALICE collaboration 39 Contents 1 Introduction 1 2 Experimental setup and data analysis 3 2.1 Event and track selection 4 2.2 Identiﬁcation of π±, K± and p+p 4 2.3 Reconstruction of K0 S and Λ + Λ 5 2.4 Reconstruction of φ-mesons 6 2.5 Flow analysis techniques 7 3 Systematic uncertainties 8 4 Results and discussion 13 4.1 Centrality and pT dependence of ﬂow coeﬃcients 13 4.2 Scaling properties 18 4.3 Comparison with model calculations 20 4.4 Shape evolution of vn(pT) as function of centrality 21 4.5 Comparison with vn of identiﬁed particles at √sNN = 2.76 TeV 27 5 Summary 28 The ALICE collaboration 39 Contents 1 Introduction 1 2 Experimental setup and data analysis 3 2.1 Event and track selection 4 2.2 Identiﬁcation of π±, K± and p+p 4 2.3 Reconstruction of K0 S and Λ + Λ 5 2.4 Reconstruction of φ-mesons 6 2.5 Flow analysis techniques 7 3 Systematic uncertainties 8 4 Results and discussion 13 4.1 Centrality and pT dependence of ﬂow coeﬃcients 13 4.2 Scaling properties 18 4.3 Comparison with model calculations 20 4.4 Shape evolution of vn(pT) as function of centrality 21 4.5 Comparison with vn of identiﬁed particles at √sNN = 2.76 TeV 27 5 Summary 28 The ALICE collaboration 39 JHEP09(2018)006 The ALICE collaboration E-mail: ALICE-publications@cern.ch Abstract: The elliptic (v2), triangular (v3), and quadrangular (v4) ﬂow coeﬃcients of π±, K±, p + p, Λ + Λ, K0 S, and the φ-meson are measured in Pb-Pb collisions at √sNN = 5.02 TeV. Results obtained with the scalar product method are reported for the rapidity range \|y\| < 0.5 as a function of transverse momentum, pT, at diﬀerent collision centrality intervals between 0–70%, including ultra-central (0–1%) collisions for π±, K±, and p + p. For pT < 3 GeV/c, the ﬂow coeﬃcients exhibit a particle mass dependence. At intermediate transverse momenta (3 < pT < 8–10 GeV/c), particles show an approximate grouping according to their type (i.e., mesons and baryons). The φ-meson v2, which tests both particle mass dependence and type scaling, follows p + p v2 at low pT and π± v2 at intermediate pT. The evolution of the shape of vn(pT) as a function of centrality and harmonic number n is studied for the various particle species. Flow coeﬃcients of π±, K±, and p + p for pT < 3 GeV/c are compared to iEBE-VISHNU and MUSIC hydrody- namical calculations coupled to a hadronic cascade model (UrQMD). The iEBE-VISHNU calculations describe the results fairly well for pT < 2.5 GeV/c, while MUSIC calculations reproduce the measurements for pT < 1 GeV/c. A comparison to vn coeﬃcients measured in Pb-Pb collisions at √sNN = 2.76 TeV is also provided. Keywords: Heavy Ion Experiments Keywords: Heavy Ion Experiments Keywords: Heavy Ion Experiments ArXiv ePrint: 1805.04390 https://doi.org/10.1007/JHEP09(2018)006 Open Access, Copyright CERN, for the beneﬁt of the ALICE Collaboration. Article funded by SCOAP3. 1 Introduction Ultra-relativistic heavy-ion collisions are used to study the properties of the quark-gluon plasma (QGP), a state of deconﬁned quarks and gluons expected at high temperatures or baryon densities [1]. Measurements of anisotropies in particle azimuthal distributions rel- ative to the collision symmetry planes at the Relativistic Heavy Ion Collider (RHIC) [2–5] and the Large Hadron Collider (LHC) [6–8] have shown that the produced hot and dense matter behaves as a strongly-interacting QGP. Comparisons to predictions from hydro- dynamic models indicate that the QGP has a shear viscosity to entropy density ratio (η/s) close to the theoretical lower limit from the anti-de Sitter/conformal ﬁeld theory (AdS/CFT) correspondence of 1/4π for ℏ= kB = 1 [9]. Azimuthal anisotropies in particle production relative to the collision symmetry planes, often referred to as anisotropic ﬂow, arise from the asymmetry in the initial geometry of the collision combined with the initial inhomogeneities of the system’s energy density [10]. Anisotropic ﬂow depends on the equation of state and transport coeﬃcients of the system, – 1 – such as η/s and bulk viscosity to entropy density ratio (ζ/s). Its magnitude is quantiﬁed via the coeﬃcients vn in a Fourier decomposition of the particle azimuthal distribution [11] E d3N dp3 = 1 2π d2N pTdpTdy 1 + 2 ∞ X n=1 vn cos[n(ϕ −Ψn)] ! , (1.1) (1.1) where E is the energy, p the momentum, pT the transverse momentum, ϕ the azimuthal angle, η the pseudorapidity of the particle, and Ψn the n-th harmonic symmetry plane angle. The second order ﬂow coeﬃcient v2, called elliptic ﬂow, is the largest contribution to the asymmetry of non-central collisions because of the almond-like geometry of the overlap region between the colliding nuclei in the plane perpendicular to the beam direction. The third-order ﬂow coeﬃcient v3, named triangular ﬂow, is generated by ﬂuctuations in the initial distribution of nucleons and gluons in the overlap region [12–15]. The fourth- order ﬂow coeﬃcient v4, called quandrangular ﬂow, is generated both by initial geometry, ﬂuctuations, and is in addition sensitive to the non-linear hydrodynamic response of the medium [16, 17]. It has been shown that higher-order ﬂow coeﬃcients are more sensitive to η/s than v2 [18, 19]. JHEP09(2018)006 In addition to probing η/s and ζ/s, anisotropic ﬂow constrains the initial spatial density (e.g. energy and entropy density), freeze-out conditions of the system, and parti- cle production mechanisms in diﬀerent pT regions. 1 Introduction Stronger constraints are achieved by studying anisotropic ﬂow of identiﬁed particles. To guide interpretation of the results in the context of these processes, three kinematic ‘regions of interest’ are deﬁned in the pT- diﬀerential vn measurements, vn(pT). For pT ≲3 GeV/c, anisotropic ﬂow is a remnant of the collective dynamics during the hydrodynamic expansion of the system. The interplay between the isotropic expansion (radial ﬂow) and anisotropic ﬂow leads to a characteristic mass ordering of vn(pT) [20–28], meaning that heavier particles have smaller vn(pT). At intermediate pT (3 ≲pT ≲8 GeV/c), the values of vn for diﬀerent particles tend to separate mesons and baryons [27–33]. The ﬂow of baryons is larger than that of mesons in this pT range, supporting the hypothesis of hadronization through quark coalescence [34], where it is assumed that the invariant spectrum of produced particles is proportional to the product of the spectra of their constituents [35, 36]. However, the scaling only holds approximately at RHIC [32] and at the level of ±20% in Pb-Pb collisions at √sNN = 2.76 TeV [27, 28]. This behaviour is also qualitatively consistent with a scenario in which particle produc- tion includes interactions of jet fragments with bulk matter [37, 38]. For pT ≳8 GeV/c, anisotropic ﬂow is generated when hard partons that propagate through the system lose energy via (multiple) scattering and gluon radiation [39, 40], resulting in vn that remain non-zero up to very high pT [41–44]. Anisotropic ﬂow of identiﬁed particles is an important observable when studying the characteristics of the QGP. However, since particles can scatter and be regenerated in be- tween the chemical and kinetic freeze-out of a collision (the hadronic phase), information about the QGP phase imprinted in vn(pT) can be altered by late-stage interactions and resonance decays, which can aﬀect both vn and ⟨pT⟩[45], leading to a deviation in mass or- dering in vn(pT) at low pT [46]. The φ-meson has been suggested as a particularly sensitive probe of the early collision phase as its production rate via regeneration in the hadronic – 2 – phase is negligible [47] and it is theorized to have a low hadronic cross section [48–50], making it insensitive to the dissipative eﬀects of the hadronic phase of the collision (al- though it should be noted that there is no consensus on the exact value of the cross section between the φ-meson and nucleons in heavy-ion collisions [51–54]). 1 Introduction Recent experimental studies [27, 55, 56] suggest that the φ-meson may be more sensitive to the hadronic phase than anticipated. In this article, we present measurements of pT-diﬀerential elliptic, triangular, and quadrangular ﬂow coeﬃcients of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson in Pb-Pb collisions at √sNN = 5.02 TeV, extending greatly, and improving in precision upon, the previous measurements of identiﬁed particle vn in Pb-Pb collisions at √sNN = 2.76 TeV as carried out by ALICE [27, 28, 33]. The results are reported for a wide range of particle transverse momenta within the rapidity range \|y\| < 0.5 at diﬀerent collision centralities between 0–70% range. To isolate the fraction of anisotropic ﬂow that is generated by initial- state ﬂuctuations rather than geometry, the ﬂow coeﬃcients are also studied in ultra-central collisions (0–1% collision centrality). Centrality estimates the degree of overlap between the two colliding nuclei and is expressed as percentiles of the inelastic hadronic cross section, with low percentage values corresponding to head-on collisions. The measurements are performed using the scalar product method [57–59] with a (pseudo-)rapidity gap of \|∆η\| > 2.0 between the identiﬁed particles under study and the charged reference particles. The ﬂow coeﬃcients are measured separately for particles and anti-particles and are found to be compatible within the statistical uncertainties for most pT and centrality intervals. Any residual diﬀerences are included in the systematic uncertainties, and vn denotes the average between results for particles and anti-particles. JHEP09(2018)006 This paper is organized as follows. Analysis details, particle identiﬁcation, reconstruc- tion methods, and ﬂow measurement techniques are outlined in section 2. The evaluation of systematic uncertainties is discussed in section 3. The ﬂow coeﬃcients of π±, K±, p+p (v2, v3, and v4), Λ+Λ, K0 S (v2 and v3), and the φ-meson (v2) are reported and compared to model calculations in section 4. Finally, the results are summarized in section 5. 2.1 Event and track selection The data sample recorded by ALICE during the 2015 LHC Pb-Pb run at √sNN = 5.02 TeV is used for this analysis. The minimum-bias trigger requires signals in both V0A and V0C detectors. An oﬄine event selection is applied to remove beam-induced background (i.e. beam-gas events) and pileup events. The former is rejected utilizing the V0 and ZDC timing information. The remaining contribution of such interactions is found to be smaller than 0.02% [63]. Pileup events, which constitute about 0.25% of the recorded sample, are removed by comparing multiplicity estimates from the V0 detector to those of tracking detectors at mid-rapidity, exploiting the diﬀerence in readout times between the systems. The fraction of pileup events left after applying the dedicated pileup removal criteria is found to be negligible. The primary vertex position is determined from tracks reconstructed in the ITS and TPC as described in ref. [63]. Only events with a primary vertex position within ±10 cm from the nominal interaction point along the beam direction are used in the analysis. Approximately 67 × 106 Pb-Pb events in the 0–70% centrality interval pass these selection criteria. Centrality is estimated from the energy deposition measured in the V0 detector [69]. JHEP09(2018)006 Charged-particle tracks, used to measure the vn of π±, K±, p+p and the φ-meson, are reconstructed using the ITS and TPC within \|η\| < 0.8 and 0.5 < pT < 16.0 GeV/c with a track-momentum resolution better than 4% for the considered range [63]. Additional quality criteria are used to reduce the contamination from secondary charged particles (i.e., particles originating from weak decays, γ-conversions, and secondary hadronic interactions in the detector material) and fake tracks (random associations of space points). Only tracks with at least 70 space points, out of a maximum of 159, with a χ2 per degree-of-freedom for the track ﬁt lower than 2, are accepted. Moreover, each track is required to cross at least 70 TPC pad rows and to be reconstructed from at least 80% of the number of expected TPC space points, in addition to having at least one hit in the two innermost layers of the ITS. Furthermore, tracks with a distance of closest approach (DCA) to the reconstructed event vertex smaller than 2 cm in the longitudinal direction (z) and (0.0105 + 0.0350 (pT c/GeV)−1.1) cm in the transverse plane (xy) are selected. 2.1 Event and track selection Relevant selection criteria for tracks used for the reconstruction of K0 S and Λ+Λ are given in section 2.3. 2 Experimental setup and data analysis ALICE [60–62] is a dedicated heavy-ion experiment at the LHC optimized to study the properties of strongly interacting matter produced in heavy-ion collisions. A full overview of the detector layout and its performance can be found in [62, 63]. The main subsystems used in this analysis are the Inner Tracking System (ITS) [64], Time Projection Chamber (TPC) [65], Time Of Flight detector (TOF) [66], and V0 [67]. The ITS, TPC, and TOF detectors cover full azimuth within pseudorapidity range \|η\| < 0.9 and lie within a homoge- neous magnetic ﬁeld of up to 0.5 T. The ITS consists of six layers of silicon detectors used for tracking and vertex reconstruction. The TPC is the main tracking detector and is also used to identify particles via speciﬁc ionization energy loss, dE/dx. The TOF in conjunc- tion with the timing information from the T0 detector [68] provide particle identiﬁcation based on ﬂight time. The T0 is made up of two arrays of Cherenkov counters T0C and T0A, located at -3.3 < η < -3.0 and 4.5 < η < 4.9, respectively. Two scintillator arrays – 3 – (V0), which cover the pseudorapidity ranges −3.7 < η < −1.7 (V0C) and 2.8 < η < 5.1 (V0A), are used for triggering, event selection, and the determination of centrality [69] and Qn-vectors (see section 2.5). Both V0 detectors are segmented in four rings in the radial direction with each ring divided into eight sectors in the azimuthal direction. In addition, two tungsten-quartz neutron Zero Degree Calorimeters (ZDCs), installed 112.5 meters from the interaction point on each side, are used for event selection. 2.2 Identiﬁcation of π±, K± and p+p Particle identiﬁcation is performed using the speciﬁc ionization energy loss, dE/dx, mea- sured in the TPC and the time of ﬂight obtained from the TOF. A truncated-mean proce- dure is used to estimate the dE/dx (where the 40% highest-charge clusters are discarded), which yields a dE/dx resolution around 6.5% in the 0–5% centrality class [63]. At least 70 – 4 – clusters are used for the dE/dx estimation. The TOF measures the time that a particle needs to travel from the primary vertex to the detector itself with a time resolution of ≈80 ps [63]. The start time for the TOF measurement is provided by the T0 detector or from a combinatorial algorithm which uses the particle arrival times at the TOF detector itself [63, 66]. Expressing the diﬀerence between the expected dE/dx and the time of ﬂight for π±, K± and p+p, and the measured signals in both TPC and TOF, in units of the standard deviations from the most probable value for both detectors (nσTPC, nσTOF), and applying a selection on the number of accepted nσ, allows for particle identiﬁcation on a track-by- track basis. The TPC dE/dx of diﬀerent particle species are separated by at least 4σ for pT < 0.7 GeV/c, while in the relativistic rise region of the dE/dx (pT > 2 GeV/c) particle identiﬁcation is still possible but only on a statistical basis [63]. The TOF detector provides 3σ separation between π± and K± for pT < 2.5 GeV/c, and between K± and p+p for pT < 4 GeV/c [63]. JHEP09(2018)006 The information from the TPC and TOF is combined using a quadratic sum nσPID = q nσ2 TPC + nσ2 TOF for 0.5 < pT ≤4 GeV/c. Particles are selected by requir- ing nσPID < 3 for each species. The smallest nσPID is used to assign the identity when the selection criterion is fulﬁlled by more than one species. When measuring p+p vn(pT), only p are considered for pT < 2 GeV/c to exclude secondary protons from detector material. At high transverse momenta (pT > 4 GeV/c), K± cannot reliably be identiﬁed. Identiﬁcation of π± and p+p for pT > 4 GeV/c is done utilizing the TPC dE/dx signal only. Pions (protons) are selected from the upper (lower) part of the expected pion (proton) dE/dx distribution. 2.2 Identiﬁcation of π±, K± and p+p For example, proton selection typically varies in the range from 0 to −3σTPC or from −1.5σTPC to −4.5σTPC depending on the momentum. Secondary contamination from weak decays, studied using the procedure outlined in [70], decreases from about 30% to 5% for p+p in the pT range 0.7–4.0 GeV/c and from about 5% to 0.5% for π± in the pT range 0.5–4.0 GeV/c, while it is negligible for K±. The vn coeﬃcients are not corrected for these contaminations; their eﬀect on vn is at maximum ≈8%, for p + p v2 at pT < 1 GeV/c for central collisions, and negligible for K±, π± vn. The contamination from other particle species is below 3% and 20% at pT > 4.0 GeV/c for π± and p+p, respectively, and contamination from fake tracks is negligi- ble. The vn results are reported for 0.5 < pT < 16.0(12.0, 6.0) GeV/c for π± v2 (v3, v4), 0.7 < pT < 16.0(12.0, 6.0) GeV/c for p+p v2 (v3, v4), and 0.5 < pT < 4.0 GeV/c for K± vn, all within \|y\| < 0.5. 2.3 Reconstruction of K0 S and Λ + Λ The K0 S and Λ+Λ are reconstructed in the K0 S →π+ + π−and Λ →p + π−(Λ →p + π+) channels with branching ratios of 69.2% [71] and 63.9% [71] respectively. Reconstruction of K0 S and Λ+Λ is based on identifying secondary vertices from which two oppositely-charged particles originate, called V0s. Topological selection criteria pertaining to the shape of the V0 decay can be imposed, as well as requirements on the species identity of the decay products (called daughter particles). – 5 – The V0 candidates are selected to have an invariant mass between 0.4 and 0.6 GeV/c2 and 1.07 and 1.17 GeV/c2 for K0 S and Λ+Λ, respectively. The invariant mass of the V0 is calculated based on the assumption that the daughter particles are either a π+π−pair, or a pπ−(pπ+) pair. The daughter particles have been identiﬁed over the entire pT range using the TPC following the nσ approach detailed in section 2.2 (\|nσTPC\| < 3). The daughter tracks were reconstructed within \|η\| < 0.8, while the criteria on the number of TPC space points, the χ2 per TPC space point per degree-of-freedom, the number of crossed TPC pad rows, and the percentage of the expected TPC space points used to reconstruct a track are identical to those applied for primary particles. In addition, the minimum DCA of daughter tracks to the primary vertex is 0.1 cm. Furthermore, the maximum DCA of daughter tracks to the secondary vertex is 0.5 cm to ensure that they are products of the same decay. JHEP09(2018)006 To reject secondary vertices arising from decays into more than two particles, the cosine of the pointing angle θp is required to be larger than 0.998. This angle is deﬁned as the angle between the momentum-vector of the V0 assessed at its decay position and the line connecting the V0 decay vertex to the primary vertex and has to be close to 0 as a result of momentum conservation. In addition, the V0 candidates are only accepted when they are produced at a distance between 5 and 100 cm from the nominal primary vertex in the radial direction. The lower value is chosen to avoid any bias from the eﬃciency loss when secondary tracks are being wrongly matched to clusters in the ﬁrst layer of the ITS. 2.3 Reconstruction of K0 S and Λ + Λ To assess the systematic uncertainty related to contaminations from Λ+Λ and electron- positron pairs coming from γ-conversions to the K0 S sample, a selection in the Armenteros- Podolanski variables [72] is applied for the K0 S candidates, rejecting ones with q ≤\|α\|/5. Here q is the momentum projection of the positively charged daughter track in the plane perpendicular to the V0 momentum and α = (p+ L −p− L )/(p+ L + p− L ), with p± L the projection of the positive or negative daughter tracks’ momentum onto the momentum of the V0. To obtain the pT-diﬀerential yield of K0 S and Λ+Λ (which, together with background yields, are used for the vn extraction cf. eq. (2.3)), invariant mass distributions at var- ious pT intervals are parametrized as a sum of a Gaussian distribution and a second- order polynomial function. The latter is introduced to account for residual contaminations (background yield) that are present in the K0 S and Λ+Λ signals after the topological and daughter track selections. The K0 S and Λ+Λ yields are extracted by integration of the Gaus- sian distribution. Obtained yields have not been corrected for feed-down from higher mass baryons (Ξ±, Ω±) as earlier studies have shown that these have a negligible eﬀect on the measured vn [27]. The vn(pT) results are reported within \|y\| < 0.5 and 0.5 < pT < 10 GeV/c for K0 S and 0.8 < pT < 10 GeV/c for Λ+Λ. 2.4 Reconstruction of φ-mesons The φ-meson is reconstructed in the φ →K++K−channel with a branching ratio of 48.9% [71]. Its reconstruction proceeds by ﬁrst identifying all primary K± tracks in an event, following the procedure for primary charged K± outlined in section 2.2. The K± identiﬁcation criterion nσPID < 3 is chosen as it improves the signiﬁcance of the φ-meson yield, while retaining a suﬃcient reconstruction eﬃciency. The vector sums of all pos- – 6 – sible K± pairs are called φ-meson candidates, the yield of which is obtained as function of invariant mass MK+K−in various pT intervals. The pT-diﬀerential φ-meson yield is obtained by ﬁrst subtracting a background yield from the candidate yield. This back- ground yield is estimated using an event-mixing technique [73], in which K± from diﬀerent collisions are paired into background tracks, and is normalized to the candidate yield for 1.04 < MK+K−< 1.09 GeV/c2. Collisions with similar characteristics (vertex position, centrality) are used for this mixing. To obtain the pT-diﬀerential yield of φ-mesons, the invariant mass distributions of the candidate yield is, after the aforementioned subtrac- tion, parametrized as a sum of a Breit-Wigner distribution and a second-order polynomial function, the latter introduced to account for residual contaminations. The φ-meson yields are extracted by integration of the Breit-Wigner distribution and, together with back- ground yields, used for the vn extraction (see eq. (2.3)). The v2(pT) results are reported for 0.9 < pT < 6.5 GeV/c within \|y\| < 0.5. JHEP09(2018)006 2.5 Flow analysis techniques The ﬂow coeﬃcients vn are measured using the scalar product method [57–59], written as The ﬂow coeﬃcients vn are measured using the scalar product method [57–59], written as vn{SP} = ⟨⟨un,kQ∗ n⟩⟩ ,s ⟨QnQA∗ n ⟩⟨QnQB∗ n ⟩ ⟨QAn QB∗ n ⟩ , (2.1) (2.1) where un,k = exp(inϕk) is the unit ﬂow vector of the particle of interest k with azimuthal angle ϕk, Qn is the event ﬂow vector, and n is the harmonic number. Brackets ⟨· · · ⟩denote an average over all events, the double brackets ⟨⟨· · · ⟩⟩an average over all particles in all events, and ∗the complex conjugate. The vector Qn is calculated from the azimuthal distribution of the energy deposition measured in the V0A. Its x and y components are given by Qn,x = X j wj cos(nϕj), Qn,y = X j wj sin(nϕj), (2.2) (2.2) where the sum runs over the 32 channels j of the V0A detector, ϕj is the azimuthal angle of channel j deﬁned by the geometric center, and wj is the amplitude measured in channel j. The vectors QA n and QB n are determined from the azimuthal distribution of the energy deposition measured in the V0C and the azimuthal distribution of the tracks reconstructed in the ITS and TPC, respectively. The amplitude measured in each channel of the V0C (32 channels as for the V0A) is used as weight in the case of QA n , while unity weights are applied for QB n . Tracks used for QB n are selected following the procedure for primary charged tracks outlined in section 2.1 for 0.2 < pT < 5.0 GeV/c. In order to account for a non-uniform detector response, the components of the Qn and QA n vectors are recalibrated using a recentering procedure (i.e. subtraction of the Qn-vector averaged over many events from the Qn-vector of each event) [74]. The large gap in pseudorapidity between un,k and Qn (\|∆η\| > 2.0) greatly suppresses short-range correlations unrelated to the azimuthal asymmetry in the initial geometry, commonly referred to as ‘non-ﬂow’. These correlations largely come from the inter-jet correlations and resonance decays. 2.5 Flow analysis techniques – 7 – 1.02 1.04 ) 2 c Counts / (1.7 MeV/ 0 10000 20000 30000 40000 , \|y\| < 0.5 c < 3.0 GeV/ - K + K T p 2.7 < ) bg N + sig N yield ( - + K + K ) sig N ( - + K + K → φ Breit-Wigner fit 20-30% ALICE = 5.02 TeV NN s Pb − Pb ) 2 c (GeV/ - K + K M 1.02 1.04 2 tot v 0.34 0.36 0.38 fit tot 2 v Figure 1. (Colour online) Illustration of reconstruction and v2 measurement for the φ-meson. The reconstruction of the φ-meson and extraction of N sig and N bg are shown in the upper panel. A ﬁt of eq. (2.3) to data is presented in the lower panel. JHEP09(2018)006 Figure 1. (Colour online) Illustration of reconstruction and v2 measurement for the φ-meson. The reconstruction of the φ-meson and extraction of N sig and N bg are shown in the upper panel. A ﬁt of eq. (2.3) to data is presented in the lower panel. The vn of the K0 S, Λ+Λ, and φ-meson cannot directly be measured using eq. (2.1) as K0 S, Λ+Λ and the φ-meson cannot be identiﬁed on a particle-by-particle basis. Therefore, the vtot n of V0s and φ-meson candidates is measured as function of both invariant mass, Md+d−, and candidate pT. This vtot n can be written [75] as the weighted sum of vn(pT) of the particle of interest, vsig n , and that of background tracks, vbg n (Md+d−), as vtot n (Md+d−) = vsig n Nsig Nsig + Nbg (Md+d−) + vbg n (Md+d−) Nbg Nsig + Nbg (Md+d−), (2.3) (2.3) where signal and background yields Nsig and Nbg are obtained for each pT interval from the K0 S, Λ+Λ and φ-meson reconstruction procedures outlined in sections 2.3 and 2.4. The formalism of eq. (2.1) is used to measure vtot n (Md+d−), vsig n is obtained by parametrizing vbg n (Md+d−) as a second-order polynomial function and ﬁtting eq. (2.3) to the data. Figure 1 illustrates this procedure for the φ-meson, showing the invariant mass spectrum of the φ- meson in the upper panel, and a ﬁt of eq. (2.3) to vtot 2 (Md+d−) data in the lower panel. 3 Systematic uncertainties The systematic uncertainties on vn fall into the following categories: those arising from event selection, those arising from charged particle tracking, uncertainties in particle iden- tiﬁcation, uncertainties in V0 ﬁnding, and those coming from the extraction of vn(pT). For pT ≤4 GeV/c, a pT-dependent systematic uncertainty is assigned to v2, v3, and v4 of π±, K±, p+p, Λ+Λ, K0 S and the φ-meson. Per measured point, the diﬀerence between the nominal measurement and a variation on the nominal measurement is calculated. If this diﬀerence between the nominal data point and the systematic variation is signiﬁcant (where signiﬁcance is evaluated based on the recommendations in [76]), it is considered to – 8 – be a systematic uncertainty. When various checks are performed to quantify the eﬀect of one systematic uncertainty (e.g. using three diﬀerent centrality estimators to estimate the uncertainty in centrality determination), the maximum signiﬁcant deviation that is found between the nominal measurement and the systematic variations is assigned as a systematic uncertainty. For each particle species, a pT-independent average uncertainty is reported for pT > 4 GeV/c in order to suppress sensitivity to statistical ﬂuctuations. The uncertainty is obtained by ﬁtting a zeroth-order polynomial to the signiﬁcant pT-dependent relative uncertainties. The systematic uncertainties are evaluated (if applicable) for each particle species, vn(pT) and centrality intervals. A quadratic sum of the systematic uncertainties from the independent sources is reported as ﬁnal systematic uncertainty on the measurements. An overview of the magnitude of the relative systematic uncertainties per particle species is given in tables 1, 2, and 3 for v2, v3, and v4, respectively. JHEP09(2018)006 Event selection. The nominal event selection criteria and centrality determination are discussed in section 2.1. 3 Systematic uncertainties Event selection criteria are varied by (i) changing the default centrality estimator from energy deposition in the V0 scintillator to either an estimate based on the number of hits in the ﬁrst or second layer of the ITS; (ii) performing the vn analysis of π±, K±, and p+p in 1% wide centrality intervals to test the eﬀect of multiplicity ﬂuctuations (a test not possible for K0 S, Λ+Λ v3); (iii) not rejecting events with tracks caused by pileup or imposing a stricter than default pileup rejection by requiring a tighter correlation between the V0 and central barrel multiplicities; (iv) requiring the reconstructed primary vertex of a collision to lie alternatively within ±12 cm and ±5 cm from the nominal interaction point along the beam axis; (v) analyzing events recorded under diﬀerent magnetic ﬁeld polarities independently; (v) analyzing events recorded at diﬀerent collision rates independently. Charged particle tracking. The nominal charged particle track selection criteria are outlined in section 2.1. Charged particle track selection criteria are varied by (i) requiring the third layer of the ITS to be part of the track reconstruction rather than the ﬁrst two layers only; (ii) using only tracks that have at least three hits per track in the ITS, comple- mented by tracks without hits in the ﬁrst two layers of the ITS (in which case the primary interaction vertex is used as an additional constraint for the momentum determination); (iii) changing the requirement on the minimum number of TPC space points that are used in the reconstruction from 70 to 60, 80, and 90; (iv) an additional systematic uncertainty is evaluated combining the following checks of the track quality: rejecting tracks that are reconstructed close to the sector boundaries of the TPC to which the sensitive pad rows do not extend, varying the minimum number of crossed TPC pad rows from 70 to 120, and requesting at least 90% instead of 80% of the expected TPC space points to reconstruct a track. Variations (i) and (ii) are referred to as tracking mode in tables 1, 2, and 3. Particle identiﬁcation. The nominal particle identiﬁcation approach for π±, K±, and p+p is outlined in section 2.2. 3 Systematic uncertainties Particle identiﬁcation criteria are varied by (i) changing the minimum number of clusters in the TPC that are used to estimate the dE/dx from 70 to 60, 80, and 90; (ii) rejecting tracks that satisfy the particle identiﬁcation criterion for – 9 – JHEP09(2018)006 Error source π± K± p+p K0 S Λ+Λ φ Vertex position 0–1% 0–1% 0–2% 0–2% 0–4% 1–6% 1% wide centrality intervals 0–3% 0–4% 0–4% Centrality estimator 0–3% 0–2% 0–3% 0–4% 0–5% 1–5% Magnetic ﬁeld polarity 0–2% 0–1% 0–2% 0–3% 0–3% 1–3% Interaction rate 0–2% 0–1% 1–2% negl negl negl Pileup rejection 0–1% 0–1% 0–2% 0–1% 0–2% 0–1% Tracking mode 0–4% 0–8% 0–10% 0–5% Number of TPC space points 0–2% 0–2% 0–2% 0–4% 0–2% negl Track quality 0–3% 0–2% 0–3% 0–4% 0–3% negl Particle identiﬁcation purity 0–5% 0–7% 0–5% 0–3% 0–8% 0–6% Number of TPC clusters used for dE/dx 0–6% 0–5% 0–5% 0–5% 0% negl Exclusive particle identiﬁcation 0–2% 0–3% 0–3% Decay vertex (radial position) 0–10% 0–11% Armenteros-Podolanski variables 0–2% DCA decay products to primary vertex 0–3% 0–5% DCA between decay products 0–2% 0–7% Pointing angle cos θp 0–4% 0–9% Minimum pT of daughter tracks 0–4% 0–5% Peak shape negl negl negl Residual background in yield negl negl negl Event mixing 1–3% Positive and negative rapidities 0–3% 0–2% 0–2% 0–4% 0–7% negl Opposite charges 0–2% 0–2% 0–2% Channel removal from V0A 0–5% 0–5% 0–8% 0–3% 0–5% 0–4% vn from V0A or V0C 0–2% 0–2% 0–2% negl negl negl vbg n parametrization negl negl negl vtot n ﬁt ranges 0–1% 0–2% 0–1% Error source π± K± p+p K0 S Λ+Λ φ Vertex position 0–1% 0–1% 0–2% 0–2% 0–4% 1–6% 1% wide centrality intervals 0–3% 0–4% 0–4% Centrality estimator 0–3% 0–2% 0–3% 0–4% 0–5% 1–5% Magnetic ﬁeld polarity 0–2% 0–1% 0–2% 0–3% 0–3% 1–3% Interaction rate 0–2% 0–1% 1–2% negl negl negl Pileup rejection 0–1% 0–1% 0–2% 0–1% 0–2% 0–1% Tracking mode 0–4% 0–8% 0–10% 0–5% Number of TPC space points 0–2% 0–2% 0–2% 0–4% 0–2% negl Track quality 0–3% 0–2% 0–3% 0–4% 0–3% negl Particle identiﬁcation purity 0–5% 0–7% 0–5% 0–3% 0–8% 0–6% Number of TPC clusters used for dE/dx 0–6% 0–5% 0–5% 0–5% 0% negl Exclusive particle identiﬁcation 0–2% 0–3% 0–3% Decay vertex (radial position) 0–10% 0–11% Armenteros-Podolanski variables 0–2% DCA decay products to primary vertex 0–3% 0–5% DCA between decay products 0–2% 0–7% Pointing angle cos θp 0–4% 0–9% Minimum pT of daughter tracks 0–4% 0–5% Peak shape negl negl negl Residual background in yield negl negl negl Event mixing 1–3% Positive and negative rapidities 0–3% 0–2% 0–2% 0–4% 0–7% negl Opposite charges 0–2% 0–2% 0–2% Channel removal from V0A 0–5% 0–5% 0–8% 0–3% 0–5% 0–4% vn from V0A or V0C 0–2% 0–2% 0–2% negl negl negl vbg n parametrization negl negl negl vtot n ﬁt ranges 0–1% 0–2% 0–1% Table 1. 3 Systematic uncertainties Summary of systematic uncertainties for the v2 of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson. The uncertainties depend on pT and centrality range; minimum and maximum values are listed here. Empty ﬁelds indicate that a given check does not apply to the particle of interest. If an uncertainty has been tested but cannot be resolved within statistical precision, the ﬁeld is marked negl for negligible. Horizontal lines are used to separate the diﬀerent categories of systematic uncertainties as explained in section 3. more than one particle species simultaneously for pT < 4 GeV/c; (iii) varying the particle identiﬁcation criterion from nσPID < 3 to nσPID < 1, nσPID < 2, and nσPID < 4; (iv) varying the nσ ranges that are used for particle identiﬁcation for p > 4 GeV/c JHEP09(2018)006 Table 1. Summary of systematic uncertainties for the v2 of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson. The uncertainties depend on pT and centrality range; minimum and maximum values are listed here. Empty ﬁelds indicate that a given check does not apply to the particle of interest. If an uncertainty has been tested but cannot be resolved within statistical precision, the ﬁeld is marked negl for negligible. Horizontal lines are used to separate the diﬀerent categories of systematic uncertainties as explained in section 3. more than one particle species simultaneously for pT < 4 GeV/c; (iii) varying the particle identiﬁcation criterion from nσPID < 3 to nσPID < 1, nσPID < 2, and nσPID < 4; (iv) varying the nσTPC ranges that are used for particle identiﬁcation for pT > 4 GeV/c. The V0 ﬁnding and φ-meson reconstruction. The nominal V0 ﬁnding strategy is described in section 2.3. The V0 ﬁnding criteria fall into two categories: topological re- quirements on the V0s themselves, and selection imposed on their daughter tracks. 3 Systematic uncertainties These criteria are varied by (i) requiring a minimum pT of the V0 daughter tracks of 0.2 GeV/c; – 10 – Error source π± K± p+p K0 S Λ+Λ Vertex position 0–2% 0–1% 0–2% 0–3% 0–9% 1% wide centrality intervals 0–2% 0–2% 0–2% Centrality estimator 0–2% 0–2% 0–2% 0–4% 0–9% Magnetic ﬁeld polarity 0–2% 0–1% 0–3% 0–3% 0–3% Interaction rate 1–2% 1–2% 1–3% negl negl Pileup rejection 0–2% 0–1% 0–3% 0–1% 0–2% Tracking mode 0–3% 1–5% 0–10% Number of TPC space points 0–1% 0–2% 0–5% 0–3% 0–6% Track quality 1–3% 1–2% 1–3% 0–3% 0–6% Particle identiﬁcation purity 0–4% 1–3% 0–10% 0–4% 0–4% Number of TPC clusters used for dE/dx 0–5% 0–5% 0–5% Exclusive particle identiﬁcation 0–1% 0–2% 0–1% Decay vertex (radial position) 0–9% 0–11% Armenteros-Podolanski variables 0–4% DCA decay products to primary vertex 0–3% 0–5% DCA between decay products 0–5% 0–8% Pointing angle cos θp 0–5% 0–1% Minimum pT of daughter tracks 0–4% negl Peak shape negl negl Residual background in yield negl negl Positive and negative rapidities 0–2% 0–1% 0–3% 0–5% 0–4% Opposite charges 0–2% 0–2% 0–2% vn from V0A or V0C 0–2% 0–1% 0–2% 0–4% 0–3% Channel removal from V0A 0–8% 1–8% 1–8% 0–4% 0–5% vbg n parametrization negl negl vtot n ﬁt ranges 0–2% 0–2% T bl 2 S f t ti t i ti f th f ± K± Λ Λ d K0 Th JHEP09(2018)006 Table 2. Summary of systematic uncertainties for the v3 of π±, K±, p+p, Λ+Λ, and K0 S. The uncertainties depend on pT and centrality range; minimum and maximum values are listed here. Empty ﬁelds indicate that a given check does not apply to the particle of interest. If an uncertainty has been tested but cannot be resolved within statistical precision, the ﬁeld is marked negl for negligible. Horizontal lines are used to separate the diﬀerent categories of systematic uncertainties as explained in section 3. 3 Systematic uncertainties and 0.3 cm; (vii) varying the number of clusters in the TPC that are used to estimate the dE/dx of the V0 daughter tracks from 70 to 60 and 90; (viii) varying the particle identiﬁca- tion criterion of the V0 daughter tracks from \|nσTPC\| < 3 to \|nσTPC\| < 1 and \|nσTPC\| < 4; (ix) changing the minimum value of the cos θp from 0.998 to 0.98; (x) varying the minimum radial distance to the primary vertex at which the V0 can be produced from 5 cm to 1 cm and 15 cm; (xi) varying the maximum radial distance to the beam pipe at which the V0 can be produced from 100 cm to 50 cm and 150 cm; (xii) the contamination from Λ+Λ decays and γ-conversions to the K0 S sample is checked by only selecting V0 daughter tracks with a dE/dx value 2σ away from the expected electron dE/dx, eﬀectively excluding electrons, and limiting the value of the Armenteros-Podolanski variables α and q. and 0.3 cm; (vii) varying the number of clusters in the TPC that are used to estimate the dE/dx of the V0 daughter tracks from 70 to 60 and 90; (viii) varying the particle identiﬁca- tion criterion of the V0 daughter tracks from \|nσTPC\| < 3 to \|nσTPC\| < 1 and \|nσTPC\| < 4; (ix) changing the minimum value of the cos θp from 0.998 to 0.98; (x) varying the minimum radial distance to the primary vertex at which the V0 can be produced from 5 cm to 1 cm and 15 cm; (xi) varying the maximum radial distance to the beam pipe at which the V0 can be produced from 100 cm to 50 cm and 150 cm; (xii) the contamination from Λ+Λ decays and γ-conversions to the K0 S sample is checked by only selecting V0 daughter tracks with a dE/dx value 2σ away from the expected electron dE/dx, eﬀectively excluding electrons, and limiting the value of the Armenteros-Podolanski variables α and q. 3 Systematic uncertainties (ii) changing the requirement on the minimum number of TPC space points that are used in the reconstruction of the V0 daughter tracks form 70 to 60 and 80; (iii) varying the minimum number of TPC padrows crossed by the V0 daughter tracks from 70 to 60 and 80; (iv) requesting at least 90% or 70% instead of 80% of the expected TPC space points to reconstruct the V0 daughter tracks; (v) changing the maximum DCA of the V0 daugh- ter tracks to the secondary vertex from 0.5 cm to 0.3 cm and 0.7 cm; (vi) changing the minimum DCA of the V0 daughter tracks to the primary vertex from 0.1 cm to 0.05 cm – 11 – Error source π± K± p+p Vertex position 1–3% 1–3% 1–3% 1% wide centrality intervals 0–1% 0–1% 0–1% Centrality estimator 1–3% 1–3% 2–3% Magnetic ﬁeld polarity 1–2% 1–3% 1–3% Interaction rate 1–2% 2–3% 2–3% Pileup rejection 0–2% 1–2% 2–3% Tracking mode 0–2% 1–5% 1–10% Number of TPC space points 0–1% 0–1% 0–1% Track quality 3–4% 2–3% 3–4% Particle identiﬁcation purity 1–4% 2–4% 2–5% Number of TPC clusters used for dE/dx 0–2% 0–1% 0–1% Exclusive particle identiﬁcation 0–1% 0–2% 0–1% Positive and negative rapidities 1–3% 1–2% 2–3% Opposite charges 2–3% 2–3% 2–3% vn from V0A or V0C 1–3% 2–4% 2–4% Channel removal from V0A 6–14% 6–14% 5–15% JHEP09(2018)006 Table 3. Summary of systematic uncertainties for the v4 of π±, K±, and p+p. The uncertainties depend on pT and centrality range; minimum and maximum values are listed here. Horizontal lines are used to separate the diﬀerent categories of systematic uncertainties as explained in section 3. Table 3. Summary of systematic uncertainties for the v4 of π±, K±, and p+p. The uncertainties depend on pT and centrality range; minimum and maximum values are listed here. Horizontal lines are used to separate the diﬀerent categories of systematic uncertainties as explained in section 3. 3 Systematic uncertainties The yield extraction, as explained in section 2.3 for the K0 S and Λ+Λ, and section 2.4 for the φ-meson, is varied by: (i) using a third-order polynomial as parametrization of residual background in the invariant mass spectra; (ii) using for the φ-meson a Voigtian distribution (a convolution of a Gaussian distribution and Breit-Wigner distribution, where the width of the Breit-Wigner distribution is set to the natural width of the φ-meson, allowing for the Gaussian distribution to describe the smearing of the φ-meson width due to the detector resolution) for the parametrization of the φ-meson invariant mass yield; using for the K0 S and Λ+Λ a sum of two Gaussian distributions with the same mean for the parametrization of the K0 S, Λ+Λ invariant mass yield; (iii, for the φ-meson only) using the yield of like- sign kaon pairs, in which two kaons with equal charge from the same event are used as candidate, for background yield description instead of event mixing. – 12 – 0 2 4 6 8 10 12 14 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± π ) c (GeV/ T p 0 2 4 6 8 10 12 14 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 φ 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% ± K ) c (GeV/ T p 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 0 S K 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 p p+ ) c (GeV/ T p 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 Λ + Λ Figure 2. (Colour online) Centrality dependence of v2(pT) for π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 2. (Colour online) Centrality dependence of v2(pT) for π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Extraction of the vn(pT). 4 Results and discussion The ﬂow coeﬃcients v2, v3, and v4 of identiﬁed particles are presented for various centrality classes in section 4.1; scaling properties are discussed in section 4.2. Comparisons to various model calculations, studies on the shape evolution of vn(pT) with centrality, and comparisons to vn measured at √sNN = 2.76 TeV are shown in sections 4.3, 4.4, and 4.5, respectively. 3 Systematic uncertainties The nominal approach of measuring vn(pT) is outlined in section 2.5, and is varied by: (i) performing ﬂow analysis for π±, K±, and p+p for positive and negative charges independently; (ii) performing ﬂow analysis for positive and negative rapidities independently; (iii) performing ﬂow analysis for π±, K±, and p+p in 1% centrality intervals and merging the result rather than measuring in wider centrality intervals directly; (iv) suppressing the signal from a speciﬁc V0A channel in the evaluation of the Qn-vector (see eq. (2.2)), which, on average, measures a lower energy deposition with respect to the ones reported by the other channels from the same ring; (v) performing ﬂow analysis with the Qn-vector calculated from the V0A or V0C separately; (vi) testing various Md+d−inter- vals over which vbg n (Md+d−) is ﬁtted; (vii) testing the assumption made on vbg n by changing the parametrization from a second-order polynomial to a ﬁrst-order polynomial function. 4.1 Centrality and pT dependence of ﬂow coeﬃcients 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 ± π ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 p p+ 0 2 4 6 8 10 0 0.05 0.1 ± K ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 0 S K 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 Λ + Λ Figure 3. (Colour online) Centrality dependence of v3(pT) for π±, K±, p+p, Λ+Λ, and K0 S. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 ± π ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 p p+ 0 2 4 6 8 10 0 0.05 0.1 ± K ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 0 S K 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 Λ + Λ Figure 3. (Colour online) Centrality dependence of v3(pT) for π±, K±, p+p, Λ+Λ, and K0 S. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 3. (Colour online) Centrality dependence of v3(pT) for π±, K±, p+p, Λ+Λ, and K0 S. Statistical and systematic uncertainties are shown as bars and boxes, respectively. ) c (GeV/ T p 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0.15 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± π ) c (GeV/ T p 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 0.15 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% ± K ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 p p+ Figure 4. (Colour online) Centrality dependence of v4(pT) for π±, K±, and p+p. 4.1 Centrality and pT dependence of ﬂow coeﬃcients Figure 2 shows the v2(pT) of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for various centrality classes in the range 0–70%. For the π±, K± and p+p, measurements performed in ultra- central collisions (0–1%) are also presented. For the φ-meson, the results are reported in the 5–60% centrality range, where v2 can be measured accurately. The magnitude of v2 increases strongly with decreasing centrality up to the 40–50% centrality interval for all particle species. This evolution is expected, since the eccentricity of the overlap zone of the – 13 – 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 ± π ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 p p+ 0 2 4 6 8 10 0 0.05 0.1 ± K ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 0 S K 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 Λ + Λ Figure 3. (Colour online) Centrality dependence of v3(pT) for π±, K±, p+p, Λ+Λ, and K0 S. Statistical and systematic uncertainties are shown as bars and boxes, respectively. ) c (GeV/ T p 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0.15 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± π ) c (GeV/ T p 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 0.15 0-1% 0-5% 5-10% 10-20% 20-30% 30-40% 40-50% ± K ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 p p+ Figure 4. (Colour online) Centrality dependence of v4(pT) for π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.1 Centrality and pT dependence of ﬂow coeﬃcients This behaviour can be explained by the centrality dependence of radial ﬂow combined with the parton density, which will be detailed in section 4.4. JHEP09(2018)006 Figure 5 presents the evolution of vn(pT) of diﬀerent particle species for various central- ity classes. In the most central collisions, initial nucleon-density ﬂuctuations are expected to be the main contributor to the generation of vn. For the 0–1% centrality interval, v3 is the dominant ﬂow coeﬃcient for 1.5 < pT < 6.0 GeV/c, 2.0 < pT < 4 GeV/c, and 2.5 < pT < 6 GeV/c for π±, K±, and p+p, respectively. Furthermore, v4 becomes equal to v2 at pT ≈2.0 GeV/c (2.2, 2.5) for π± (K±, p+p), after which it increases gradually and reaches a magnitude similar to v3 at around 3.5 GeV/c. A similar trend is observed in the 0–5% centrality class for all particle species. However, the crossing between ﬂow coeﬃcients (the pT value at which they reach a similar magnitude), which also depends on the particle mass, takes place at diﬀerent pT values than for the 0–1% centrality interval. This dependence of the crossing between diﬀerent ﬂow coeﬃcients can be attributed to the interplay of elliptic, triangular, and quadrangular ﬂow with radial ﬂow. Upwards of 5% collision centrality, v2 is larger than v3 and v4, conﬁrming the hypothesis that collision geometry dominates the generation of ﬂow coeﬃcients. Figure 6 shows the v2(pT) of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson in a given centrality interval arranged into panels of various centrality classes, which allows for fur- ther illustration of the interplay between elliptic and radial ﬂow. For pT < 2–3 GeV/c, v2 of the diﬀerent particle species is mass-ordered, meaning that lighter particles have a larger v2 than heavier particles at the same pT. This behaviour is indicative of strong radial ﬂow which imposes an equal, isotropic velocity boost to all particles in addition to the anisotropic expansion of the medium [20–22]. For 3 < pT < 8–10 GeV/c, particles are grouped according to their number of constituent quarks, which supports the hypothesis of particle production via quark coalescence [34]. Particle type scaling and mass order- ing are most directly tested by the φ-meson v2, as its mass is close to the proton mass. Figure 6 demonstrates that the φ-meson v2 follows proton v2 at low pT, but pion v2 at intermediate pT in all centrality classes. 4.1 Centrality and pT dependence of ﬂow coeﬃcients Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 4. (Colour online) Centrality dependence of v4(pT) for π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. colliding nuclei increases for peripheral collisions and v2 scales approximately linearly with eccentricity [77]. For more peripheral collisions (i.e. 50–60% and 60–70%), the value of v2 is smaller than in the previous centrality intervals for all particle species except the φ-meson. This suggests that the system has a shorter lifetime in more peripheral collisions, which does not allow for the generation of large v2 [78]. Furthermore, the reduced contribution of eccentricity ﬂuctuations and hadronic interactions might play an important role in these centrality ranges [79]. A non-zero, positive v2 is found in the 0–1% centrality interval for pT < 6 GeV/c for π±, K±, and p+p, which mostly reﬂects the contribution from event-by- event ﬂuctuations in the initial nucleon and gluon density as the system shape is almost spherical at vanishing impact parameter. The third-order ﬂow coeﬃcent v3 is generated by inhomogeneities in the initial nu- cleon and gluon density and not by the collision geometry [12–15], while v4 arises from initial collision geometry, ﬂuctuations, and the non-linear hydrodynamic response of the medium [16, 17]. Higher-order ﬂow harmonics are more sensitive to transport coeﬃcients than v2 [15], as the dampening eﬀect of η/s leads to a stronger decrease of these coeﬃ- cients [18, 19]. Figures 3 and 4 present the v3(pT) of π±, K±, p+p, Λ+Λ, and K0 S and – 14 – v4(pT) of π±, K±, and p+p for various centrality classes in the 0–50% range. Statistical precision limits extending the v4 measurement to more peripheral collisions or carrying it out for Λ+Λ, K0 S, and the φ-meson. Non-zero, positive v3 and v4 are observed for particle species throughout the entire pT ranges up to ≈8 GeV/c. Unlike v2, the coeﬃcients v3 and v4 increase weakly from ultra-central to peripheral collisions. This observation illustrates that higher-order ﬂow coeﬃcients are mainly generated by event-by-event ﬂuctuations in the initial nucleon and gluon density. All ﬂow coeﬃcients increase monotonically with increasing pT up to 3–4 GeV/c where a maximum is reached. The position of this maximum depends on centrality and particle species as it takes place at higher pT for heavier particles for various centrality classes. 4.1 Centrality and pT dependence of ﬂow coeﬃcients (Colour online) The evolution of vn(pT) of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systemat ertainties are shown as bars and boxes, respectively. 4.1 Centrality and pT dependence of ﬂow coeﬃcients 0 2 4 6 8 10 0 0.05 0.1 0-1% ± π 0 2 4 6 8 10 0 0.05 0.1 0-5% ± π 0 2 4 6 8 10 0 0.05 0.1 0.15 5-10% ± π 0 2 4 6 8 10 \|>2} η ∆ {2, \| n v 0 0.1 0.2 10-20% ± π 0 2 4 6 8 10 0 0.1 0.2 20-30% ± π 0 2 4 6 8 10 0 0.1 0.2 0.3 30-40% ± π ) c (GeV/ T p 0 2 4 6 8 10 0 0.1 0.2 0.3 40-50% ± π 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 0.15 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 ± K ) c (GeV/ T p 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 ± K 0 2 4 6 8 10 0 0.05 0.1 p p+ 0 2 4 6 8 10 0 0.05 0.1 p p+ 0 2 4 6 8 10 0 0.05 0.1 0.15 p p+ 0 2 4 6 8 10 0 0.1 0.2 p p+ 0 2 4 6 8 10 0 0.1 0.2 p p+ 0 2 4 6 8 10 0 0.1 0.2 0.3 p p+ ) c (GeV/ T p 0 2 4 6 8 10 0 0.1 0.2 0.3 p p+ 2 v 3 v 4 v 0 2 4 6 8 0 0.05 0.1 0 S K 0 2 4 6 8 0 0.05 0.1 0.15 0 S K 0 2 4 6 8 0 0.1 0.2 0 S K 0 2 4 6 8 0 0.1 0.2 0 S K 0 2 4 6 8 0 0.1 0.2 0.3 0 S K ) c (GeV/ T p 0 2 4 6 8 0 0.1 0.2 0.3 0 S K = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 2 4 6 8 0 0.05 0.1 Λ + Λ 0 2 4 6 8 0 0.05 0.1 0.15 Λ + Λ 0 2 4 6 8 0 0.1 0.2 Λ + Λ 0 2 4 6 8 0 0.1 0.2 Λ + Λ 0 2 4 6 8 0 0.1 0.2 0.3 Λ + Λ ) c (GeV/ T p 0 2 4 6 8 0 0.1 0.2 0.3 Λ + Λ Figure 5. 4.1 Centrality and pT dependence of ﬂow coeﬃcients The crossing between meson and baryon v2, which depends on the particle species, happens at higher pT values for central than peripheral collisions as a result of the larger radial ﬂow in the former. Lastly, it is seen that the v2 of baryons is higher than that of mesons up to pT ≈10 GeV/c, indicating that particle type dependence of v2 persists up to high pT. For pT > 10 GeV/c, where v2 depends only weakly on transverse momentum, the magnitude of p+p v2 is compatible with that for – 15 – gure 5. (Colour online) The evolution of vn(pT) of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. 4.1 Centrality and pT dependence of ﬂow coeﬃcients Statistical and systematic certainties are shown as bars and boxes respectively 0 2 4 6 8 10 0 0.05 0.1 0-1% ± π 0 2 4 6 8 10 0 0.05 0.1 0-5% ± π 0 2 4 6 8 10 0 0.05 0.1 0.15 5-10% ± π 0 2 4 6 8 10 \|>2} η ∆ {2, \| n v 0 0.1 0.2 10-20% ± π 0 2 4 6 8 10 0 0.1 0.2 20-30% ± π 0 2 4 6 8 10 0 0.1 0.2 0.3 30-40% ± π ) c (GeV/ T p 0 2 4 6 8 10 0 0.1 0.2 0.3 40-50% ± π 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.05 0.1 0.15 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 ± K 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 ± K ) c (GeV/ T p 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 ± K 0 2 4 6 8 10 0 0.05 0.1 p p+ 0 2 4 6 8 10 0 0.05 0.1 p p+ 0 2 4 6 8 10 0 0.05 0.1 0.15 p p+ 0 2 4 6 8 10 0 0.1 0.2 p p+ 0 2 4 6 8 10 0 0.1 0.2 p p+ 0 2 4 6 8 10 0 0.1 0.2 0.3 p p+ ) c (GeV/ T p 0 2 4 6 8 10 0 0.1 0.2 0.3 p p+ 2 v 3 v 4 v 0 2 4 6 8 0 0.05 0.1 0 S K 0 2 4 6 8 0 0.05 0.1 0.15 0 S K 0 2 4 6 8 0 0.1 0.2 0 S K 0 2 4 6 8 0 0.1 0.2 0 S K 0 2 4 6 8 0 0.1 0.2 0.3 0 S K ) c (GeV/ T p 0 2 4 6 8 0 0.1 0.2 0.3 0 S K = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 2 4 6 8 0 0.05 0.1 Λ + Λ 0 2 4 6 8 0 0.05 0.1 0.15 Λ + Λ 0 2 4 6 8 0 0.1 0.2 Λ + Λ 0 2 4 6 8 0 0.1 0.2 Λ + Λ 0 2 4 6 8 0 0.1 0.2 0.3 Λ + Λ ) c (GeV/ T p 0 2 4 6 8 0 0.1 0.2 0.3 Λ + Λ gure 5. 4.1 Centrality and pT dependence of ﬂow coeﬃcients (Colour online) The evolution of vn(pT) of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systema uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 – 16 – 0 2 4 6 8 10 12 14 \|>2} η ∆ {2, \| 2 v 0 0.05 0.1 0.15 0-1% ± π ± K p p+ φ 0 S K Λ + Λ 0 2 4 6 8 10 12 14 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 10-20% ) c (GeV/ T p 0 2 4 6 8 10 12 14 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 40-50% 0 2 4 6 8 10 12 14 0 0.05 0.1 0.15 0-5% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 20-30% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 50-60% 0 2 4 6 8 10 12 14 0 0.05 0.1 0.15 5-10% 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 30-40% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0 0.1 0.2 0.3 60-70% Figure 6. (Colour online) The pT-diﬀerential v2 of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 6. (Colour online) The pT-diﬀerential v2 of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. π± within statistical and systematic uncertainties. Furthermore, the nuclear modiﬁcation factor in this high pT region is found to be the same for the two particle species within uncertainties [80]. Figures 7 and 8 present the v3(pT) and v4(pT) for diﬀerent particle species in a given centrality interval. Both v3 and v4 show a clear mass ordering at pT < 2–3 GeV/c, con- ﬁrming the interplay between triangular and quadrangular ﬂow and radial ﬂow. For 3 < pT < 8 GeV/c, particles are grouped into mesons and baryons and, analogous to the trend of v2 in this pT region, the ﬂow of baryons is larger than that of mesons. 4.1 Centrality and pT dependence of ﬂow coeﬃcients (Colour online) The pT-diﬀerential v3 of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes Statistical and systematic uncertainties are shown as bars and boxes respectively 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 0-1% ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 10-20% 0 2 4 6 8 10 0 0.05 0.1 0.15 0-5% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 20-30% 0 2 4 6 8 10 0 0.05 0.1 0.15 5-10% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 30-40% ± π ± K p p+ 0 S K Λ + Λ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 40-50% Figure 7. (Colour online) The pT-diﬀerential v3 of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0-1% ) c (GeV/ T p 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0.15 10-20% 0 1 2 3 4 5 0 0.05 0.1 0-5% ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 20-30% 0 1 2 3 4 5 0 0.05 0.1 5-10% ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 30-40% ± π ± K p p+ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 40-50% Figure 8. (Colour online) The pT-diﬀerential v4 of π±, K±, and p+p for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.1 Centrality and pT dependence of ﬂow coeﬃcients 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 0-1% ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 10-20% 0 2 4 6 8 10 0 0.05 0.1 0.15 0-5% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 20-30% 0 2 4 6 8 10 0 0.05 0.1 0.15 5-10% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 30-40% ± π ± K p p+ 0 S K Λ + Λ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 40-50% Figure 7. (Colour online) The pT-diﬀerential v3 of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 7. (Colour online) The pT-diﬀerential v3 of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0-1% ) c (GeV/ T p 0 1 2 3 4 5 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0.15 10-20% 0 1 2 3 4 5 0 0.05 0.1 0-5% ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 20-30% 0 1 2 3 4 5 0 0.05 0.1 5-10% ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 30-40% ± π ± K p p+ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 0 0.05 0.1 0.15 40-50% Figure 8. (Colour online) The pT-diﬀerential v4 of π±, K±, and p+p for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 8. (Colour online) The pT-diﬀerential v4 of π±, K±, and p+p for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.1 Centrality and pT dependence of ﬂow coeﬃcients The crossing between meson and baryon v3 and v4 also exhibits a centrality and particle mass dependence. Figures 6 and 7 also show a comparison between K± and K0 S v2 and v3 as a function of pT for various centrality classes. A diﬀerence in vn(pT) is found between the K± and K0 S measurements: the magnitude of K0 S vn is systematically smaller than the magnitude of K± vn. This diﬀerence in vn exhibits no pT dependence, but changes with centrality for v2. For 0.8 < pT < 4.0 GeV/c, the diﬀerence in v2 ranges from 7% ± 3.5%(syst) ± 0.7%(stat) in the most central collisions to 1.5% ± 1.5%(syst) ± 0.4%(stat) in peripheral collisions. In the same kinematic range, a deviation in v3 of 6.5% ± 5%(syst) ± 1.7%(stat) is found in the most central collisions and of 6% ± 4.5%(syst) ± 1%(stat) in peripheral collisions. This diﬀerence is similar in magnitude and centrality dependence as the one reported by ALICE in Pb-Pb collisions at √sNN = 2.76 TeV in [27]. – 17 – 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 0-1% ) c (GeV/ T p 0 2 4 6 8 10 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0.15 10-20% 0 2 4 6 8 10 0 0.05 0.1 0.15 0-5% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 20-30% 0 2 4 6 8 10 0 0.05 0.1 0.15 5-10% ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 30-40% ± π ± K p p+ 0 S K Λ + Λ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.05 0.1 0.15 40-50% Figure 7. 4.2 Scaling properties To test the hypothesis of particle production via quark coalescence [34], which would lead to a grouping of vn of mesons and baryons at intermediate pT, both vn and pT are divided by the number of constituent quarks (nq) independently for each particle species. The v2/nq, v3/nq, and v4/nq of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson, plotted as a function of pT/nq, are reported in ﬁgures 9, 10, and 11 for various centrality classes. For pT/nq > 1 GeV/c, the scaling is only approximate. To quantify the degree to which the measurements deviate from the nq scaling, the pT/nq dependence of vn/nq has been divided by a cubic spline ﬁt to the p+p vn/nq. In the region where quark coalescence is hypothesized to be the dominant process (≈1 < pT/nq < 3 GeV/c) [34, 81], a deviation from the exact scaling of ± 20% is found for v2 for central collisions, which decreases to ±15% for the most peripheral collisions. For higher harmonics, a ±20% deviation is found for all centrality classes. This deviation is in agreement with earlier observations [27, 28, 32]. – 18 – 0 1 2 3 4 5 6 7 q n / \|>2} η ∆ {2, \| 2 v 0 0.02 0.04 0.06 0-1% ± π ± K p p+ φ 0 S K Λ + Λ 0 1 2 3 4 5 6 7 q n / \|>2} η ∆ {2, \| 2 v 0 0.05 0.1 10-20% ) c (GeV/ q n / T p 0 1 2 3 4 5 6 7 q n / \|>2} η ∆ {2, \| 2 v 0 0.05 0.1 40-50% 0 1 2 3 4 5 6 7 0 0.02 0.04 0.06 0-5% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 1 2 3 4 5 6 7 0 0.05 0.1 20-30% ) c (GeV/ q n / T p 0 1 2 3 4 5 6 7 0 0.05 0.1 50-60% 0 1 2 3 4 5 6 7 0 0.02 0.04 0.06 5-10% 0 1 2 3 4 5 6 7 0 0.05 0.1 30-40% ) c (GeV/ q n / T p 0 1 2 3 4 5 6 7 0 0.05 0.1 60-70% Figure 9. 4.2 Scaling properties (Colour online) The pT/nq dependence of v2/nq of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 9. (Colour online) The pT/nq dependence of v2/nq of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 q n / \|>2} η ∆ {2, \| 3 v 0 0.02 0.04 0.06 0-1% ) c (GeV/ q n / T p 0 1 2 3 4 5 q n / \|>2} η ∆ {2, \| 3 v 0 0.02 0.04 0.06 10-20% 0 1 2 3 4 5 0 0.02 0.04 0.06 0-5% ) c (GeV/ q n / T p 0 1 2 3 4 5 0 0.02 0.04 0.06 20-30% 0 1 2 3 4 5 0 0.02 0.04 0.06 5-10% ) c (GeV/ q n / T p 0 1 2 3 4 5 0 0.02 0.04 0.06 30-40% ± π ± K p p+ 0 S K Λ + Λ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ q n / T p 0 1 2 3 4 5 0 0.02 0.04 0.06 40-50% Figure 10. (Colour online) The pT/nq dependence of v3/nq of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 10. (Colour online) The pT/nq dependence of v3/nq of π±, K±, p+p, Λ+Λ, and K0 S for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.2 Scaling properties – 19 – 0 0.5 1 1.5 2 2.5 3 q n / \|>2} η ∆ {2, \| 4 v 0 0.02 0.04 0.06 0-1% ) c (GeV/ q n / T p 0 0.5 1 1.5 2 2.5 3 q n / \|>2} η ∆ {2, \| 4 v 0 0.02 0.04 0.06 10-20% 0 0.5 1 1.5 2 2.5 3 0 0.02 0.04 0.06 0-5% ) c (GeV/ q n / T p 0 0.5 1 1.5 2 2.5 3 0 0.02 0.04 0.06 20-30% 0 0.5 1 1.5 2 2.5 3 0 0.02 0.04 0.06 5-10% ) c (GeV/ q n / T p 0 0.5 1 1.5 2 2.5 3 0 0.02 0.04 0.06 30-40% ± π ± K p p+ = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ q n / T p 0 0.5 1 1.5 2 2.5 3 0 0.02 0.04 0.06 40-50% Figure 11. (Colour online) The pT/nq dependence of v4/nq of π±, K±, and p+p for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 11. (Colour online) The pT/nq dependence of v4/nq of π±, K±, and p+p for various centrality classes. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.3 Comparison with model calculations To test the validity of the hydrodynamic description of the QGP evolution, the vn measure- ments in the 0–5%, 10–20% and 40–50% centrality intervals are compared to hydrodynam- ical calculations in ﬁgures 12, 13, and 14 for π±, K±, and p+p, respectively. Predictions from MUSIC [82] and iEBE-VISHNU [83] simulations are depicted by the diﬀerent coloured curves. The ﬁrst calculation is based on MUSIC [84], an event-by-event 3+1 dimensional viscous hydrodynamic model, coupled to a hadronic cascade model (UrQMD) [85, 86], which allows the inﬂuence of the hadronic phase on the anisotropic ﬂow to be studied for diﬀerent particle species. The IP-Glasma model [87, 88] is used to simulate the initial conditions of the collision. MUSIC uses a starting time for the hydrodynamic evolution of τ0 = 0.4 fm/c, a switching temperature between the macroscopic hydrodynamic description and the microscopic transport evolution of Tsw = 145 MeV, a value of η/s = 0.095, and a temperature dependent ζ/s. The second calculation employs the iEBE-VISHNU hybrid model [89], which is an event-by-event version of the VISHNU hybrid model [90], and cou- ples 2+1 dimensional viscous hydrodynamics VISH2+1 [78] to UrQMD. The TRENTo [91] and AMPT [92] models are used to describe the initial conditions. For both conﬁgura- tions, τ0 = 0.6 fm/c and Tsw = 148 MeV are set from [93], where these values have been obtained utilizing Bayesian statistics from a simultaneous ﬁt of ﬁnal charged-particle den- sity, mean transverse momentum, and integrated ﬂow coeﬃcients vn in Pb-Pb collisions at √sNN = 2.76 TeV. The temperature-dependent η/s and ζ/s extracted in [93] are used for TRENTo initial conditions, while η/s = 0.08 and ζ/s = 0 are taken for AMPT. Figures 12, 13, and 14 show that the hydrodynamical calculations qualitatively re- produce the vn measurements. The diﬀerences between the data points and models are visualized in ﬁgures 12, 13, and 14 as the ratios of the measured vn to a ﬁt to the theoretical calculations. The iEBE-VISHNU calculations using AMPT initial conditions describe the pT-diﬀerential vn of π±, K±, and p+p more accurately than TRENTo based and MUSIC calculations for pT > 1 GeV/c. 4.3 Comparison with model calculations 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 2 v 0.05 0.1 0.15 0.2 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± π 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 3 v 0.05 0.1 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 4 v 0.02 0.04 0.06 0.08 0-5% 0.5 1 1.5 2 2.5 0.02 0.04 0.06 0.08 10-20% 0.5 1 1.5 2 2.5 0.02 0.04 0.06 0.08 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 model data 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 JHEP09(2018)006 Figure 12. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of π± for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. between π± and K± vn and iEBE-VISHNU calculations for pT < 2 GeV/c, while p+p vn is described fairly well up to pT = 3 GeV/c. The TRENTo based predictions follow π± and K± vn up to slightly lower transverse momenta (pT <1–2 GeV/c) and to pT < 3 GeV/c for p+p, depending on the considered centrality interval. 4.3 Comparison with model calculations The MUSIC calculations are in agreement with the measured vn for pT < 1 GeV/c in central collisions, however they overestimate v2 at lower pT in more peripheral collisions. 4.3 Comparison with model calculations Using AMPT initial conditions, there is good agreement – 20 – 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 2 v 0.05 0.1 0.15 0.2 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± π 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 3 v 0.05 0.1 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 4 v 0.02 0.04 0.06 0.08 0-5% 0.5 1 1.5 2 2.5 0.02 0.04 0.06 0.08 10-20% 0.5 1 1.5 2 2.5 0.02 0.04 0.06 0.08 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 model data 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 Figure 12. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of π± for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. 4.4 Shape evolution of vn(pT) as function of centrality The evolution of the shape of vn(pT) as function of centrality is quantiﬁed by taking the ratio of vn(pT) in a given centrality interval to the vn(pT) measured in the 20–30% centrality interval vn(pT)ratio to 20−30% = vn(pT) vn(pT)\|20−30% vn\|20−30% vn , (4.1) (4.1) where the second fraction on the right-hand side of the equation serves as a normalization factor which is constructed from the pT-integrated vn values obtained in the 20–30% cen- trality interval (vn\|20−30%) and the centrality interval of interest (vn). Centrality-dependent – 21 – – 21 – 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 2 v 0.05 0.1 0.15 0.2 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± K 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 3 v 0.05 0.1 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 4 v 0 0.02 0.04 0.06 0.08 0-5% 0.5 1 1.5 2 2.5 0 0.02 0.04 0.06 0.08 10-20% 0.5 1 1.5 2 2.5 0 0.02 0.04 0.06 0.08 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 model data 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 Figure 13. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of K± for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. 4.4 Shape evolution of vn(pT) as function of centrality The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 2 v 0.05 0.1 0.15 0.2 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± K 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 0.15 0.2 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 3 v 0.05 0.1 0-5% 0.5 1 1.5 2 2.5 0.05 0.1 10-20% 0.5 1 1.5 2 2.5 0.05 0.1 40-50% 0.5 1 1.5 2 2.5 model data 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 0.5 1 0.5 1 1.5 2 2.5 \|>2} η ∆ {2, \| 4 v 0 0.02 0.04 0.06 0.08 0-5% 0.5 1 1.5 2 2.5 0 0.02 0.04 0.06 0.08 10-20% 0.5 1 1.5 2 2.5 0 0.02 0.04 0.06 0.08 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 model data 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 ) c (GeV/ T p 0.5 1 1.5 2 2.5 0.5 1 1.5 JHEP09(2018)006 Figure 13. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of K± for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. variations in the shape of vn(pT) will present themselves as deviations from unity of the observed vn(pT)ratio to 20−30%. The shape evolution of elliptic and triangular ﬂow is shown in ﬁgures 15 and 16 for π±, K±, p+p, and inclusive charged particles (the latter taken from [44]). 4.4 Shape evolution of vn(pT) as function of centrality For inclusive charged particles, variations in shape of about 10% are observed for pT < 3 GeV/c, which increase to about 30% for pT < 6 GeV/c. The shape evolution of v2(pT) shows diﬀerent trends for π±, K±, and p+p. While π± v2(pT)ratio to 20−30% follows inclusive charged particle over the considered pT range, the elliptic ﬂow of p+p (K±) varies between 20% (10%) to 250% (55%) at low pT from most central to peripheral collisions. The variations are more pronounced for v3(pT), in particular for central collisions. The mass dependence found in the shape evolution of both v2 and v3 for pT < 4 GeV/c can be attributed to variations of the magnitude of radial ﬂow and quark density, both being larger for central than peripheral collisions. Radial ﬂow has a stronger eﬀect on the vn of heavier particles than that of lighter particles at low pT, while the quark density inﬂuences the peak value of vn(pT) in the coalescence model picture [35, 36, 94]. For pT > 4 GeV/c, the shape evolution shows little (if any) particle type dependence. 4.4 Shape evolution of vn(pT) as function of centrality 0.811.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.811.21.41.61.822.22.42.62.8 0 0.1 0.2 0.3 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| p p+ 10-20% 0.811.21.41.61.822.22.42.62.8 0 0.1 0.2 0.3 40-50% 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.811.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0-5% 0.811.21.41.61.822.22.42.62.8 0 0.05 0.1 10-20% 0.811.21.41.61.822.22.42.62.8 0 0.05 0.1 40-50% 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 11.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0-5% 0.8 11.21.41.61.822.22.42.62.8 0 0.05 0.1 10-20% 0.8 11.21.41.61.822.22.42.62.8 0 0.05 0.1 40-50% ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 0.5 1 1.5 ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 0.5 1 1.5 ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 0.5 1 1.5 JHEP09(2018)006 Figure 14. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of p+p for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. The shape evolution of v2(pT) for π±, K±, and p+p is compared to calculations from the MUSIC and iEBE-VISHNU hybrid models in ﬁgure 17. Both models describe the shape evolution for p+p over the pT range 0.7 < pT < 3 GeV/c. 4.4 Shape evolution of vn(pT) as function of centrality – 22 – 0.811.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 2 v 0 0.1 0.2 0.3 ALICE MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0-5% 0.811.21.41.61.822.22.42.62.8 0 0.1 0.2 0.3 = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| p p+ 10-20% 0.811.21.41.61.822.22.42.62.8 0 0.1 0.2 0.3 40-50% 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.811.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 3 v 0 0.05 0.1 0-5% 0.811.21.41.61.822.22.42.62.8 0 0.05 0.1 10-20% 0.811.21.41.61.822.22.42.62.8 0 0.05 0.1 40-50% 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 1 1.5 0.8 11.21.41.61.822.22.42.62.8 \|>2} η ∆ {2, \| 4 v 0 0.05 0.1 0-5% 0.8 11.21.41.61.822.22.42.62.8 0 0.05 0.1 10-20% 0.8 11.21.41.61.822.22.42.62.8 0 0.05 0.1 40-50% ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 model data 0.5 1 1.5 ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 0.5 1 1.5 ) c (GeV/ T p 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 0.5 1 1.5 Figure 14. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of p+p for the 0–5%, 10–20%, and 40–50% centrality classes compared to hydrodynamical calculations from MUSIC model using IP-Glasma initial conditions (magenta) [82] and the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. The uncertainties of the hydrodynamical calculations are depicted by the thickness of the curves. The ratios of the measured vn to a ﬁt to the hydrodynamical calculations are also presented for clarity. 4.4 Shape evolution of vn(pT) as function of centrality 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.5 1 1.5 0-5% 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 1 1.2 1.4 10-20% ) c (GeV/ T p 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.5 1 1.5 2 2.5 40-50% 0 2 4 6 8 10 12 14 0.5 1 1.5 5-10% 0 2 4 6 8 10 12 14 0.8 1 1.2 1.4 20-30% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0.5 1 1.5 2 2.5 50-60% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± h ± π ± K p p+ 0 2 4 6 8 10 12 14 0.8 1 1.2 1.4 30-40% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0.5 1 1.5 2 2.5 60-70% JHEP09(2018)006 Figure 15. (Colour online) Centrality dependence of v2(pT)ratio to 20−30% for π±, K±, p+p, and inclusive charged particles [44]. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 6 7 8 ratio to 20-30% \|>2} η ∆ {2, \| 3 v 0.5 1 1.5 0-5% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 ratio to 20-30% \|>2} η ∆ {2, \| 3 v 0.5 1 1.5 20-30% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 1 2 3 4 5 6 7 8 0.5 1 1.5 5-10% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 0.5 1 1.5 30-40% 0 1 2 3 4 5 6 7 8 0.5 1 1.5 10-20% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 0.5 1 1.5 40-50% Figure 16. (Colour online) Centrality dependence of v3(pT)ratio to 20−30% for π±, K±, p+p, and inclusive charged particles [44]. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.4 Shape evolution of vn(pT) as function of centrality The iEBE-VISHNU model reproduces the shape evolution for π± and K± for pT < 1.5 GeV/c. Calculations from the MUSIC model deviate strongly from the observed shape evolution for π± and K± in peripheral collisions. As quark density depends on centrality, the maximum vn is expected to be found at higher pT in more central collisions. To further quantify this aspect of the shape evo- lution of vn(pT), the pT of π±, p+p, Λ+Λ, and K0 S where v2(pT) and v3(pT) reach a maximum, divided by number of constituent quarks nq, is reported in ﬁgure 18 as a func- tion of centrality. The φ-meson and K± are not included since the kinematic range and granularity of the measurements do not allow for a reliable extraction of a maximum. The left panel of ﬁgure 18 shows that the pT/nq at which v2(pT) reaches a maximum, denoted as pT\|vmax 2 /nq, decreases with increasing centrality percentile for collision central- ities larger than 5–10%, following the expectations from the hypothesis of hadronization – 23 – 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.5 1 1.5 0-5% 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 1 1.2 1.4 10-20% ) c (GeV/ T p 0 2 4 6 8 10 12 14 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.5 1 1.5 2 2.5 40-50% 0 2 4 6 8 10 12 14 0.5 1 1.5 5-10% 0 2 4 6 8 10 12 14 0.8 1 1.2 1.4 20-30% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0.5 1 1.5 2 2.5 50-60% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ± h ± π ± K p p+ 0 2 4 6 8 10 12 14 0.8 1 1.2 1.4 30-40% ) c (GeV/ T p 0 2 4 6 8 10 12 14 0.5 1 1.5 2 2.5 60-70% Figure 15. (Colour online) Centrality dependence of v2(pT)ratio to 20−30% for π±, K±, p+p, and inclusive charged particles [44]. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.4 Shape evolution of vn(pT) as function of centrality Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 0.9 1 1.1 1.2 0-5% 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 1 1.2 0-5% ± π ± K p p+ ) c (GeV/ T p 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 1 1.5 0-5% MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0.5 1 1.5 2 2.5 0.8 1 1.2 10-20% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 10-20% 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 40-50% 0.5 1 1.5 2 2.5 0.8 1 1.2 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 40-50% 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 0.9 1 1.1 1.2 0-5% 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 1 1.2 0-5% ± π ± K p p+ ) c (GeV/ T p 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 1 1.5 0-5% MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0.5 1 1.5 2 2.5 0.8 1 1.2 10-20% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 10-20% 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 40-50% 0.5 1 1.5 2 2.5 0.8 1 1.2 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 40-50% Figure 17. (Colour online) Centrality dependence of v2(pT)ratio to 20−30% for π± (upper pan- els), K± (middle panels), and p+p (lower panels) compared to hydrodynamical calculations from the MUSIC model using IP-Glasma initial conditions (magenta) [82], the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. 4.4 Shape evolution of vn(pT) as function of centrality Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 17. (Colour online) Centrality dependence of v2(pT)ratio to 20−30% for π± (upper pan- els), K± (middle panels), and p+p (lower panels) compared to hydrodynamical calculations from the MUSIC model using IP-Glasma initial conditions (magenta) [82], the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. centrality (%) 0 10 20 30 40 50 60 70 ) c (GeV/ q n / max 2 v of T p 0.5 1 1.5 2 ± π p p+ 0 S K Λ + Λ \| < 0.5 y = 5.02 TeV \| NN s Pb − ALICE Pb centrality (%) 0 5 10 15 20 25 30 35 40 45 50 ) c (GeV/ q n / max 3 v of T p 0.5 1 1.5 2 ± π p p+ 0 S K Λ + Λ \| < 0.5 y = 5.02 TeV \| NN s Pb − ALICE Pb Figure 18. (Colour online) Centrality dependence of pT\|vmax 2 /nq (left) and pT\|vmax 3 /nq (right) divided by number of constituent quarks, nq, for π±, p+p, Λ+Λ, and K0 S. Points are slightly shifted along the horizontal axis for better visibility in both panels. Statistical and systematic uncertainties are shown as bars and boxes, respectively. centrality (%) 0 10 20 30 40 50 60 70 ) c (GeV/ q n / max 2 v of T p 0.5 1 1.5 2 ± π p p+ 0 S K Λ + Λ \| < 0.5 y = 5.02 TeV \| NN s Pb − ALICE Pb ) c (GeV/ q n / max 3 v of T p Figure 18. (Colour online) Centrality dependence of pT\|vmax 2 /nq (left) and pT\|vmax 3 /nq (right) divided by number of constituent quarks, nq, for π±, p+p, Λ+Λ, and K0 S. Points are slightly shifted along the horizontal axis for better visibility in both panels. Statistical and systematic uncertainties are shown as bars and boxes, respectively. p+p pT\|vmax 2 /nq, which is slightly lower in the 0–20% centrality range. 4.4 Shape evolution of vn(pT) as function of centrality 0 1 2 3 4 5 6 7 8 ratio to 20-30% \|>2} η ∆ {2, \| 3 v 0.5 1 1.5 0-5% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 ratio to 20-30% \|>2} η ∆ {2, \| 3 v 0.5 1 1.5 20-30% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0 1 2 3 4 5 6 7 8 0.5 1 1.5 5-10% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 0.5 1 1.5 30-40% 0 1 2 3 4 5 6 7 8 0.5 1 1.5 10-20% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 0.5 1 1.5 40-50% Figure 16. (Colour online) Centrality dependence of v3(pT)ratio to 20−30% for π±, K±, p+p, and inclusive charged particles [44]. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 16. (Colour online) Centrality dependence of v3(pT)ratio to 20−30% for π±, K±, p+p, and inclusive charged particles [44]. Statistical and systematic uncertainties are shown as bars and boxes, respectively. through coalescence. The systematic uncertainties as presented in ﬁgure 18 have been evaluated directly on pT\|vmax n /nq to accurately take into account that some systematic un- certainties can be point-by-point correlated in pT. In the 0–5% centrality interval, there is a hint of a lower pT\|vmax 2 /nq than in the 5–10% centrality class for all particle species. 4.4 Shape evolution of vn(pT) as function of centrality The observed pT\|vmax 2 /nq is compatible among all particle species with the exception of the – 24 – 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 0.9 1 1.1 1.2 0-5% 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 0.8 1 1.2 0-5% ± π ± K p p+ ) c (GeV/ T p 0.5 1 1.5 2 2.5 ratio to 20-30% \|>2} η ∆ {2, \| 2 v 1 1.5 0-5% MUSIC + IP-Glasma IC + UrQMD iEBE-VISHNU + AMPT IC + UrQMD iEBE-VISHNU + TRENTo IC + UrQMD 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| 0.5 1 1.5 2 2.5 0.8 1 1.2 10-20% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 10-20% 0.5 1 1.5 2 2.5 0.8 0.9 1 1.1 1.2 40-50% 0.5 1 1.5 2 2.5 0.8 1 1.2 40-50% ) c (GeV/ T p 0.5 1 1.5 2 2.5 1 1.5 40-50% Figure 17. (Colour online) Centrality dependence of v2(pT)ratio to 20−30% for π± (upper pan- els), K± (middle panels), and p+p (lower panels) compared to hydrodynamical calculations from the MUSIC model using IP-Glasma initial conditions (magenta) [82], the iEBE-VISHNU hybrid model using AMPT (orange) or TRENTo (cyan) initial conditions [83]. Statistical and systematic uncertainties of the data points are shown as bars and boxes, respectively. centrality (%) 0 10 20 30 40 50 60 70 ) c (GeV/ q n / max 2 v of T p 0.5 1 1.5 2 ± π p p+ 0 S K Λ + Λ \| < 0.5 y = 5.02 TeV \| NN s Pb − ALICE Pb centrality (%) 0 5 10 15 20 25 30 35 40 45 50 ) c (GeV/ q n / max 3 v of T p 0.5 1 1.5 2 ± π p p+ 0 S K Λ + Λ \| < 0.5 y = 5.02 TeV \| NN s Pb − ALICE Pb Figure 18. (Colour online) Centrality dependence of pT\|vmax 2 /nq (left) and pT\|vmax 3 /nq (right) divided by number of constituent quarks, nq, for π±, p+p, Λ+Λ, and K0 S. Points are slightly shifted along the horizontal axis for better visibility in both panels. 4.4 Shape evolution of vn(pT) as function of centrality The right panel of ﬁgure 18 presents pT\|vmax 3 /nq, which shows, within the large uncertainties, a weak (if any) centrality dependence for π± and K0 S and no centrality dependence for p+p and Λ+Λ. The pT\|vmax 3 /nq is the same for the diﬀerent particle species within uncertainties. – 25 – 0 2 4 6 8 10 /GeV) c ( T p / 1/2 \| \|>2} η ∆ {2, \| 2 v \| 0 0.2 0.4 0.6 0.8 0-5% ± h ± π ± K p p+ φ 0 S K Λ + Λ ) c (GeV/ T p 0 2 4 6 8 10 /GeV) c ( T p / 1/3 \| \|>2} η ∆ {2, \| 3 v \| 0 0.2 0.4 0.6 0.8 0-5% 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 10-20% 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 40-50% ) c (GeV/ T p 0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 40-50% Figure 19. (Colour online) Centrality dependence of \|vn\|1/n/pT of inclusive charged particles [44], π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for n = 2 (upper panels) and n = 3 (lower panels). Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 19. (Colour online) Centrality dependence of \|vn\|1/n/pT of inclusive charged particles [44], π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson for n = 2 (upper panels) and n = 3 (lower panels). Statistical and systematic uncertainties are shown as bars and boxes, respectively. In the scenario of ideal hydrodynamics, vn is a power law function of the radial ex- pansion velocity of the medium [95, 96] so that vn ∝pTn up to pT ∼M for particles with mass M. Figure 19 shows \|vn\|1/n/pT as function of pT for n = 2 and n = 3 in various centrality intervals for inclusive charged particles [44], π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson (n = 2 only). When vn ∝pTn, the observable \|vn\|1/n/pT should be a constant. 4.4 Shape evolution of vn(pT) as function of centrality For π± and the inclusive charged particles, the vn ∝pTn scaling is broken both for v2 and v3 for all centrality intervals, as is also hypothesized in [97]. It should be noted however that the kinematic constraints imposed on the measurement preclude testing the scaling hypothesis in the full relevant momentum region. The scaling holds up to pT ≈1 GeV/c for K± and K0 S, and up to pT ≈2 GeV/c for p+p, Λ+Λ, and the φ-meson for the 0–5% and 10–20% centrality intervals. Similar qualitative observations are found in the three hydrodynamical calculations [82, 83]. If vn indeed exhibits a power law dependence on pTn, ratios of the form of v1/n n /v1/m m are pT-independent. Previous measurements at RHIC [98, 99] and the LHC [100, 101] have shown that the ratios v1/n n /v1/m m show little to no pT dependence up to about 6 GeV/c independent of the harmonic n and m for peripheral and semi-central collisions. However, a pT dependence is observed for central collisions, which might be due to ﬂuctuations in the initial geometry [99]. The ratios v3/\|v2\|3/2, v4/\|v2\|4/2, and v4/\|v3\|4/3, which probe the same scaling but are in practice more sensitive, are shown in ﬁgures 20, 21, and 22, respectively. For each ﬁgure, vn/\|vm\|n/m is shown for inclusive charged particles [44], π±, K± and p+p in various centrality intervals. For v3/\|v2\|3/2 and v4/\|v2\|4/2, no obvious pT dependence is found for inclusive charged particles between 5–50% collision centrality. For the 0–5% centrality class, the ratios are ﬂat for pT < 3 GeV/c and rise monotonically for higher momenta. No particle type dependence of the ratios is found for pT > 1.5 GeV/c, below which the ratios for p+p vn rise. This rise of the p+p vn ratios can be attributed to an increase of radial ﬂow which aﬀects the independent harmonics diﬀerently. 4.4 Shape evolution of vn(pT) as function of centrality For the ratio – 26 – 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 2 4 6 0-5% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 1 2 20-30% 0 1 2 3 4 5 6 7 8 2 4 6 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 30-40% 0 1 2 3 4 5 6 7 8 2 4 6 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 40-50% Figure 20. (Colour online) Centrality dependence of v3/\|v2\|3/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 10 20 0-5% ) c (GeV/ T p 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 1 2 3 20-30% 0 1 2 3 4 5 10 20 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 30-40% 0 1 2 3 4 5 10 20 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 40-50% Figure 21. (Colour online) Centrality dependence of v4/\|v2\|4/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.4 Shape evolution of vn(pT) as function of centrality 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 10 20 0-5% ) c (GeV/ T p 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 1 2 3 20-30% 0 1 2 3 4 5 10 20 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 30-40% 0 1 2 3 4 5 10 20 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 40-50% Figure 21. (Colour online) Centrality dependence of v4/\|v2\|4/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 10 20 0-5% ) c (GeV/ T p 0 1 2 3 4 5 4/2 \| 2 v /\| 4 v 1 2 3 20-30% 0 1 2 3 4 5 10 20 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 30-40% 0 1 2 3 4 5 10 20 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 40-50% Figure 21. (Colour online) Centrality dependence of v4/\|v2\|4/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 21. (Colour online) Centrality dependence of v4/\|v2\|4/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. v4/\|v3\|4/3, no pT dependence is observed over the full centrality range. Large statistical uncertainties do not allow conclusions to be drawn on the behaviour of p+p vn in the v4/\|v3\|4/3 ratio. 4.4 Shape evolution of vn(pT) as function of centrality 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 2 4 6 0-5% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 1 2 20-30% 0 1 2 3 4 5 6 7 8 2 4 6 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 30-40% 0 1 2 3 4 5 6 7 8 2 4 6 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 40-50% Figure 20. (Colour online) Centrality dependence of v3/\|v2\|3/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 2 4 6 0-5% ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 3/2 \| 2 v /\| 3 v 1 2 20-30% 0 1 2 3 4 5 6 7 8 2 4 6 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 30-40% 0 1 2 3 4 5 6 7 8 2 4 6 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 6 7 8 1 2 40-50% JHEP09(2018)006 Figure 20. (Colour online) Centrality dependence of v3/\|v2\|3/2 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. Figure 20. (Colour online) Centrality dependence of v3/\|v2\| for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. 4.5 Comparison with vn of identiﬁed particles at √sNN = 2.76 TeV The transport properties and initial condition models can be further constrained by studying the energy dependence of anisotropic ﬂow. Figure 23 presents the v2(pT), v3(pT), and v4(pT) of π±, K±, and p+p compared to ALICE measurements performed at √sNN = 2.76 TeV [28]. The vn coeﬃcients at √sNN = 2.76 TeV have been measured using the scalar product method, taking the particle of interest under study and the charged reference particles from diﬀerent, non-overlapping pseudorapidity regions between \|η\| < 0.8. Assuming no – 27 – 0 1 2 3 4 5 4/3 \| 3 v /\| 4 v 1 2 3 0-5% ) c (GeV/ T p 0 1 2 3 4 5 4/3 \| 3 v /\| 4 v 1 2 3 20-30% 0 1 2 3 4 5 1 2 3 5-10% ± h ± π ± K p p+ ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 30-40% 0 1 2 3 4 5 1 2 3 10-20% = 5.02 TeV NN s Pb − ALICE Pb \| < 0.5 y\| ) c (GeV/ T p 0 1 2 3 4 5 1 2 3 40-50% Figure 22. (Colour online) Centrality dependence of v4/\|v3\|4/3 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. JHEP09(2018)006 Figure 22. (Colour online) Centrality dependence of v4/\|v3\|4/3 for inclusive charged particles [44], π±, K±, and p+p. Statistical and systematic uncertainties are shown as bars and boxes, respectively. anisotropic ﬂow in minimum bias pp collisions at the same collision energy, the non-ﬂow contributions are estimated from minimum bias pp collisions and subtracted from the mea- sured vn coeﬃcients. Ratios of the measurements presented in this paper to a cubic spline ﬁt to the ones performed at √sNN = 2.76 TeV are given in the ﬁgure for each presented centrality interval and ﬂow coeﬃcient. The uncertainties in these ratios are obtained by summing the statistical and systematic uncertainties on the independent measurements in quadrature, and propagating the obtained uncertainties as uncorrelated. An increase of radial ﬂow with increasing collision energy is expected to lead to a suppression of vn at low pT, an eﬀect which would be most pronounced for heavier particles. 4.5 Comparison with vn of identiﬁed particles at √sNN = 2.76 TeV Although a possible suppression of p+p vn at √sNN = 5.02 TeV can be seen between 1 ≲pT ≲3 GeV/c in central collisions and additionally for v2(pT) of π± and K± at the same centrality interval, the precision of the results does not allow for conclusions to be drawn as the measurements at diﬀerent collision energies are compatible within uncertainties. Figure 24 shows the v2(pT) of Λ+Λ, K0 S, and the φ-meson compared to ALICE measure- ments performed at √sNN = 2.76 TeV [27], where the v2 coeﬃcients at √sNN = 2.76 TeV have been measured using the scalar product method with an \|∆η\| > 0.9 gap to suppress non-ﬂow. No diﬀerences are observed between the K0 S and Λ+Λ v2(pT) measured at two diﬀerent collision energies. The strongly improved precision of the φ-meson measurement at √sNN = 5.02 TeV, both in terms of statistical uncertainty and granularity in pT, shows that the v2(pT) follows a mass ordering at low pT and groups with mesons after pT≈3 GeV/c for all centrality intervals. 5 Summary 1 2 3 4 5 2 v 0 0.1 0.2 0.3 = 5.02 TeV NN s \|>2} η ∆ {2, \| n v ± π ± K p p+ = 2.76 TeV NN s {AA-pp} n v ± π ± K p p+ 0-5% 1 2 3 4 5 0 0.1 0.2 0.3 \| < 0.5 y Pb \| − ALICE Pb 10-20% 1 2 3 4 5 0 0.1 0.2 0.3 40-50% 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 3 v 0 0.05 0.1 0.15 0-5% 1 2 3 4 5 0 0.05 0.1 0.15 10-20% 1 2 3 4 5 0 0.05 0.1 0.15 40-50% 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 4 v 0 0.05 0.1 0.15 0-5% 1 2 3 4 5 0 0.05 0.1 0.15 10-20% 1 2 3 4 5 0 0.05 0.1 0.15 40-50% ) c (GeV/ T p 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 JHEP09(2018)006 Figure 23. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of π±, K±, and p+p compared to ALICE measurements performed in Pb-Pb collisions at √sNN = 2.76 TeV (coloured bands) [28] for the 0–5%, 10–20%, and 40–50% centrality classes. For the measurements at √sNN = 5.02 TeV, statistical and systematic uncertainties are shown as bars and boxes, respectively. For the measurements at √sNN = 2.76 TeV, the thickness of the bands corresponds to the quadratic sum of statistical and systematic uncertainties. The ratios of measurements at √sNN = 5.02 TeV to a cubic spline ﬁt to the measurements at √sNN = 2.76 TeV are also presented for clarity. sion of these measurements provide constraints for initial-state ﬂuctuations and transport coeﬃcients of the medium. The magnitude of vn increases with decreasing centrality up to the 40–50% centrality interval for all particle species. 5 Summary In summary, the elliptic, triangular, and quadrangular ﬂow coeﬃcients of π±, K±, p+p, Λ+Λ, K0 S, and the φ-meson have been measured in Pb-Pb collisions at √sNN = 5.02 TeV over a broad range of transverse momentum and in various centrality ranges. The preci- – 28 – 1 2 3 4 5 2 v 0 0.1 0.2 0.3 = 5.02 TeV NN s \|>2} η ∆ {2, \| n v ± π ± K p p+ = 2.76 TeV NN s {AA-pp} n v ± π ± K p p+ 0-5% 1 2 3 4 5 0 0.1 0.2 0.3 \| < 0.5 y Pb \| − ALICE Pb 10-20% 1 2 3 4 5 0 0.1 0.2 0.3 40-50% 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 3 v 0 0.05 0.1 0.15 0-5% 1 2 3 4 5 0 0.05 0.1 0.15 10-20% 1 2 3 4 5 0 0.05 0.1 0.15 40-50% 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 0.5 1 1.5 1 2 3 4 5 4 v 0 0.05 0.1 0.15 0-5% 1 2 3 4 5 0 0.05 0.1 0.15 10-20% 1 2 3 4 5 0 0.05 0.1 0.15 40-50% ) c (GeV/ T p 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 Figure 23. (Colour online) The pT-diﬀerential v2 (top), v3 (middle), and v4 (bottom) of π±, K±, and p+p compared to ALICE measurements performed in Pb-Pb collisions at √sNN = 2.76 TeV (coloured bands) [28] for the 0–5%, 10–20%, and 40–50% centrality classes. For the measurements at √sNN = 5.02 TeV, statistical and systematic uncertainties are shown as bars and boxes, respectively. For the measurements at √sNN = 2.76 TeV, the thickness of the bands corresponds to the quadratic sum of statistical and systematic uncertainties. The ratios of measurements at √sNN = 5.02 TeV to a cubic spline ﬁt to the measurements at √sNN = 2.76 TeV are also presented for clarity. 5 Summary This increase is stronger for v2 than for v3 and v4, which indicates that collision geometry dominates the generation of elliptic ﬂow while higher ﬂow coeﬃcients are mainly generated by event-by-event ﬂuctua- tions in the initial nucleon and gluon densities. This interpretation is also supported by the non-zero, positive vn found in the 0–1% centrality interval. In most central collisions (i.e. 0–1% and 0–5%), v3 and v4 reach a similar magnitude as v2 at diﬀerent pT values depending on particle mass, after which they increase gradually. For pT < 3 GeV/c, the vn coeﬃcients show a mass ordering consistent with an interplay between anisotropic ﬂow and the isotropic expansion (radial ﬂow) of the collision system. In this transverse mo- mentum range, the iEBE-VISHNU hydrodynamical calculations describe the measured vn of π±, K±, and p+p fairly well for pT < 2.5 GeV/c, while MUSIC reproduces the measure- ments for pT < 1 GeV/c. It should be noted that neither of the presented hydrodynamical models is able to fully describe the measurements. At intermediate transverse momenta – 29 – 1 2 3 4 5 2 v 0 0.1 0.2 0.3 = 5.02 TeV NN s \|>2} η ∆ {2, \| 2 v 0 S K Λ + Λ φ = 2.76 TeV NN s \|>0.9} η ∆ {2, \| 2 v 0 S K Λ + Λ φ 0-5% 1 2 3 4 5 0 0.1 0.2 0.3 \| < 0.5 y Pb \| − ALICE Pb 10-20% 1 2 3 4 5 0 0.1 0.2 0.3 40-50% ) c (GeV/ T p 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 Figure 24. (Colour online) The pT-diﬀerential v2 of Λ+Λ, K0 S, and the φ-meson compared to AL- ICE measurements performed in Pb-Pb collisions at √sNN = 2.76 TeV (coloured bands) [27] for the 0–5%, 10–20%, and 40–50% centrality classes. For the measurements at √sNN = 5.02 TeV, statisti- cal and systematic uncertainties are shown as bars and boxes, respectively. For the measurements at √sNN = 2.76 TeV, the thickness of the bands corresponds to the quadratic sum of statistical and systematic uncertainties. 5 Summary The ratios of measurements at √sNN = 5.02 TeV to a cubic spline ﬁt to the measurements at √sNN = 2.76 TeV are also presented for clarity. 1 2 3 4 5 2 v 0 0.1 0.2 0.3 = 5.02 TeV NN s \|>2} η ∆ {2, \| 2 v 0 S K Λ + Λ φ = 2.76 TeV NN s \|>0.9} η ∆ {2, \| 2 v 0 S K Λ + Λ φ 0-5% 1 2 3 4 5 0 0.1 0.2 0.3 \| < 0.5 y Pb \| − ALICE Pb 10-20% 1 2 3 4 5 0 0.1 0.2 0.3 40-50% ) c (GeV/ T p 1 2 3 4 5 2.76 TeV 5.02 TeV 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 ) c (GeV/ T p 1 2 3 4 5 0.5 1 1.5 JHEP09(2018)006 Figure 24. (Colour online) The pT-diﬀerential v2 of Λ+Λ, K0 S, and the φ-meson compared to AL- ICE measurements performed in Pb-Pb collisions at √sNN = 2.76 TeV (coloured bands) [27] for the 0–5%, 10–20%, and 40–50% centrality classes. For the measurements at √sNN = 5.02 TeV, statisti- cal and systematic uncertainties are shown as bars and boxes, respectively. For the measurements at √sNN = 2.76 TeV, the thickness of the bands corresponds to the quadratic sum of statistical and systematic uncertainties. The ratios of measurements at √sNN = 5.02 TeV to a cubic spline ﬁt to the measurements at √sNN = 2.76 TeV are also presented for clarity. gure 24. (Colour online) The pT-diﬀerential v2 of Λ+Λ, K0 S, and the φ-meson compared to AL- (3 < pT < 8–10 GeV/c), particles show an approximate grouping by the number of con- stituent quarks at the level of ±20% for all ﬂow coeﬃcients in the 0–50% centrality range. The φ-meson v2, which tests both particle mass dependence and type scaling, follows p+p v2 at low pT and π± v2 at intermediate pT. The baryon vn has a magnitude larger than that of mesons for pT < 8–10 GeV/c, indicating that the particle type dependence persists up to high pT. For pT > 10 GeV/c, the v2 of p+p is compatible with that of π± within uncertainties. 5 Summary The shape evolution of v2(pT) as function of centrality shows diﬀerent trends for π±, K±, and p+p and varies between 20% (10%) to 250% (55%) for p+p (K±) at low pT from most central to peripheral collisions; variations are more pronounced for v3(pT), in particular for central collisions. Ratios v3/\|v2\|3/2 and v4/\|v2\|4/2 are ﬂat for pT < 3 GeV/c and rise monotonically for higher momenta for the 0–5% centrality class. No particle type dependence of the ratios is found for pT > 1.5 GeV/c, below which the ratios for p+p vn rise, which can be attributed to an increase of radial ﬂow which aﬀects the indepen- dent harmonics diﬀerently. For the ratio v4/\|v3\|4/3, no pT dependence is observed over the full centrality range. The measurements are compatible with those performed in Pb-Pb collisions at √sNN = 2.76 TeV within uncertainties. Acknowledgments Alikhanyan National Science Laboratory (Yerevan Physics Insti- tute) Foundation (ANSL), State Committee of Science and World Federation of Scientists (WFS), Armenia; Austrian Academy of Sciences and Nationalstiftung f¨ur Forschung, Tech- nologie und Entwicklung, Austria; Ministry of Communications and High Technologies, National Nuclear Research Center, Azerbaijan; Conselho Nacional de Desenvolvimento Cient´ıﬁco e Tecnol´ogico (CNPq), Universidade Federal do Rio Grande do Sul (UFRGS), Financiadora de Estudos e Projetos (Finep) and Funda¸c˜ao de Amparo `a Pesquisa do Es- tado de S˜ao Paulo (FAPESP), Brazil; Ministry of Science & Technology of China (MSTC), National Natural Science Foundation of China (NSFC) and Ministry of Education of China (MOEC), China; Ministry of Science and Education, Croatia; Ministry of Education, Youth and Sports of the Czech Republic, Czech Republic; The Danish Council for Independent Research — Natural Sciences, the Carlsberg Foundation and Danish National Research Foundation (DNRF), Denmark; Helsinki Institute of Physics (HIP), Finland; Commissariat `a l’Energie Atomique (CEA) and Institut National de Physique Nucl´eaire et de Physique des Particules (IN2P3) and Centre National de la Recherche Scientiﬁque (CNRS), France; Bundesministerium f¨ur Bildung, Wissenschaft, Forschung und Technologie (BMBF) and GSI Helmholtzzentrum f¨ur Schwerionenforschung GmbH, Germany; General Secretariat for Research and Technology, Ministry of Education, Research and Religions, Greece; Na- tional Research, Development and Innovation Oﬃce, Hungary; Department of Atomic En- ergy Government of India (DAE), Department of Science and Technology, Government of India (DST), University Grants Commission, Government of India (UGC) and Council of Scientiﬁc and Industrial Research (CSIR), India; Indonesian Institute of Science, Indone- sia; Centro Fermi - Museo Storico della Fisica e Centro Studi e Ricerche Enrico Fermi and Istituto Nazionale di Fisica Nucleare (INFN), Italy; Institute for Innovative Science and Technology, Nagasaki Institute of Applied Science (IIST), Japan Society for the Promotion of Science (JSPS) KAKENHI and Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan; Consejo Nacional de Ciencia (CONACYT) y Tecnolog´ıa, through Fondo de Cooperaci´on Internacional en Ciencia y Tecnolog´ıa (FONCICYT) and Direcci´on General de Asuntos del Personal Academico (DGAPA), Mexico; Nederlandse Or- ganisatie voor Wetenschappelijk Onderzoek (NWO), Netherlands; The Research Council of Norway, Norway; Commission on Science and Technology for Sustainable Development in the South (COMSATS), Pakistan; Pontiﬁcia Universidad Cat´olica del Per´u, Peru; Ministry of Science and Higher Education and National Science Centre, Poland; Korea Institute of Science and Technology Information and National Research Foundation of Korea (NRF), Republic of Korea; Ministry of Education and Scientiﬁc Research, Institute of Atomic Physics and Romanian National Agency for Science, Technology and Innovation, Roma- nia; Joint Institute for Nuclear Research (JINR), Ministry of Education and Science of the Russian Federation and National Research Centre Kurchatov Institute, Russia; Min- istry of Education, Science, Research and Sport of the Slovak Republic, Slovakia; National Research Foundation of South Africa, South Africa; Centro de Aplicaciones Tecnol´ogicas y Desarrollo Nuclear (CEADEN), Cubaenerg´ıa, Cuba and Centro de Investigaciones En- Worldwide LHC Computing Grid (WLCG) collaboration. Acknowledgments The ALICE Collaboration ac- knowledges the following funding agencies for their support in building and running the ALICE detector: A. I. Acknowledgments The ALICE Collaboration would like to thank all its engineers and technicians for their invaluable contributions to the construction of the experiment and the CERN accelerator teams for the outstanding performance of the LHC complex. The ALICE Collaboration gratefully acknowledges the resources and support provided by all Grid centres and the – 30 – Worldwide LHC Computing Grid (WLCG) collaboration. The ALICE Collaboration ac- knowledges the following funding agencies for their support in building and running the ALICE detector: A. I. Acknowledgments erg´eticas, Medioambientales y Tecnol´ogicas (CIEMAT), Spain; Swedish Research Council (VR) and Knut & Alice Wallenberg Foundation (KAW), Sweden; European Organization for Nuclear Research, Switzerland; National Science and Technology Development Agency (NSDTA), Suranaree University of Technology (SUT) and Oﬃce of the Higher Education Commission under NRU project of Thailand, Thailand; Turkish Atomic Energy Agency (TAEK), Turkey; National Academy of Sciences of Ukraine, Ukraine; Science and Tech- nology Facilities Council (STFC), United Kingdom; National Science Foundation of the United States of America (NSF) and United States Department of Energy, Oﬃce of Nu- clear Physics (DOE NP), United States of America. JHEP09(2018)006 Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. Acknowledgments Alikhanyan National Science Laboratory (Yerevan Physics Insti- tute) Foundation (ANSL), State Committee of Science and World Federation of Scientists (WFS), Armenia; Austrian Academy of Sciences and Nationalstiftung f¨ur Forschung, Tech- nologie und Entwicklung, Austria; Ministry of Communications and High Technologies, National Nuclear Research Center, Azerbaijan; Conselho Nacional de Desenvolvimento Cient´ıﬁco e Tecnol´ogico (CNPq), Universidade Federal do Rio Grande do Sul (UFRGS), Financiadora de Estudos e Projetos (Finep) and Funda¸c˜ao de Amparo `a Pesquisa do Es- tado de S˜ao Paulo (FAPESP), Brazil; Ministry of Science & Technology of China (MSTC), National Natural Science Foundation of China (NSFC) and Ministry of Education of China (MOEC), China; Ministry of Science and Education, Croatia; Ministry of Education, Youth and Sports of the Czech Republic, Czech Republic; The Danish Council for Independent Research — Natural Sciences, the Carlsberg Foundation and Danish National Research Foundation (DNRF), Denmark; Helsinki Institute of Physics (HIP), Finland; Commissariat `a l’Energie Atomique (CEA) and Institut National de Physique Nucl´eaire et de Physique des Particules (IN2P3) and Centre National de la Recherche Scientiﬁque (CNRS), France; Bundesministerium f¨ur Bildung, Wissenschaft, Forschung und Technologie (BMBF) and GSI Helmholtzzentrum f¨ur Schwerionenforschung GmbH, Germany; General Secretariat for Research and Technology, Ministry of Education, Research and Religions, Greece; Na- tional Research, Development and Innovation Oﬃce, Hungary; Department of Atomic En- ergy Government of India (DAE), Department of Science and Technology, Government of India (DST), University Grants Commission, Government of India (UGC) and Council of Scientiﬁc and Industrial Research (CSIR), India; Indonesian Institute of Science, Indone- sia; Centro Fermi - Museo Storico della Fisica e Centro Studi e Ricerche Enrico Fermi and Istituto Nazionale di Fisica Nucleare (INFN), Italy; Institute for Innovative Science and Technology, Nagasaki Institute of Applied Science (IIST), Japan Society for the Promotion of Science (JSPS) KAKENHI and Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan; Consejo Nacional de Ciencia (CONACYT) y Tecnolog´ıa, through Fondo de Cooperaci´on Internacional en Ciencia y Tecnolog´ıa (FONCICYT) and Direcci´on General de Asuntos del Personal Academico (DGAPA), Mexico; Nederlandse Or- ganisatie voor Wetenschappelijk Onderzoek (NWO), Netherlands; The Research Council of Norway, Norway; Commission on Science and Technology for Sustainable Development in the South (COMSATS), Pakistan; Pontiﬁcia Universidad Cat´olica del Per´u, Peru; Ministry of Science and Higher Education and National Science Centre, Poland; Korea Institute of Science and Technology Information and National Research Foundation of Korea (NRF), Republic of Korea; Ministry of Education and Scientiﬁc Research, Institute of Atomic Physics and Romanian National Agency for Science, Technology and Innovation, Roma- nia; Joint Institute for Nuclear Research (JINR), Ministry of Education and Science of the Russian Federation and National Research Centre Kurchatov Institute, Russia; Min- istry of Education, Science, Research and Sport of the Slovak Republic, Slovakia; National Research Foundation of South Africa, South Africa; Centro de Aplicaciones Tecnol´ogicas y Desarrollo Nuclear (CEADEN), Cubaenerg´ıa, Cuba and Centro de Investigaciones En- JHEP09(2018)006 – 31 – erg´eticas, Medioambientales y Tecnol´ogicas (CIEMAT), Spain; Swedish Research Council (VR) and Knut & Alice Wallenberg Foundation (KAW), Sweden; European Organization for Nuclear Research, Switzerland; National Science and Technology Development Agency (NSDTA), Suranaree University of Technology (SUT) and Oﬃce of the Higher Education Commission under NRU project of Thailand, Thailand; Turkish Atomic Energy Agency (TAEK), Turkey; National Academy of Sciences of Ukraine, Ukraine; Science and Tech- nology Facilities Council (STFC), United Kingdom; National Science Foundation of the United States of America (NSF) and United States Department of Energy, Oﬃce of Nu- clear Physics (DOE NP), United States of America. References [1] S.A. Bass, M. Gyulassy, H. Stoecker and W. Greiner, Signatures of quark gluon plasma formation in high-energy heavy ion collisions: A Critical review, J. Phys. G 25 (1999) R1 [hep-ph/9810281] [INSPIRE]. [2] BRAHMS collaboration, I. Arsene et al., Quark gluon plasma and color glass condensate at RHIC? The Perspective from the BRAHMS experiment, Nucl. Phys. A 757 (2005) 1 [nucl-ex/0410020] [INSPIRE]. [3] PHENIX collaboration, K. Adcox et al., Formation of dense partonic matter in relativistic nucleus-nucleus collisions at RHIC: Experimental evaluation by the PHENIX collaboration, Nucl. Phys. A 757 (2005) 184 [nucl-ex/0410003] [INSPIRE]. [4] B.B. Back et al., The PHOBOS perspective on discoveries at RHIC, Nucl. Phys. A 757 (2005) 28 [nucl-ex/0410022] [INSPIRE]. [4] B.B. Back et al., The PHOBOS perspective on discoveries at RHIC, Nucl. Phys. A 757 (2005) 28 [nucl-ex/0410022] [INSPIRE]. [5] STAR collaboration, J. Adams et al., Experimental and theoretical challenges in the search for the quark gluon plasma: The STAR Collaboration’s critical assessment of the evidence from RHIC collisions, Nucl. Phys. A 757 (2005) 102 [nucl-ex/0501009] [INSPIRE]. [5] STAR collaboration, J. Adams et al., Experimental and theoretical challenges in the search for the quark gluon plasma: The STAR Collaboration’s critical assessment of the evidence from RHIC collisions, Nucl. Phys. A 757 (2005) 102 [nucl-ex/0501009] [INSPIRE]. [6] ALICE collaboration, Elliptic ﬂow of charged particles in Pb-Pb collisions at 2.76 TeV, Phys. Rev. Lett. 105 (2010) 252302 [arXiv:1011.3914] [INSPIRE]. [7] ATLAS collaboration, Measurement of the pseudorapidity and transverse momentum dependence of the elliptic ﬂow of charged particles in lead-lead collisions at √sNN = 2.76 TeV with the ATLAS detector, Phys. Lett. B 707 (2012) 330 [arXiv:1108.6018] [INSPIRE]. [8] CMS collaboration, Measurement of the elliptic anisotropy of charged particles produced in PbPb collisions at √sNN = 2.76 TeV, Phys. Rev. C 87 (2013) 014902 [arXiv:1204.1409] [INSPIRE]. [9] P. Kovtun, D.T. Son and A.O. Starinets, Viscosity in strongly interacting quantum ﬁeld theories from black hole physics, Phys. Rev. Lett. 94 (2005) 111601 [hep-th/0405231] [INSPIRE]. [10] J.-Y. Ollitrault, Anisotropy as a signature of transverse collective ﬂow, Phys. Rev. D 46 (1992) 229 [INSPIRE]. – 32 – [11] S. Voloshin and Y. Zhang, Flow study in relativistic nuclear collisions by Fourier expansion of Azimuthal particle distributions, Z. Phys. C 70 (1996) 665 [hep-ph/9407282] [INSPIRE]. [12] R.S. Bhalerao and J.-Y. Ollitrault, Eccentricity ﬂuctuations and elliptic ﬂow at RHIC, Phys. Lett. B 641 (2006) 260 [nucl-th/0607009] [INSPIRE]. [13] PHOBOS collaboration, B. References Alver et al., Importance of correlations and ﬂuctuations on the initial source eccentricity in high-energy nucleus-nucleus collisions, Phys. Rev. C 77 (2008) 014906 [arXiv:0711.3724] [INSPIRE]. [14] B. Alver and G. Roland, Collision geometry ﬂuctuations and triangular ﬂow in heavy-ion collisions, Phys. Rev. C 81 (2010) 054905 [Erratum ibid. C 82 (2010) 039903] [arXiv:1003.0194] [INSPIRE]. JHEP09(2018)006 JHEP09(2018)006 [15] B.H. Alver, C. Gombeaud, M. Luzum and J.-Y. Ollitrault, Triangular ﬂow in hydrodynamics and transport theory, Phys. Rev. C 82 (2010) 034913 [arXiv:1007.5469] [INSPIRE]. [16] D. Teaney and L. Yan, Non linearities in the harmonic spectrum of heavy ion collisions with ideal and viscous hydrodynamics, Phys. Rev. C 86 (2012) 044908 [arXiv:1206.1905] [INSPIRE]. [17] ALICE collaboration, Linear and non-linear ﬂow modes in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Lett. B 773 (2017) 68 [arXiv:1705.04377] [INSPIRE]. [18] D. Teaney and L. Yan, Triangularity and Dipole Asymmetry in Heavy Ion Collisions, Phys. Rev. C 83 (2011) 064904 [arXiv:1010.1876] [INSPIRE]. [19] G.-Y. Qin, H. Petersen, S.A. Bass and B. M¨uller, Translation of collision geometry ﬂuctuations into momentum anisotropies in relativistic heavy-ion collisions, Phys. Rev. C 82 (2010) 064903 [arXiv:1009.1847] [INSPIRE]. [20] P. Huovinen, P.F. Kolb, U.W. Heinz, P.V. Ruuskanen and S.A. Voloshin, Radial and elliptic ﬂow at RHIC: Further predictions, Phys. Lett. B 503 (2001) 58 [hep-ph/0101136] [INSPIRE]. [21] D. Teaney, J. Lauret and E.V. Shuryak, Flow at the SPS and RHIC as a quark gluon plasma signature, Phys. Rev. Lett. 86 (2001) 4783 [nucl-th/0011058] [INSPIRE]. [22] S.A. Voloshin, Transverse radial expansion and directed ﬂow, Phys. Rev. C 55 (1997) R1630 [nucl-th/9611038] [INSPIRE]. [23] C. Shen, U. Heinz, P. Huovinen and H. Song, Radial and elliptic ﬂow in Pb+Pb collisions at the Large Hadron Collider from viscous hydrodynamic, Phys. Rev. C 84 (2011) 044903 [arXiv:1105.3226] [INSPIRE]. [24] E877 collaboration, J. Barrette et al., Proton and pion production relative to the reaction plane in Au + Au collisions at AGS energies, Phys. Rev. C 56 (1997) 3254 [nucl-ex/9707002] [INSPIRE]. [25] STAR collaboration, J. Adams et al., Particle type dependence of azimuthal anisotropy and nuclear modiﬁcation of particle production in Au + Au collisions at √sNN = 200 GeV, Phys. Rev. Lett. 92 (2004) 052302 [nucl-ex/0306007] [INSPIRE]. [26] PHENIX collaboration, S.S. Adler et al., Elliptic ﬂow of identiﬁed hadrons in Au+Au collisions at √sNN = 200 GeV, Phys. Rev. Lett. 91 (2003) 182301 [nucl-ex/0305013] [INSPIRE]. References – 33 – [27] ALICE collaboration, Elliptic ﬂow of identiﬁed hadrons in Pb-Pb collisions at √sNN = 2.76 TeV, JHEP 06 (2015) 190 [arXiv:1405.4632] [INSPIRE]. [28] ALICE collaboration, Higher harmonic ﬂow coeﬃcients of identiﬁed hadrons in Pb-Pb collisions at √sNN = 2.76 TeV, JHEP 09 (2016) 164 [arXiv:1606.06057] [INSPIRE]. [29] STAR collaboration, B.I. Abelev et al., Mass, quark-number and √sNN dependence of the second and fourth ﬂow harmonics in ultra-relativistic nucleus-nucleus collisions, Phys. Rev. C 75 (2007) 054906 [nucl-ex/0701010] [INSPIRE]. [30] STAR collaboration, B.I. Abelev et al., Partonic ﬂow and phi-meson production in Au + Au collisions at √sNN = 200 GeV, Phys. Rev. Lett. 99 (2007) 112301 [nucl-ex/0703033] [INSPIRE]. JHEP09(2018)006 [31] PHENIX collaboration, A. Adare et al., Scaling properties of azimuthal anisotropy in Au+Au and Cu+Cu collisions at √sNN = 200 GeV, Phys. Rev. Lett. 98 (2007) 162301 [nucl-ex/0608033] [INSPIRE]. [32] PHENIX collaboration, A. Adare et al., Deviation from quark-number scaling of the anisotropy parameter v2 of pions, kaons and protons in Au+Au collisions at √sNN = 200 GeV, Phys. Rev. C 85 (2012) 064914 [arXiv:1203.2644] [INSPIRE]. [33] ALICE collaboration, Anisotropic ﬂow of charged hadrons, pions and (anti-)protons measured at high transverse momentum in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Lett. B 719 (2013) 18 [arXiv:1205.5761] [INSPIRE]. [34] D. Moln´ar and S.A. Voloshin, Elliptic ﬂow at large transverse momenta from quark coalescence, Phys. Rev. Lett. 91 (2003) 092301 [nucl-th/0302014] [INSPIRE]. [35] H. Sato and K. Yazaki, On the coalescence model for high-energy nuclear reactions, Phys. Lett. B 98 (1981) 153 [INSPIRE]. [36] C.B. Dover, U.W. Heinz, E. Schnedermann and J. Zim´anyi, Relativistic coalescence model for high-energy nuclear collisions, Phys. Rev. C 44 (1991) 1636 [INSPIRE]. [37] K. Werner, I. Karpenko, M. Bleicher, T. Pierog and S. Porteboeuf-Houssais, Jets, Bulk Matter and their Interaction in Heavy Ion Collisions at Several TeV, Phys. Rev. C 85 (2012) 064907 [arXiv:1203.5704] [INSPIRE]. [38] K. Werner, V2 Scaling in PbPb Collisions at 2.76 TeV, arXiv:1205.3379 [INSPIRE]. [39] R. Baier, Y.L. Dokshitzer, S. Peigne and D. Schiﬀ, Induced gluon radiation in a QCD medium, Phys. Lett. B 345 (1995) 277 [hep-ph/9411409] [INSPIRE]. [40] M. Gyulassy and M. Plumer, Jet Quenching in Dense Matter, Phys. Lett. B 243 (1990) 432 [INSPIRE]. [41] ATLAS collaboration, Measurement of the Azimuthal Angle Dependence of Inclusive Jet Yields in Pb+Pb Collisions at √sNN = 2.76 TeV with the ATLAS detector, Phys. Rev. Lett. 111 (2013) 152301 [arXiv:1306.6469] [INSPIRE]. References [42] ALICE collaboration, Azimuthal anisotropy of charged jet production in √sNN = 2.76 TeV Pb-Pb collisions, Phys. Lett. B 753 (2016) 511 [arXiv:1509.07334] [INSPIRE]. [43] CMS collaboration, Azimuthal anisotropy of charged particles with transverse momentum up to 100 GeV/c in PbPb collisions at √sNN = 5.02 TeV, Phys. Lett. B 776 (2018) 195 [arXiv:1702.00630] [INSPIRE]. – 34 – [44] ALICE collaboration, Energy dependence and ﬂuctuations of anisotropic ﬂow in Pb-Pb collisions at √sNN = 5.02 and 2.76 TeV, JHEP 07 (2018) 103 [arXiv:1804.02944] [INSPIRE]. [45] T. Hirano, U.W. Heinz, D. Kharzeev, R. Lacey and Y. Nara, Hadronic dissipative eﬀects on elliptic ﬂow in ultrarelativistic heavy-ion collisions, Phys. Lett. B 636 (2006) 299 [nucl-th/0511046] [INSPIRE]. [46] S. Takeuchi, K. Murase, T. Hirano, P. Huovinen and Y. Nara, Eﬀects of hadronic rescattering on multistrange hadrons in high-energy nuclear collisions, Phys. Rev. C 92 (2015) 044907 [arXiv:1505.05961] [INSPIRE]. [47] STAR collaboration, B.I. Abelev et al., Measurements of φ-meson production in relativistic heavy-ion collisions at RHIC, Phys. Rev. C 79 (2009) 064903 [arXiv:0809.4737] [INSPIRE]. JHEP09(2018)006 JHEP09(2018)006 [48] A. Shor, ϕ-meson production as a probe of the quark gluon plasma, Phys. Rev. Lett. 54 (1985) 1122 [INSPIRE]. [49] K. Haglin, Collision rates for ρ-, ω- and φ-mesons at nonzero temperature, Nucl. Phys. A 584 (1995) 719 [nucl-th/9410028] [INSPIRE]. [50] W. Smith and K.L. Haglin, Collision broadening of the ϕ meson in baryon rich hadronic matter, Phys. Rev. C 57 (1998) 1449 [nucl-th/9710026] [INSPIRE]. [51] L. ´Alvarez-Ruso and V. Koch, ϕ-meson propagation in a hot hadronic gas, Phys. Rev. C 65 (2002) 054901 [nucl-th/0201011] [INSPIRE]. [52] T. Ishikawa et al., φ photo-production from Li, C, Al and Cu nuclei at Eγ = 1.5–2.4 GeV, Phys. Lett. B 608 (2005) 215 [nucl-ex/0411016] [INSPIRE]. [53] H.J. Behrend et al., Photoproduction of φ-mesons at small t values, Phys. Lett. B 56 (1975) 408 [INSPIRE]. [54] P. Muhlich, T. Falter, C. Greiner, J. Lehr, M. Post and U. Mosel, Photoproduction of φ-mesons from nuclei, Phys. Rev. C 67 (2003) 024605 [nucl-th/0210079] [INSPIRE]. [55] ALICE collaboration, K∗(892)0 and φ(1020) production in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. C 91 (2015) 024609 [arXiv:1404.0495] [INSPIRE]. [56] HADES collaboration, J. Adamczewski-Musch et al., Deep sub-threshold φ production in Au+Au collisions, Phys. Lett. B 778 (2018) 403 [arXiv:1703.08418] [INSPIRE]. [57] STAR collaboration, C. References Adler et al., Elliptic ﬂow from two and four particle correlations in Au+Au collisions at √sNN = 130 GeV, Phys. Rev. C 66 (2002) 034904 [nucl-ex/0206001] [INSPIRE]. [58] S.A. Voloshin, A.M. Poskanzer and R. Snellings, Collective phenomena in non-central nuclear collisions, Landolt-Bornstein 23 (2010) 293 [arXiv:0809.2949] [INSPIRE]. [59] M. Luzum and J.-Y. Ollitrault, Eliminating experimental bias in anisotropic-ﬂow measurements of high-energy nuclear collisions, Phys. Rev. C 87 (2013) 044907 [arXiv:1209.2323] [INSPIRE]. [60] ALICE collaboration, ALICE: Physics performance report, volume I, J. Phys. G 30 (2004) 1517 [INSPIRE]. [61] ALICE collaboration, ALICE: Physics performance report, volume II, J. Phys. G 32 (2006) 1295 [INSPIRE]. – 35 – [62] ALICE collaboration, The ALICE experiment at the CERN LHC, 2008 JINST 3 S08002 [INSPIRE]. [63] ALICE collaboration, Performance of the ALICE Experiment at the CERN LHC, Int. J. Mod. Phys. A 29 (2014) 1430044 [arXiv:1402.4476] [INSPIRE]. [64] ALICE collaboration, Alignment of the ALICE Inner Tracking System with cosmic-ray tracks, 2010 JINST 5 P03003 [arXiv:1001.0502] [INSPIRE]. [65] J. Alme et al., The ALICE TPC, a large 3-dimensional tracking device with fast readout for ultra-high multiplicity events, Nucl. Instrum. Meth. A 622 (2010) 316 [arXiv:1001.1950] [INSPIRE]. JHEP09(2018)006 [66] A. Akindinov et al., Performance of the ALICE Time-Of-Flight detector at the LHC, Eur. Phys. J. Plus 128 (2013) 44 [INSPIRE]. [67] ALICE collaboration, Performance of the ALICE VZERO system, 2013 JINST 8 P10016 [arXiv:1306.3130] [INSPIRE]. [68] M. Bondila et al., ALICE T0 detector, IEEE Trans. Nucl. Sci. 52 (2005) 1705. [69] ALICE collaboration, Centrality determination of Pb-Pb collisions at √sNN = 2.76 TeV with ALICE, Phys. Rev. C 88 (2013) 044909 [arXiv:1301.4361] [INSPIRE]. [70] ALICE collaboration, Centrality dependence of π, K, p production in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. C 88 (2013) 044910 [arXiv:1303.0737] [INSPIRE]. [71] Particle Data Group collaboration, C. Patrignani et al., Review of Particle Physics, Chin. Phys. C 40 (2016) 100001 [INSPIRE]. [72] J. Podolanski and R. Armenteros, III. Analysis of V-events, Philos. Mag. 45 (1954) 13. [73] G.I. Kopylov, Like particle correlations as a tool to study the multiple production mechanism, Phys. Lett. B 50 (1974) 472 [INSPIRE]. [74] I. Selyuzhenkov and S. Voloshin, Eﬀects of non-uniform acceptance in anisotropic ﬂow measurement, Phys. Rev. C 77 (2008) 034904 [arXiv:0707.4672] [INSPIRE]. [75] N. Borghini and J.Y. Ollitrault, Azimuthally sensitive correlations in nucleus-nucleus collisions, Phys. Rev. C 70 (2004) 064905 [nucl-th/0407041] [INSPIRE]. [76] R. References Barlow, Systematic errors: Facts and ﬁctions, in Advanced Statistical Techniques in Particle Physics. Proceedings, Conference, Durham, U.K., March 18–22, 2002, pp. 134–144 (2002) [hep-ex/0207026] [INSPIRE]. [77] F.G. Gardim, F. Grassi, M. Luzum and J.-Y. Ollitrault, Mapping the hydrodynamic response to the initial geometry in heavy-ion collisions, Phys. Rev. C 85 (2012) 024908 [arXiv:1111.6538] [INSPIRE]. [78] H. Song and U.W. Heinz, Suppression of elliptic ﬂow in a minimally viscous quark-gluon plasma, Phys. Lett. B 658 (2008) 279 [arXiv:0709.0742] [INSPIRE]. [79] H. Song, S. Bass and U.W. Heinz, Spectra and elliptic ﬂow for identiﬁed hadrons in 2.76A TeV Pb + Pb collisions, Phys. Rev. C 89 (2014) 034919 [arXiv:1311.0157] [INSPIRE]. [80] ALICE collaboration, Centrality dependence of the nuclear modiﬁcation factor of charged pions, kaons and protons in Pb-Pb collisions at √sNN = 2.76 TeV, Phys. Rev. C 93 (2016) 034913 [arXiv:1506.07287] [INSPIRE]. – 36 – [81] V. Greco, C.M. Ko and P. Levai, Parton coalescence at RHIC, Phys. Rev. C 68 (2003) 034904 [nucl-th/0305024] [INSPIRE]. [82] S. McDonald, C. Shen, F. Fillion-Gourdeau, S. Jeon and C. Gale, Hydrodynamic predictions for Pb+Pb collisions at 5.02 TeV, Phys. Rev. C 95 (2017) 064913 [arXiv:1609.02958] [INSPIRE]. [83] W. Zhao, H.-j. Xu and H. Song, Collective ﬂow in 2.76 A TeV and 5.02 A TeV Pb+Pb collisions, Eur. Phys. J. C 77 (2017) 645 [arXiv:1703.10792] [INSPIRE]. [84] B. Schenke, S. Jeon and C. Gale, Elliptic and triangular ﬂow in event-by-event (3+1)D viscous hydrodynamics, Phys. Rev. Lett. 106 (2011) 042301 [arXiv:1009.3244] [INSPIRE]. JHEP09(2018)006 JHEP09(2018)006 [85] S.A. Bass et al., Microscopic models for ultrarelativistic heavy ion collisions, Prog. Part. Nucl. Phys. 41 (1998) 255 [nucl-th/9803035] [INSPIRE]. [86] M. Bleicher et al., Relativistic hadron hadron collisions in the ultrarelativistic quantum molecular dynamics model, J. Phys. G 25 (1999) 1859 [hep-ph/9909407] [INSPIRE]. [87] B. Schenke, P. Tribedy and R. Venugopalan, Event-by-event gluon multiplicity, energy density and eccentricities in ultrarelativistic heavy-ion collisions, Phys. Rev. C 86 (2012) 034908 [arXiv:1206.6805] [INSPIRE]. [88] B. Schenke, P. Tribedy and R. Venugopalan, Fluctuating Glasma initial conditions and ﬂow in heavy ion collisions, Phys. Rev. Lett. 108 (2012) 252301 [arXiv:1202.6646] [INSPIRE]. [89] C. Shen, Z. Qiu, H. Song, J. Bernhard, S. Bass and U. Heinz, The iEBE-VISHNU code package for relativistic heavy-ion collisions, Comput. Phys. Commun. 199 (2016) 61 [arXiv:1409.8164] [INSPIRE]. [90] H. Song, S.A. Bass and U. Heinz, Viscous QCD matter in a hybrid hydrodynamic+Boltzmann approach, Phys. Rev. C 83 (2011) 024912 [arXiv:1012.0555] [INSPIRE]. References [91] J.S. Moreland, J.E. Bernhard and S.A. Bass, Alternative ansatz to wounded nucleon and binary collision scaling in high-energy nuclear collisions, Phys. Rev. C 92 (2015) 011901 [arXiv:1412.4708] [INSPIRE]. [92] Z.-W. Lin, C.M. Ko, B.-A. Li, B. Zhang and S. Pal, A Multi-phase transport model for relativistic heavy ion collisions, Phys. Rev. C 72 (2005) 064901 [nucl-th/0411110] [INSPIRE]. [93] J.E. Bernhard, J.S. Moreland, S.A. Bass, J. Liu and U. Heinz, Applying Bayesian parameter estimation to relativistic heavy-ion collisions: simultaneous characterization of the initial state and quark-gluon plasma medium, Phys. Rev. C 94 (2016) 024907 [arXiv:1605.03954] [INSPIRE]. [94] R.J. Fries, V. Greco and P. Sorensen, Coalescence Models For Hadron Formation From Quark Gluon Plasma, Ann. Rev. Nucl. Part. Sci. 58 (2008) 177 [arXiv:0807.4939] [INSPIRE]. [95] P.M. Dinh, N. Borghini and J.-Y. Ollitrault, Eﬀects of HBT correlations on ﬂow measurements, Phys. Lett. B 477 (2000) 51 [nucl-th/9912013] [INSPIRE]. [96] N. Borghini and J.-Y. Ollitrault, Momentum spectra, anisotropic ﬂow and ideal ﬂuids, Phys. Lett. B 642 (2006) 227 [nucl-th/0506045] [INSPIRE]. – 37 – [97] B.H. Alver, C. Gombeaud, M. Luzum and J.-Y. Ollitrault, Triangular ﬂow in hydrodynamics and transport theory, Phys. Rev. C 82 (2010) 034913 [arXiv:1007.5469] [INSPIRE]. [98] STAR collaboration, J. Adams et al., Azimuthal anisotropy at RHIC: The First and fourth harmonics, Phys. Rev. Lett. 92 (2004) 062301 [nucl-ex/0310029] [INSPIRE]. [99] PHENIX collaboration, A. Adare et al., Elliptic and hexadecapole ﬂow of charged hadrons in Au+Au collisions at √sNN = 200 GeV, Phys. Rev. Lett. 105 (2010) 062301 [arXiv:1003.5586] [INSPIRE]. [99] PHENIX collaboration, A. Adare et al., Elliptic and hexadecapole ﬂow of charged hadrons in Au+Au collisions at √sNN = 200 GeV, Phys. Rev. Lett. 105 (2010) 062301 [arXiv:1003.5586] [INSPIRE]. [100] ATLAS collaboration, Measurement of the azimuthal anisotropy for charged particle production in √sNN = 2.76 TeV lead-lead collisions with the ATLAS detector, Phys. Rev. C 86 (2012) 014907 [arXiv:1203.3087] [INSPIRE]. JHEP09(2018)006 [101] CMS collaboration, Studies of azimuthal dihadron correlations in ultra-central PbPb collisions at √sNN = 2.76 TeV, JHEP 02 (2014) 088 [arXiv:1312.1845] [INSPIRE]. – 38 – The ALICE collaboration The ALICE collaboration S. Acharya139 , F.T.-. Acosta22 , D. Adamov´a94 , J. Adolfsson81 , M.M. Aggarwal98 , G. Aglieri Rinella36 , M. Agnello33 , N. Agrawal49 , Z. Ahammed139 , S.U. Ahn77 , S. Aiola144 , A. Akindinov65 , M. Al-Turany104 , S.N. Alam139 , D.S.D. Albuquerque120 , D. Aleksandrov88 , B. Alessandro59 , R. Alfaro Molina73 , Y. Ali16 , A. Alici11 ,54 ,29 , A. Alkin3 , J. Alme24 , T. Alt70 , L. Altenkamper24 , I. Altsybeev138 , M.N. Anaam7 , C. Andrei48 , D. Andreou36 , H.A. Andrews108 , A. Andronic142 ,104 , M. Angeletti36 , V. Anguelov102 , C. Anson17 , T. Antiˇci´c105 , F. Antinori57 , P. Antonioli54 , R. Anwar124 , N. Apadula80 , L. Aphecetche112 , H. Appelsh¨auser70 , S. Arcelli29 , R. Arnaldi59 , O.W. Arnold103 ,115 , I.C. Arsene23 , S. Acharya139 , F.T.-. Acosta22 , D. Adamov´a94 , J. Adolfsson81 , M.M. Aggarwal98 , G. Aglieri Rinella36 , M. Agnello33 , N. Agrawal49 , Z. Ahammed139 , S.U. Ahn77 , S. Aiola144 , A. Akindinov65 , M. Al-Turany104 , S.N. Alam139 , D.S.D. Albuquerque120 , D. Aleksandrov88 , B. Alessandro59 , R. Alfaro Molina73 , Y. Ali16 , A. Alici11 ,54 ,29 , A. Alkin3 , J. Alme24 , T. Alt70 , L. Altenkamper24 , I. Altsybeev138 , M.N. Anaam7 , C. Andrei48 , D. Andreou36 , H.A. Andrews108 , A. Andronic142 ,104 , M. Angeletti36 , V. Anguelov102 , C. Anson17 , T. Antiˇci´c105 , F. Antinori57 , P. Antonioli54 , R. Anwar124 , N. Apadula80 , L. Aphecetche112 , H. Appelsh¨auser70 , S. Arcelli29 , R. Arnaldi59 , O.W. Arnold103 ,115 , I.C. Arsene23 , H. Appelsh¨auser70 , S. Arcelli29 , R. Arnaldi59 , O.W. Arnold103 ,115 , I.C. Arsene23 , JHEP09(2018)006 M. Arslandok102 , B. Audurier112 , A. Augustinus36 , R. Averbeck104 , M.D. Azmi18 , A. Badala Y.W. Baek61 ,42 , S. Bagnasco59 , R. Bailhache70 , R. Bala99 , A. Baldisseri134 , M. Ball44 , C. Baral86 , A.M. Barbano28 , R. Barbera30 , F. Barile53 , L. Barioglio28 , G.G. Barnaf¨oldi143 , S. Barnby93 , V. Barret131 , P. Bartalini7 , K. Barth36 , E. Bartsch70 , N. Bastid131 , S. Basu141 , Batigne112 , B. Batyunya76 , P.C. Batzing23 , J.L. Bazo Alba109 , I.G. Bearden89 , H. Beck102 , N. Bianchi52 , J. Bielˇc´ık39 , J. Bielˇc´ıkov´a94 , A. The ALICE collaboration Guerzoni29 , M. Guittiere112 , K. Gulbrandsen89 , T. Gunji129 , A. Gupta99 , R. Gupta99 , I.B. Guzman2 , R. Haake36 , M.K. Habib104 , C. Hadjidakis62 , H. Hamagaki82 , G. Hamar143 , M. Hamid7 , J.C. Hamon133 , R. Hannigan117 , M.R. Haque64 , J.W. Harris144 , A. Harton12 , H. Hassan79 , R. Hannigan117 , M.R. Haque64 , J.W. Harris144 , A. Harton12 , H. Hassan79 , D. Hatzifotiadou54 ,11 , S. Hayashi129 , S.T. Heckel70 , E. Hellb¨ar70 , H. Helstrup38 , D. Hatzifotiadou54 ,11 , S. Hayashi129 , S.T. Heckel70 , E. Hellb¨ar70 , H. Helstrup38 A. Herghelegiu48 , E.G. Hernandez2 , G. Herrera Corral10 , F. Herrmann142 , K.F. Hetland38 45 36 126 133 103 3 g g , G , G C , , , T.E. Hilden45 , H. Hillemanns36 , C. Hills126 , B. Hippolyte133 , B. Hohlweger103 , D. Horak39 , S. Hornung104 , R. Hosokawa130 ,79 , J. Hota67 , P. Hristov36 , C. Huang62 , C. Hughes127 , P. Huhn70 , T.J. Humanic19 , H. Hushnud107 , N. Hussain43 , T. Hussain18 , D. Hutter41 , D.S. Hwang21 , J.P. Iddon126 , S.A. Iga Buitron71 , R. Ilkaev106 , M. Inaba130 , M. Ippolitov88 , M.S. Islam107 , M. Ivanov104 , V. Ivanov96 , V. Izucheev91 , B. Jacak80 , N. Jacazio29 , P.M. Jacobs80 , M.B. Jadhav49 , S. Jadlovska114 , J. Jadlovsky114 , S. Jaelani64 , hnke119 ,115 , M.J. Jakubowska140 , M.A. Janik140 , C. Jena86 , M. Jercic97 , O. Jevons108 , R.T. Jimenez Bustamante104 , M. Jin124 , P.G. Jones108 , A. Jusko108 , P. Kalinak66 , T. Karavicheva63 , P. Karczmarczyk36 , E. Karpechev63 , U. Kebschull75 , R. Keidel47 , T. Karavicheva63 , P. Karczmarczyk36 , E. Karpechev63 , U. Kebschull75 , R. Keidel47 , D.L.D. Keijdener64 , M. Keil36 , B. Ketzer44 , Z. Khabanova90 , A.M. Khan7 , S. Khan18 , S.A. Khan139 , A. Khanzadeev96 , Y. Kharlov91 , A. Khatun18 , A. Khuntia50 , S.A. Khan139 , A. Khanzadeev96 , Y. Kharlov91 , A. Khatun18 , A. Khuntia50 , M. Kielbowicz116 , B. Kileng38 , B. Kim130 , D. Kim145 , D.J. Kim125 , E.J. Kim14 , H. Kim145 , S. Kim42 , J. Kim102 , M. Kim61 ,102 , S. Kim21 , T. Kim145 , T. Kim145 , S. Kirsch41 , Kisel41 , S. The ALICE collaboration Bilandzic115 ,103 , G. Biro143 , R. Biswas4 , S. Biswas4 , J.T. Blair117 , D. Blau88 , C. Blume70 , G. Boca136 , F. Bock36 , A. Bogdanov92 , N. Bianchi52 , J. Bielˇc´ık39 , J. Bielˇc´ıkov´a94 , A. Bilandzic115 ,103 , G. Biro143 , R. Biswas4 , Butt16 , J.T. Buxton19 , J. Cabala114 , D. Caﬀarri90 , H. Caines144 , A. Caliva104 , E. Calvo , R.S. Camacho2 , P. Camerini27 , A.A. Capon111 , F. Carena36 , W. Carena36 , – 39 – A. Festanti31 ,36 , V.J.G. Feuillard102 , J. Figiel116 , M.A.S. Figueredo119 , S. Filchagin106 Festanti31 ,36 , V.J.G. Feuillard102 , J. Figiel116 , M.A.S. Figueredo119 , S. Filchagin106 , D. Finogeev63 , F.M. Fionda24 , G. Fiorenza53 , F. Flor124 , M. Floris36 , S. Foertsch74 , D. Finogeev63 , F.M. Fionda24 , G. Fiorenza53 , F. Flor124 , M. Floris36 , S. Foertsch74 g P. Foka104 , S. Fokin88 , E. Fragiacomo60 , A. Francescon36 , A. Francisco112 , U. Frankenfeld104 G.G. Fronze28 , U. Fuchs36 , C. Furget79 , A. Furs63 , M. Fusco Girard32 , J.J. Gaardhøje89 , M. Gagliardi28 , A.M. Gago109 , K. Gajdosova89 , M. Gallio28 , C.D. Galvan118 , P. Ganoti84 , C. Garabatos104 , E. Garcia-Solis12 , K. Garg30 , C. Gargiulo36 , P. Gasik115 ,103 , E.F. Gauger11 B. Gay Ducati72 , M. Germain112 , J. Ghosh107 , P. Ghosh139 , S.K. Ghosh4 , P. Gianotti52 , P. Giubellino104 ,59 , P. Giubilato31 , P. Gl¨assel102 , D.M. Gom´ez Coral73 , A. Gomez Ramirez75 , P. Giubellino104 ,59 , P. Giubilato31 , P. Gl¨assel102 , D.M. Gom´ez Coral73 , A. Gomez Ramirez75 , V G l 104 P G ´l Z 2 S G b 41 L G¨ li h116 S G 37 V. Grabski73 , L.K. Graczykowski140 , K.L. Graham108 , L. Greiner80 , A. Grelli64 , C. Grigoras36 , V. Grabski73 , L.K. Graczykowski140 , K.L. Graham108 , L. Greiner80 , A. Grelli64 , C. Grigoras36 , 92 1 76 104 33 abski73 , L.K. Graczykowski140 , K.L. Graham108 , L. Greiner80 , A. Grelli64 , C. Grigoras36 , igoriev92 , A. Grigoryan1 , S. Grigoryan76 , J.M. Gronefeld104 , F. Grosa33 , JHEP09(2018)006 JHEP09(2018)006 J.F. Grosse-Oetringhaus36 , R. Grosso104 , R. Guernane79 , B. Guerzoni29 , M. Guittiere112 , J.F. Grosse-Oetringhaus36 , R. Grosso104 , R. Guernane79 , B. The ALICE collaboration Kiselev65 , A. Kisiel140 , J.L. Klay6 , C. Klein70 , J. Klein36 ,59 , C. Klein-B¨osing142 , Klewin102 , A. Kluge36 , M.L. Knichel36 , A.G. Knospe124 , C. Kobdaj113 , M. Kofarago143 , M.K. K¨ohler102 , T. Kollegger104 , N. Kondratyeva92 , E. Kondratyuk91 , A. Konevskikh63 , M. Konyushikhin141 , O. Kovalenko85 , V. Kovalenko138 , M. Kowalski116 , I. Kr´alik66 , A. Kravˇc´akov´a40 , L. Kreis104 , M. Krivda66 ,108 , F. Krizek94 , M. Kr¨uger70 , E. Kryshen96 , A. Kravˇc´akov´a40 , L. Kreis104 , M. Krivda66 ,108 , F. Krizek94 , M. Kr¨uger70 , E. Kryshen96 , M. Krzewicki41 , A.M. Kubera19 , V. Kuˇcera94 ,61 , C. Kuhn133 , P.G. Kuijer90 , J. Kumar49 , L. Kumar98 , S. Kumar49 , S. Kundu86 , P. Kurashvili85 , A. Kurepin63 , A.B. Kurepin63 , A. Kuryakin106 , S. Kushpil94 , J. Kvapil108 , M.J. Kweon61 , Y. Kwon145 , S.L. La Pointe41 , P. La Rocca30 , Y.S. Lai80 , I. Lakomov36 , R. Langoy122 , K. Lapidus144 , A. Lardeux23 , P. Larionov52 , E. Laudi36 , R. Lavicka39 , R. Lea27 , L. Leardini102 , S. Lee145 , F. Lehas90 , S. Lehner111 , J. Lehrbach41 , R.C. Lemmon93 , I. Le´on Monz´on118 , P. L´evai143 , X. Li13 , X.L. Li7 , J. Lien122 , R. Lietava108 , B. Lim20 , S. Lindal23 , V. Lindenstruth41 , S.W. Lindsay126 , C. Lippmann104 , M.A. Lisa19 , V. Litichevskyi45 , A. Liu80 , H.M. Ljunggren81 , W.J. Llope141 , D.F. Lodato64 , V. Loginov92 , C. Loizides95 ,80 , P. Loncar37 , X. Lopez131 , E. L´opez Torres9 , A. Lowe143 , P. Luettig70 , J.R. Luhder142 , M. Lunardon31 , G. Luparello60 , M. Lupi36 , A. Maevskaya63 , M. Mager36 , S.M. Mahmood23 , A. Maire133 , R.D. Majka144 , M. Malaev96 , Q.W. Malik23 , L. Malinina76 , iii, D. Mal’Kevich65 , P. Malzacher104 , A. Mamonov106 , – 40 – Manko88 , F. Manso131 , V. Manzari53 , Y. Mao7 , M. Marchisone128 ,74 ,132 , J. Mareˇs68 , V. Manko88 , F. Manso131 , V. Manzari53 , Y. Mao7 , M. Marchisone128 ,74 ,132 , J. Mareˇs G.V. Margagliotti27 , A. Margotti54 , J. Margutti64 , A. Mar´ın104 , C. Markert117 , G.V. Margagliotti27 , A. Margotti54 , J. Margutti64 , A. Mar´ın104 , C. The ALICE collaboration Markert117 , g g g g M. Marquard70 , N.A. Martin104 , P. Martinengo36 , J.L. Martinez124 , M.I. Mart´ınez2 JHEP09(2018)006 JHEP09(2018)006 Nikulin96 , F. Noferini11 ,54 , P. Nomokonov76 , G. Nooren64 , J.C.C. Noris2 , J. Norman79 , A. Nyanin88 , J. Nystrand24 , H. Oh145 , A. Ohlson102 , J. Oleniacz140 , A.C. Oliveira Da Silva11 H. Oliver144 , J. Onderwaater104 , C. Oppedisano59 , R. Orava45 , M. Oravec114 , A. Ortiz lasquez71 , A. Oskarsson81 , J. Otwinowski116 , K. Oyama82 , Y. Pachmayer102 , V. Pacik89 , D. Pagano137 , G. Pai´c71 , P. Palni7 , J. Pan141 , A.K. Pandey49 , S. Panebianco134 , V. Papikyan1 , P. Pareek50 , J. Park61 , J.E. Parkkila125 , S. Parmar98 , A. Passfeld142 , S.P. Pathak124 , R.N. Patra139 , B. Paul59 , H. Pei7 , T. Peitzmann64 , X. Peng7 , L.G. Pereira72 , S.P. Pathak124 , R.N. Patra139 , B. Paul59 , H. Pei7 , T. Peitzmann64 , X. Peng7 , L.G. Pereira V. Petr´aˇcek39 , M. Petrovici48 , C. Petta30 , R.P. Pezzi72 , S. Piano60 , M. Pikna15 , P. Pillot112 V. Petr´aˇcek39 , M. Petrovici48 , C. Petta30 , R.P. Pezzi72 , S. Piano60 , M. Pikna15 , P. Pillot112 , O.D.L. Pimentel89 , O. Pinazza54 ,36 , L. Pinsky124 , S. Pisano52 , D.B. Piyarathna124 , L.O.D.L. Pimentel89 , O. Pinazza54 ,36 , L. Pinsky124 , S. Pisano52 , D.B. Piyarath M. P losko´n80 , M. Planinic97 , F. Pliquett70 , J. Pluta140 , S. Pochybova143 , M. P losko´n80 , M. Planinic97 , F. Pliquett70 , J. Pluta140 , S. Pochybova143 , P.L.M. Podesta-Lerma118 , M.G. Poghosyan95 , B. Polichtchouk91 , N. Poljak97 , P.L.M. Podesta-Lerma118 , M.G. Poghosyan95 , B. Polichtchouk91 , N. Poljak97 W. Poonsawat113 , A. Pop48 , H. Poppenborg142 , S. Porteboeuf-Houssais131 , V. P W. Poonsawat113 , A. Pop48 , H. Poppenborg142 , S. Porteboeuf-Houssais131 , V. Pozdniakov76 , W. Poonsawat113 , A. Pop48 , H. Poppenborg142 , S. Porteboeuf-Houssais131 , V. Pozdniakov76 , S.K. Prasad4 , R. Preghenella54 , F. Prino59 , C.A. Pruneau141 , I. Pshenichnov63 , M. Puccio28 , V. Punin106 , J. Putschke141 , S. Raha4 , S. Rajput99 , J. Rak125 , A. Rakotozaﬁndrabe134 , L. Ramello34 , F. Rami133 , R. Raniwala100 , S. Raniwala100 , S.S. R¨as¨anen45 , B.T. Rascanu70 , V. Ratza44 , I. The ALICE collaboration Ravasenga33 , K.F. Read127 ,95 , K. Redlich85 , v, A. Rehman24 , P. Reichelt70 , F. Reidt36 , X. Ren7 , R. Renfordt70 , A. Reshetin63 , J.-P. Revol11 , K. Reygers102 , V. Riabov96 , T. Richert64 ,81 , M. Richter23 , P. Riedler36 , W. Riegler36 , F. Riggi30 , C. Ristea69 , S.P. Rode50 , M. Rodr´ıguez Cahuantzi2 , K. Røed23 , R. Rogalev91 , E. Rogochaya76 , D. Rohr36 , D. R¨ohrich24 , P.S. Rokita140 , F. Ronchetti52 , E.D. Rosas71 , K. Roslon140 , P. Rosnet131 , A. Rossi31 , A. Rotondi136 , F. Roukoutakis84 , C. Roy133 , P. Roy107 , O.V. Rueda71 , R. Rui27 , B. Rumyantsev76 , A. Rustamov87 , E. Ryabinkin88 , Y. Ryabov96 , A. Rybicki116 , S. Saarinen45 , S. Sadhu139 , S. Sadovsky91 , K. ˇSafaˇr´ık36 , S.K. Saha139 , B. Sahoo49 , P. Sahoo50 , R. Sahoo50 , S. Sahoo67 , P.K. Sahu67 , J. Saini139 , S. Sakai130 , M.A. Saleh141 , S. Sambyal99 , V. Samsonov96 ,92 , A. Sandoval73 , A. Sarkar74 , D. Sarkar139 , N. Sarkar139 , P. Sarma43 , M.H.P. Sas64 , E. Scapparone54 , F. Scarlassara31 , B. Schaefer95 , H.S. Scheid70 , C. Schiaua48 , R. Schicker102 , C. Schmidt104 , H.R. Schmidt101 , M.O. Schmidt102 , M. Schmidt101 , N.V. Schmidt95 ,70 , J. Schukraft36 , Y. Schutz36 ,133 , K. Schwarz104 , K. Schweda104 , G. Scioli29 , E. Scomparin59 , M. ˇSefˇc´ık40 , J.E. Seger17 , Y. Sekiguchi129 , D. Sekihata46 , Prasad4 , R. Preghenella54 , F. Prino59 , C.A. Pruneau141 , I. Pshenichnov63 , M. Puccio28 , unin106 J Putschke141 S Raha4 S Rajput99 J Rak125 A Rakotozaﬁndrabe134 S.K. Prasad4 , R. Preghenella54 , F. Prino59 , C.A. Pruneau141 , I. Pshenichnov63 , M. Puccio28 Prasad4 , R. Preghenella54 , F. Prino59 , C.A. Pruneau141 , I. Pshenichnov63 , M. Puccio28 , unin106 , J. Putschke141 , S. Raha4 , S. Rajput99 , J. Rak125 , A. Rakotozaﬁndrabe134 , V. Punin106 , J. Putschke141 , S. Raha4 , S. Rajput99 , J. Rak125 , A. Rakotozaﬁndrabe134 , L. Ramello34 , F. Rami133 , R. Raniwala100 , S. Raniwala100 , S.S. R¨as¨anen45 , B.T. Rascanu70 , V. Ratza44 , I. Ravasenga33 , K.F. Read127 ,95 , K. Redlich85 , v, A. Rehman24 , P. Reichelt70 , F. Reidt36 , X. Ren7 , R. The ALICE collaboration Sharma98 , , y , , g , Sharma99 , M. Sharma99 , N. Sharma98 , A.I. Sheikh139 , K. Shigaki46 , M. Shimomura83 , Shirinkin65 , Q. Shou7 ,110 , K. Shtejer28 , Y. Sibiriak88 , S. Siddhanta55 , K.M. Sielewicz36 , T. Siemiarczuk85 , D. Silvermyr81 , G. Simatovic90 , G. Simonetti36 ,103 , R. Singaraju139 , Singh86 , R. Singh99 , V. Singhal139 , T. Sinha107 , B. Sitar15 , M. Sitta34 , T.B. Skaali23 , Slupecki125 , N. Smirnov144 , R.J.M. Snellings64 , T.W. Snellman125 , J. Song20 , F. Soramel31 , S. Sorensen127 , F. Sozzi104 , I. Sputowska116 , J. Stachel102 , I. Stan69 , P. Stankus95 , S. Sorensen127 , F. Sozzi104 , I. Sputowska116 , J. Stachel102 , I. Stan69 , P. Stankus95 , C. Suire62 , M. Suleymanov16 , M. Suljic36 ,27 , R. Sultanov65 , M. ˇSumbera94 , C. Suire62 , M. Suleymanov16 , M. Suljic36 ,27 , R. Sultanov65 , M. ˇSumbera94 , JHEP09(2018)006 J. Takahashi120 , G.J. Tambave24 , N. Tanaka130 , M. Tarhini112 , M. Tariq18 , M.G. Tarzila48 A. Tauro36 , G. Tejeda Mu˜noz2 , A. Telesca36 , C. Terrevoli31 , B. Teyssier132 , D. Thakur50 , S. Thakur139 , D. Thomas117 , F. Thoresen89 , R. Tieulent132 , A. Tikhonov63 , A.R. Timmins12 S. Thakur139 , D. Thomas117 , F. Thoresen89 , R. Tieulent132 , A. Tikhonov63 , A.R. Timmins124 , A T i 70 N T il k 63 M T i52 S R T 118 S T i th 50 S T l 28 akur139 , D. Thomas117 , F. Thoresen89 , R. Tieulent132 , A. Tikhonov63 , A.R. Timm ia70 , N. Topilskaya63 , M. Toppi52 , S.R. Torres118 , S. Tripathy50 , S. Trogolo28 , A. Toia70 , N. Topilskaya63 , M. Toppi52 , S.R. Torres118 , S. Tripathy50 , S. Trogolo28 , G. Trombetta35 , L. Tropp40 , V. Trubnikov3 , W.H. Trzaska125 , T.P. Trzcinski140 , B.A. Trzeciak64 , T. Tsuji129 , A. Tumkin106 , R. Turrisi57 , T.S. Tveter23 , K. Ullaland24 , E.N. Umaka124 , A. Uras132 , G.L. Usai26 , A. Utrobicic97 , M. Vala114 , J.W. Van Hoorne36 , M. van Leeuwen64 , P. Vande Vyvre36 , D. Varga143 , A. Vargas2 , M. Vargyas125 , R. Varma M. Vasileiou84 , A. Vasiliev88 , A. Vauthier79 , O. V´azquez Doce103 ,115 , V. The ALICE collaboration Vechernin138 , A.M. Veen64 , E. Vercellin28 , S. Vergara Lim´on2 , L. Vermunt64 , R. Vernet8 , R. V´ertesi143 , L. Vickovic37 , J. Viinikainen125 , Z. Vilakazi128 , O. Villalobos Baillie108 , A. Villatoro Tello2 , A. Vinogradov88 , T. Virgili32 , V. Vislavicius89 ,81 , A. Vodopyanov76 , M.A. V¨olkl101 , K. Voloshin65 , S.A. Voloshin141 , G. Volpe35 , B. von Haller36 , I. Vorobyev115 ,103 , D. Voscek D. Vranic104 ,36 , J. Vrl´akov´a40 , B. Wagner24 , H. Wang64 , M. Wang7 , Y. Watanabe130 , M. Weber , S.G. Weber , A. Wegrzynek , D.F. Weiser , S.C. Wenzel , J.P. Wessels , U. Westerhoﬀ142 , A.M. Whitehead123 , J. Wiechula70 , J. Wikne23 , G. Wilk85 , J. Wilkinson54 , G.A. Willems142 ,36 , M.C.S. Williams54 , E. Willsher108 , B. Windelband102 , W.E. Witt127 , 7 78 46 46 7 79 130 20 U. Westerhoﬀ142 , A.M. Whitehead123 , J. Wiechula70 , J. Wikne23 , G. Wilk85 , J. Wilkinson54 , Westerhoﬀ142 , A.M. Whitehead123 , J. Wiechula70 , J. Wikne23 , G. Wilk85 , J. Wilkinson54 , A. Willems142 ,36 , M.C.S. Williams54 , E. Willsher108 , B. Windelband102 , W.E. Witt127 , J.H. Yoon61 , V. Yurchenko3 , V. Zaccolo59 , A. Zaman16 , C. Zampolli36 , H.J.C. Zanoli119 N. Zardoshti108 , A. Zarochentsev138 , P. Z´avada68 , N. Zaviyalov106 , H. Zbroszczyk140 , , , , y , y , M. Zhalov96 , X. Zhang7 , Y. Zhang7 , Z. Zhang7 ,131 , C. Zhao23 , V. Zherebchevskii138 , N. Zhigareva65 , D. Zhou7 , Y. Zhou89 , Z. Zhou24 , H. Zhu7 , J. Zhu7 , Y. Zhu7 , A. Zichichi29 ,11 , M.B. Zimmermann36 , G. Zinovjev3 , J. Zmeskal111 , S. Zou7 N. Zhigareva65 , D. Zhou7 , Y. Zhou89 , Z. Zhou24 , H. Zhu7 , J. Zhu7 , Y. Zhu7 , A. Zichichi29 ,1 i Deceased ii Dipartimento DET del Politecnico di Torino, Turin, Italy iii M.V. Lomonosov Moscow State University, D.V. Skobeltsyn Institute of Nuclear, Physics, Moscow, Russia iv Department of Applied Physics, Aligarh Muslim University, Aligarh, India v Institute of Theoretical Physics, University of Wroclaw, Poland 1 A.I. Alikhanyan National Science Laboratory (Yerevan Physics Institute) Foundation, Yerevan, Armenia 1 A.I. The ALICE collaboration Renfordt70 , A. Reshetin63 , J.-P. Revol11 , K. Reygers102 , V. Riabov96 , F. Reidt36 , X. Ren7 , R. Renfordt70 , A. Reshetin63 , J.-P. Revol11 , K. Reygers102 , V. Riabov96 , T. Richert64 ,81 , M. Richter23 , P. Riedler36 , W. Riegler36 , F. Riggi30 , C. Ristea69 , S.P. Rode50 , M. Rodr´ıguez Cahuantzi2 , K. Røed23 , R. Rogalev91 , E. Rogochaya76 , D. Rohr36 , D. R¨ohrich24 , P.S. Rokita140 , F. Ronchetti52 , E.D. Rosas71 , K. Roslon140 , P. Rosnet131 , A Rossi31 A Rotondi136 F Roukoutakis84 C Roy133 P Roy107 O V Rueda71 R Rui27 F. Reidt36 , X. Ren7 , R. Renfordt70 , A. Reshetin63 , J.-P. Revol11 , K. Reygers102 , V. Riabov96 64 81 23 36 36 30 69 T. Richert64 ,81 , M. Richter23 , P. Riedler36 , W. Riegler36 , F. Riggi30 , C. Ristea69 , S.P. Rode50 , M. Rodr´ıguez Cahuantzi2 , K. Røed23 , R. Rogalev91 , E. Rogochaya76 , D. Rohr36 , D. R¨ohrich24 , P.S. Rokita140 , F. Ronchetti52 , E.D. Rosas71 , K. Roslon140 , P. Rosnet131 , A. Rossi31 , A. Rotondi136 , F. Roukoutakis84 , C. Roy133 , P. Roy107 , O.V. Rueda71 , R. Rui27 , B. Rumyantsev76 , A. Rustamov87 , E. Ryabinkin88 , Y. Ryabov96 , A. Rybicki116 , S. Saarinen45 , S. Sadhu139 , S. Sadovsky91 , K. ˇSafaˇr´ık36 , S.K. Saha139 , B. Sahoo49 , P. Sahoo50 , R. Sahoo50 , S. Sahoo67 , P.K. Sahu67 , J. Saini139 , S. Sakai130 , M.A. Saleh141 , S. Sambyal99 , V. Samsonov96 ,92 , A. Sandoval73 , A. Sarkar74 , D. Sarkar139 , N. Sarkar139 , P. Sarma43 , M.H.P. Sas64 , E. Scapparone54 , F. Scarlassara31 , B. Schaefer95 , H.S. Scheid70 , C. Schiaua48 , R. Schicker102 , C. Schmidt104 , H.R. Schmidt101 , M.O. Schmidt102 , M. Schmidt101 , N.V. Schmidt95 ,70 , J. Schukraft36 , Y. Schutz36 ,133 , K. Schwarz104 , K. Schweda104 , G. Scioli29 , E. Scomparin59 , M. ˇSefˇc´ık40 , J.E. Seger17 , Y. Sekiguchi129 , D. Sekihata46 , – 41 – Selyuzhenkov104 ,92 , K. Senosi74 , S. Senyukov133 , E. Serradilla73 , P. Sett49 , A. Sevcenco69 , A. Shabanov63 , A. Shabetai112 , R. Shahoyan36 , W. Shaikh107 , A. Shangaraev91 , A. 5 Budker Institute for Nuclear Physics, Novosibirsk, Russia The ALICE collaboration Alikhanyan National Science Laboratory (Yerevan Physics Institute) Foundation, Yerevan, Armenia 2 Benem´erita Universidad Aut´onoma de Puebla, Puebla, Mexico 3 Bogolyubov Institute for Theoretical Physics, National Academy of Sciences of Ukraine, Kiev, Ukraine 4 Bose Institute, Department of Physics and Centre for Astroparticle Physics and Space Science (CAPSS), Kolkata, India 4 Bose Institute, Department of Physics and Centre for Astroparticle Physics and Space Science (CAPSS), Kolkata, India 5 Budker Institute for Nuclear Physics, Novosibirsk, Russia 5 Budker Institute for Nuclear Physics, Novosibirsk, Russia 5 Budker Institute for Nuclear Physics, Novosibirsk, Russia – 42 – 6 California Polytechnic State University, San Luis Obispo, California, United States 7 Central China Normal University, Wuhan, China 7 Central China Normal University, Wuhan, China 8 Centre de Calcul de l’IN2P3, Villeurbanne, Lyon, France 9 Centro de Aplicaciones Tecnol´ogicas y Desarrollo Nuclear (CEADEN), Havana, Cub Centro de Investigaci´on y de Estudios Avanzados (C ( ) 11 Centro Fermi - Museo Storico della Fisica e Centro Studi e Ricerche “Enrico Fermi’, Rome, Ital 12 Chicago State University, Chicago, Illinois, United States 13 China Institute of Atomic Energy, Beijing, China 14 Chonbuk National University, Jeonju, Republic of Korea 15 Comenius University Bratislava, Faculty of Mathematics, Physics and Informatics, Bratislava, Slovakia JHEP09(2018)006 17 Creighton University, Omaha, Nebraska, United States 18 Department of Physics, Aligarh Muslim University, Aligarh, India 19 Department of Physics, Ohio State University, Columbus, Ohio, United States 20 Department of Physics, Pusan National University, Pusan, Republic of Korea epartment of Physics, Pusan National University, Pu 21 Department of Physics, Sejong University, Seoul, Republic of Korea 22 Department of Physics, University of California, Berkeley, California, United States epartment of Physics, University of Oslo, Oslo, Norw 24 Department of Physics and Technology, University of Bergen, Bergen, Norway 25 Dipartimento di Fisica dell’Universit`a ‘La Sapienza’ and Sezione INFN, Ro 26 Dipartimento di Fisica dell’Universit`a and Sezione INFN, Cagliari, Italy 27 Dipartimento di Fisica dell’Universit`a and Sezione INFN, Trieste, Italy Dipartimento di Fisica dell’Universit`a and Sezione 28 Dipartimento di Fisica dell’Universit`a and Sezione INFN, Turin, Italy 29 Dipartimento di Fisica e Astronomia dell’Universit`a and Sezione INFN, Bologna, Italy ipartimento di Fisica e Astronomia dell’Universit`a an 31 Dipartimento di Fisica e Astronomia dell’Universit`a and Sezione INFN, Padova, Italy 32 Dipartimento di Fisica ‘E.R. The ALICE collaboration Caianiello’ dell’Universit`a and Gruppo Collegato INFN, Sale 33 Dipartimento DISAT del Politecnico and Sezione INFN, Turin, Italy 34 Dipartimento di Scienze e Innovazione Tecnologica dell’Universit`a del P INFN Sezione di Torino Alessandria Italy 34 Dipartimento di Scienze e Innovazione Tecnologica dell’Universit`a del Piemonte Orientale and 4 Dipartimento di Scienze e Innovazione Tecnologic INFN Sezione di Torino, Alessandria, Italy INFN Sezione di Torino, Alessandria, Italy 35 Dipartimento Interateneo di Fisica ‘M. Merlin’ and Sezione INFN, Bari, Italy 36 European Organization for Nuclear Research (CERN), Geneva, Switzerland 37 Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split, Split, Croatia 38 Faculty of Engineering and Science, Western Norway University of Applied Sciences, Bergen, Norway 39 Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University in Prague, Prague, Czech Republic 40 Faculty of Science, P.J. ˇSaf´arik University, Koˇsice, Slovakia of Science, P.J. ˇSaf´arik University, Koˇsice, Slovakia 40 Faculty of Science, P.J. The ALICE collaboration ˇSaf´arik University, Koˇsice, Slovakia 41 Frankfurt Institute for Advanced Studies, Johann Wolfgang Goethe-Universit¨at Frankfurt, Frankfurt, Germany 42 Gangneung-Wonju National University, Gangneung, Republic of Korea 43 Gauhati University, Department of Physics, Guwahati, India 44 Helmholtz-Institut f¨ur Strahlen- und Kernphysik, Rheinische Friedrich-Wilhelms-Universit¨at Bonn Bonn, Germany 45 Helsinki Institute of Physics (HIP), Helsinki, Finland 46 45 Helsinki Institute of Physics (HIP), Helsinki, Finland 46 Hiroshima University, Hiroshima, Japan 46 Hiroshima University, Hiroshima, Japan 47 Hochschule Worms, Zentrum f¨ur Technologietransfer und Telekommunikation (ZTT), Worms, Germany 48 Horia Hulubei National Institute of Physics and Nuclear Engineering, Bucharest, Romania Horia Hulubei National Institute of Physics and Nuclear Engineering, Bucharest, Romania 49 Indian Institute of Technology Bombay (IIT), Mumbai, India 49 Indian Institute of Technology Bombay (IIT), Mumbai, India – 43 – 50 Indian Institute of Technology Indore, Indore, India 51 Indonesian Institute of Sciences, Jakarta, Indonesia 52 INFN, Laboratori Nazionali di Frascati, Frascati, Italy 53 INFN, Sezione di Bari, Bari, Italy 54 INFN, Sezione di Bologna, Bologna, Italy 55 INFN, Sezione di Cagliari, Cagliari, Italy 56 INFN, Sezione di Catania, Catania, Italy 57 INFN, Sezione di Padova, Padova, Italy 58 INFN, Sezione di Roma, Rome, Italy 59 INFN, Sezione di Torino, Turin, Italy 60 INFN, Sezione di Trieste, Trieste, Italy 61 Inha University, Incheon, Republic of Korea JHEP09(2018)006 62 Institut de Physique Nucl´eaire d’Orsay (IPNO), Institut National de Physique Nucl´eaire et de Physique des Particules (IN2P3/CNRS), Universit´e de Paris-Sud, Universit´e Paris-Saclay, Orsay, France 63 Institute for Nuclear Research, Academy of Sciences, Moscow, Russia 64 Institute for Subatomic Physics, Utrecht University/Nikhef, Utrecht, Netherlands 65 Institute for Theoretical and Experimental Physics, Moscow, Russia 67 Institute of Physics, Bhubaneswar, India 67 Institute of Physics, Bhubaneswar, India 68 Institute of Physics of the Czech Academy of Sciences, Prague, Czech Republic 68 Institute of Physics of the Czech Academy 69 Institute of Space Science (ISS), Bucharest, Romania 69 Institute of Space Science (ISS), Bucharest, Romania ( ) 70 Institut f¨ur Kernphysik, Johann Wolfgang Goethe-Universit¨at Frankfurt, Frankfurt, Germany stitut f¨ur Kernphysik, Johann Wolfgang Goethe-Univ 71 Instituto de Ciencias Nucleares, Universidad Nacional Aut´onoma de M´exico, Mexico City, Mexic 72 Instituto de F´ısica, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil 73 Instituto de F´ısica, Universidad Nacional Aut´onoma de M´exico, Mexico City, Mexic 74 74 iThemba LABS, National Research Foundation, Somerset West, South Africa iThemba LABS, National Research Foundation, Som 75 Johann-Wolfgang-Goethe Universit¨at Frankfurt Institut f¨ur Informatik, Fachbereich Informatik und 75 Johann-Wolfgang-Goethe Universitat Frankfurt Institut fur Informatik, Fachbereich Informatik und Mathematik, Frankfurt, Germany fg g f f f , f Mathematik, Frankfurt, Germany Mathematik, Frankfurt, Germany Mathematik, Frankfurt, Germany 76 Joint Institute for Nuclear Research (JINR), Dubna, Russia 77 Korea Institute of Science and Technology Information, Daejeon, Republic of Korea 78 78 KTO Karatay University, Konya, Turkey 78 KTO Karatay University, Konya, Turkey 79 Laboratoire de Physique Subatomique et de Cosmologie, Universit´e Grenoble-Alpes, CNRS-IN2P3, Grenoble, France 80 Lawrence Berkeley National Laboratory, Berkeley, California, United States Lund University Department of Physics, Division of Particle Physics, Lund, Sweden 81 Lund University Department of Physics, Division of Particle Physics, Lund, Sweden 82 Nagasaki Institute of Applied Science, Nagasaki, Japan 83 Nara Women’s University (NWU), Nara, Japan 84 National and Kapodistrian University of Athens, School of Science, Department of Physics, Athens, Greece 85 National Centre for Nuclear Research, Warsaw, Poland ational Centre for Nuclear Research, Warsaw, Polan 86 National Institute of Science Education and Research, HBNI, Jatni, India 86 National Institute of Science Education and Research, HBNI, Jatni, India 87 National Nuclear Research Center, Baku, Azerbaijan 88 National Research Centre Kurchatov Institute, Moscow, Russia 89 Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 89 Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 90 Nikhef, National institute for subatomic physics, Amsterdam, Netherlands 1 NRC Kurchatov Institute IHEP, Protvino, Russia 92 NRNU Moscow Engineering Physics Institute, Moscow, Russia 2 NRNU Moscow Engineering Physics Institute, M 93 Nuclear Physics Group, STFC Daresbury Laboratory, Daresbury, United Kingdom 94 Nuclear Physics Institute of the Czech Academy of Sciences, ˇReˇz u Prahy, Czech Republi 95 Oak Ridge National Laboratory, Oak Ridge, Tennessee, United States 96 Petersburg Nuclear Physics Institute, Gatchina, Russia – 44 – 97 Physics department, Faculty of science, University of Zagreb, Zagreb, Croatia 97 Physics department, Faculty of science, University of Zagreb, Zagreb, Croatia 98 Physics Department, Panjab University, Chandigarh, India 99 Physics Department, University of Jammu, Jammu, India 100 Physics Department, University of Rajasthan, Jaipur, India Physikalisches Institut, Eberhard-Karls-Universit¨at T 102 Physikalisches Institut, Ruprecht-Karls-Universit¨at Heidelberg, Heidelberg, Germany 103 Physik Department, Technische Universit¨at M¨unchen, Munich, Germany 104 Research Division and ExtreMe Matter Institute EMMI, GSI Helmholtzzentrum f¨ur Schwerionenforschung GmbH Darmstadt Germany 104 Research Division and ExtreMe Matter Institute EMM Schwerionenforschung GmbH, Darmstadt, Germany 104 Research Division and ExtreMe Matter Institute EMMI, GSI Helmholtzzentrum f¨ur Schwerionenforschung GmbH, Darmstadt, Germany Research Division and ExtreMe Matter Institute EMMI, GS Schwerionenforschung GmbH, Darmstadt, Germany 105 Rudjer Boˇskovi´c Institute, Zagreb, Croatia 106 Russian Federal Nuclear Center (VNIIEF), Sarov, Russia 7 Saha Institute of Nuclear Physics, Kolkata, India JHEP09(2018)006 108 School of Physics and Astronomy, University of Birmingham, Birmingham, United Kingdom 108 School of Physics and Astronomy, University of Birmingham, Birming 8 School of Physics and Astronomy, University of B 109 Secci´on F´ısica, Departamento de Ciencias, Pontiﬁcia Universidad Cat´olica del Per´u, Lima, Peru 110 Shanghai Institute of Applied Physics, Shanghai, China 111 Stefan Meyer Institut f¨ur Subatomare Physik (SMI), Vienna, Austria Stefan Meyer Institut f¨ur Subatomare Physik (SMI), 112 SUBATECH, IMT Atlantique, Universit´e de Nantes, CNRS-IN2P3, Nantes, France 113 Suranaree University of Technology, Nakhon Ratchasima, Thailand 4 Technical University of Koˇsice, Koˇsice, Slovakia 115 Technische Universit¨at M¨unchen, Excellence Cluster ‘Universe’, Munich, German 5 Technische Universit¨at M¨unchen, Excellence Clus 16 The Henryk Niewodniczanski Institute of Nuclear Physics, Polish Academy of Sciences, Cracow, Poland 116 The Henryk Niewodniczanski Institute of Nuclear Physics, Polish Academy of Sciences, Cracow, Poland 17 The University of Texas at Austin, Austin, Texas, United States 117 The University of Texas at Austin, Austin, Texas, United States 118 Universidad Aut´onoma de Sinaloa, Culiac´an, Mexico 119 Universidade de S˜ao Paulo (USP), S˜ao Paulo, Brazil 120 Universidade Estadual de Campinas (UNICAMP), Campinas, Brazil 121 Universidade Federal do ABC, Santo Andre, Brazil 121 Universidade Federal do ABC, Santo Andre, Brazil 122 University College of Southeast Norway, Tonsberg, Norway 123 University of Cape Town, Cape Town, South Africa 124 University of Houston, Houston, Texas, United States 125 University of Jyv¨askyl¨a, Jyv¨askyl¨a, Finland 126 University of Liverpool, Department of Physics Oliver Lodge Laboratory, Liverpool, United Kingdom 126 University of Liverpool, Department of Physics Oliver Lodge Labora 126 University of Liverpool, Department of Physics Oliver Lodge Laboratory, Liverpool, United Kingdom 127 U i it f T K ill T U it d St t 127 University of Tennessee, Knoxville, Tennessee, United States 128 University of the Witwatersrand, Johannesburg, South Africa 128 University of the Witwatersrand, Johannesburg, South Africa versity of the Witwatersrand, Johannesburg, South A 129 University of Tokyo, Tokyo, Japan 130 University of Tsukuba, Tsukuba, Japan 131 Universit´e Clermont Auvergne, CNRS/IN2P3, LPC, Clermont-Ferrand, France 131 Universit´e Clermont Auvergne, CNRS/IN 131 Universit´e Clermont Auvergne, CNRS/IN2P3, LPC, Clermont-Ferrand, France 132 Universit´e de Lyon, Universit´e Lyon 1, CNRS/IN2P3, IPN-Lyon, Villeurbanne, Lyon, France 132 Universit´e de Lyon, Universit´e Lyon 1, CNRS/IN2P3, IPN-Lyon, Villeurbanne 133 Universit´e de Strasbourg, CNRS, IPHC UMR 7178, F-67000 Strasbourg, France, Strasbourg, France 133 Universit´e de Strasbourg, CNRS, IPHC UMR 7178, F-67000 Strasbourg, France, Strasbourg, Fran g, , , g, , g, 134 Universit´e Paris-Saclay Centre d’ ´Etudes de Saclay (CEA), IRFU, Department de Physique Nucl´eaire (DPhN), Saclay, France 134 Universit´e Paris-Saclay Centre d’ ´Etudes de Saclay (CEA), IRFU, Department de Physique Nucl´eaire (DPhN), Saclay, France 135 Universit`a degli Studi di Foggia, Foggia, Italy 135 Universit`a degli Studi di Foggia, Foggia, Italy 136 Universit`a degli Studi di Pavia, Pavia, Italy 137 Universit`a di Brescia, Brescia, Italy 138 V. Mathematik, Frankfurt, Germany Fock Institute for Physics, St. Petersburg State University, St. Petersburg, Russia 139 138 V. Fock Institute for Physics, St. Petersburg State University, St. Petersburg, Russia 139 Variable Energy Cyclotron Centre, Kolkata, India Warsaw University of Technology, Warsaw, Poland 140 Warsaw University of Technology, Warsaw, Poland 141 141 Wayne State University, Detroit, Michigan, United States 141 Wayne State University, Detroit, Michigan, United States Wayne State University, Detroit, Michigan, United S 142 Westf¨alische Wilhelms-Universit¨at M¨unster, Institut f¨ur Kernphysik, M¨unster, Germany 142 Westf¨alische Wilhelms-Universit¨at M¨unster, Institut f¨ur Ke 143 Wigner Research Centre for Physics, Hungarian Academy of Sciences, Budapest, Hungary 143 Wigner Research Centre for Physics, Hungarian Academy 144 Yale University, New Haven, Connecticut, United States 145 Yonsei University, Seoul, Republic of Korea – 45 –
https://openalex.org/W3185549813	https://www.frontiersin.org/articles/10.3389/fclim.2021.705229/pdf	English	null	Advancing Relevance, Credibility, Legitimacy, and Effectiveness as a Heuristic for Local-Parallel Scenarios	Frontiers in climate	2,021	cc-by	7,715	Advancing Relevance, Credibility, Legitimacy, and Effectiveness as a Heuristic for Local-Parallel Scenarios Nicholas A. Cradock-Henry 1* and Bob Frame 2 1 Landscape Governance & Policy, Manaaki Whenua Landcare Research, Lincoln, New Zealand, 2 Gateway Antarctica, University of Canterbury, Christchurch, New Zealand The parallel scenario process provides a framework for developing plausible scenarios of future conditions. Combining greenhouse gas emissions, social and economic trends, and policy responses, it enables researchers and policy makers to consider global-scale interactions, impacts and implications of climate change. Increasingly, researchers are developing extended scenarios, based on this framework, and incorporating them into adaptation planning and decision-making processes at the local level. To enable the identiﬁcation of possible impacts and assess vulnerability, these local-parallel scenarios must successfully accommodate diverse knowledge systems, multiple values, and competing priorities including both “top down” modeling and “bottom-up” participatory processes. They must link across scales, to account for the ways in which global changes affect and inﬂuence decision-making in local places. Due to the growing use of scenarios, there is value in assessing these developments using criteria or, more speciﬁcally, heuristics that may be implicitly acknowledged rather than formally monitored and evaluated. In this Perspective, we reﬂect on various contributions regarding the value of heuristics and propose the adoption of current deﬁnitions for Relevance, Credibility, and Legitimacy for guiding local scenario development as the most useful as well as using Effectiveness for evaluation purposes. We summarize the internal trade-offs (personal time, clarity-complexity, speed-quality, push-pull) and the external stressors (equity and the role of science in society) that inﬂuence the extent to which heuristics are used as “rules of thumb,” rather than formal assessment. These heuristics may help reﬁne the process of extending the parallel scenario framework to the local and enable cross-case comparisons. Edited by: Marina Baldissera Pacchetti, University of Leeds, United Kingdom Edited by: Marina Baldissera Pacchetti, University of Leeds, United Kingdom Reviewed by: Sam Grainger, Maynooth University, Ireland Hemen Mark Butu, Kyungpook National University, South Korea *Correspondence: Nicholas A. Cradock-Henry cradockhenryn@ landcareresearch.co.nz Specialty section: This article was submitted to Climate Risk Management, a section of the journal Frontiers in Climate Keywords: boundary work, climate change, cross-scale, integrated assessment models, local-parallel scenarios, multi-scale, research evaluation, SSPs Received: 04 May 2021 Accepted: 10 June 2021 Published: 02 July 2021 Keywords: boundary work, climate change, cross-scale, integrated assessment models, local-parallel scenarios, multi-scale, research evaluation, SSPs PERSPECTIVE published: 02 July 2021 doi: 10.3389/fclim.2021.705229 PERSPECTIVE Citation: In this, the three interrelated parallel pathways; the RCPs, SSPs and SPAs explore the impact climate change will have on social-ecological systems, the degree to which mitigation and adaptation policies can avoid and reduce those risks, and the costs and beneﬁts of various policy mixes (Ebi et al., 2014). An emerging trend in impacts, vulnerability, and adaptation research, therefore, is to improve the links between the global and sub-national (hereafter local) level by extending the parallel scenario framework and incorporating outputs into applied adaptation decision-making processes (Campos et al., 2016; Cradock-Henry et al., 2018, 2020; Aguiar et al., 2020; Schmitt Olabisi et al., 2020). Such extended SSPs have been developed for a range of settings and problems, including speciﬁc sectors and activities such as agriculture and forestry (Daigneault et al., 2019; Mitter et al., 2020; Lehtonen et al., 2021), and scales and places (Frame et al., 2018; Lino et al., 2019; Chen et al., 2020; Gomes et al., 2020; Pedde et al., 2021). The relevance, credibility, and legitimacy heuristic (hereafter RCL) has been associated with desired attributes for information at the boundary between science and policy communities. It has been used extensively in the literature on adaptation (and climate change more generally), due in part to its origins in assessing the usability of seasonal climate forecasts for decision-making (Cash et al., 2006). We use RCL as our anchor point from the literature (e.g., Cash et al., 2002, 2003, 2006; Sarkki et al., 2014; Belcher et al., 2016, 2019; Cash and Belloy, 2020). We ﬁnd this workable in practice at the local level, especially when there is a need for something that, while academically rigorous, can be easily understood by non-technical, on-the-ground practitioners. Or, to phrase it diﬀerently, we see the use of a heuristic to be of greater practical beneﬁt than a formal evaluation methodology (Nalau et al., 2021). However, as Elsawah and colleagues point out, while conceptual papers such as that by Cash et al. (2006) are often quoted, their “recommendations are rarely used beyond the point of acknowledging that they exist” (Elsawah et al., 2020, p. 13). Our aim here is to consider the various formulations of criteria and attributes in the literature and propose a simple, reproducible formulation that can be used across local case studies. Citation: The multi-scale and systemic nature of climate risk requires greater consideration of the ways in which responses to climate impacts and anticipated risks can be aﬀected and inﬂuenced by conditions at the global, regional, and national scales (Simpson et al., 2021). The parallel scenario framework is a sophisticated, global-scale architecture involving representative concentration pathways (RCPs) of greenhouse gas emissions, shared socio-economic pathways (SSPs), and shared policy assumptions (SPAs) (Ebi et al., 2014). Since 2014 the framework has been used to develop Cradock-Henry NA and Frame B (2021) Advancing Relevance, Credibility, Legitimacy, and Effectiveness as a Heuristic for Local-Parallel Scenarios. Front. Clim. 3:705229. doi: 10.3389/fclim.2021.705229 July 2021 \| Volume 3 \| Article 705229 Frontiers in Climate \| www.frontiersin.org Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame agreed practice regarding methods for downscaling the SSPs” and more detailed Integrated Assessment Models are needed (Pereira et al., 2021; Rosen, 2021). long-term futures, providing insight into the potential eﬀects of climate change on social-ecological systems, the eﬀectiveness of adaptation and mitigation, and the policies necessary to reduce climate-related risks (O’Neill et al., 2014). The framework provides a set of boundary conditions for constructing internally consistent, plausible representations of diverse futures. Elaborating futures scenarios enables researchers, policymakers, and practitioners to explore interactions and feedback mechanisms between large-scale drivers of global change, and to identify and assess possible pathways for change (O’Neill et al., 2020). The rapid growth in the application and development of decision-making tools and processes for adaptation is prompting reﬂection on the value of heuristics. These are seen as a “branch of study” that seeks to “understand the methods and rules of discovery and invention” (Pólya, 1990). Heuristics can expedite conceptual and methodological development by stimulating thinking. In this essay we use the word “heuristic” to refer to a rule of thumb. Following Starﬁeld et al. (1994), “a heuristic is a plausible or reasonable approach that has often proved to be useful.” In so doing, we build on and extend recent work in the ﬁeld, focusing on one of the most common heuristics in sustainability and climate science: relevance, credibility and legitimacy. This global scenario architecture provides a versatile and ﬂexible structure that can accommodate diverse applications, at diﬀerent scales, and provide insight into potential impacts and implications. Citation: Extending the basic architecture of the global parallel scenarios developed by the climate change research community to the local level also continues the trend in adaptation research, of researchers working with stakeholders—including communities and regions, policymakers and practitioners— to co-produce knowledge (Bremer and Meisch, 2017). Co- production processes seek to better understand local conditions, assess current and anticipated impacts and implications, and explore adaptation options (Ford et al., 2014; Boon et al., 2019; Cradock-Henry et al., 2020; Hill et al., 2020; Cradock-Henry, 2021). These local-parallel scenarios typically combine elements of top-down and bottom-up data derived from probabilistic or econometric models, or through interviews and other participatory methods, respectively. In this Perspective we reﬂect on developing and applying local-parallel scenarios as part of adaptation planning. Findings have been generated inductively based on our own experience, and deductively from a review of the literature. We begin by summarizing the process of nesting the local in the global (section Nesting the Local in the Global), before describing the various RCL formulations (section What Is Meant by Relevance, Credibility, and Legitimacy, and What Is Eﬀective?). In section Tradeoﬀs and Stressors in Developing Local-Parallel Scenarios we discuss how these criteria can accommodate internal and external stressors when working at the science– policy interface. We conclude by proposing how developing local-parallel scenarios might use these criteria most eﬀectively. Local scenarios can improve understanding of the types and magnitude of change, explore sensitivities, and evaluate ways of managing risks. Often these scenarios are used as part of an adaptive planning or pathways process that begins with a comprehensive understanding of the current situation, and then bounds future uncertainty within a manageable set of conditions (Cradock-Henry et al., 2018; Frame et al., 2018; Aguiar et al., 2020). However, development of these scenarios assumes seamless ways to coordinate and apply the frameworks from the global through the regional and national to the local while accommodating new directions. However, as O’Neill et al. (2020, p. 1,079) highlight, “At present, there is no commonly Frontiers in Climate \| www.frontiersin.org WHAT IS MEANT BY RELEVANCE, CREDIBILITY, AND LEGITIMACY, AND WHAT IS EFFECTIVE? Cash et al. (2002, 2003, 2006) established the key concepts of “credibility, salience and legitimacy” as attributes for information at the boundary between science and policy communities. These acknowledged the science and policy interface as a complex terrain requiring skilful navigation, the dynamics of which, within a rapidly changing world, are becoming increasingly challenging (Cash and Belloy, 2020). The terms evolved as a heuristic means to evaluate the boundary between research and policy without necessarily delving into the politics of the situation, or to challenge underlying assumptions (Preston et al., 2015; Nalau et al., 2021). Such research is eﬀectively transdisciplinary, where high-level modeling is likely to be of limited value and datasets are likely to be incomplete or inconsistent (Carlsen et al., 2016; Bosomworth and Gaillard, 2019; Cradock-Henry et al., 2020). The global-scale parallel scenario framework, however, is unable to model the localized eﬀects of climate change (Ebi et al., 2014; O’Neill et al., 2014). Precipitation, timing, and intensity of weather events, the role of local geography, or the speciﬁc socio-economic factors that aﬀect local decisions on adaptation and mitigation in regions and communities therefore need to be elaborated on and bounded in other ways (O’Neill et al., 2020; Pereira et al., 2021). At the regional or local level climate scenarios have tended to fall into two broad categories. The ﬁrst involves emulating the parallel process by collecting and reﬁning expert data and projections into relatively complex scenarios for speciﬁc regions. These scenarios are then used with planners, policy makers, and others to synthesize large amounts of scientiﬁc data, compare and contrast policy options, and inform decision-making. The second approach uses more community-development-type approaches by working with local communities to co-create scenarios that prioritize local knowledge and memories and community aspirations (Mistry et al., 2014). An emerging third way is the use of local socioeconomic and climate scenarios as a tool for exploring plausible future conditions and how these may inﬂuence adaptation strategies (Nilsson et al., 2017; Zandersen et al., 2019; Reimann et al., 2021). These local-parallel scenarios use the basic architecture of the global framework to provide a set of boundary conditions. The combination of emissions, policy mixes, and socioeconomic pathways is contextualized for local conditions through stakeholder knowledge and experience. The resulting narratives represent alternative trends, with a loose or soft linkage to national and/or global conditions (Lino et al., 2019). NESTING THE LOCAL IN THE GLOBAL There are many examples of adaptation planning and decision- making ranging from adaptation pathways to resilience and vulnerability assessment. Here we restrict ourselves to the July 2021 \| Volume 3 \| Article 705229 2 Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame growing use of scenarios at the local level, and the corresponding increase in case studies (Nilsson et al., 2017; Lino et al., 2019; Zandersen et al., 2019; Butler et al., 2020; Lehtonen et al., 2021). Scenarios are narratives describing plausible future worlds. They are a strategic planning method developed to make ﬂexible and robust long-term plans in response to complex and uncertain futures. Scenarios were initially developed by mid-nineteenth century European military intelligence specialists, but since the 1960s they have been used in a variety of contexts and scales, including business and trade (Berkhout et al., 2002), conservation and development (Peterson et al., 2003; Daconto and Sherpa, 2010; Pereira et al., 2021), community development (Rawluk and Godber, 2011), and adaptive infrastructure management (Hamilton et al., 2013). Due to the uncertainties surrounding the magnitude and eﬀects of climate change, natural variability, and the extent to which human societies will adopt mitigation and adaptation, scenarios are used extensively to explore the eﬀects of certain decisions on climate change. be eﬀective at the local scale, it has to successfully engage and negotiate with local concerns (Cradock-Henry et al., 2020; Cradock-Henry and Frame, 2021). Also, as discussed next, useful heuristics are needed that provide guidance on the extent to which this has been achieved. Frontiers in Climate \| www.frontiersin.org External Stressors Cash and Belloy (2020) describe four external stressors in the global context which address criticism that RCL does not fully address socio-political aspects, distributive justice, and the rapidly changing knowledge–action landscape. First there is the challenge of engaging with quite diﬀerent forms of knowledge when working across scales, and the need to ensure trust is created in the process. Second there are equity issues, which are both urgent and complex and include ethical dimensions, populations with existing vulnerabilities, issues of privilege, as well as historically disadvantaged populations. Third there is the degradation of the role of science in society and the trust placed in science in the “post-truth” world where trust in science has become corroded with an increasing emphasis on personal or political preferences. This is most clearly exhibited through perspectives such as climate change denialism and resistance to evidence-based responses to the COVID- 19 pandemic (Jasanoﬀ, 2021). Finally, there are issues related to the production of knowledge through digital technologies across multiple platforms, including ease of access to shared information, which blurs the citizen–science boundary, and the ways in which social media is used by institutions, community groups, and businesses to inﬂuence opinions. • Relevance: the importance, signiﬁcance, and usefulness of the research problem, objectives, processes, and ﬁndings to the problem context • Credibility: the research ﬁndings are robust and the sources of knowledge are dependable—this includes clear demonstration of the adequacy of the data and the methods used to procure the data, including clearly presented and logical interpretation of ﬁndings • Legitimacy: the research process is perceived as fair and ethical—this encompasses the ethical and fair representation of all involved, and the appropriate and genuine inclusion and consideration of diverse participants, values, interests, and perspectives. Belcher et al. (2016) also included the principle of eﬀectiveness. This is, in this case, an assessment criterion to be considered ex ante at the proposal stage, with actual eﬀectiveness determined ex post through the use of appropriate assessment tools. Thus, eﬀectiveness deﬁnes the extent to which research generates knowledge and stimulates actions that address the problem and contribute to solutions and innovations. As a result, Belcher et al. (2019) placed this eﬀectiveness heuristic outside the bounds of the adaptation processes, and we do not include eﬀectiveness to assess adaptation processes prior to their application. External Stressors While these originated through consideration of global developments, we seek here to consider them speciﬁcally in relation to local scenario processes. That is to say, we consider the use of the RCL heuristics primarily through the ﬁrst of these dynamics—working across scales. We then look at how this aﬀects issues of equity and science in a post-truth society and with digital transformations only playing a relatively minor role. We do so based on our experiences with impacts, vulnerability, and adaptation at the local level in Aotearoa—New Zealand (New Zealand) that will be highly contextual and place-speciﬁc, but also seek to provide a more general perspective (Cradock-Henry et al., 2018, 2020; Frame et al., 2018; Ausseil et al., 2019; Cradock-Henry and Frame, 2021). WHAT IS MEANT BY RELEVANCE, CREDIBILITY, AND LEGITIMACY, AND WHAT IS EFFECTIVE? In other words, “credibility, salience, and legitimacy” provide criteria that link the processes of developing climate change information and its usefulness within the transdisciplinary research world, and that other world experienced by end-users, including policymakers. While the heuristic has been widely used to describe the science–policy boundary (White et al., 2010; Kunseler et al., 2015; Dannevig and Hovelsrud, 2016; Cash and Belloy, 2020), various alternative formulations have been proposed. “Salient” was considered to be analogous to “relevant,” which led to use of the term CRELE (credible, relevant, and legitimate), which has morphed into relevant, credible, and legitimate, resulting in the acronym, RCL. This, as discussed by others, is what appears to be gaining traction and which we adopt as the preferred terminology (Belcher et al., 2016, 2019; Dunn and Laing, 2017). Other terms have also been proposed and include, for example, iterativity (Lemos and Morehouse, 2005; Dilling and Lemos, 2011), deﬁned as “a continuous multi-directional interaction that goes beyond simple repetition, building on previous practices, learning from success and failure, and fostering evaluation itself among all participants at the interface and between science-policy interfaces and external audiences” (Sarkki et al., 2015, p. 507) which led to CRELE + IT. While this is useful in terms of transdisciplinary research, it can, for use in local scenarios, be seen as an essential requirement absorbed into the overall concept of eﬀectiveness, as discussed later. Dunn and Laing (2017) suggested that CRELE is not suitable to describe policymakers’ needs because it is more focused on information supply rather than information demand. They recommended the use of applicability, comprehensiveness, timing, and accessibility. This, and to a lesser extent CRELE, was criticized by Tangney Developing and applying such scenarios involves developing quite speciﬁc artifacts, such as narratives or other representations of plausible future conditions. These, in turn, may challenge established norms and values, and cut across other place-speciﬁc issues. For local stakeholders at least, these have as much, if not more, importance in the short-term decision-making on such issues as infrastructure investments, the viability of primary production, and employment (Cradock-Henry et al., 2018). Consequently, for any scenario development process to July 2021 \| Volume 3 \| Article 705229 3 Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame • The clarity–complexity trade-oﬀbetween simple, strong, clear messages (relevance) vs. WHAT IS MEANT BY RELEVANCE, CREDIBILITY, AND LEGITIMACY, AND WHAT IS EFFECTIVE? thorough treatment of uncertainties and systemic dimensions (credibility and legitimacy) (2017) as ﬂawed, though with the latter acknowledging the criteria that should be prioritized for the use of evidence in decision-making (Hansson and Polk, 2018). • The clarity–complexity trade-oﬀbetween simple, strong, clear messages (relevance) vs. thorough treatment of uncertainties and systemic dimensions (credibility and legitimacy) To this lexicon, Maier et al. (2016) added terms to describe multiple plausible futures: deep uncertainty, global/local uncertainty and volatility, uncertainty, complexity, and ambiguity. Interestingly, with reference to the development and application of the parallel scenarios, O’Neill et al. (2020) use the term “credible, reproducible, and consistent methods for the use of the SSPs across scales” (p. 6), which perhaps lacks enough rigor. All of this suggests that a formal and broadly accepted formulation would be helpful as the process of extending scenarios gains momentum globally. • The speed–quality trade-oﬀ between timely and rapid responses to policy needs (relevance) vs. time-consuming quality assessment (credibility) and/or consensus building (legitimacy) To this lexicon, Maier et al. (2016) added terms to describe multiple plausible futures: deep uncertainty, global/local uncertainty and volatility, uncertainty, complexity, and ambiguity. Interestingly, with reference to the development and application of the parallel scenarios, O’Neill et al. (2020) use the term “credible, reproducible, and consistent methods for the use of the SSPs across scales” (p. 6), which perhaps lacks enough rigor. All of this suggests that a formal and broadly accepted formulation would be helpful as the process of extending scenarios gains momentum globally. • The speed–quality trade-oﬀ between timely and rapid responses to policy needs (relevance) vs. time-consuming quality assessment (credibility) and/or consensus building (legitimacy) • The push–pull trade-oﬀbetween following strong policy demand (relevance) and more supply-oriented research strategies to enable identiﬁcation of emerging issues or development of innovative solutions (credibility and legitimacy). Sarkki et al. (2014) also identiﬁed issues relating to trust and inclusion of other worldviews, which are considered as external stressors. The systematic review by Belcher et al. (2016), and its subsequent reﬁnement (Hansson and Polk, 2018), led to deﬁnitions for the principles and criteria for assessing the quality of transdisciplinary research. We propose the use of their working deﬁnitions for the development of extended parallel scenarios at the local level, as follows: Internal Tradeoffs Sarkki et al. (2014) used empirical data to identify and explore four internal trade-oﬀs at the science–policy interface, which they described as: • The personal time trade-oﬀbetween the commitment required by those involved to cover the highly complex, multi-faceted terrain of adaptation vs. commitment to an existing discipline or process Equity Cash and Belloy (2020) describe four aspects of inequity in contemporary society: historically disadvantaged populations; Frontiers in Climate \| www.frontiersin.org July 2021 \| Volume 3 \| Article 705229 4 Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame delivered by experts telling local people what they need to do. In other words, the role of the external knowledge provider is perceived as privileged, with the potential to dominate the local. While this has been a topic within the science and technology studies ﬁeld for several decades (e.g., Wynne, 1991, 1992), it is not always widely understood in practice by biophysical researchers or local authorities. Failure to accommodate this, or to have resources available to address this, could lead to early failures, which then place climate change as even less of a legitimate issue. This is exacerbated in the post-truth world, as highlighted through populist campaigns such as climate change denialism (Harvey et al., 2018; Bloomﬁeld and Tillery, 2019; Bowden et al., 2019; Kovaka, 2021) and resistance to immunization programmes for the COVID-19 virus (Jaiswal et al., 2020; Uscinski et al., 2020). those with existing vulnerabilities; those in post-disaster recovery; over-arching ethical considerations of human and non-human ecology. Of these, the historically disadvantaged and those with existing vulnerabilities are, currently, the most pertinent to our discussion. In the New Zealand context we propose that the ﬁrst relates solely to the indigenous population, with the latter covering other disadvantaged groups. Aotearoa—New Zealand is a bi-cultural nation, with indigenous Maori retaining governance rights and management responsibilities for ancestral land through Te Tiriti o Waitangi (the Maori version of the Treaty of Waitangi, 1840)—New Zealand’s founding document, framing the relationship between Maori and the Crown. In addition to their lands, many iwi (Maori tribes) have commercial agribusiness and forestry interests. They also live in or near coastal margins, which are likely to be exposed to the eﬀects of climate change. Ideally, scenarios would be developed using kaupapa-based M¯aori methodologies (Smith, 2012); i.e., designed by and for M¯aori, addressing M¯aori concerns, conducted predominantly by M¯aori researchers and based on M¯aori cultural values. In practice, available technical capability and capacity can, currently, be a limiting factor. To ensure a M¯aori perspective, any local consultation process must involve iwi or hapu as an equal partner. This cannot be considered an optional add-on but must be seen as a central component. Equity If this is not undertaken in a sincere and comprehensive manner, then success in achieving a legitimate result is greatly diminished. ) In the context of developing local scenarios through participatory and co-production processes, this can result in the wider project being perceived by some locals as an extension of “government,” with researchers being seen as “suits from the city.” This has led, in our experience, to being advised against mentioning speciﬁc topics (in the rural New Zealand context this includes the use of toxins for predator management and proposed bans on fossil fuel mining) that can be particularly contested and divisive. Until recently, work on adaptation in New Zealand—particularly with farming communities—was fraught, with stakeholders’ conﬂating mitigation and adaptation (Reisinger et al., 2011; Cradock-Henry, 2017) and opposition to taxes on carbon or greenhouse gas emissions (Cooper and Rosin, 2014). Attitudes, trust in science, and knowledge sharing have become politically polarized, with people rejecting scientiﬁc evidence that misaligns with their personal or political preferences, requiring greater sensitivity and diplomacy when managing participatory deliberation and analysis. It is essential therefore, to acknowledge and reﬂect on one’s role in the scenario process. Assuming the role of the “honest broker” (Pielke, 2007; Sarkki et al., 2020) in scenario development can help navigate the tension between scenario types: what could happen vs. what should happen (Börjeson et al., 2006). Our experience also suggests it is both critical and diﬃcult to negotiate and balance the power dynamics between ostensibly equal stakeholders (Cradock-Henry et al., in press). As with geopolitics, there are strong players and silent players. For example, there can be wealthy landowners with extensive business interests, positions in local-body politics, and strong, long-term family connections who may be able to exert a disproportionate inﬂuence on decision-making processes (sometimes over-stated as “oligarchs”). This authority needs to be made relatively transparent and the opportunity for alternative perspectives enabled. If “oligarchs” are potentially over-privileged, then there is an equal risk of under-privileging local, “silent,” voices. These are groups or individuals who are potentially less able to form coherent and consistent opinions than those endowed with politically relevant resources. Exclusion of silent voices reduces the relevance and legitimacy of local scenarios. Science in Society Expert-led scenario development tends to prioritize the knowledge and involvement of professionals over local communities and is usually expensive and resource intensive. However, experts can introduce information and identify opportunities not readily available to local communities. Nonetheless, scenarios can be perceived as derived from and Equity Consideration therefore also needs to be given to how participatory scenario processes unfold; for example, who participates, and to what extent local hierarchies of social diﬀerence (age, gender, class, ethnicity, etc.) shape how scenarios are facilitated, and the actions that emerge out of them (Stirling, 2008). Participatory scenario development processes are more likely to reﬂect local people’s experiences and aspirations. These should include important knowledge missed by experts, and include multiple stressors, and through this process build local agency to enact appropriate and socially acceptable solutions (Bohunovsky et al., 2011; Mistry et al., 2014; Wesche and Armitage, 2014; Flynn et al., 2018; Guaita García et al., 2020). However, they are often not spatially explicit, can miss important expert knowledge, are not robust enough and connected to national/global drivers, and can be resource intensive for participants (Guaita García et al., 2020). This tension between local and professional knowledges and experiences needs to be carefully facilitated. Frontiers in Climate \| www.frontiersin.org REFERENCES Boon, W. P. C., Hessels, L. K., and Horlings, E. (2019). Knowledge co-production in protective spaces: case studies of two climate adaptation projects. Reg. Environ. Change 19, 1935–1947. doi: 10.1007/s10113-019-01517-4 Aguiar, A. P. D., Collste, D., Harmáˇcková, Z. V., Pereira, L., Selomane, O., Galafassi, D., et al. (2020). Co-designing global target-seeking scenarios: a cross-scale participatory process for capturing multiple perspectives on pathways to sustainability. Glob. Environ. Change 65:102198. doi: 10.1016/j.gloenvcha.2020.102198 Börjeson, L., Höjer, M., Dreborg, K.-H., Ekvall, T., and Finnveden, G. (2006). Scenario types and techniques: towards a user’s guide. Futures 38, 723–739. doi: 10.1016/j.futures.2005.12.002 Bosomworth, K., and Gaillard, E. (2019). Engaging with uncertainty and ambiguity through participatory ‘Adaptive Pathways’ approaches: scoping the literature. Environ. Res. Lett. 14:093007. doi: 10.1088/1748-9326/ab3095 Ausseil, A. G. E., Daigneault, A. J., Frame, B., and Teixeira, E. I. (2019). Towards an integrated assessment of climate and socio-economic change impacts and implications in New Zealand. Environ. Model. Softw. 119, 1–20. doi: 10.1016/j.envsoft.2019.05.009 Bowden, V., Nyberg, D., and Wright, C. (2019). Planning for the past: local temporality and the construction of denial in climate change adaptation. Glob. Environ. Change 57:101939. doi: 10.1016/j.gloenvcha.2019.101939 Belcher, B. M., Claus, R., Davel, R., and Ramirez, L. F. (2019). Linking transdisciplinary research characteristics and quality to eﬀectiveness: a comparative analysis of ﬁve research-for-development projects. Environ. Sci. Policy 101, 192–203. doi: 10.1016/j.envsci.2019. 08.013 Bremer, S., and Meisch, S. (2017). Co-production in climate change research: reviewing diﬀerent perspectives. Wiley Interdiscipl. Rev. Clim. Change 8:e482. doi: 10.1002/wcc.482 Butler, J. R. A., Bergseng, A. M., Bohensky, E., Pedde, S., Aitkenhead, M., and Hamden, R. (2020). Adapting scenarios for climate adaptation: practitioners’ perspectives on a popular planning method. Environ. Sci. Policy 104, 13–19. doi: 10.1016/j.envsci.2019.10.014 Belcher, B. M., Rasmussen, K. E., Kemshaw, M. R., and Zornes, D. A. (2016). Deﬁning and assessing research quality in a transdisciplinary context. Res. Eval. 25, 1–17. doi: 10.1093/reseval/rvv025 Berkhout, F., Hertin, J., and Jordan, A. (2002). Socio-economic futures in climate change impact assessment: using scenarios as ‘learning machines.’ Glob. Environ. Change 12, 83–95. doi: 10.1016/S0959-3780(02) 00006-7 Campos, I., Vizinho, A., Coelho, C., Alves, F., Truninger, M., Pereira, C., et al. (2016). Participation, scenarios and pathways in long-term planning for climate change adaptation. Plan. Theor. Pract. 17, 537–556. doi: 10.1080/14649357.2016.1215511 Bloomﬁeld, E. F., and Tillery, D. (2019). The circulation of climate change denial online: rhetorical and networking strategies on Facebook. Environ. Commun. 13, 23–34. ACKNOWLEDGMENTS We acknowledge the support of colleagues in the SLMACC project Robust Responses. Thanks are also due to the Editor and Reviewers for their astute comments. CONCLUDING COMMENTS Climate change adaptation processes are one possible intervention in the complex assemblage of climate risk management. However, they are deeply enmeshed in wider July 2021 \| Volume 3 \| Article 705229 Frontiers in Climate \| www.frontiersin.org 5 Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame social contexts, including the exercise of authority and agency, irrespective of scale. The arguments about this at the global and national level are well-rehearsed (O’Neill et al., 2020; Rosen, 2021) and we propose, these are equally valid at the local level. social contexts, including the exercise of authority and agency, irrespective of scale. The arguments about this at the global and national level are well-rehearsed (O’Neill et al., 2020; Rosen, 2021) and we propose, these are equally valid at the local level. imply a casual approach. Only by consistently applying these heuristics can we counter Elsawah et al.’s critique (Elsawah et al., 2020, p. 13) that such concepts are developed but not used. DATA AVAILABILITY STATEMENT For the global parallel-scenario process to connect to local contexts in culturally meaningful ways, scenarios should be formally reviewed during their development to assess the extent to which they are relevant, credible, and legitimate for local stakeholders, and not just for researchers. They need to be made relevant by ensuring expert knowledge is contextualized for local concerns, conﬂicts, and aspirations. The processes cannot privilege the expert over the local and must provide credible and legitimate ways in which the future can be plausibly negotiated. We propose that the consistent use of heuristics, rather than formal deﬁnitions, for assessment criteria are important, and that the deﬁnitions produced by Belcher et al. (2016) and restated earlier provide a means of creating consistency between case studies. Inevitably these will be subject to review over time, as they have already. However, a consistency of approach in practical examples appears the most useful next step. The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s. FUNDING This research was funded by the Ministry for Primary Industries (NZ) through the Sustainable Land Management and Climate Change (SLMACC) programme and the Ministry of Business and Innovation—Resilience to Nature’s Challenges Kia Manawaroa— Ng¯a Akina o Te Ao Turoa National Science Challenge (Contract No. C05X1909). It is the local’s inherent complexity that makes the political so critical to an overarching understanding of adaptation options. Accommodation of this complexity must address not just the rigors of sound research but also the trade-oﬀs and stressors described. While absolute deﬁnitions for criteria have an obvious attractiveness, the dynamic around the terms used and the concerns their use seeks to address (see Cash and Belloy, 2020 for a deeper discussion) reinforce that these should be seen as heuristics which, as demonstrated so clearly by Nalau et al. (2021) must retain a high degree of ﬂexibility. That does not, however, AUTHOR CONTRIBUTIONS NC-H and BF: conceptualization, writing—reviewing, and editing. NC-H: funding acquisition. All authors contributed to the article and approved the submitted version. REFERENCES Applying scenario planning to park and tourism management in Sagarmatha National Park, Khumbu, Nepal. Mount. Res. Dev. 30, 103–112. doi: 10.1659/MRD-JOURNAL-D-09-00047.1 Lehtonen, H. S., Aakkula, J., Fronzek, S., Helin, J., Hildén, M., Huttunen, S., et al. (2021). Shared socioeconomic pathways for climate change research in Finland: co-developing extended SSP narratives for agriculture. Reg. Environ. Change 21:7. doi: 10.1007/s10113-020-01734-2 Daigneault, A., Johnston, C., Korosuo, A., Baker, J. S., Forsell, N., Prestemon, J. P., et al. (2019). Developing detailed Shared Socioeconomic Pathway (SSP) narratives for the Global Forest Sector. J. For. Econ. 34, 7–45. doi: 10.1561/112.00000441 Lemos, M. C., and Morehouse, B. J. (2005). The co-production of science and policy in integrated climate assessments. Glob. Environ. Change 15, 57–68. doi: 10.1016/j.gloenvcha.2004.09.004 Dannevig, H., and Hovelsrud, G. K. (2016). Understanding the need for adaptation in a natural resource dependent community in Northern Norway: issue salience, knowledge and values. Clim. Change 135, 261–275. doi: 10.1007/s10584-015-1557-1 Lino, J., Rohat, G., Kirshen, P., and Dao, H. (2019). Extending the shared socioeconomic pathways at the city scale to inform future vulnerability assessments — the case of Boston, Massachusetts. J. Extr. Even. 6:2050009. doi: 10.1142/S2345737620500098 Dilling, L., and Lemos, M. C. (2011). Creating usable science: opportunities and constraints for climate knowledge use and their implications for science policy. Glob. Environ. Change 21, 680–689. doi: 10.1016/j.gloenvcha.2010.11.006 Maier, H. R., Guillaume, J. H. A., van Delden, H., Riddell, G. A., Haasnoot, M., and Kwakkel, J. H. (2016). An uncertain future, deep uncertainty, scenarios, robustness and adaptation: how do they ﬁt together? Environ. Model. Softw. 81, 154–164. doi: 10.1016/j.envsoft.2016.03.014 Dunn, G., and Laing, M. (2017). Policy-makers perspectives on credibility, relevance and legitimacy (CRELE). Environ. Sci. Policy 76, 146–152. doi: 10.1016/j.envsci.2017.07.005 Ebi, K. L., Kram, T., van Vuuren, D. P., O’Neill, B. C., and Kriegler, E. (2014). A new toolkit for developing scenarios for climate change research and policy analysis. Environ. Sci. Policy Sustain. Dev. 56, 6–16. doi: 10.1080/00139157.2014.881692 Mistry, J., Tschirhart, C., Verwer, C., Glastra, R., Davis, O., Jaﬀerally, D., et al. (2014). Our common future? Cross-scalar scenario analysis for social– ecological sustainability of the Guiana Shield, South America. Environ. Sci. Policy 44, 126–148. doi: 10.1016/j.envsci.2014.05.007 Elsawah, S., Hamilton, S. H., Jakeman, A. J., Rothman, D., Schweizer, V., Trutnevyte, E., et al. (2020). Scenario processes for socio-environmental systems analysis of futures: a review of recent eﬀorts and a salient research agenda for supporting decision making. Sci. Total Environ. 729:138393. REFERENCES Decis. 33, 89–103. doi: 10.1007/s10669-012-9424-3 Cooper, M. H., and Rosin, C. (2014). Absolving the sins of emission: the politics of regulating agricultural greenhouse gas emissions in New Zealand. J. Rural Stud. 36, 391–400. doi: 10.1016/j.jrurstud.2014.06.008 Hansson, S., and Polk, M. (2018). Assessing the impact of transdisciplinary research: the usefulness of relevance, credibility, and legitimacy for understanding the link between process and impact. Res. Eval. 27, 132–144. doi: 10.1093/reseval/rvy004 Cradock-Henry, N. (2021). Linking the social, economic, and agroecological: a resilience framework for dairy farming. Ecol. Soc. 26:3. doi: 10.5751/ES-12122-260103 Cradock-Henry, N. A. (2017). New Zealand kiwifruit growers’ vulnerability to climate and other stressors. Reg. Environ. Change 17, 245–259. doi: 10.1007/s10113-016-1000-9 Harvey, J. A., van den Berg, D., Ellers, J., Kampen, R., Crowther, T. W., Roessingh, P., et al. (2018). Internet blogs, polar bears, and climate-change denial by proxy. Bioscience 68, 281–287. doi: 10.1093/biosci/bix133 Cradock-Henry, N. A., Blackett, P., Connolly, J., Frame, B., Teixera, E., Johnstone, P., et al. (in press). Adaptation pathways: guidelines and a worked example of a participatory planning process in the agricultural sector. Elementa. Hill, R., Walsh, F. J., Davies, J., Sparrow, A., Mooney, M., Wise, R. M., et al. (2020). Knowledge co-production for Indigenous adaptation pathways: transform post-colonial articulation complexes to empower local decision-making. Glob. Environ. Change 65:102161. doi: 10.1016/j.gloenvcha.2020.102161 Environ. Change 65:102161. doi: 10.1016/j.gloenvcha.2020.10216 Cradock-Henry, N. A., Blackett, P., Hall, M., Johnstone, P., Teixeira, E., and Wreford, A. (2020). Climate adaptation pathways for agriculture: insights from a participatory process. Environ. Sci. Policy 107, 66–79. doi: 10.1016/j.envsci.2020.02.020 Jaiswal, J., LoSchiavo, C., and Perlman, D. C. (2020). Disinformation, misinformation and inequality-driven mistrust in the time of COVID- 19: lessons unlearned from AIDS denialism. AIDS Behav. 24, 2776–2780. doi: 10.1007/s10461-020-02925-y Cradock-Henry, N. A., and Frame, B. (2021). Balancing scales: enhancing local applications of adaptation pathways. Environ. Sci. Policy 121, 42–48. doi: 10.1016/j.envsci.2021.04.001 Jasanoﬀ, S. (2021). Humility in the anthropocene. Globalizations 1–15. doi: 10.1080/14747731.2020.1859743 Cradock-Henry, N. A., Frame, B., Preston, B. L., Reisinger, A., and Rothman, D. S. (2018). Dynamic adaptive pathways in downscaled climate change scenarios. Clim. Change 150, 333–341. doi: 10.1007/s10584-018-2270-7 Kovaka, K. (2021). Climate change denial and beliefs about science. Synthese 198, 2355–2374. doi: 10.1007/s11229-019-02210-z Kunseler, E.-M., Tuinstra, W., Vasileiadou, E., and Petersen, A. C. (2015). The reﬂective futures practitioner: balancing salience, credibility and legitimacy in generating foresight knowledge with stakeholders. Futures 66, 1–12. doi: 10.1016/j.futures.2014.10.006 Daconto, G., and Sherpa, L. N. (2010). REFERENCES doi: 10.1080/17524032.2018.1527378 Carlsen, H., Lempert, R., Wikman-Svahn, P., and Schweizer, V. (2016). Choosing small sets of policy-relevant scenarios by combining vulnerability and diversity approaches. Environ. Model. Softw. 84, 155–164. doi: 10.1016/j.envsoft.2016.06.011 Bohunovsky, L., Jäger, J., and Omann, I. (2011). Participatory scenario development for integrated sustainability assessment. Reg. Environ. Change 11, 271–284. doi: 10.1007/s10113-010-0143-3 Cash, D., Clark, W. C., Alcock, F., Dickson, N., Eckley, N., and Jäger, J. (2002). Salience, Credibility, Legitimacy and Boundaries: Linking Research, July 2021 \| Volume 3 \| Article 705229 6 Frontiers in Climate \| www.frontiersin.org Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame research focusing on the built environment in northern Canada. Mitig. Adapt. Strateg. Glob. Change 20, 1267–1288. doi: 10.1007/s11027-014-9543-x Assessment and Decision Making. Cambridge, MA: John F. Kennedy School of Government, Harvard University. y Cash, D. W., and Belloy, P. G. (2020). Salience, credibility and legitimacy in a rapidly shifting world of knowledge and action. Sustainability 12:7376. doi: 10.3390/su12187376 Frame, B., Lawrence, J., Ausseil, A.-G., Reisinger, A., and Daigneault, A. (2018). Adapting global shared socio-economic pathways for national and local scenarios. Clim. Risk Manag. 21, 39–51. doi: 10.1016/j.crm.2018.05.001 Cash, D. W., Borck, J. C., and Patt, A. G. (2006). Countering the loading-dock approach to linking science and decision making comparative analysis of El Niño/Southern Oscillation (ENSO) forecasting systems. Sci. Technol. Hum. Val. 31, 465–494. doi: 10.1177/0162243906287547 Gomes, L. C., Bianchi, F. J. J. A., Cardoso, I. M., Schulte, R. P. O., Arts, B. J. M., and Fernandes Filho, E. I. (2020). Land use and land cover scenarios: an interdisciplinary approach integrating local conditions and the global shared socioeconomic pathways. Land Use Policy 97:104723. doi: 10.1016/j.landusepol.2020.104723 Cash, D. W., Clark, W. C., Alcock, F., Dickson, N. M., Eckley, N., Guston, D. H., et al. (2003). Knowledge systems for sustainable development. Proc. Natl. Acad. Sci. U.S.A. 100, 8086–8091. doi: 10.1073/pnas.1231332100 Guaita García, N., Martínez Fernández, J., and Fitz, C. (2020). Environmental scenario analysis on natural and social-ecological systems: a review of methods, approaches and applications. Sustainability 12:7542. doi: 10.3390/su12187542 Chen, H., Matsuhashi, K., Takahashi, K., Fujimori, S., Honjo, K., and Gomi, K. (2020). Adapting global shared socio-economic pathways for national scenarios in Japan. Sustain. Sci. 15, 985–1000. doi: 10.1007/s11625-019-00780-y Hamilton, M. C., Thekdi, S. A., Jenicek, E. M., Harmon, R. S., Goodsite, M. E., Case, M. P., et al. (2013). Case studies of scenario analysis for adaptive management of natural resource and infrastructure systems. Environ. Syst. REFERENCES Decolonizing Methodologies: Research and Indigenous Peoples, 2nd Edn. London: Zed Books. Starﬁeld, A. M., Smith, K. J., and Bleloch, A. L. (1994). How to Model It: Problem Solving for the Computer Age. Edina, MN: Burgess International. Pereira, L., Kuiper, J. J., Selomane, O., Aguiar, A. P. D., Asrar, G. R., Bennett, E. M., et al. (2021). Advancing a toolkit of diverse futures approaches for global environmental assessments. Ecosyst. People 17, 191–204. doi: 10.1080/26395916.2021.1901783 Stirling, A. (2008). “Opening up” and “closing down”: power, participation, and pluralism in the social appraisal of technology. Sci. Technol. Hum. Val. 33, 262–294. doi: 10.1177/0162243907311265 Tangney, P. (2017). What use is CRELE? A response to Dunn and Laing. Environ. Sci. Policy 77, 147–150. doi: 10.1016/j.envsci.2017.08.012 Peterson, G. D., Cumming, G. S., and Carpenter, S. R. (2003). Scenario planning: a tool for conservation in an uncertain world. Conserv. Biol. 17, 358–366. doi: 10.1046/j.1523-1739.2003.01491.x Uscinski, J. E., Enders, A. M., Klofstad, C. A., Seelig, M. I., Funchion, J. R., Everett, C., et al. (2020). Why do people believe COVID-19 conspiracy theories? Harvard Kennedy School (HKS) Misinf. Rev. doi: 10.37016/mr- 2020-015 Pielke, R. A. Jr. (2007). The Honest Broker: Making Sense of Science in Policy and Politics. Cambridge: Cambridge University Press. Wesche, S. D., and Armitage, D. R. (2014). Using qualitative scenarios to understand regional environmental change in the Canadian North. Reg. Environ. Change 14, 1095–1108. doi: 10.1007/s10113-013-0537-0 Pólya, G. (1990). How to Solve It: A New Aspect of Mathematical Method, 2nd Edn. London: Penguin Books. Preston, B. L., Mustelin, J., and Maloney, M. C. (2015). Climate adaptation heuristics and the science/policy divide. Mitig. Adapt. Strateg. Glob. Change 20, 467–497. doi: 10.1007/s11027-013-9503-x White, D. D., Wutich, A., Larson, K. L., Gober, P., Lant, T., and Senneville, C. (2010). Credibility, salience, and legitimacy of boundary objects: water managers’ assessment of a simulation model in an immersive decision theater. Sci. Public Policy 37, 219–232. doi: 10.3152/030234210X497726 Rawluk, A., and Godber, A. (2011). Widening the scope of scenario planning in small communities: a case study use of an alternative method. Ecol. Soc. 16:11. doi: 10.5751/ES-03852-160111 Sci. Public Policy 37, 219–232. doi: 10.3152/030234210X497726 Wynne, B. (1991). Knowledges in context. Sci. Technol. Hum. Val. 16, 111–121. doi: 10.1177/016224399101600108 Reimann, L., Vollstedt, B., Koerth, J., Tsakiris, M., Beer, M., and Vafeidis, A. T. (2021). Extending the Shared Socioeconomic Pathways (SSPs) to support local adaptation planning—a climate service for Flensburg, Germany. Futures 127:102691. REFERENCES doi: 10.1016/j.scitotenv.2020.138393 Mitter, H., Techen, A.-K., Sinabell, F., Helming, K., Schmid, E., Bodirsky, B. L., et al. (2020). Shared socio-economic pathways for european agriculture and food systems: the eur-agri-SSPs. Glob. Environ. Change 65:102159. doi: 10.1016/j.gloenvcha.2020.102159 j Flynn, M., Ford, J. D., Pearce, T., and Harper, S. L. (2018). Participatory scenario planning and climate change impacts, adaptation and vulnerability research in the Arctic. Environ. Sci. Policy 79, 45–53. doi: 10.1016/j.envsci.2017. 10.012 Nalau, J., Torabi, E., Edwards, N., Howes, M., and Morgan, E. (2021). A critical exploration of adaptation heuristics. Clim. Risk Manag. 32:100292. doi: 10.1016/j.crm.2021.100292 Ford, J. D., Champalle, C., Tudge, P., Riedlsperger, R., Bell, T., and Sparling, E. (2014). Evaluating climate change vulnerability assessments: a case study of Nilsson, A. E., Bay-Larsen, I., Carlsen, H., van Oort, B., Bjørkan, M., Jylhä, K., et al. (2017). Towards extended shared socioeconomic pathways: July 2021 \| Volume 3 \| Article 705229 Frontiers in Climate \| www.frontiersin.org 7 Advancing Heuristics for Local-Parallel Scenarios Cradock-Henry and Frame a combined participatory bottom-up and top-down methodology with results from the Barents region. Glob. Environ. Change 45, 124–132. doi: 10.1016/j.gloenvcha.2017.06.001 Sarkki, S., Tinch, R., Niemelä, J., Heink, U., Waylen, K., Timaeus, J., et al. (2015). Adding ‘iterativity’ to the credibility, relevance, legitimacy: a novel scheme to highlight dynamic aspects of science–policy interfaces. Environ. Sci. Policy 54, 505–512. doi: 10.1016/j.envsci.2015.02.016 O’Neill, B. C., Carter, T. R., Ebi, K., Harrison, P. A., Kemp-Benedict, E., Kok, K., et al. (2020). Achievements and needs for the climate change scenario framework. Nat. Clim. Change 10, 1074–1084. doi: 10.1038/s41558-020-00952-0 Schmitt Olabisi, L., Ugochukwu Onyeneke, R., Prince Choko, O., Nwawulu Chiemela, S., Liverpool-Tasie, L. S. O., Ifeyinwa Achike, A., et al. (2020). Scenario planning for climate adaptation in agricultural systems. Agriculture 10:274. doi: 10.3390/agriculture10070274 O’Neill, B. C., Kriegler, E., Riahi, K., Ebi, K. L., Hallegatte, S., Carter, T. R., et al. (2014). A new scenario framework for climate change research: the concept of shared socioeconomic pathways. Clim. Change 122:387–400. doi: 10.1007/s10584-013-0905-2 Simpson, N. P., Mach, K. J., Constable, A., Hess, J., Hogarth, R., Howden, M., et al. (2021). A framework for complex climate change risk assessment. One Earth 4, 489–501. doi: 10.1016/j.oneear.2021.03.005 Pedde, S., Harrison, P. A., Holman, I. P., Powney, G. D., Lofts, S., Schmucki, R., et al. (2021). Enriching the Shared Socioeconomic Pathways to co-create consistent multi-sector scenarios for the UK. Sci. Total Environ. 756:143172. doi: 10.1016/j.scitotenv.2020.143172 Smith, L. T. (2012). Frontiers in Climate \| www.frontiersin.org REFERENCES doi: 10.1016/j.futures.2020.102691 Wynne, B. (1992). Misunderstood misunderstanding: social identities and public uptake of science. Public Underst. Sci. 3, 281–304. doi: 10.1088/0963-6625/1/3/004 Zandersen, M., Hyytiäinen, K., Meier, H. E. M., Tomczak, M. T., Bauer, B., Haapasaari, P. E., et al. (2019). Shared socio-economic pathways extended for the Baltic Sea: exploring long-term environmental problems. Reg. Environ. Change 19, 1073–1086. doi: 10.1007/s10113-018-1453-0 Reisinger, A., Wratt, D., Allan, S., and Larsen, H. (2011). “The role of local government in adapting to climate change: lessons from New Zealand,” in Climate Change Adaptation in Developed Nations, Advances in Global Change Research, eds J. D. Ford and L. Berrang-Ford (Dordrecht: Springer Netherlands), 303–319. doi: 10.1007/978-94-007-0567-8_22 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Rosen, R. A. (2021). Why the shared socioeconomic pathway framework has not been useful for improving climate change mitigation policy analysis. Technol. Forecast. Soc. Change 166:120611. doi: 10.1016/j.techfore.2021.120611 Sarkki, S., Heikkinen, H. I., Komu, T., Partanen, M., Vanhanen, K., and Lépy, É. (2020). How boundary objects help to perform roles of science arbiter, honest broker, and issue advocate. Sci. Public Policy 47, 161–171. doi: 10.1093/scipol/scz055 Copyright © 2021 Cradock-Henry and Frame. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Sarkki, S., Niemelä, J., Tinch, R., van den Hove, S., Watt, A., and Young, J. (2014). Balancing credibility, relevance and legitimacy: a critical assessment of trade-oﬀs in science–policy interfaces. Sci. Public Policy 41, 194–206. doi: 10.1093/scipol/sct046 July 2021 \| Volume 3 \| Article 705229 Frontiers in Climate \| www.frontiersin.org 8
W4386045150.txt	https://zenodo.org/records/8272838/files/Science%20Of%20Rahu%20Kaal%20_%20Lunar%20Astro.pdf	en		Science Of Rahu Kaal	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	1,332	8/22/23, 4:29 PM Science Of Rahu Kaal \| Lunar Astro Science of Rahu Kaal 57 Comments / English Blogs / By Deepanshu Giri We all have heard this term from our elders to not perform anything during the Rahu kaal as every day for 90 minutes there is a Rahu kaal – A time period of 90 minutes where we should not buy new things or start anything new instead, we should focus on spiritual activities. As working with Panchang for the last several years and observing various patterns on a daily basis, which we also shared in our telegram group, Rahu kaal standing true to its name was always the mystery, but to solve any mystery of the universe, we need to experience it, We need to find ways to measure the subtle changes which are happening around us as this whole universe. Rahu Kaal is calculated by dividing the time of Sunrise- Sunset into eight different parts, and based on each day, this time period is given; based on local sunrise and sunset time, this time will differ, and it is best to check local panchang for the Rahu Kaalam. Monday: 7:30 a.m.–9:00 a.m. (2nd part) Tuesday: 3:00 p.m.–4:30 p.m. (7th part) https://lunarastro.org/science-of-rahu-kaal/ 1/20 8/22/23, 4:29 PM Science Of Rahu Kaal \| Lunar Astro Wednesday: 12:00 p.m.–1:30 p.m. (5th part) Thursday: 1:30 p.m.–3:00 p.m. (6th part) Friday: 10:30 a.m.–12:00 p.m. (4th part) Saturday: 9:00 a.m.–10:30 a.m. (3rd part) Sunday: 4:30 p.m.–6:00 p.m. (8th part) My question to myself was, Why is this time period given to Rahu as any time allocated to Rahu in the panchang is when there is a void of energy? You are unsure of anything. This is the Zero time period of the universe, such as Ashtami and Amavas tithi in panchang belongs to Rahu, and this is the tithi of indecisiveness, A tithi of a black hole. Still, these tithis are the best remedies for life’s chronic problems. To understand this, we need first to understand why there are only seven days a week, not 5 or 10- as if it would come to me, I would choose to have four days in a week or, like Romans, had eight days in a week, or Egyptians had ten days in their weekly calendar but since Hindu calendar was based on the movement of planets and all rituals are designed around this calendar so the humans can uplift themselves and gather energy from space using the movement of planets, As human DNA corresponds to different planets in different months and energy of planets and this is what our ancient rishis called as Graha Shanti. This whole world works on a rhythm from human life to the Stock market. This rhythm is superior to any other cycle; as https://lunarastro.org/science-of-rahu-kaal/ 2/20 8/22/23, 4:29 PM Science Of Rahu Kaal \| Lunar Astro on Twitter till now, without any technical parameter and against all odds, only on the basis of this cycle, I was able to point out 15 stocks and from the last three years, over ten multi-baggers. Still, we can use the same cycle and use it for the whole universe. Coming back to the number of days in a week confining to seven, given we have seven planets in this universe and each day is named after them starting from Sunday. Moreover, the 28-day moon cycle is divided into four parts- New Moon- Full Moon- Half Moon – So the Moon will be in Kendra precisely 90 degrees from the New Moon origin after seven days. The day New Moon is rising and after every seven days, the same weekday will be repeated and based on this, you can make predictions. Such as, any month which has 5 Saturdays will be extremely tough, and oil/steel/machinery prices will rise in this month; a month with five Mondays will bring emotional turmoil but will be suitable for FMCG sectors and Car Sales, but every day when Sunrise- we calculate the tithi based on the distance between Sun and Moon and this Tithi is the emotion and relation between not only Sun and Moon but also between – Atman and Mann. Atman is someone who is deeply buried inside us but cannot communicate until and unless we go into deep isolation and listen to our own sound- All the Beej Mantras are nothing but the sounds of Atman, and we chant them to get into resonance with our own selves, but since we are busy in the outer world, we use our shell which helps to communicate which is Maan- the mirror image of Atman. Most people in this world operate in the reverse direction as they communicate from Mann to Atman as I need Car, House, and Job as these all are liked by Mann and communicated to Atman with a belief system that this can make me happy while the happiest people on earth are those who communicate from inside to outside clearly telling the world- This is what I really require to be happy as this is my soul desire. The Book on Atmakarka is the sole attempt, so anyone who reads it can understand the deeper meaning of their soul desire and be able to predict patterns and life choices based on Atmakarka as every work we do in this life and which becomes legendary is only due to one reason that we believed in it so much as it came out of our deep-rooted knowledge of Atman, which we were carrying from several lifetimes. Coming back to Tithi – which governs our emotions, our feelings and the way we communicate from Atman to Mann- what message are we carrying, how are we feeling in a day? What our Atman is trying to communicate as if you see Rahu Kaal- it is working in alternate groups and pair. Morning Group Evening Group Moon Sun https://lunarastro.org/science-of-rahu-kaal/ 3/20 8/22/23, 4:29 PM Science Of Rahu Kaal \| Lunar Astro Saturday Tuesday Wednesday Friday Thursday Every day, based on which planet is ruling the day, there is a solar energy flare disturbance, and this flare disturbance happens, which blocks the communication of the Sun and Moon; it is a grahan happening at a minor level of 22.5 degrees. – This 90 mins is completely scientific as even the ISSInternational space station, which is a man-made satellite is rotating at a speed of 90 mins/rotation as this is not by choice but by the necessity Similarly, Rahu/Ketu, which are satellites of the Moon, blocks energy to Moon for 90 mins creating darkness in the absence of communication. One thing you will notice during the dasa of Rahu or people who have prominent Rahu in their chart is the way these people speak; there is a lot of air movement and clarity in their speech, just like the language of snakes, a lot of hissing sounds, cleaning saliva which is coming out of the mouth from both ends as there is an absence of Atman, darkness comes in the head, and this can lead to all the wrong decisions, but this is also the time period which gives you time to connect to your own self and the realm of Rahu. This is the best time to perform any remedies related to materialistic things. Solving problems of life, removing any chronic issues in life, and making offerings to ancestors this is the best time to it; one of the remedies you can try today is to offer Milk and honey for seven days regularly to the bottom of any tree. You can see the results of which will be your life will start to change for the better as suddenly events will happen which will be surprising for you. ← Previous Post 57 thoughts on “Science of Rahu Kaal” SUMANGLA VEGAD 20 AUGUST 2023 AT 08:49 Wow, Sir this is an amazing article with a complete details about Rahu Kaal and also thanks for sharing the remedy too 😊🙏 Reply https://lunarastro.org/science-of-rahu-kaal/ 4/20
https://openalex.org/W2773775165	https://europepmc.org/articles/pmc5818128?pdf=render	English	null	Foods, Dietary Patterns and Occupational Class and Leukocyte Telomere Length in the Male Population	American journal of men's health	2,017	cc-by	10,719	Article Article https://doi.org/10.1177/1557988317743385 American Journal of Men’s Health 2018, Vol. 12(2) 479–492 © The Author(s) 2017 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1557988317743385 journals.sagepub.com/home/ajmh Keywords Keywords foods groups, dietary patterns, occupational class, serum lipids, telomere length Received May 23, 2017; revised September 17, 2017; accepted September 25, 2017 Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us sagepub com/en-us/nam/open-access-at-sage) 1Department of Environmental Health Engineering, School of Public Health, Arak University of Medical Sciences, Arak, Iran 2Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran 3Department of Research Methodology and Data Analysis, Institute for Environmental Research (IER), Tehran University of Medical Sciences, Tehran, Iran 4Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran 5Center for Air Pollution Research (CAPR), Institute for Environmental Research (IER), Tehran University of Medical Sciences, Tehran, Iran 6Blood Transfusion Research Center, High Institute for Research and Education in Transfusion Medicine, Tehran, Iran Corresponding Authors: Masud Yunesian, Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences, Enghelab St., Tehran, Iran. Email: yunesian@tums.ac.ir Parvin Mehdipour, Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Enghelab St., Tehran, Iran. Email: mehdiporar@tums.ac.ir 1Department of Environmental Health Engineering, School of Public Health, Arak University of Medical Sciences, Arak, Iran 2Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran 3Department of Research Methodology and Data Analysis, Institute for Environmental Research (IER), Tehran University of Medical Sciences, Tehran, Iran 4Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran Abstract Telomeres contain TTAGGG repetitive sequences and are located at the end of human chromosomes. Telomere dysfunction is associated with some age-related and chronic diseases, but its relationship with foods, dietary patterns, and occupational class in the young male population is not yet known. In this cross-sectional study, 300 healthy men, residents of Tehran, aged 25–40 years were enrolled from January to December 2016. We employed a cross-sectional study of 300 healthy people, residents of Tehran, aged 25-40 years. A food frequency questionnaire was used to obtain food intakes of all participants and converted into actual food intake (g/day). The principal components analysis was used to determine dietary patterns and other demographic characteristics. Leukocyte telomere length (TL) was measured by quantitative real-time polymerase chain reaction (PCR) to measure number of telomere repeat copy number (T) to the relative number of 36B4 copies (S) (T/S ratio). T/S in office-workers, waste recyclers, and other workers were 1.22 ± 0.4, 1.08 ± 0.3, and 1.094 ± 0.34, respectively. The results of multivariate linear regression adjusted for age, body mass index (BMI), and smoking were showed that whole grains (β = 0.02; p = .05), refined grains, fruits and vegetables, fish and dairy products were associated with an increase in log-T/S, but consumption of nuts and seeds (β = −0.00072; p = .06), meats (β = −0.00043; p = .9), produced meats (β = −0.00238; p = .03), oils and solid fats (β = −0.00146; p = .03) had a negative relationship with log-T/S in all studied occupational classes. A positive relationship was reported between the healthy (β = 0.017; p = .2) and traditional dietary pattern (β = 0.012; p = .4) with log-T/S, but western pattern identified negative relationship (β = −0.004; p = .7). Adherence to a healthy (with consumption whole grains, refined grains, dairy, and cereals) and then traditional pattern with increased consumption of fruits, vegetables and whole grains, fish and dairy products are necessary to prevent TL destruction in all studied occupational classes. Foods, Dietary Patterns and Occupational Class and Leukocyte Telomere Length in the Male Population Behrooz Karimi1,2, Ramin Nabizadeh2, Masud Yunesian2,3, Parvin Mehdipour4, Noushin Rastkari5, and Afsaneh Aghaie6 Corresponding Authors: Email: yunesian@tums.ac.ir Parvin Mehdipour, Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Enghelab St., Tehran, Iran. Email: mehdiporar@tums.ac.ir Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). American Journal of Men’s Health 12(2) 480 Telomeres are complex nuclei-proteins located at the ends of eukaryotic chromosome, formed by TTAGGG repeti- tive sequence (Hoxha et al., 2009). Telomere length (TL) is considered as a reliable biomarker of biological aging and age-related chronic diseases (Harte et al., 2012). The loop structures of telomeres are to maintain cellular stabil- ity and genomic integrity by preventing irregular recombi- nation of chromosome, DNA degradation by exonuclease, end-to-end fusion, and loss of specific genes due to DNA replication repeat (Karimi, Yunesian, Nabizadeh, Mehdipour, & Aghaie, 2016). Leukocyte telomere length (TL) in the human body is dynamic and typically ranges between 10 and 15 kb (Blackburn, Greider, & Szostak, 2006) and in each cell division cycle about 50 to 200 bp of the TL is progressively shortened, due to the lack of telo- meres elongation mechanisms and ends of chromosome replication problem (Karimi et al., 2016). Short TL in leu- kocyte indicates the risk of cardiovascular and respiratory diseases (Brouilette et al., 2007). Oxidative stress, inflam- mation, and aging are reported to be associated with TL shortening (Epel et al., 2004). In this condition, cell divi- sion capacity is highly increased and TL becomes too short in several stages of cell proliferation. Cells with shorter TL lost the ability of the division and aging, thereby leading to apoptosis and preventing replication of cells (Hohensinner, Goronzy, & Weyand, 2014; Joshu et al., 2015). effect of the job on oxidative stress and TL, this study investigates the effect of food intake, dietary patterns, and serum lipids on TL in three occupational classes in a pop- ulation-based study. Waste recycling with excessive exposure to a variety of pollutants and higher levels of oxidative markers in the blood resulted in the comparison of the effect of dietary patterns in this job with office-workers that are less exposed to pollutants (Ni, Huang, Wang, Zhang, & Wu, 2014). Questionnaire Study In this study, 300 healthy men of 25 to 40 years old were selected from three occupational classes according to International Standard Classification of Occupations (ISCO) with minor modifications as follows (Porta et al., 2008): (a) managers, professionals, employees, students, elementary occupations such as housework, self-employment, and so on (as office-workers); (b) solid waste-recyclers; and (c) other occupational groups such as service workers, industrial work- ers, agricultural and fishery workers. The lifestyle factors such as obesity, BMI, smoking, and higher blood and cellular lipids may be associated with both telomere biology and oxidative stress (Buxton et al., 2011; García-Calzón et al., 2014). The associations between foods, dietary patterns and leukocyte TL in pre- vious epidemiological studies have been conflicted (Cassidy et al., 2010; Crous-Bou et al., 2014; Nettleton, Diez-Roux, Jenny, Fitzpatrick, & Jacobs, 2008; Tiainen et al., 2012). Observation and intervention studies identi- fied that adherence to a healthy and Mediterranean dietary patterns has numerous advantages such as reduced chronic diseases and increased chances of health in old age and is associated with longer TL (Strandberg et al., 2011), but other studies did not report this relationship between TL and food groups (Crous-Bou et al., 2014; Gu et al., 2015). Some food components have antioxidant and anti-inflammatory properties, thus, preventing some cancer effects (Shen et al., 2009). Intake of foods contain- ing antioxidants was associated with longer TL (García- Calzón, Moleres, et al., 2015). The questionnaire is made up of two parts. The first part is about the demographic information with focus on the residence, workplace, and other confounding factors that might affect TL, such as smoking. The second part is related to the food intake and dietary pattern by food fre- quency questionnaire (FFQ). The participants’ demo- graphic, age, length of stay in the city, current address, education, previous illnesses, occupation and occupational exposure, and pesticide mixture exposure were asked, and BMI through the measurement of weight and height (kg/ m2) was calculated and categorized in three groups: 25 (normal), 25 to 30 (overweight), and ≥30 kg/m2 (obese) (World Health Organization, 2013). Smoking status was classified into four groups: never-smokers, second-hand smokers, former smokers, and current smokers. The cumu- lative dose of smoking was computed by the following equation: pack-years = smoked years × packs per day. Corresponding Authors: Therefore the aim of this study was to investigate the effect of food intake and dietary pattern and other confounding factors which affect TL in three classes of occupations (waste-recyclers, office-workers, and other workers) in the urban areas of Tehran. Dietary Analysis Concerning conflicting reports on the effect of the food groups on TL in different communities (Cassidy et al., 2010; Crous-Bou et al., 2014; Nettleton et al., 2008; Tiainen et al., 2012), a part of this conflict can be related to occupational type (Shiels et al., 2011). The type of occupation is one of the major risk factors associated with age-related diseases (Shiels et al., 2011). Regarding the The semiquantitative FFQ designed in the Tehran Lipid and Glucose Cohort Study was employed for assessing the dietary patterns and usual dietary intakes during the past 12 months (Hosseini Esfahani, Asghari, Mirmiran, & Azizi, 2010). The FFQ comprised 148 food items in addition to a standard size of each nutrient and its validity and reliability 481 Karimi et al. the same conditions were included for primer-dimer study and correct gene amplification process. Melting curve analysis was used for evaluating the property and verifying the specificity of each run. The standard curve to evaluate the PCR efficiency in each run was utilized by serially diluting one reference DNA sample with deion- ized water to make six concentrations of DNA ranging from 1.56 to 50 ng/ml. The TL reference value was deter- mined by mixing DNA of 10 randomly selected DNA samples. The TL for each sample was estimated by deter- mining the ratio of the number of telomere repeat copy number (T) to the relative number of 36B4 copies (S) with respect to the same reference DNA sample. The results were expressed in terms of T/S ratios. have been confirmed in 13 districts of Tehran (Hosseini Esfahani et al., 2010). The consumption frequency of each food was questioned from individuals in the last year. Frequency of consumption of any food in terms of fre- quency of usage per day, week, or month was questioned. The values stated for each food were converted into actual food intake (g/day) using the guide scales. Blood Samples and DNA Extraction After completing the questionnaires, 10 cc blood samples were taken. Quickly, the serum content of about half of the blood samples was separated by centrifuge (5 min at 4000 rpm), kept in glass vials, and maintained in −20°C before testing. Whole blood samples were collected in ethylenedi- aminetetraacetic acid (EDTA) tubes and stored at −20°C before use. Genomic DNA was extracted from 2 ml of each blood sample by salting-out method (MWer, Dykes, & Polesky, 1988). The concentration of extracting DNA and its quantity and purity was assayed using Nanodrop (Thermo Scientific, Wilmington, DE, USA) and by consid- ering the A260/A280 ratio. All extracted DNA was stored at −70°C until use. Statistical Analyses The χ2 test was applied to compare the frequency of qualita- tive independent variables (smoking, education, and years of occupation) in all occupational classes. Furthermore, Analysis of variance (ANOVA) was used for quantitative variables’ [height, weight, BMI, age, residence in Tehran, place of residence (years), mean number of pack-years, occupational class and concentration of TC, TG, PL, and total serum lipids (TSL)] to examine the mean difference between the tertiles of T/S in three occupational groups. Lipid Analysis Serum samples were assessed on the first day of admis- sion of serum lipid levels. Total cholesterol (TC) and tri- glycerides (TG) were measured by the kits purchased from the PARS-Azmoon Company. Per the study of Bernert et al. 2007 and the existence of a linear relation- ship between the phospholipids and TC, phospholipids (PL) concentrations were obtained from PL = (0.766 × TC) + 62.3 mg/dl (Bernert, Turner, Patterson, & Needham, 2007). Finally, the total serum lipids (TSL) were esti- mated by the following equation (Bernert et al., 2007): TSL = (1.494 × TC ) + TG + PL mg/dl. The dietary pattern was determined by factor analysis and then dietary patterns and other effecting factors will be selected by principal components analysis (PCA) with varimax rotation. Afterwards, resulting factors were assessed based on Eigenvalues ≥1 and Scree-plot graph. By observing the factor loading of food items and based on previous studies, four factors extracted were named as healthy, western, traditional, and vegetarian dietary pat- terns, respectively. The factor loadings of each food pat- terns were reported in Supplementary Table 1. Multiple linear regression (MLR) model considering T/S as the response variable was used to investigate the relationship between dependent variables and T/S. Per Shapiro–Wilk test, T/S was not normally distributed (p = .00), as a result log-T/S was applied. Two unadjusted and adjusted models were examined based on age, BMI, and smoking status (data of unadjusted model not reported). MLR coefficients were reported for the three job catego- ries (office-workers, waste-recyclers, and other occupa- tional class) based on serum lipids, food group intake, dietary patterns, and other confounding factors. The mean ± SD, 95% confidence interval (95% CI), and median of predicted T/S from MLR and generalized additive model (GAM) were reported. All statistical tests were two-sided, and p-values less than .05 were statistically significant. All statistical analyses were carried out using statistical software R 3.2.2 and SPSS 22. Relative Telomere Length The relative TL was determined with a real-time PCR method-based telomere assay previously described by Cawthon (2002). In summary, real-time PCR was done using SYBR Premix Ex TaqTM kit (Takara) and the primer concentrations for the telomere were 270 nM Tel1 [5’-CGG TTT(GTTTGG)5GTT-3’] and 900 nM Tel2 [5’- GGC TTG(CCTTAC)5CCT-3’] and for single-copy gene primers, 300 nM for 36B4u [5’-CCCATTCTATCATCA ACGGGTACAA-3’] and 500 nM for 36B4d [5’-CAG CAAGTGGGAAGGTGTAATCC-3’] were used. The reaction was performed three times in duplicate wells for each sample using 25 ng/ml of DNA. In each run, three no-template controls alongside other samples and under Table 1. Characteristics of the Study Participants by Tertiles of T/S. Relative Telomere Length Short, mean ± SD Middle, mean ± SD Longest, mean ± SD pa Work Office- workers Waste recycling Other workers Office- workers Waste recycling Other workers Office- workers Waste recycling Other workers T/S 0.75 ± 0.2 0.77 ± 0.2 0.78 ± 0.15 1.13 ± 0.12 1.11 ± 0.1 1.1 ± 0.11 1.67 ± 0.28 1.65 ± 0.3 1.61 ± 0.3 .007 Height (cm) 180.66 ± 6.24 180.97 ± 7.4 178 ± 8.6 175.08 ± 8 173.6 ± 8.86 174.92 ± 8.5 175.45 ± 9.6 175.5 ± 13.2 170 ± 1 .49 Weight (kg) 83.14 ± 17.5 83.63 ± 12.4 80.45 ± 14.5 78.38 ± 8.6 76.11 ± 9.62 79.32 ± 12.4 80.7 ± 14.6 70.57 ± 13.3 73.16 ± 1 .35 BMI (kg/m2) 25.32 ± 4.6 25.73 ± 4.8 25.74 ± 6 25.69 ± 3.36 25.51 ± 4.56 26.11 ± 4.9 26.17 ± 4.16 22.93 ± 3.46 25.53 ± 5 .54 Age (years) 33.45 ± 5.18 35.06 ± 3.6 35.75 ± 4.2 33.29 ± 4.4 32.81 ± 3.46 34.27 ± 3.8 31.33 ± 3.96 31.5 ± 3.7 32.1 ± 4.8 .02 Residence (years) 18.2 ± 9.1 14.69 ± 9.3 14.59 ± 8 17.83 ± 10.5 10.19 ± 5.38 11.88 ± 7.4 17.17 ± 13.2 6.13 ± 5.4 16.3 ± 12.57 .00 Live in place (years) 12.45 ± 9.25 9.72 ± 6.8 10.97 ± 6.6 13.12 ± 1 6.37 ± 3.26 8.43 ± 5.2 11.1 ± 11.62 5.13 ± 4.2 10.05 ± 11 .00 Mean number of pack-years 23 ± 15.47 40.6 ± 14.3 38.75 ± 13 55.83 ± 5.5 38.15 ± 9.81 37.79 ± 18.9 86.18 ± 12 29.14 ± 7.3 57.5 ± 37.46 .07 TC (mg/dl) 205.05 ± 25.78 203.12 ± 26 200.83 ± 25.7 176.98 ± 29.1 185.28 ± 25.47 186.6 ± 23.1 162.7 ± 22.62 158.57 ± 20.8 170.5 ± 26.2 .04 TG (mg/dl) 172.58 ± 40.95 161.12 ± 30 160.21 ± 32.6 145.05 ± 26.7 145.78 ± 31.52 150 ± 23.3 128.2 ± 47.7 107.68 ± 32.5 124.4 ± 45.4 .79 PL (mg/dl) 216 ± 22.5 216.2 ± 20.6 204.9 ± 21.6 200.56 ± 20.3 208.39 ± 18.78 206.1 ± 17 188.06 ± 2 182.15 ± 14.7 194.8 ± 22.6 .06 TSL (mg/dl) 694.91 ± 79.35 680.77 ± 65.8 665.16 ± 72 610.01 ± 52.6 625.86 ± 74.4 628.54 ± 65 559.4 ± 76.7 526.74 ± 54.7 574 ± 82.5 .24 Note. Relative Telomere Length BMI = body mass index; TC = total cholesterol; TG = triglycerides; PL = phospholipids, TSL = total serum lipids. a l f ANOVA b ffi k li d h k 482 Karimi et al. 483 Figure 1. (a) Spearman correlation test, (b) telomere length with age, (c) quartiles of TSL, and (d) occupational class. TSL = total serum lipids. Figure 1. (a) Spearman correlation test, (b) telomere length with age, (c) quartiles of TSL, and (d) occupational class. TSL = total serum lipids. office-workers, waste-recyclers, and other workers were 612 ± 83, 628 ± 84.5, and 629 ± 77 mg/dl, respectively (Table 1). With increasing TC, telomere length was sig- nificantly reduced in all occupational classes (p = .04). Spearman correlation analysis revealed an inverse rela- tionship between age (r = −.26), weight (r = −.22), BMI (r = −.03), years of residence (r =−.18), TSL concentra- tion (r = −.57), and T/S in all occupational class. Shortening of T/S with age, occupational class, and quar- tile of TSL is reported in Figure 1a. With increase in TC, TG, PL, and TSL, telomere length was significantly reduced in all occupational classes (p = .00). Spearman correlation analysis reported an inverse relationship between age (r = −.26), weight (r =−.22), BMI (r = −.03), years of residence(r = −.18), TSL concentration (r = −.57), and T/S in all occupational classes. Shortening of T/S with age (Figure 1b), occupational class (Figure 1c), and quartile of TSL is presented in Figure 1d. Results The average T/S ratio in the study population was 1.13 ± 0.36. Figure 1 reports the average T/S ratio of three occu- pational classes (p = .007). Longer T/S reports more desirable performance status of the body’s cells. The average age of office-workers, waste-recyclers, and other workers were 32.6 ± 4.5, 33.35 ± 3.7, and 34.3 ± 4.3 years, respectively. Increase in age was associated with shortening of T/S during blood sampling (p = .02) (Table 1). Approximately 17–24% of the total studied popula- tion were smokers in three occupational classes. But no statistical difference was reported between the three groups in terms of smoking (Table 2). The average T/S of smokers in all groups was 1.16 ± 0.39, that in waste-recyclers, office-workers, and other workers were 1.26 ± 0.34, 1.11 ± 0.42, and 1.1± 0.41, respectively. T/S ratio in never-smokers in all groups was reached (1.09 ± 0.34). T/S ratio of never-smokers in waste-recyclers, office-workers, and other workers were 1.1769 ± 0.42, 1.016 ± 0.24, and 1.08± 0.32, respectively. The average concentrations of total serum lipids (TSL) in Per PCA and Scree-plot graph, four dominant dietary patterns of the study were determined (Figure 2a and b). The four patterns stated above explained a total of 64.5% of total variance and the first dietary pattern had the American Journal of Men’s Health 12(2) 484 Table 2. Characteristics of the Qualitative Variables of Participants. Work, n (%) pa Total Office-workers Waste recycling Others Smoking status (%) Never smoke 42 (46.7) 43 (43) 35 (40.7) − 120 (43.5) Second-hand smokers 15 (16.7) 17 (17) 20 (23.3) 52 (18.8) Previously smoking 15 (16.7) 16 (16) 16 (18.6) 47 (17) Smokers 18 (20) 24 (24) 15 (17.4) 57 (20.7) Total 90 (100) 100 (100) 86 (100) .83 276 (100) Education (%) − Under diploma 4 (4.1) 21 (20.8) 13 (13.7) 38 (13) Diploma 27 (27.8) 44 (43.6) 36 (37.9) 107 (36.5) Graduate 33 (34) 27 (26.7) 38 (40) 98 (33.4) Postgraduate 33 (34) 9 (8.9) 8 (8.4) 50 (17.1) Total 97 (100) 101 (100) 95 (100) .00 293 (100) Years of occupation − <2 16 (16.8) 9 (9) 12 (12.8) 37 (12.8) 2–5 36 (37.9) 61 (61) 32 (34) 129 (44.6) 6–10 22 (23.2) 27 (27) 35 (37.2) 84 (29.1) 11–20 21 (22.1) 3 (3) 15 (16) 40 (13.5) Total 95 (100) 100 (100) 94 (100) .00 289 (100) Note. Results a p-values from χ2 squared for categorical variables. Table 2. Characteristics of the Qualitative Variables of Participants. Note. a p-values from χ2 squared for categorical variables. Note. a p-values from χ2 squared for categorical variables. Figure 2. PCA-based analysis of data. (a) Scree-plot (b and c); PCA was applied to separation occupational class. Multiple linear regression (MLR) model in food groups identified that whole grains (p = .05), refined grains, fruits and vegetables, fish and dairy products were associated with positive correlation with longer T/S; con- sumption of other food compounds such as meat, pro- duced meats, oils, and fats had a negative relationship with T/S (Table 3). A similar trend was observed in other occupational classes. The results of coefficients and significant value of MLR model, between T/S and height, weight, BMI, age, residence in Tehran, years of place lived in, years of occu- pation, pack-years, TC, TG, PL, TSL, and dietary pat- terns (Healthy, Western, Traditional and Vegetarian) are presented in Table 4. R² adjusted value for the overall model is .684. Predicted values of T/S from MLR and GAM are reported in Table 5. Figure 2. PCA-based analysis of data. (a) Scree-plot (b and c); PCA was applied to separation occupational class. Discussion highest contribution. In a healthy diet, whole grains, refined grains, dairy, and cereals had the highest factor- loading, respectively. In the western pattern, animal and solid fat, fish and poultry, salt and spice, processed meats, sweets, dessert, sugar, and refined cereals had the highest factor-loading, respectively. In the traditional pattern, it was fruits and vegetables, whole grains, and dairy prod- ucts and in the fourth dietary pattern (vegetarian diet) fruits and vegetables had the highest factor-loading, and other foods were not consumed. This is the first study that evaluates the relationship between the TL and foods, dietary patterns, and occupa- tional groups in a young male population. Three hundred young male aged 25 to 40 years were compared with each other after grouping in terms of occupational class. Only healthy subjects with no disease and having the least effect on the results of TL were selected. Thus, the effects of aging and chronic diseases on TL were eliminated. Numerous studies have reported that shorter telomeres 485 Karimi et al. Table 3. Multiple Linear Regressions Between log-T/S and Intake of Different Food Groups. All Office-workers Waste-recyclers Other workers β p-value β p-value β p-value β p-value (Intercept) 0.4169 .04 −0.7184 .13 0.7089 0 0.6467 .28 Whole grains 0.02312 .05 0.000015 .9 0.0277 .04 0.04701 .08 Refined grains 0.000245 .87 −0.005 .17 0.0021 .22 0.003694 .27 Vegetables and fruits 0.000099 .86 0.00064 .53 0.00062 .38 0.000675 .6 Fish products 0.00016 .24 0.00107 .01 0.0002 .16 0.000196 .52 Dairy products 0.000091 .61 −0.0002 .63 0.000121 .55 −0.000084 .84 Nuts seeds −0.00072 .06 −0.00036 .69 −0.00085 .04 −0.00042 .68 Meats −0.00043 .92 −0.00868 .27 −0.008 .12 −0.00067 .94 Produced meats −0.00238 .03 −0.00238 .3 −0.003 .02 −0.00174 .41 Liquid oils −0.00034 .55 −0.0016 .3 −0.00015 .8 −0.00173 .29 Solid fats −0.00146 .03 −0.00036 .83 −0.0015 .04 −0.00345 .08 Adjusted Model for age (at time of blood sample collection for telomere), education, BMI, smoking status. Table 3. Multiple Linear Regressions Between log-T/S and Intake of Different Food Groups. Adjusted Model for age (at time of blood sample collection for telomere), education, BMI, smoking status. Table 4. Multiple Linear Regressions Between log-T/S and Dietary Pattern and Other Covariate. Discussion All Office-workers Waste-recyclers Other workers β p-value β p-value β p-value β p-value (Intercept) 1.717 .25 5.667 .005 3.324 .23 2.637 .59 Height (cm) −0.00683 .43 −0.03394 .0065 −0.01779 .26 −0.0089 .74 Weight (kg) 0.001655 .86 0.03051 .012 0.01094 .55 0.0056 .84 BMI (kg/m2) −0.00642 .82 −0.08675 .016 −0.04386 .44 −0.0175 .85 Age (years) −0.00145 .67 0.0108 .01 0.009595 .23 −0.0057 .61 Residence (years) −0.00027 .89 −0.0022 .29 −0.00101 .87 −0.0052 .49 Live in a place (years) −0.00143 .62 −0.00622 .06 −0.00782 .23 0.0081 .46 Occupational class 0.037436 .01 0.000017 .99 0.01372 .71 0.02494 .60 Mean number of pack-years 0.000399 .18 0.000709 .00 0.001271 .56 0.00082 .70 TC −0.00216 .0 −0.00029 .78 0.000691 .61 −0.00387 .07 TG −0.00146 .002 −0.00175 .01 −0.00309 .006 −0.00138 .41 PL 0.00035 .58 0.000418 .46 0.0000384 .98 0.00008 .97 TSL −0.0011 .002 −0.00132 .00 −0.00103 .00 −0.00088 .00 Healthy dietary pattern 0.017133 .21 0.002123 .91 −0.00204 .91 0.0281 .59 Western dietary pattern −0.00424 .79 −0.05059 .13 −0.01646 .57 −0.016 .78 Traditional dietary pattern 0.012512 .45 −0.04081 .17 0.04068 .16 0.01306 .79 Vegetarian diet 0.012989 .33 −0.02224 .14 0.0252 .52 −0.0021 .95 Note. BMI = body mass index; TC = total cholesterol; TG = triglycerides; PL = phospholipids; TSL = total serum lipids. Adjusted model for age (at time of blood sample collection for telomere), education, BMI, smoking status. Table 4. Multiple Linear Regressions Between log-T/S and Dietary Pattern and Other Covariate. Note. BMI = body mass index; TC = total cholesterol; TG = triglycerides; PL = phospholipids; TSL = total serum lipids Adjusted model for age (at time of blood sample collection for telomere), education, BMI, smoking status. were associated with age and chronic diseases like dys- lipidemia (Aulinas et al., 2015), hypertension (Paik, Kang, Cho, & Shin, 2016), and diabetes (Zhao et al., 2014). Telomere shortening was increased with increase in age, BMI and weight, residence years, and total serum lipids (TSL) (Lynch et al., 2017). TL is dynamic and con- tinuously changing, and its length can be changed in both directions during life (Hovatta et al., 2012). Longer TL was seen after a definite period of observation and inter- vention studies (Nordfjäll et al., 2009; Svenson et al., 2013). Per the study of Svenson et al. (Svenson et al., 2013), this TL shortening is dependent on the original length of TL. Discussion Similar to this study and according to the study of Lee, a direct relationship was identified between the consumption of fruits and TL in middle-aged and adults. In another study, in men, a direct relationship was identified between the consumption of fruits and TL, but this relationship was not identified in women. Another study reported the consumption of the fish, fruits and vegetables are positively associated with TL. Similar to this study and per the study of Lee et al., a direct relation- ship was identified between the consumption of fruits and TL in middle-aged and adults (Lee et al., 2015). In another study, a direct relationship was reported between the con- sumption of fruits and TL in men, but this relationship was not identified in women (Tiainen et al., 2012). Another study reported that the consumption of fish, fruits and vegetables are positively associated with TL (Bethancourt, 2014). But some studies such as Gu et al. (Gu et al., 2015) and several other studies reported no significant relationship between vegetables and/or fruits and TL (Gu et al., 2015; Tiainen et al., 2012). Participants in these studies were mainly from certain races (white and Caucasian) and the average age of most of them in these studies such as the study of Chan was very high (Chan, Woo, Suen, Leung, & Tang, 2010). Thus, the results cannot be generalized for all. Vegetables and fruits are the major sources of flavonoids, which have antioxi- dant property (Freitas-Simoes, Ros, & Sala-Vila, 2016). Nutrients in fruits and vegetables can affect TL through mechanisms that are involved in cellular functions such as DNA repair and maintenance of chromosomes, pre- vention of methylation of DNA, prevention of inflamma- tion, oxidative stress, and increased telomerase activity In the current study, the type of food intake, diet, and job affected TL. Among all food compounds, greater con- sumption of grains, fruit and vegetables, fish and dairy products (only in waste recycle) were related to longer TL and, meat, and processed meat were associated with shorter TL (Table 3). Per other studies, whole grains, veg- etables, and nuts were inversely associated with age and total mortality-related diseases (Jaiswal McEligot, Largent, Ziogas, Peel, & Anton-Culver, 2006; Knoops et al., 2004; Steffen et al., 2003). This relationship reveals the effect of diet on the aging process of cells (Cherkas et al., 2006). Discussion In most studies, TL is reduced by aging (Brümmendorf et al., 2001; Hohensinner et al., 2014; Robertson et al., 2000), but in the study of Strindberg et al., age was not associated with TL (Strandberg et al., 2011). In addition, obesity, BMI, and smoking had a direct relationship with shorter TL (Strandberg et al., 2011). Although the use of a broad range of age in telo- mere studies is common (Aviv, Valdes, & Spector, 2006), American Journal of Men’s Health 12(2) 486 Table 5. Predicted T/S Values From Dietary Pattern and Other Covariate by MLR and GAM. Predicted values from LRM Predicted values from GAM Mean ± SD 95% CI Median Mean ± SD 95% CI Median Total 1.090 ± 0.26 [0.994, 1.204] 1.055 1.092 ± 0.20 [0.917, 1.304] 1.081 Office-workers 1.201 ± 0.31 [0.979, 1.473] 1.146 1.196 ± 0.24 [0.920, 1.560] 1.199 Waste-recyclers 1.069 ± 0.38 [0.747, 1.535] 1.037 1.042 ± 0.19 [0.857, 1.271] 1.068 Other workers 1.046 ± 0.24 [0.858, 8.538] 1.002 1.074 ± 0.26 [0.846, 1.356] 1.048 Note. MLR = multiple linear regression; GAM = generalized additive model. T/S Values From Dietary Pattern and Other Covariate by MLR and GAM. Table 5. Predicted T/S Values From Dietary Pattern and Other Covariate by MLR and GAM. Note. MLR = multiple linear regression; GAM = generalized additive model. had the highest relationship with TL (Lee, Jun, Yoon, Shin, & Baik, 2015; Marcon et al., 2012). Per a case- control study having more than 1,600 participants, a healthy lifestyle with fruits and vegetables in diet, in addition to physical activity, low BMI, and lack of smoking were associated with a longer TL (Mirabello et al., 2009). In a cohort study, a direct relationship was observed between vegetables’ consumption and TL length in women which is similar to the results of this study (Tiainen et al., 2012). In a case-control study, this relationship was observed between 300 gastric-cancer patients and 416 controls (Hou et al., 2009). Similar to this study and according to the study of Lee, a direct relationship was identified between the consumption of fruits and TL in middle-aged and adults. In another study, in men, a direct relationship was identified between the consumption of fruits and TL, but this relationship was not identified in women. Another study reported the consumption of the fish, fruits and vegetables are positively associated with TL. Discussion Similar to this study and per the study of Lee et al., a direct relation- ship was identified between the consumption of fruits and TL in middle-aged and adults (Lee et al., 2015). In another study, a direct relationship was reported between the con- sumption of fruits and TL in men, but this relationship was not identified in women (Tiainen et al., 2012). Another study reported that the consumption of fish, fruits and vegetables are positively associated with TL (Bethancourt, 2014). But some studies such as Gu et al. (Gu et al., 2015) and several other studies reported no significant relationship between vegetables and/or fruits and TL (Gu et al., 2015; Tiainen et al., 2012). Participants in these studies were mainly from certain races (white and Caucasian) and the average age of most of them in these studies such as the study of Chan was very high (Chan, Woo, Suen, Leung, & Tang, 2010). Thus, the results cannot be generalized for all. Vegetables and fruits are the major sources of flavonoids, which have antioxi- dant property (Freitas-Simoes, Ros, & Sala-Vila, 2016). Nutrients in fruits and vegetables can affect TL through mechanisms that are involved in cellular functions such as DNA repair and maintenance of chromosomes, pre- vention of methylation of DNA, prevention of inflamma- tion, oxidative stress, and increased telomerase activity these studies may have been biased and reported TL was not correct. With longer age, obesity, inflammatory reac- tions, oxidative stress, and other affected risk factors on TL shortening, such as smoking and chronic diseases are usually increased. Per previous reports, higher concentra- tion of blood lipids was associated with TL shortening and atherosclerosis (O’Donnell et al., 2008; Okuda et al., 2002). had the highest relationship with TL (Lee, Jun, Yoon, Shin, & Baik, 2015; Marcon et al., 2012). Per a case- control study having more than 1,600 participants, a healthy lifestyle with fruits and vegetables in diet, in addition to physical activity, low BMI, and lack of smoking were associated with a longer TL (Mirabello et al., 2009). In a cohort study, a direct relationship was observed between vegetables’ consumption and TL length in women which is similar to the results of this study (Tiainen et al., 2012). In a case-control study, this relationship was observed between 300 gastric-cancer patients and 416 controls (Hou et al., 2009). Discussion All of these such as TL are age-related diseases. In an experimental study on mice, it was identified that there was reduction in TL when mice were fed with beef for 4 weeks, and dose–response relationship was also found (O’Callaghan et al., 2012). The principal mechanism of the biological relationship between foods and TL is vague, and inflammatory reactions and oxidative path- ways may be involved (García-Calzón, Zalba, et al., 2015; Szebeni et al., 2014). Nevertheless, this rela- tionship was not identified in several other studies (Crous-Bou et al., 2014; Gu et al., 2015; Marcon et al., 2012). Although, in most of these studies, the sample size was large and there were different ethnicities, the studied populations were mostly women and/or elderly men. In this study, an inverse relationship was identified between fat intake and TL. Tiainen et al. reported that men significantly had a shorter TL when total fat, SFA, ( ) Although TL is under the control of genetic character- istics and individual differences (Graakjaer et al., 2004), it is also affected by environmental factors and job (Fujishiro, Diez-Roux, Landsbergis, Jenny, & Seeman, 2013). Work-related stress occurs when the demands of the work environment exceed the worker’s capacity (Ahola et al., 2012). The side effect of this type of oxida- tive stress is cell aging (Epel et al., 2004). TL shortening is a part of the response to chronic stress or psychological stress by the body’s cells (Ahola et al., 2012). In both unadjusted and adjusted models based on the age, BMI, and smoking, waste-recycling significantly had shorter TL than office-workers. Per other studies, the degree of inflammation and oxidative status is different among dif- ferent ethnic and occupational groups (Needham et al., 2013). TL shortening in waste recycling can be caused by high job stress (Fujishiro et al., 2013). Population-based studies indicated that waste recycling had higher levels of oxidative stress markers in the blood compared to other people in the community, due to excessive exposure to pollutants (Chen, Dietrich, Huo, & Ho, 2011). DNA oxi- dative damage indices such as 8-hydroxy-2-deoxyguano- sine and the presence of leukocytes in urine and systemic oxidative stress indices, including blood malondialde- hyde, lipid-peroxide, and total-urinary biopyrrins in waste-recyclers’ blood was very high (Yoshida et al., 2003). Discussion TL shortening is the best index of biological aging of cells. Thus, proper consumption patterns that are influenced by lifestyle can be associated with TL and its related parameters such as inflammation, chronic dis- eases, and mortality (Lin, Epel, & Blackburn, 2012). Whole grains are considered as a useful component in a healthy diet. The results of this study and that of Cassidy et al. reported that whole and refined grains were associ- ated with longer TL (A. Cassidy et al., 2010). Consumption of whole grains and plant foods due to the present of complex carbohydrates was related to the reduction of blood lipids, obesity, diabetes, cardiovascular disease and enhance the systemic inflammation (Lopez-Garcia et al., 2004; Poppitt et al., 2002). Thus, grains as a useful com- ponent in a healthy diet with vegetables, fruits, and smaller amounts of dairy and meat have beneficial health effects on the entire community and improve TL. But in some cross-sectional studies, whole and refined grains in the diet had no positive effect on TL (García-Calzón, Moleres, et al., 2015; Gu et al., 2015). This study and several others report a positive effect of vegetables and fruits on longer TL, and a direct relation- ship was observed between TL and the consumption of vegetables, root, pepper, and carrot among the vegetables 487 Karimi et al. (Gu et al., 2015). Anti-inflammatory and antioxidant properties of the materials can reduce the loss of telomere sequences (Gu et al., 2015). (Gu et al., 2015). Anti-inflammatory and antioxidant properties of the materials can reduce the loss of telomere sequences (Gu et al., 2015). and butter were consumed (Tiainen et al., 2012). In Song et al.’s study, butter consumption was inversely associ- ated with TL after adjusting the model based on the age (Song et al., 2013). Nettleton et al. in his study, identified that a dietary pattern with fat and processed meat was inversely associated with TL (Nettleton et al., 2008). But in the study of Macron et al., this relationship was not found (Marcon et al., 2012). In the study of Farzaneh et al. it was identified that comprehensive changes in men’s lifestyle, including reduction of fatty foods, intake of herbal food, consumption of omega-3 supplements, soy, vitamins C and E over a period of 3 months increased telomerase activity and finally resulted in longer TL (Farzaneh-Far et al., 2010). Discussion In the study of Cassidy et al., an inverse relationship was identified between fat intake of polyunsaturated fatty acids (PUFA) and TL in women (Cassidy et al., 2010). In another study, reduced low- density lipoprotein plasma concentration was associated with increased telomerase activity and reduced psycho- logical distress (Ornish et al., 2008). Thus, the intake of total fat and saturated fat is associated with shorter TL. Long-chain PUFA diet, which has antioxidant and anti- inflammatory properties, is useful to prevent the destruc- tion of TL (Freitas-Simoes et al., 2016). Per Table 3, a negative relationship was identified between the consumption of meat, dairy products (only in office-workers), processed-meat, and TL. Several other studies reported no significant relationship between dairy consumption and TL (Chan et al., 2010; Marcon et al., 2012; Nettleton et al., 2008). Majority of these studies were conducted in different ethnic groups and in both female and male genders. In another study on children, dairy consumption had no effect on TL (García-Calzón, Moleres, et al., 2015). In the study of Song et al., milk consumption was negatively related to TL in healthy women, and fat cheese consumption compared with low- fat cheese was associated with shorter TL (Song et al., 2013). In contrast to these results and similar to this study, Lee demonstrated that higher consumption of dairy prod- ucts was associated with an increase in TL (Lee et al., 2015). Gu et al., 2015 revealed that dairy consumption was associated with an increase in TL in the non-Hispanic white population (Gu et al., 2015). Meat products are probably one of the highest sources of saturated fatty acids (SFA) (Givens, 2009). Lee et al. reported that the consumption of less red meat and processed meat and higher intake of grains in Korean middle-aged and adults were associated with longer TL (Lee et al., 2015). Another study on 840 adults indicated that the consumption of processed meat was inversely associated with TL (Nettleton et al., 2008). In the study of Betancourt et al., processed meat, grilled meat, oil, and beverages were associated with shorter TL (Bethancourt, 2014). Per other studies, high intake of processed meat leads to the forma- tion of inflammatory mediator compounds and increases the risk of numerous cancers such as breast cancer (Cho et al., 2006), diabetes (Pan et al., 2011) and cardiovascu- lar diseases (CVD) (Micha, Wallace, & Mozaffarian, 2010). Discussion Low concentrations of tocopherol antioxidants (Ekor & Odewabi, 2014) and higher levels of inflamma- tory indices (Kuijer, Sluiter, & Frings-Dresen, 2010) such as IL-6 were identified in their blood (Kuijer et al., 2010). In this study, an inverse relationship was identified between fat intake and TL. Tiainen et al. reported that men significantly had a shorter TL when total fat, SFA, American Journal of Men’s Health 12(2) 488 Diet and lifestyle can cause inflammation, oxidative stress, and increased psychological pressure, and all these factors affect TL (Crous-Bou et al., 2014). Other nonbio- logical mechanisms and confounding factors or exposure to environmental pollutants’ may affect TL shortening in waste-recyclers. Diet and lifestyle can cause inflammation, oxidative stress, and increased psychological pressure, and all these factors affect TL (Crous-Bou et al., 2014). Other nonbio- logical mechanisms and confounding factors or exposure to environmental pollutants’ may affect TL shortening in waste-recyclers. consumption of vegetarian foods was related to longer TL in healthy women (Crous-Bou et al., 2014). Furthermore, per Boccardi et al. study, adherence to the Mediterranean (Healthy) pattern was associated with a longer TL when compared to the other patterns (Boccardi et al., 2013). With the aim of fixing the problems of previous stud- ies, in this study, the male’s age of individuals was selected in the range of 25 to 40 years and several con- founding factors affecting TL were eliminated. As a result, the statistical power for detecting the relationships between dietary factors and telomere length was increased. Another strong point is the examination of the effect of dietary patterns on three different occupational groups. All samples were population-based and TL was measured using real-time PCR in which a small amount of DNA was required compared to the Southern blot method and it can be used in most of the epidemiological studies (Boccardi et al., 2013; Crous-Bou et al., 2014; Gu et al., 2015). Evidence suggests that the food health depends on the total diet composition, rather than its components, and synergistic-interactions may exist between different foods. Per PCA of food data, four healthy, western, tradi- tional, and vegetarian dietary patterns were identified. Per Table 4, a positive relationship was identified between the healthy dietary pattern and TL, and a negative rela- tionship was identified between western pattern and TL in all occupational groups. Discussion The results revealed that the traditional pattern and herbal pattern, respectively were associated with increased and reduced TL. This negative relationship between TL and western pattern can be attributed to the consumption of low healthy foods and high levels of undesirable foods (such as produced meats and high fat content). Intake of saturated fatty acids in the western pattern is associated with insulin resistance and metabolic diseases (Isharwal, Misra, Wasir, & Nigam, 2009). In previous studies, a healthy or Mediterranean dietary pattern has been associated with reduced obesity and longer TL (Boccardi et al., 2013; Crous-Bou et al., 2014; Gu et al., 2015). The beneficial effects of healthy pattern are due to the variety of vitamins, minerals, phy- tochemicals, and fiber. High consumption of vegetables reduces visceral fat and risk factors for type 2 diabetes (Mozaffarian et al., 2010). Dairy intake is associated with reduced insulin resistance and dyslipidemia (Mozaffarian et al., 2010). The results of a cohort study revealed a strong relationship between a healthy dietary pattern and reduced markers of inflammation in adults (Greenlee, Arbuckle, & Chyou, 2003). In Nettleton et al.’s study, western pattern with higher consumption of fats and pro- cessed meat was associated with shorter TL (Nettleton et al., 2008). In Lopez et al.’s study, western pattern with high consumption of processed meats, sweets, chips, and refined grains was associated with increased inflamma- tory markers (Lopez-Garcia et al., 2004). During 10 years follow-up of the samples by Lee et al. using FFQ, it was observed that prudent pattern with high intake of whole grains, seafood, legumes, vegetables, and seaweed was associated with longer TL and western pattern with high consumption of refined grains, processed meat, and bev- erages was associated with a shorter TL (Lee et al., 2015). Another study with 520 participants after 5 years of mon- itoring reported a negative relationship between Western pattern, by shortening TL and inflammatory reactions in the elderly and those prone to cardiovascular diseases (García-Calzón, Moleres, et al., 2015). Mediterranean i h l i k f d f d h References Ahola, K., Sirén, I., Kivimäki, M., Ripatti, S., Aromaa, A., Lönnqvist, J., & Hovatta, I. (2012). Work-related exhaus- tion and telomere length: A population-based study. PLoS one, 7(7), e40186. Chen, A., Dietrich, K. N., Huo, X., & Ho, S.-m. (2011). Developmental neurotoxicants in e-waste: An emerg- ing health concern. Environmental Health Perspectives, 119(4), 431–438. Aulinas, A., Ramírez, M.-J., Barahona, M.-J., Valassi, E., Resmini, E., Mato, E., … Webb, S. M. (2015). Dyslipidemia and chronic inflammation markers are correlated with telo- mere length shortening in Cushing’s Syndrome. PLoS one, 10(3), e0120185. Cherkas, L. F., Aviv, A., Valdes, A. M., Hunkin, J. L., Gardner, J. P., Surdulescu, G. L., … Spector, T. D. (2006). The effects of social status on biological aging as measured by white-blood-cell telomere length. Aging Cell, 5(5), 361– 365. Aviv, A., Valdes, A. M., & Spector, T. D. (2006). Human telo- mere biology: Pitfalls of moving from the laboratory to epi- demiology. International Journal of Epidemiology, 35(6), 1424–1429. Cho, E., Chen, W. Y., Hunter, D. J., Stampfer, M. J., Colditz, G. A., Hankinson, S. E., & Willett, W. C. (2006). Red meat intake and risk of breast cancer among premenopausal women. Archives of Internal Medicine, 166(20), 2253–2259. Bernert, J. T., Turner, W. E., Patterson, D. G., & Needham, L. L. (2007). Calculation of serum “total lipid” concentrations for the adjustment of persistent organohalogen toxicant mea- surements in human samples. Chemosphere, 68(5), 824–831. Crous-Bou, M., Fung, T. T., Prescott, J., Julin, B., Du, M., Sun, Q., & …De Vivo, I. (2014). Mediterranean diet and telomere length in nurses’ health study: Population based cohort study. BMJ, 349, g6674. Bethancourt, H. J. (2014). Association between longitudinally assessed dietary composition and blood telomere length among young adult Filipinos (Master Thesis). University of Washington, Seattle, WA. Ekor, M., & Odewabi, A. O. (2014). Occupational exposure to municipal solid wastes and development of toxic neuropa- thies: Possible role of nutrient supplementation, comple- mentary and alternative medicines in chemoprevention. Chinese Journal of Integrative Medicine, 20(9), 643–653. Blackburn, E. H., Greider, C. W., & Szostak, J. W. (2006). Telomeres and telomerase: The path from maize, Tetrahymena and yeast to human cancer and aging. Nature Medicine, 12(10), 1133–1138. Epel, E. S., Blackburn, E. H., Lin, J., Dhabhar, F. S., Adler, N. E., Morrow, J. D., & Cawthon, R. M. (2004). Accelerated telomere shortening in response to life stress. Proceedings Boccardi, V., Esposito, A., Rizzo, M. Acknowledgments This work was supported by the Ph.D. Programs Foundation of Public Health, Tehran University of Medical Sciences, Iran (Grant No. 9211150001). Brümmendorf, T. H., Rufer, N., Holyoake, T. L., Maciejewski, J., Barnett, M. J., Eaves, C. J., … Lansdorp, P. M. (2001). Telomere length dynamics in normal individuals and in patients with hematopoietic stem cell-associated disor- ders. Annals of the New York Academy of Sciences, 938(1), 293–304. Conclusion In conclusion, change and intervention in a diet is a pow- erful tool for preventing and delaying chronic diseases. With a thorough understanding of the mechanisms affect- ing the cellular age, inflammation, oxidative stress, apop- tosis, and their relationship with TL, the prevention of numerous diseases can be addressed. The results of this study demonstrated that an inverse relationship was iden- tified between TL and weight, BMI, age, and TSL. High serum lipid concentration may be associated with sys- temic inflammation and atherosclerosis and may lead to oxidative stress, resulting in telomere shortening. Adherence to a healthy diet with increased consumption of fruits, vegetables, whole grains, and white meat (fish) is necessary to prevent TL destruction and increase life span. According to this study, adherence to a healthy and then traditional diet was significantly associated with an increase in TL in all occupational groups. But the western dietary pattern was associated with shorter TL. Further well-design cohort studies with larger sample size are necessary to survey the effect of the consumption of red and processed meat and dairy products on TL. This study was cross-sectional and no time relation- ship has been examined between food groups and TL. Hence, TL reduction due to food intake over the time is unknown. Although the validity and reliability of this method have been investigated in several studies (Hosseini Esfahani et al., 2010), data were collected by using FFQ, and food intake is dependent on the mind of an individual. Therefore, the possibility of error existed. In the next cohort studies, the use of biological markers of nutrient intake to assess the food intake is recom- mended. Also, investigating the relationship between 489 Karimi et al. maintenance and health status among elderly. PLoS one, 8(4), e62781. dietary factors with telomerase activity and markers of inflammation will provide further information. dietary factors with telomerase activity and markers of inflammation will provide further information. Brouilette, S. W., Moore, J. S., McMahon, A. D., Thompson, J. R., Ford, I., Shepherd, J., … Group, W. o. S. C. P. S. (2007). Telomere length, risk of coronary heart disease, and statin treatment in the West of Scotland Primary Prevention Study: A nested case-control study. The Lancet, 369(9556), 107–114. Author Contributions B.K. and M.Y. designed the study, participated in the data col- lection, and wrote the main manuscript. M.Y. and R.N. per- formed the statistical analysis and assisted in the data collection and approved the final version. P.M., N.R., and A.A. partici- pated in its design and helped to draft the manuscript and con- tributed to the raw data collection. All authors reviewed and approved the final manuscript. Buxton, J. L., Walters, R. G., Visvikis-Siest, S., Meyre, D., Froguel, P., & Blakemore, A. I. (2011). Childhood obesity is associated with shorter leukocyte telomere length. The Journal of Clinical Endocrinology & Metabolism, 96(5), 1500–1505. Declaration of Conflicting Interests Cassidy, A., De Vivo, I., Liu, Y., Han, J., Prescott, J., Hunter, D. J., & Rimm, E. B. (2010). Associations between diet, life- style factors, and telomere length in women. The American Journal of Clinical Nutrition, 91(5), 1273–1280. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding Cassidy, A., De Vivo, I., Liu, Y., Han, J., Prescott, J., Hunter, D. J., & Rimm, E. B. (2010). Associations between diet, lifestyle factors, and telomere length in women. American Journal of Clinical Nutrition, 91(5), 1273–1280. doi:10.3945/ajcn.2009.28947 The author(s) disclosed receipt of the following financial sup- port for the research, authorship, and/or publication of this arti- cle: The author(s) received financial support from Award Number 9211150001 from the Tehran University of Medical Sciences for the research of this article. Cawthon, R. M. (2002). Telomere measurement by quantitative PCR. Nucleic Acids Research, 30(10), e47–e47. Chan, R., Woo, J., Suen, E., Leung, J., & Tang, N. (2010). Chinese tea consumption is associated with longer telo- mere length in elderly Chinese men. British Journal of Nutrition, 103(01), 107–113. References R., Marfella, R., Barbieri, M., & Paolisso, G. (2013). Mediterranean diet, telomere American Journal of Men’s Health 12(2) 490 Hosseini Esfahani, F., Asghari, G., Mirmiran, P., & Azizi, F. (2010). Reproducibility and relative validity of food group intake in a food frequency questionnaire developed for the Tehran lipid and glucose study. Journal of Epidemiology, 20(2), 150–158. of the National Academy of Sciences of the United States of America, 101(49), 17312–17315. of the National Academy of Sciences of the United States of America, 101(49), 17312–17315. Farzaneh-Far, R., Lin, J., Epel, E. S., Harris, W. S., Blackburn, E. H., & Whooley, M. A. (2010). Association of marine omega-3 fatty acid levels with telomeric aging in patients with coronary heart disease. JAMA, 303(3), 250–257. Hou, L., Savage, S. A., Blaser, M. J., Perez-Perez, G., Hoxha, M., Dioni, L., … Lissowska, J. (2009). Telomere length in peripheral leukocyte DNA and gastric cancer risk. Cancer Epidemiology and Prevention Biomarkers, 18(11), 3103– 3109. Freitas-Simoes, T.-M., Ros, E., & Sala-Vila, A. (2016). Nutrients, foods, dietary patterns and telomere length: Update of epidemiological studies and randomized trials. Metabolism, 65(4), 406–415. Hovatta, I., de Mello, V. D., Kananen, L., Lindström, J., Eriksson, J. G., Ilanne-Parikka, P., … Uusitupa, M. (2012). Leukocyte telomere length in the Finnish diabetes preven- tion study. PLoS one, 7(4), e34948. Fujishiro, K., Diez-Roux, A. V., Landsbergis, P. A., Jenny, N. S., & Seeman, T. (2013). Current employment status, occupational category, occupational hazard exposure and job stress in relation to telomere length: The Multiethnic Study of Atherosclerosis (MESA). Occupational and Environmental Medicine, 70(8), 552–560. Hoxha, M., Dioni, L., Bonzini, M., Pesatori, A. C., Fustinoni, S., Cavallo, D., … Schwartz, J. (2009). Association between leukocyte telomere shortening and exposure to traffic pollution: A cross-sectional study on traffic officers and indoor office workers. Environmental Health, 8(1), 1. García-Calzón, S., Gea, A., Razquin, C., Corella, D., Lamuela- Raventos, R. M., Martínez, J. A., … Marti, A. (2014). Longitudinal association of telomere length and obesity indices in an intervention study with a Mediterranean diet: The PREDIMED-NAVARRA trial. International Journal of Obesity, 38(2), 177–182. Isharwal, S., Misra, A., Wasir, J., & Nigam, P. (2009). Diet & insulin resistance: A review & Asian Indian perspective. Indian Journal of Medical Research, 129(5), 485–499. García-Calzón, S., Moleres, A., Martínez-González, M. A., Martínez, J. A., Zalba, G., & Marti, A. (2015). References Dietary total antioxidant capacity is associated with leukocyte telomere length in a children and adolescent population. Clinical Nutrition, 34(4), 694–699. Jaiswal McEligot, A., Largent, J., Ziogas, A., Peel, D., & Anton-Culver, H. (2006). Dietary fat, fiber, vegetable, and micronutrients are associated with overall survival in postmenopausal women diagnosed with breast cancer. Nutrition and Cancer, 55(2), 132–140. Joshu, C. E., Peskoe, S. B., Heaphy, C. M., Kenfield, S. A., Van Blarigan, E. L., Mucci, L. A., … Lee, T. K. (2015). Prediagnostic obesity and physical inactivity are associ- ated with shorter telomere length in prostate stromal cells. Cancer Prevention Research, 8(8), 737–742. García-Calzón, S., Zalba, G., Ruiz-Canela, M., Shivappa, N., Hébert, J. R., Martínez, J. A., … Marti, A. (2015). Dietary inflammatory index and telomere length in subjects with a high cardiovascular disease risk from the PREDIMED- NAVARRA study: Cross-sectional and longitudinal analy- ses over 5 y. The American Journal of Clinical Nutrition, 102(4), 897–904. Karimi, B., Yunesian, M., Nabizadeh, R., Mehdipour, P., & Aghaie, A. (2016). Is leukocyte telomere length related with lung cancer risk?: A meta-analysis. Iranian Biomedical Journal, 21(3), 142–153. Givens, I. (2009). Animal nutrition and lipids in animal prod- ucts and their contribution to human intake and health. Nutrients, 1(1), 71–82. Knoops, K. T., de Groot, L. C., Kromhout, D., Perrin, A.-E., Moreiras-Varela, O., Menotti, A., & Van Staveren, W. A. (2004). Mediterranean diet, lifestyle factors, and 10-year mortality in elderly European men and women: The HALE project. JAMA, 292(12), 1433–1439. Graakjaer, J., Pascoe, L., Der-Sarkissian, H., Thomas, G., Kolvraa, S., Christensen, K., & Londoño-Vallejo, J. A. (2004). The relative lengths of individual telomeres are defined in the zygote and strictly maintained during life. Aging Cell, 3(3), 97–102. Kuijer, P., Sluiter, J. K., & Frings-Dresen, M. H. (2010). Health and safety in waste collection: Towards evidence-based worker health surveillance. American Journal of Industrial Medicine, 53(10), 1040–1064. Greenlee, A. R., Arbuckle, T. E., & Chyou, P.-H. (2003). Risk factors for female infertility in an agricultural region. Epidemiology, 14(4), 429–436. Lee, J., Jun, N., Yoon, D., Shin, C., & Baik, I. (2015). Association between dietary patterns in the remote past and telomere length. European Journal of Clinical Nutrition,69(9), 1048–1052. Gu, Y., Honig, L. S., Schupf, N., Lee, J. H., Luchsinger, J. A., Stern, Y., & Scarmeas, N. (2015). Mediterranean diet and leukocyte telomere length in a multi-ethnic elderly popula- tion. Age, 37(2), 1–13. References The association between leukocyte telomere length and cigarette smoking, dietary and physical variables, and risk of prostate cancer. Aging Cell, 8(4), 405–413. Poppitt, S. D., Keogh, G. F., Prentice, A. M., Williams, D. E., Sonnemans, H. M., Valk, E. E., … Wareham, N. J. (2002). Long-term effects of ad libitum low-fat, high-carbohydrate diets on body weight and serum lipids in overweight sub- jects with metabolic syndrome. The American Journal of Clinical Nutrition, 75(1), 11–20. Mozaffarian, D., Cao, H., King, I. B., Lemaitre, R. N., Song, X., Siscovick, D. S., & Hotamisligil, G. k. S. (2010). Trans- palmitoleic acid, metabolic risk factors, and new-onset diabetes in US adults: A cohort study. Annals of Internal Medicine, 153(12), 790–799. Porta, M., de Basea, M. B., Benavides, F. G., López, T., Fernandez, E., Marco, E., … Group, P. I. S. (2008). Differences in serum concentrations of organochlorine compounds by occupational social class in pancreatic can- cer. Environmental Research, 108(3), 370–379. MWer, S., Dykes, D., & Polesky, H. (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Research, 16, 1215. Robertson, J. D., Gale, R. E., Wynn, R. F., Dougal, M., Linch, D. C., Testa, N. G., & Chopra, R. (2000). Dynamics of telomere shortening in neutrophils and T lymphocytes during ageing and the relationship to skewed X chromosome inactivation patterns. British Journal of Haematology, 109(2), 272–279. Needham, B. L., Adler, N., Gregorich, S., Rehkopf, D., Lin, J., Blackburn, E. H., & Epel, E. S. (2013). Socioeconomic status, health behavior, and leukocyte telomere length in the National Health and Nutrition Examination Survey, 1999–2002. Social Science & Medicine, 85, 1–8. Shen, J., Gammon, M. D., Terry, M. B., Wang, Q., Bradshaw, P., Teitelbaum, S. L., … Santella, R. M. (2009). Telomere length, oxidative damage, antioxidants and breast cancer risk. International Journal of Cancer, 124(7), 1637–1643. Nettleton, J. A., Diez-Roux, A., Jenny, N. S., Fitzpatrick, A. L., & Jacobs, D. R. (2008). Dietary patterns, food groups, and telomere length in the Multi-Ethnic Study of Atherosclerosis (MESA). The American Journal of Clinical Nutrition, 88(5), 1405–1412. Shiels, P. G., McGlynn, L. M., MacIntyre, A., Johnson, P. C., Batty, G. D., Burns, H., … McConnachie, A. (2011). Accelerated telomere attrition is associated with relative household income, diet and inflammation in the pSoBid cohort. PLoS one, 6(7), e22521. Ni, W., Huang, Y., Wang, X., Zhang, J., & Wu, K. (2014). References Lin, J., Epel, E., & Blackburn, E. (2012). Telomeres and life- style factors: Roles in cellular aging. Mutation Research/ Fundamental and Molecular Mechanisms of Mutagenesis, 730(1), 85–89. Harte, A. L., da Silva, N. F., Miller, M. A., Cappuccio, F. P., Kelly, A., O’Hare, J. P., … Alokail, M. (2012). Telomere length attrition, a marker of biological senescence, is inversely correlated with triglycerides and cholesterol in South Asian males with type 2 diabetes mellitus. Experimental Diabetes Research, 2012, 1–7. Lopez-Garcia, E., Schulze, M. B., Fung, T. T., Meigs, J. B., Rifai, N., Manson, J. E., & Hu, F. B. (2004). Major dietary patterns are related to plasma concentrations of markers of inflammation and endothelial dysfunction. The American Journal of Clinical Nutrition, 80(4), 1029–1035. Hohensinner, P. J., Goronzy, J. J., & Weyand, C. M. (2014). Telomere dysfunction, autoimmunity and aging. Aging and Disease, 2(6), 524–537. 491 Karimi et al. World Health Organization. (2013). BMI classification 2013. Retrieved June 30, 2014, from: apps. who. int/bmi/index.jsp Lynch, S. M., Mitra, N., Ravichandran, K., Mitchell, J., Spangler, E., Zhou, W., … Stowe, R. (2017). Telomere length and neighborhood circumstances: Evaluating biological response to unfavorable exposures. Cancer Epidemiology Biomarkers & Prevention, 26(4), 553–560. Ornish, D., Lin, J., Daubenmier, J., Weidner, G., Epel, E., Kemp, C., … Carroll, P. R. (2008). Increased telomer- ase activity and comprehensive lifestyle changes: A pilot study. The Lancet Oncology, 9(11), 1048–1057. Marcon, F., Siniscalchi, E., Crebelli, R., Saieva, C., Sera, F., Fortini, P., … Palli, D. (2012). Diet-related telomere short- ening and chromosome stability. Mutagenesis, 27(1), 49– 57. Paik, J. K., Kang, R., Cho, Y., & Shin, M.-J. (2016). Association between genetic variations affecting mean telomere length and the prevalence of hypertension and coronary heart dis- ease in Koreans. Clinical Nutrition Research, 5(4), 249–260. Micha, R., Wallace, S. K., & Mozaffarian, D. (2010). Red and processed meat consumption and risk of incident coronary heart disease, stroke, and diabetes mellitus a systematic review and meta-analysis. Circulation, 121(21), 2271– 2283. Pan, A., Sun, Q., Bernstein, A. M., Schulze, M. B., Manson, J. E., Willett, W. C., & Hu, F. B. (2011). Red meat con- sumption and risk of type 2 diabetes: 3 cohorts of US adults and an updated meta-analysis. The American Journal of Clinical Nutrition, 94(4), 1088–1096. Mirabello, L., Huang, W. Y., Wong, J. Y., Chatterjee, N., Reding, D., David Crawford, E., … Savage, S. A. (2009). References Associations of neonatal lead, cadmium, chromium and nickel co-exposure with DNA oxidative damage in an electronic waste recycling town. Science of the Total Environment, 472, 354–362. Song, Y., You, N. C. Y., Song, Y. Q., Kang, M. K., Hou, L. F., Wallace, R., … Liu, S. M. (2013). Intake of small- to-medium-chain saturated fatty acids is associated with peripheral leukocyte telomere length in postmeno- pausal women. Journal of Nutrition, 143(6), 907–914. doi:10.3945/jn.113.175422 Nordfjäll, K., Svenson, U., Norrback, K.-F., Adolfsson, R., Lenner, P., & Roos, G. (2009). The individual blood cell telomere attrition rate is telomere length dependent. PLoS Genet, 5(2), e1000375. O’Callaghan, N. J., Toden, S., Bird, A. R., Topping, D. L., Fenech, M., & Conlon, M. A. (2012). Colonocyte telomere shortening is greater with dietary red meat than white meat and is attenuated by resistant starch. Clinical Nutrition, 31(1), 60–64. Steffen, L. M., Jacobs, D. R., Stevens, J., Shahar, E., Carithers, T., & Folsom, A. R. (2003). Associations of whole-grain, refined-grain, and fruit and vegetable consumption with risks of all-cause mortality and incident coronary artery disease and ischemic stroke: The Atherosclerosis Risk in Communities (ARIC) study. The American Journal of Clinical Nutrition, 78(3), 383–390. O’Donnell, C. J., Demissie, S., Kimura, M., Levy, D., Gardner, J. P., White, C., … Cupples, L. A. (2008). Leukocyte telo- mere length and carotid artery intimal medial thickness the Framingham heart study. Arteriosclerosis, Thrombosis, and Vascular Biology, 28(6), 1165–1171. Strandberg, T. E., Saijonmaa, O., Tilvis, R. S., Pitkälä, K. H., Strandberg, A. Y., Miettinen, T. A., & Fyhrquist, F. (2011). Association of telomere length in older men with mortality and midlife body mass index and smoking. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 66(7), 815–820. Okuda, K., Bardeguez, A., Gardner, J. P., Rodriguez, P., Ganesh, V., Kimura, M., … Aviv, A. (2002). Telomere length in the newborn. Pediatric Research, 52(3), 377–381. American Journal of Men’s Health 12(2) 492 telomere length and its relation to food and nutrient intake in an elderly population. European Journal of Clinical Nutrition, 66(12), 1290–1294. Svenson, U., Grönlund, E., Söderström, I., Sitaram, R. T., Ljungberg, B., & Roos, G. (2013). Telomere length in rela- tion to immunological parameters in patients with renal cell carcinoma. PLoS one, 8(2), e55543. Yoshida, R., Ogawa, Y., Mori, I., Nakata, A., Wang, R., Ueno, S., … Hisanaga, N. (2003). References Associations between oxidative stress levels and total duration of engagement in jobs with exposure to fly ash among workers at municipal solid waste incinerators. Mutagenesis, 18(6), 533–537. Szebeni, A., Szebeni, K., DiPeri, T., Chandley, M. J., Crawford, J. D., Stockmeier, C. A., & Ordway, G. A. (2014). Shortened telomere length in white matter oligodendro- cytes in major depression: Potential role of oxidative stress. International Journal of Neuropsychopharmacology, 17(10), 1579–1589. g ( ) Zhao, J., Zhu, Y., Lin, J., Matsuguchi, T., Blackburn, E., Zhang, Y., … Howard, B. V. (2014). Short leukocyte telomere length predicts risk of diabetes in American Indians: The strong heart family study. Diabetes, 63(1), 354–362. Tiainen, A.-M., Männistö, S., Blomstedt, P., Moltchanova, E., Perälä, M., Kaartinen, N., … Eriksson, J. (2012). Leukocyte
https://openalex.org/W4250601694	http://vital.lib.tsu.ru/vital/access/services/Download/koha:000719686/SOURCE1	Russian	null	МИКРОМЕХАНИЧЕСКАЯ МОДЕЛЬ ЭВОЛЮЦИИ МЕЗОСКОПИЧЕСКОГО ДЕФОРМАЦИОННОГО РЕЛЬЕФА В α-ТИТАНЕ	Международная конференция "Физическая мезомеханика. Материалы с многоуровневой иерархически организованной структурой и интеллектуальные производственные технологии"	2,021	cc-by	371	МЕЖДУНАРОДНАЯ КОНФЕРЕНЦИЯ «Физическая мезомеханика. Материалы с многоуровневой иерархически организованной структурой и интеллектуальные производственные технологии» 6–10 сентября 2021 г. Томск, Россия DOI: 10.17223/978-5-907442-03-0-2021-370 МИКРОМЕХАНИЧЕСКАЯ МОДЕЛЬ ЭВОЛЮЦИИ МЕЗОСКОПИЧЕСКОГО ДЕФОРМАЦИОННОГО РЕЛЬЕФА В α-ТИТАНЕ 1,2Писарев М., 1,2Емельянова Е.С., 1Зиновьева О.С., 1Романова В.А. 1Институт физики прочности и материаловедения СО РАН, Томск 2Томский государственный университет, Томск Д Ц 1,2Писарев М., 1,2Емельянова Е.С., 1Зиновьева О.С., 1Романова В.А. 1Институт физики прочности и материаловедения СО РАН, Томск 2Томский государственный университет, Томск В данной работе проведено численное моделирование деформационного поведения поликристаллического титана с учетом зеренной структуры, особенностей дислокационного скольжения в гексагональных металлах по призматическим, базисным и пирамидальным системам скольжения, а также деформационного упрочнения. Деформационное поведение поликристаллического титана описывается в рамках физической теории пластичности кристаллов, где поликристалл представляется как конгломерат монокристаллов с различными кристаллографическими ориентациями. Для верификации разработанной модели была проведена серия расчетов одноосного растяжения 12 монокристаллов титана с различными ориентациями относительно оси нагружения. Оценки напряжений пластического течения, полученные в численных расчетах, сравнивались с аналитическими значениями, вычисленными согласно закону Шмида. Показано, что результаты во всех случаях сходятся с аналитическими оценками с погрешностью не более 0.5%, что говорит о корректности численной реализации. Проведены оценки вкладов различных семейств скольжения в пластическую деформацию поликристаллического титана, особое внимание уделено вкладу пирамидальных систем скольжения. Показано, что в ходе нагружения превалирует сдвиг по призматическим системам, базисные системы включаются в деформационный процесс на более поздней стадии деформации, а сдвиг по пирамидальным системам наблюдается лишь в единичных областях материала. Вместе с тем, моделирование без учета пирамидальных систем скольжения приводит к существенно завышенному уровню локальных напряжений в единичных зернах, а также к завышенным макроскопическим напряжениям. р р р Определены размеры представительного объема для исследования эволюции мезоскопического деформационного рельефа и требования к разрешению расчетных сеток. Показано, что все рассмотренные модели могут адекватно воспроизводить деформационные явления на мезоуровне, но только в ограниченном диапазоне приложенной деформации. Как только масштаб локализации пластической деформации становится сравнимым с размером модели, она больше не может корректно описывать деформационное поведение. а б в г Рис. 1. Деформационный рельеф в модельных поликристаллах разных размеров при 26% (а), 33% (б), 36% (в) и 46% (г) деформации б г в Рис. 1. Деформационный рельеф в модельных поликристаллах разных размеров при 26% (а), 33% (б), 36% (в) и 46% (г) деформации Исследование выполнено за счет гранта Российского научного фонда (проект № 20-19-00600). 578
https://openalex.org/W4390058880	https://www.mdpi.com/2073-4395/14/1/24/pdf?version=1703150726	English	null	Designing Management Strategies for Sheep Production and Bees in Dryland Pastures	Agronomy	2,023	cc-by	12,966	Citation: Caudillo, M.; Melathopoulos, A.; Prado-Tarango, D.E.; Smallman, M.; Taylor, S.A.; Ates, S. Designing Management Strategies for Sheep Production and Bees in Dryland Pastures. Agronomy 2024, 14, 24. https://doi.org/10.3390/ agronomy14010024 Keywords: lamb growth; closing dates; pasture diversity; ecosystem benefits; pollinator health agronomy agronomy agronomy agronomy Article Mia Caudillo 1, Andony Melathopoulos 2, David Eduardo Prado-Tarango 1 , Mary Smallm and Serkan Ates 1,* Mia Caudillo 1, Andony Melathopoulos 2, David Eduardo Prado-Tarango 1 , Mary Smallman 1, Sarah A. Taylor 2 and Serkan Ates 1,* , and Serkan Ates 1,* 1 Department of Animal and Rangeland Sciences, Oregon State University, Corvallis, OR 97331-5503, USA; mia.c.caudillo@gmail.com (M.C.); pradotad@oregonstate.edu (D.E.P.-T.); mary.smallman@oregonstate.edu (M.S.) 1 Department of Animal and Rangeland Sciences, Oregon State University, Corvallis, OR 97331-5503, USA; mia.c.caudillo@gmail.com (M.C.); pradotad@oregonstate.edu (D.E.P.-T.); mary.smallman@oregonstate.edu (M.S.) 1 Department of Animal and Rangeland Sciences, Oregon State University, Corvallis, OR 97331-5503, USA; mia.c.caudillo@gmail.com (M.C.); pradotad@oregonstate.edu (D.E.P.-T.); marysmallman@oregonstate edu (M S ) y g ( ) 2 Department of Horticulture, Oregon State University, Corvallis, OR 97331-5503, USA; andony.melathopoulos@oregonstate.edu (A.M.); sarah.kincaid@oregonstate.edu (S.A.T.) y g 2 Department of Horticulture, Oregon State University, Corvallis, OR 97331-5503, USA; andony.melathopoulos@oregonstate.edu (A.M.); sarah.kincaid@oregonstate.edu (S.A.T.) y p g ( ) g ( * Correspondence: serkan.ates@oregonstate.edu Abstract: Novel grazing management practices for livestock and bee health are becoming increasingly crucial in pasture-based farming systems. The effect of pasture type and spring closing dates on lamb liveweight gain, pasture production, botanical composition, bloom density and bee visitation was monitored over 2 years. Total annual dry matter yield (DMY) of diverse pastures in 2020/2021 was 8.8 t DM ha−1. This yield was greater than the DMY obtained from both simple (7.6 t DM ha−1) and legume pastures (6.6 t DM ha−1). In 2021/2022, the total annual DMY of simple (8.6 t DM ha−1) and diverse pastures (9.0 t DM ha−1) was similar. However, the legume pastures produced 27–30% less than simple and diverse pastures. In successive years, lambs grew faster in legume pastures (287, 215 g per head d−1) than diverse (207, 151 g per head d−1) and simple pastures (204, 132 g per head d−1). However, spring liveweight production (kg ha−1 day−1) from pastures did not differ due to the lower stocking density of legume pastures as compared to the other two pasture mixtures. Bloom density (flower/m2) and bee visitation (bees/min−1 m2) were 16 and 40 times greater with legume rather than simple pastures. Bloom density for diverse pastures was also relatively lower than for the legume pastures. Our findings indicated that the diversification of pastures greatly increased pasture productivity, while legume pastures provided the highest bee benefit without penalizing lamb liveweight production in spring. 1. Introduction Demand for sustainable management of pasture-based animal production has been increasing in recent years due to rising consumer trends seeking natural and healthier food sources [1]. As a result, there has been a growing emphasis on pasture-based systems that deliver not only greater animal production but also offer broader benefits to the ecosystem and the environment. The main goal when choosing pasture mixtures for a given environment is to ensure persistent and productive stands that provide high-quality forages without causing any detrimental effects to the health of grazing animals. Diverse pastures containing a variety of herbs provide high-quality forage for both grazing animals and pollinating insects, which can help in creating a pasture system that can improve livestock production and health while providing a number of ecosystem benefits. The diversification of pastures has been shown to provide greater benefits to both grazing animals [2,3] and pollinators [4,5].i Received: 8 November 2023 Revised: 5 December 2023 Accepted: 15 December 2023 Published: 21 December 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). p The Pacific Northwest (PNW) region encompassing the states of Washington, Oregon, and Idaho between the Rocky Mountains to the east and the Cascade Range to the west [6] has a total of 11,680,118 hectares of pasturelands and rangelands [7–9], and is home to a wide variety of native bee species (approximately 900) [10]. This diverse bee community https://www.mdpi.com/journal/agronomy Agronomy 2024, 14, 24. https://doi.org/10.3390/agronomy14010024 Agronomy 2024, 14, 24 25 of 43 25 of 43 provides valuable pollination services, benefitting the region immensely [11–13]. The region is also one of the largest stocks of honey bee (Apis mellifera L.) for the pollination of California almonds, which is the most valued agricultural crop pollination service in the United States [14]. The physical and climatic characteristics of the PNW are highly conducive to establish honey bee pasture and pollinator habitat conservation. However, the PNW has lagged behind other regions (such as the Great Plains) in its efforts for pollinator habitat conservation [15]. Additionally, a combination of changing land-use away from nectar- and pollen-bearing field crops (e.g., clover seed) towards crops that do not provide these resources (e.g., wine grapes, and hazelnut) and hotter and drier summers are decreasing the extent summer bee forage, which is strongly linked the success of many bee species [16–18]. Thus, there is a need to investigate management practices to promote pollinators in pastures by incorporating multiple plant species and shifting from simple pastures to diversified pastures without reducing grazing animals’ forage quality and quantity [19,20]. With the diversified and legume pastures proving to be beneficial to lamb growth and bee health, creating a dual-use pasture system that benefits them both is of great need. Even small improvements in grassland diversity enhance the delivery of pollination services [4]. Thus, it is possible to design pasture mixtures and manage them for high animal production and pasture persistence, as well as nectar and pollen for bees.i Notable examples of plant species suitable for the PNW with the potential to benefit pollinators as well as grazing animals include Trifolium repens (white clover), Medicago sativa (alfalfa) Lotus corniculatus (birdsfoot trefoil), Trifolium michelianum (balansa clover), Tri- folium pratense (red clover) as well as Cichorium intybus (chicory) [21]. 2.1. Site This study was conducted at Oregon State University Farm in Corvallis, Oregon (44◦34′ N, 123◦18′ W 78 m a.s.l.), in 2021 and 2022. All procedures were approved by the Institutional Animal Care and Use Committee (ACUP 0152). The soil in the site was primarily Amity silt loam with 0–3% slopes [28]. Soil tests conducted by Oregon State University Soils Health Laboratory indicated the site had the following conditions: organic matter, 6.4%; available P (Bray), 113 ppm; Ca, 12.5 meq/100 g; Mg, 4.5 meq/100 g; K, 280 ppm; and soil pH(w), 5.3. The combination of these species has the potential to increase forage production, animal liveweight gain, antiparasitic properties [3,22] as well as a nectar and pollen source [23]. Furthermore, pasture management can be adjusted to promote bee abundance and diversity without compromising animal production [24]. For instance, applying intermediate grazing inten- sity or closing pastures from grazing in late spring (stockpiling) can improve the floral bloom density and consequently bee visitations when the nectar source is most needed in summer [25]. Earlier closing can also help the regeneration of cool season annual legumes in dryland pastures [26,27]. y p Thus, in this study, we investigated the effects of pasture diversification and spring pasture closing dates on pasture production, lamb growth, bloom density and bee visita- tions utilizing pasture species with the potential to benefit pollinators as well as grazing animals. We hypothesized that diversified and legume pastures will provide greater lamb growth rates and increase pollen and nectar availability in summer as compared to simple pastures; and that the early closing of pastures will increase the abundance of blooms and therefore promote greater bee visitations. 2.2. Meteorological Conditions Overall, the mean air temperature followed a similar trend with long-term means (LTMs) except the mean monthly air temperatures were higher than usual in fall 2020 and colder in spring 2022 (Table 1). The rainfall was 65% less than long-term means in spring 2021, while it was 47% greater than usual in spring 2022. Agronomy 2024, 14, 24 26 of 43 Table 1. Monthly rainfall and mean daily air temperatures over two growing seasons. Months Air Temperature (◦C) Rainfall (mm) 2020/2021 2021/2022 LTM * 2020/2021 2021/2022 LTM * September 17.9 17.4 16.8 50 72 34 October 12.4 11.3 11.9 46 70 81 November 6.8 9.3 6.7 181 127 174 December 5.5 4.8 4.5 180 269 194 January 6.1 4.4 4.7 211 119 160 February 5.5 5.0 6.2 160 39 133 March 6.7 8.6 8.2 55 111 111 April 10.9 8.0 10.1 13 146 79 May 13.3 12.0 13.1 22 113 60 June 19.3 16.7 16.1 44 60 36 July 21.1 - 19.3 0 - 14 August 20.8 - 19.4 0 - 13 * LTM: Long-term means of air temperature and rainfall are for the period 1997–2018. Table 1. Monthly rainfall and mean daily air temperatures over two growing seasons. 2.3. Pasture Establishment and Experimental Design 2.3. Pasture Establishment and Experimental Design Pasture treatments were established in a 1.332 ha plot on October 1 in 2020. Prior to establishment, pasture paddocks were divided into four equal blocks to serve as replicates for the experiment. Each block was divided into 3 subplots, which were randomly allocated to a combination of (1) simple, (2) diverse and (3) legume pastures, giving a total of 12 plots, each covering an area of 0.111 hectares (30 m × 37 m.). Pastures were further divided into three equal subplots (12.3 m × 30 m.) to apply rotational grazing. These subplots also served as different closing date (early, mid and late) treatments in both years (see Section 2.4). The experimental design was a split plot design where the pasture mixtures were the main plot and the closing dates were the subplot treatments. Pasture mixtures were sown with a no-till seeder (Land Pride, Great Plain Manufacturing, Salina, KS, USA). All the legumes were inoculated with appropriate Rhizobia strains. The species sown and the seeding rates (kg ha−1) of each pasture mixture are presented in Table 2. 2.4. Grazing Management 2.4. Grazing Management Weaned, 2.5-month-old Polypay lambs (mixed sex) were blocked by weight (21.4 ± 4.04 kg) in 2021 and (25 ± 4.53 kg) in 2022, and by sex, and then randomly as- signed to treatments. A put-and-take grazing system was implemented, allowing regulator animals to move between treatment plots and be removed as needed with the changing supply of forage [29]. Each treatment had a core group of 3–6 weaned lambs (testers) with 3 spare lambs (regulators) used in a put-and-take grazing system to match feed demand with fluctuating supply. Lambs had free access to forage, mineral blocks, and clean water in troughs connected to a water supply. All pasture mixtures were rotationally grazed with a flock of lambs in each plot grazing one subplot while the other two subplots were rested. Stocking density of the plots was dependent on the availability and growth rates of forages within the plot and ranged from 27.0 to 63.1 lambs per hectare. In 2021, sixty-five lambs grazed beginning in the simple and diverse pastures on 12 April, the legume plots were delayed due to a slower forage growth, and once sufficient, began grazing on 27 April. Lambs were moved between their respective subplots every 7–10 days depending on forage availability in the subplot currently being grazed, and the availability in the subplot the lambs were intended to move into. To quantify the effect of the stockpiling strategy on the nectar production of pastures in June-August and potential of self-regeneration of annual legumes, pastures were closed on 12–14 May, 26–28 May or 8 June. Following spring grazing, pastures were destocked for stockpiling the forages and providing forage for bees until Mid-July-August. Stockpiled pastures were grazed by dry ewes in August before the fall rains activated the pasture growth and germination of self-regenerating annual legumes. In spring 2022, grazing started on 28 March and ran through 16 June. The closing dates were 27 May, 6 June, and 16 June. 2.6. Measurements 2.6.1. Pasture Dry Matter (DM) Production 2.2. Meteorological Conditions Urea (46-0-0) fertilizer was applied at a rate of 50 kg N ha−1 at sowing in fall of 2020. In spring 2021 and 2022, prior to the commencement of grazing, urea sulfate (Ureasul, Wilco © Wilco-Winfield LLC, Mt. Angel, OR, USA) fertilizer (33-0-0-12) was applied at 50 kg N ha−1. In January 2020 and December 2021, post-emergence, grass-specific herbicide (26.4—Clethodim) was applied at a 443 mL ha−1 ai. rate in legume pastures to kill the volunteer grass weeds. Table 2. Species grown in each treatment type and seeding rates used (kg ha−1) in the experiment. Species Scientific Names Variety Sowing Rate kg ha−1 Simple Diverse Legume Perennial Ryegrass (G) Lolium perenne Albion 15 10 Festulolium (G) Festulolium Perseus 5 5 Orchard Grass (G) Dactylis glomerata Quickdraw 5 3 White Clover (L) Trifolium repens Domino 4 2 3 Subterranean Clover (L) Trifolium subterraneum Campeda and Denmark 20 10 10 Red Clover (L) Trifolium pratense Dynamite 3 Balansa Clover (L) Trifolium michelianum Savi Paradana 2 2 Berseem Clover (L) Trifolium alexandrinum Frosty 2 5 Birdsfoot Trefoil (L) Lotus corniculatus Bruce 5 3.5 Bigleaf Trefoil (L) Lotus pedunculatus Not Specified 3.5 White Sweet Clover Melilotus alba Hubam 1 Agronomy 2024, 14, 24 27 of 43 Table 2. Cont. Species Scientific Names Variety Sowing Rate kg ha−1 Simple Diverse Legume Chicory (H) Cichorium intybus Antler 2 Plantain (H) Plantago lanceolata Boston 4 Yarrow (H) Achillea millefolium Not Specified 0.2 Phacelia (H) Phacelia tanacetifolia Angelia 1 TOTAL 49 46.2 31 G = Grass; L = Legume; H = Herbs. Table 2. Cont. G = Grass; L = Legume; H = Herbs. 2.5. Management of Plots Outside Spring Experimental Period Two groups of 30 mature Polypay ewes were turned onto pasture plots where each plot was only grazed one time, with simple and diverse pasture plots being grazed for two days at a time, and legume plots grazed one day at a time before moving to the next plot from 9 August and 19 August 2021. In Fall 2021, forty mature Polypay ewes were placed into simple and diverse pastures (five in each plot) and grazed until 12 November 2021. Legume plots were not grazed at this time due to a lack of forage availability. Plots were mown in summer 2022, while they were spelt in fall due to lack of forage growth. 2.6. Measurements 2.6.4. Botanical Composition and Nutritive Value of Pasture on Offer 2.6.4. Botanical Composition and Nutritive Value of Pasture on Offer A total of 30–50 pluck samples, representative of the forage available to the lambs, were hand collected in a random pattern across the subplots prior to moving the lambs into them. Pluck samples were sorted into subsamples and sorted into their respective botanical species (i.e., grass, legume, herb, weed, and dead material) and dried in an oven at 65 ◦C for 48 h. Dry weights of the pluck sample and botanical species were then used to calculate the percentage botanical composition of the samples. A well-mixed bulk sample was ground in a Wiley mill using a 1 mm stainless steel screen (Thomas/Wiley, Swedesboro, NJ, USA) for chemical analyses using the Association of Official Analytical Chemists methods [30]. The ground sample was used to perform chemical analyses. NDF and ADF were performed following the methods described by Van Soest et al. [31] using an Ankom200 Fiber Analyzer (ANKOM Technology Corp., Macedon, NY, USA). Crude protein (CP) was analyzed using a LECO N analyzer (LECO FP828, MI, USA). Ash was analyzed (method 942.05), and ether extract (EE) was analyzed (method 920.39). Total digestible nutrients (TDN%) was calculated using the following formula: TDN% = 82.38 −(0.7515 × ADF). 2.6.1. Pasture Dry Matter (DM) Production Pasture dry matter (DM) was measured using 1 m2 exclosure cages during active growth in spring, summer, and fall. No samples were collected in the winter months as there was minimal pasture growth due to cold temperatures and heavy rains. Forage growth was harvested from under each cage using a rectangular 0.25 m2 quadrat and electric clippers to cut the forage to a stubble height of 3 cm. After cutting the forage, a new area of each subplot was mowed to a stubble 3 cm height and the exclosure cage was Agronomy 2024, 14, 24 28 of 43 28 of 43 placed there for the next growing period. Forage cuts obtained from under the cages were then sub-sampled and sorted into their respective botanical species before being dried in an oven at 65 ◦C for 48 h and then weighed. Total DM production (kg DM ha−1) was divided by the number of days elapsed since the previous harvest to determine the mean daily growth rates (kg DM ha−1 day−1). 2.6.2. Lamb Production To determine LWG (g head−1 day−1), individual tester animals were weighed prior to and following each grazing period. LWG was determined by the change in weight between each live weight measurement date. Further, liveweight production (LWP) (kg ha−1 d−1) was determined by taking the liveweight gain per head of tester lambs, multiplied by the number of testers plus regulator lambs per hectare. 2.6.5. Fecal Egg Counts Individual fecal samples (~10 g) were taken from the rectum of each lamb using a powder-free latex examination glove. Fecal egg counts (Strongyle) were obtained at random from test animals in each treatment at the start of the trial on 12 April 2021, and 28 March 2022. Previously, tested animals were then sampled again at the conclusion of the trial on 6 June 2021, and 18 June 2022. Samples were taken to the Oregon State University Veterinary Diagnostic Lab and viewed for eggs per gram (EPG). 2.6.3. Pasture Mass on Offer Pre- and post-grazing pasture mass (kg DM ha−1) was measured using a rising plate meter (PM; Jenquip, Feilding, New Zealand). Walking in a zig-zag pattern, 30 measure- ments were collected in subplots prior to lambs entering the subplots during rotation to obtain the pre-grazing pasture mass. Additionally, 30 measurements were taken in the sub- plots after lambs rotated out of them to obtain the post-grazing pasture mass. Rising plate meter measurements were calibrated by regression against the herbage mass using 0.25 m2 quadrats. Calibrations of herbage pasture mass were performed initially on 6 April 2021, with 1 quadrat in each simple and diverse subplot for a total of 24 data points. In 2022, calibration cuts were performed on 22 March with 1 quadrat in each plot for a total of 12 data points. Calibration cuts were repeated 6 July, and 27 July with 2 quadrats in each subplot for a total of 72 data points. 2.6.7. Seedling Numbers As an indicator of the effects of spring grazing management on persistence of bal- ansa and subterranean clover, seedling numbers were counted in three randomly placed 0.01 m2 quadrats in each closing date plot in all three pasture types on 4 November 2020, 1 November 2021 and 1 October 2022. 2.6.6. Bloom Density and Bee Visitations Bloom density was measured using a rectangular 0.25 m2 quadrat. Subplots were divided into numbered areas and a number randomizer was used to determine each place- ment. The quadrat was placed in each subplot three times and each flower in bloom was Agronomy 2024, 14, 24 29 of 43 29 of 43 counted and organized into its respective species. In 2021, measurements were taken in each subplot weekly starting on 23 April and concluding on 27 July. In 2022, measurements were taken weekly in each subplot starting on 5 May and concluding on 3 August. Bee visi- tations were measured using a variable transect walk [32] in five- or ten-minute increments by walking back and forth in the subplots for the appointed time and counting bumble bees and honey bees when seen visiting a flower (five minutes with two people and 10 min with one person). In 2021, bee visitations were measured on a weekly basis in each closing date subplot, starting on 23 April and concluding on 27 July. In 2022, visitations were measured on a weekly basis in each legume subplot and rotating through simple and diverse subplots weekly. In 2022, bee visit observations started on 16 May and concluded on 3 August. 2.7. Statistical Analysis Annual DM production was analyzed by ANOVA in a completely randomized design with four replicates (block). Pasture growth rates, botanical composition, LWG of lambs per head (g hd−1 d−1) and per hectare (kg ha−1 d−1) were analyzed by one-way ANOVA with four replicates for every measurement period. Pasture nutritive value and herbage mass on offer values were averaged across each rotation and analyzed in a complete randomized design with four replicates. Fecal EPG values were analyzed in a complete randomized design with four replicates (block), using pastures as the experimental unit. Bloom density, bee visitations and seedling numbers were analyzed in a split–split plot design where pasture mixtures were the main plot, closing dates were the sub-plot and the sampling period was sub-subplot. No statistical analysis was performed for the bloom composition since the comparison among treatments was not possible. This was because the blooming species were not common to each treatment. Significant differences among treatment means were compared by Fisher’s protected least significant difference at p < 0.05. The computations were conducted using GENSTAT statistical software (v. 22). 3.1. Pasture Dry Matter (DM) Production Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage gr rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars repr standard error mean (SEM). Diﬀerent letters show signiﬁcant diﬀerences among pasture treatm (p < 0.05) (a). * = when ANOVA was signiﬁcant (p < 0.05) (b). Figure 1. Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage growth rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars represent standard error mean (SEM). Different letters show significant differences among pasture treatments (p < 0.05) (a). * = when ANOVA was significant (p < 0.05) (b). Figure 1. Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage gro rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars repre standard error mean (SEM). Diﬀerent letters show signiﬁcant diﬀerences among pasture treatm (p < 0.05) (a). * = when ANOVA was signiﬁcant (p < 0.05) (b). Figure 1. Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage growth rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars represent standard error mean (SEM). Different letters show significant differences among pasture treatments (p < 0.05) (a). * = when ANOVA was significant (p < 0.05) (b). Percentage of legumes, sown grasses, total forbs (legumes and herbs), dead mate and weeds for each pasture type for both 2021, and 2022 are presented in Figure 2 Percentage of legumes in both the diverse and simple pasture remained similar acr both years (Figure 2a), while the percentage of sown grass in the diverse and simple p ture also remained similar across both years (all p > 0.05; Figure 2c). The percentag total forbs in the diverse and simple pastures remained similar through 2021 and up u April of 2022 when there was an increase (p < 0.05) in the forb content of the diverse p tures. Total forbs remained highest across both years in the legume pasture and ha gradual decline from May of 2021 until July of 2022. Dead material was similar acros Percentage of legumes, sown grasses, total forbs (legumes and herbs), dead material, and weeds for each pasture type for both 2021, and 2022 are presented in Figure 2a–e. 3.1. Pasture Dry Matter (DM) Production In the 2020/2021 growing season, the total annual dry matter yield (DMY) of diverse pastures was greater (p < 0.01) than the DMY of simple grass-dominated, and legume pastures (Figure 1a). In the 2021/2022 growing season, the total annual DMY of simple and diverse pastures were similar. However, the legume pastures produced 27–30% less than simple and diverse pastures. p p In 2021, the herbage growth rates (kg ha−1 d−1) displayed a sharp increase at the beginning of April with diverse pastures being the greatest (10.1 kg ha−1 d−1) followed by simple pastures (8.8 kg ha−1 d−1) and no growth for legumes due to the delay of growth (p < 0.01; Figure 1b). In early May, the trend continues with diverse pastures having the greatest growth (80.5 kg ha−1 d−1) followed by simple pastures (76.6 kg ha−1 d−1) and then legume pastures (15.1 kg ha−1 d−1) (p < 0.01). Through the rest of the year until early December, the herbage growth continued to decline, However, the amounts per treatment were insignificant (p > 0.01). In 2022, from late March until late May, there was an increase in herbage growth among all treatments, but the total growth between treatment types was insignificant (p > 0.01). However, beginning in late May, differences between treatment types were significant (p < 0.01) where diverse pastures were greater than simple pastures, followed by legume pastures (78.5, 75.1 and 50.2 kg ha−1 d−1), respectively. In late June, herbage growth continued to increase with simple pastures having the greatest growth, Agronomy 2024, 14, 24 30 of 43 ively. In the grea 30 of 43 ively. In the grea followed by diverse pastures, and then legume pastures (91.5, 82.4, and 59.1 kg ha−1 d−1), respectively (p < 0.01). growth, followed by diverse pastures, and then legume pastures (91.5, 82.4, and 59.1 ha−1 d−1), respectively (p < 0.01). respectively (p < 0.01). Figure 1. Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage gro rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars repre Figure 1. Total annual accumulated dry matter (DM) production (a) (t ha−1 y−1) and herbage growth rates (b) (kg ha−1 d−1) for simple, diverse and legume pastures from 2010 to 2022. Bars represent p y (p ) Figure 1. 3.1. Pasture Dry Matter (DM) Production Percentage of legumes in both the diverse and simple pasture remained similar across both years (Figure 2a), while the percentage of sown grass in the diverse and simple pasture also remained similar across both years (all p > 0.05; Figure 2c). The percentage of total forbs in the diverse and simple pastures remained similar through 2021 and up until April of 2022 when there was an increase (p < 0.05) in the forb content of the diverse pastures. Total forbs remained highest across both years in the legume pasture and had a gradual decline from May of 2021 until July of 2022. Dead material was similar across all treatment types through both years (all p > 0.05; Figure 2e). The percentage of weeds did not differ between simple and diverse pastures, but legume pastures had the greatest (p < 0.05) percentage of weeds which stayed constant through May of 2022 before peaking to its highest point in July of 2022. Agronomy 2024, 14, 24 31 of 43 peaking Figure 2. Botanical composition (%) of pastures showing contributions from (a) legumes, (b) to forbs (legumes + herbs), (c) sown grasses, (d) weeds and (e) dead material in simple, diverse a legume pastures from 2021 to 2022. Error bars represent maximum SEM. Figure 2. Botanical composition (%) of pastures showing contributions from (a) legumes, (b) total forbs (legumes + herbs), (c) sown grasses, (d) weeds and (e) dead material in simple, diverse and legume pastures from 2021 to 2022. Error bars represent maximum SEM. Figure 2. Botanical composition (%) of pastures showing contributions from (a) legumes, (b) to forbs (legumes + herbs), (c) sown grasses, (d) weeds and (e) dead material in simple, diverse a legume pastures from 2021 to 2022. Error bars represent maximum SEM. Figure 2. Botanical composition (%) of pastures showing contributions from (a) legumes, (b) total forbs (legumes + herbs), (c) sown grasses, (d) weeds and (e) dead material in simple, diverse and legume pastures from 2021 to 2022. Error bars represent maximum SEM. 3.2. Lamb Production 3.2. Lamb Production In both years, the lamb growth rates (g hd−1 d−1) from legume pastures were grea (p < 0.01) than the other pasture types (Table 3). However, the LWP (kg ha−1 d−1) from three pasture types were comparable (p = 0.23). Lambs grew faster in the ﬁrst half of spr in both years. Their growth rates reduced as the season progressed. The diverse pastu provided greater (p < 0.01) lamb total liveweight production than simple pastures a legume pastures by 38 and 220 kg ha−1, respectively, in spring 2021. While the total la liveweight gains were 745, 628 and 581 kg ha-1 for diverse, simple and legume pastur respectively in 2022 the diﬀerence was not signiﬁcant (p = 0 20) In both years, the lamb growth rates (g hd−1 d−1) from legume pastures were greater (p < 0.01) than the other pasture types (Table 3). However, the LWP (kg ha−1 d−1) from all three pasture types were comparable (p = 0.23). Lambs grew faster in the first half of spring in both years. Their growth rates reduced as the season progressed. The diverse pastures provided greater (p < 0.01) lamb total liveweight production than simple pastures and legume pastures by 38 and 220 kg ha−1, respectively, in spring 2021. While the total lamb liveweight gains were 745, 628 and 581 kg ha-1 for diverse, simple and legume pastures, respectively, in 2022, the difference was not significant (p = 0.20). Agronomy 2024, 14, 24 32 of 43 Table 3. Mean liveweight gain (g hd−1 d−1) and production (kg ha−1 d−1) of lambs from simple, diverse and legume pastures over three grazing periods in 2021 and two grazing periods in 2022. 3.4. Fecal Egg Counts 3.3. Pasture Mass, Botanical Composition and Nutritive Value of Pasture on Offer 3.3. Pasture Mass, Botanical Composition and Nutritive Value of Pasture on Offer Pre-grazing pasture mass (kg DM ha−1) in spring of 2021 and 2022 is presented in Figure 3a,b. In 2021, pre-grazing herbage mass of pasture mixtures ranged from 1812 kg DM ha−1 to 2802 kg DM ha−1 and was comparable across pasture types at each rotation (all p > 0.05). In 2022, simple and diverse pastures had a similar pre-grazing herbage mass, ranging from 1740 to 1976 kg DM ha−1. However, legume pastures had approximately 300 kg DM ha−1 less (p < 0.0) herbage on offer than grass-based pastures throughout the grazing period. g g g p The proportion of forbs (legume + herbs) in herbage on offer in spring of 2021 and 2022 are presented in Figure 3c,d. In 2021, the forb content of diverse pasture was greater (p < 0.05) than that of simple pastures during rotations 1 and 3 but similar (25.7 vs. 14.5%) in rotation 2. Legume pastures had a substantially greater (p < 0.01) forb content than other pastures, ranging from 81.8% in rotation 2 to 65.3% in rotation 3. In spring 2022, the forb content of simple pastures was negligible (>2%), while diverse pastures had over 40% forb content in rotation 1. Legume pastures had the highest forb content during the same rotation period. The forb content of the simple pastures went up to 10% in the second rotation, but this was still substantially lower (p < 0.01) than the forb content in diverse and legume pastures. In rotation 3, the forb content of the legume and diverse pastures was comparable (51.1% vs. 59.4%) but greater than simple pastures which had only a 15.4% forb content. In 2021, TDN content (%) of pasture on offer remained similar (all p > 0.05) across treatment types through all rotations, except for the diverse pasture which showed a decrease in TDN in the third rotation (Figure 3e,f). In 2022, the TDN content of pastures ranged from 62.5 to 68.2% and it was relatively stable across the rotations. The TDN content of legume pastures tended to be greater than simple pastures in rotation 1 (p = 0.05) and rotation 2 (p = 0.07). All pastures had a similar (p = 0.72) TDN content in rotation 3. 3.2. Lamb Production 3.2. Lamb Production Year Period Simple Diverse Legume SEM p Liveweight gain (g hd−1 d−1) 12–27 April 220 235 - 6.3 0.19 2021 27 April–10 May 224 252 316 21.4 0.05 10 May–8 June 172 b 174 b 259 a 14.3 0.01 2022 28 March–6 May 181 c 220 b 291 a 10.6 0.01 6 May–16 June 82 79 140 21.6 0.15 Liveweight production (kg ha−1 d−1) 12–27 April 11.9 12.7 - 0.35 0.19 2021 27 April–10 May 12.1 13.6 12.1 1.66 0.76 10 May–8 June 9.3 9.4 9.9 0.80 0.83 2022 28 March–6 May 11.0 13.9 11.8 1.09 0.24 6 May–16 June 5.0 5.0 4.0 0.96 0.69 Total liveweight production (kg ha−1) 2021 12 April–8 June 666 b 704 a 446 c 38.7 0.01 2022 28 March–16 June 628 745 581 58.7 0.20 LWG = liveweight gain; LWP = Liveweight production (kg/hectare/day) LWG × stocking rate (number of animals per hectare); Total LWP = LWP × grazing duration (days); Different letters within rows show significant differences (p < 0.05). SEM = Standard error of means. 3.3. Pasture Mass, Botanical Composition and Nutritive Value of Pasture on Offer 3.4. Fecal Egg Counts Agronomy 2024, 14, 24 34 of 43 34 of 43 3.5. Bloom Density and Bee Visitations 3.5. Bloom Density and Bee Visitations Averaged across the entire grazing period, bloom density of pastures was 1.9, 8.7 and 50.4 flowers per m2 for simple, diverse, and legume pastures, respectively, in 2021 (Figure 4a,b). A pasture treatment × period interaction (p < 0.01) was detected as the bloom density for simple pastures was low and relatively stagnant as compared to diverse and legume pastures. In both 2021 and 2022, simple and diverse pastures had a similar trend for bloom density. In 2021, legume pastures had a high bloom density at the beginning of the season and began to gradually decline starting in late May until terminating in July. In 2022, the bloom density in legume pastures remained somewhat constant through- out, aside from mid-June and mid-July, where bloom density decreased (Figure 4c,d). There was a three-way interaction (p < 0.01) among pasture treatment, closing date, and period for bloom density for the grazing period following the first closing date. Neither the closing date, nor the period had any effect on simple grass pastures for bloom density. Earlier closing of legume pastures led to greater bloom density as compared to mid and late closing dates, but the difference in bloom density induced by closing dates disappeared after the first week of July. EVIEW 12 of 21 Figure 4. Bloom density as aﬀected by pasture type (a,c) and closing date (b,d) (CD). Bars repres SEM for pasture × closing date × period interaction. Figure 4. Bloom density as affected by pasture type (a,c) and closing date (b,d) (CD). Bars represe SEM for pasture × closing date × period interaction. Figure 4. Bloom density as aﬀected by pasture type (a,c) and closing date (b,d) (CD). Bars represent SEM for pasture × closing date × period interaction. Figure 4. Bloom density as affected by pasture type (a,c) and closing date (b,d) (CD). Bars represent SEM for pasture × closing date × period interaction. Bee visitation across 2021 was similar between simple and diverse pastures, with the highest (p < 0.01) visitations being displayed in the legume pastures (Figure 5a,b). A three- way interaction (p < 0.01) was detected among pasture treatment, closing date, and period for bee visitations. Bee visitations in legume pastures had a peak in early May and began Bee visitation across 2021 was similar between simple and diverse pastures, with the highest (p < 0.01) visitations being displayed in the legume pastures (Figure 5a,b). 3.4. Fecal Egg Counts In 2021, EPG counts did not differ among the three pasture mixtures (all p > 0.05) (Table 4). In 2022, EPG counts were lower (p < 0.05) for lambs grazing diverse pastures than simple pastures at the end of the grazing trial. Agronomy 2024, 14, 24 Agronomy 2024, 14, x FO 33 of 43 10 of 21 Figure 3. Pre-grazing pasture mass (a,b) and forb (legume + herb) content (c,d) and total digestib nutrients (TDN) (e,f) of pasture on oﬀer in three rotations in spring 2021 and 2022. Error bars re resent SEM. 3.4. Fecal Egg Counts I E d d d ﬀ h h ll Figure 3. Pre-grazing pasture mass (a,b) and forb (legume + herb) content (c,d) and total digestib nutrients (TDN) (e,f) of pasture on offer in three rotations in spring 2021 and 2022. Error bar represent SEM. Table 4. Total egg counts (Strongyle) of lambs grazing simple, diverse and legume pastures in sprin 2021 d i 2022 Figure 3. Pre-grazing pasture mass (a,b) and forb (legume + herb) content (c,d) and total digestible nutrients (TDN) (e,f) of pasture on oﬀer in three rotations in spring 2021 and 2022. Error bars rep- resent SEM. Figure 3. Pre-grazing pasture mass (a,b) and forb (legume + herb) content (c,d) and total digestible nutrients (TDN) (e,f) of pasture on offer in three rotations in spring 2021 and 2022. Error bars represent SEM. 3.4. Fecal Egg Counts In 2021, EPG counts did not diﬀer among the three pasture mixtures (all p > 0.05) Table 4. Total egg counts (Strongyle) of lambs grazing simple, diverse and legume pastures in spring 2021 and in 2022. 3.4. Fecal Egg Counts In 2021, EPG counts did not diﬀer among the three pasture mixtures (all p > 0.05) Table 4. Total egg counts (Strongyle) of lambs grazing simple, diverse and legume pastures in spring 2021 and in 2022. , g p ( p ) (Table 4). In 2022, EPG counts were lower (p < 0.05) for lambs grazing diverse pastures than simple pastures at the end of the grazing trial. Dates Diverse Simple Legume SEM p 12 April 2021 212 288 129 91.9 0.51 6 June 2021 92 138 115 29.6 0.58 28 March 2022 622 1030 331 285.5 0.29 18 June 2022 289 a 1247 b 621 ab 197.2 0.05 Different letters within rows show significant differences (p < 0.05). 3.5. Bloom Density and Bee Visitations A three-way interaction (p < 0.01) was detected among pasture treatment, closing date, and Agronomy 2024, 14, 24 35 of 43 35 of 43 period for bee visitations. Bee visitations in legume pastures had a peak in early May and began to decrease sharply starting in mid-May until they were comparable to simple and diverse pastures in early July. The closing date treatments did not have any effect on the bee visitations for simple and diverse pastures, but bee visitations were higher with earlier closing (first) then mid (second) and late (third) closing for legume pastures. In 2022, bee visitations were similar (p > 0.05) across treatments, with legume pastures showing a slightly higher number of visitations (Figure 5c). In 2022, no discernable differences among closing dates were observed for the bee visitations (Figure 5d). IEW 13 of Figure 5. Honey bee visits as aﬀected by pasture type (a,c) and closing date (b,d) (CD; (b,d)). represent SEM for pasture × closing date × period interaction. Figure 5. Honey bee visits as affected by pasture type (a,c) and closing date (b,d) (CD; (b,d)). Bar represent SEM for pasture × closing date × period interaction. Figure 5. Honey bee visits as aﬀected by pasture type (a,c) and closing date (b,d) (CD; (b,d)). Ba represent SEM for pasture × closing date × period interaction. Figure 5. Honey bee visits as affected by pasture type (a,c) and closing date (b,d) (CD; (b,d)). Bars represent SEM for pasture × closing date × period interaction. In 2021, balansa clover dominated the legume and diverse pastures through mi June (Figure 6a). In the legume pastures, there was then a shift to white and berseem cl ver beginning to dominate the pasture, and in late July, weeds and red clover made a appearance. In the diverse pastures, white clover was the main component of the bloom ing ﬂowers. In mid-June until mid-July, plantain and chicory bloom predominated. In th simple pasture, sub clover dominated through mid-May, when white clover then too over through mid-July. In 2022, balansa clover was the prominent clover in the legum pastures through mid-June, where there was then a shift to white clover until mid-Ju when weeds took over. However, the broadleaved weeds (predominantly Crepis spp. Toward the end of July, broadleaved weeds were the plants that predominantly contributed to the blooms. REVIEW 14 of Toward the end of July, broadleaved weeds were the plants that predominantly contributed o the blooms. Toward the end of July, broadleaved weeds were the plants that predominantly contribut to the blooms. Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloom com- position was only monitored in legume pastures in 2022. B. weeds= broadleaved weeds; BFT = birds- foot trefoil. Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloom composition was only monitored in legume pastures in 2022. B. weeds = broadleaved weeds; BFT = birdsfoot trefoil. Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloom com- position was only monitored in legume pastures in 2022. B. weeds= broadleaved weeds; BFT = birds- foot trefoil. Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloom composition was only monitored in legume pastures in 2022. B. weeds = broadleaved weeds; BFT = birdsfoot trefoil. 3.6. Closing Date Eﬀect on Persistence of Annual Legumes 3.6. Closing Date Effect on Persistence of Annual Legumes 3.5. Bloom Density and Bee Visitations an Rumex crispus) were also a major component of the blooming plants ranging approx In 2021, balansa clover dominated the legume and diverse pastures through mid- June (Figure 6a). In the legume pastures, there was then a shift to white and berseem clover beginning to dominate the pasture, and in late July, weeds and red clover made an appearance. In the diverse pastures, white clover was the main component of the blooming flowers. In mid-June until mid-July, plantain and chicory bloom predominated. In the simple pasture, sub clover dominated through mid-May, when white clover then took over through mid-July. In 2022, balansa clover was the prominent clover in the legume pastures through mid-June, where there was then a shift to white clover until mid-July when weeds took over. However, the broadleaved weeds (predominantly Crepis spp. and Rumex crispus) were also a major component of the blooming plants ranging approximately from 20 to 40%. Agronomy 2024, 14, 24 36 of 43 Toward the end of July, broadleaved weeds were the plants that predominantly contribute to the blooms. Agronomy 2024, 14, x FOR PEER REVIEW 14 o Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloom c position was only monitored in legume pastures in 2022. B. weeds= broadleaved weeds; BFT = bi foot trefoil. 3.6. Closing Date Eﬀect on Persistence of Annual Legumes The number of established sub clover plants on 4 November 2020 was 100, 146 a 208 plants per m2 in diverse, legume and simple pastures, respectively, whereas the nu ber of established balansa clover numbers was 96 and 104 in diverse and legume pastu Figure 6. Bloom composition (%) of legume (a,d), diverse (b) and simple pastures (c). Bloo composition was only monitored in legume pastures in 2022. B. weeds = broadleaved weed BFT = birdsfoot trefoil. 3.6. Closing Date Effect on Persistence of Annual Legumes The number of established sub clover plants on 4 November 2020 was 100, 146 an 208 plants per m2 in diverse, legume and simple pastures, respectively, whereas the numb of established balansa clover numbers was 96 and 104 in diverse and legume pasture Toward the end of July, broadleaved weeds were the plants that predominantly contributed to the blooms. REVIEW 14 of 2 3.6. Closing Date Eﬀect on Persistence of Annual Legumes 3.6. Closing Date Effect on Persistence of Annual Legumes The number of established sub clover plants on 4 November 2020 was 100, 146 and 208 plants per m2 in diverse, legume and simple pastures, respectively, whereas the num- ber of established balansa clover numbers was 96 and 104 in diverse and legume pastures, respectively (Table 5). The subterranean clover seedling numbers were similar in legume and di erse pastures but they were lower (p < 0 01) than simple pastures on 1 No ember The number of established sub clover plants on 4 November 2020 was 100, 146 and 208 plants per m2 in diverse, legume and simple pastures, respectively, whereas the number of established balansa clover numbers was 96 and 104 in diverse and legume pastures, respectively (Table 5). The subterranean clover seedling numbers were similar in legume Agronomy 2024, 14, 24 37 of 43 37 of 43 and diverse pastures, but they were lower (p < 0.01) than simple pastures on 1 November 2021. Early and mid-closing dates resulted in a similar number of seedlings, but late closing resulted in a fewer (p < 0.05) number of established sub clover seedlings. While the closing dates in previous spring did not affect (p = 0.90) the established number of balansa clover seedlings, the legume pasture had a substantially greater (p < 0.01) number of seedlings than diverse pastures. In fall 2022, the effect of pasture type on subterranean clover seedlings was not significant, while closing earlier on 27 May resulted in a greater (p < 0.05) than established number of seedlings than closing on 16 June. The number of established balansa clover seedlings was virtually none in diverse pastures. The effect of closing date on balansa clover seedling in legume pastures was not significant (p = 0.93). Table 5. Effects of closing dates on seedling numbers (plant m2) of subterranean and balansa clover in subsequent fall 2021 and in 2022. 3.6. Closing Date Eﬀect on Persistence of Annual Legumes 3.6. Closing Date Effect on Persistence of Annual Legumes Subterranean Clover Year Closing Date Simple Diverse Legume SEM p Values 1 TRT CD TRT × CD 12–14 May 558 217 500 2021 26–28 May 1142 425 650 128.4 0.01 0.01 0.43 8 June 908 500 617 27 May 250 300 313 2022 6 June 131 188 206 69.9 0.23 0.05 0.54 16 June 150 50 269 Balansa Clover 12–14 May - 167 2300 2021 26–28 May - 125 2531 215.3 0.01 0.90 0.80 8 June - 196 2406 27 May - 413 2022 6 June - 456 97.5 0.93 16 June - 456 1 TRT = Treatments; SEM = Standard error of means; CD = Closing dates. 4.1. Forage Production Total annual forage production was strongly influenced by the erratic weather con- ditions over the two-year study period. Specifically, in spring of 2021, there was a 65% decrease in rainfall when compared to the LTM. In contrast, in spring of 2022, there were wetter conditions, with a 46% increase in rainfall when compared to the LTM. Furthermore, slightly warmer and wetter conditions in fall of 2021 induced favorable conditions for germination and growth for the self-regenerating annual legumes. This in turn helped legumes with higher forage production in early spring compared to the previous year, allowing for the earlier commencement of the grazing season. In contrast, the late onset of rainfall in fall of 2022 resulted in negligible forage growth for that season and could not support fall grazing.i The value of diversification of pastures using pasture forbs was particularly high- lighted in the drier-than-usual spring conditions of 2021. However, the same benefit was not as obvious in a wet spring when simple and diverse pastures had a similar DMY in 2022. This was possibly due to the greater summer growth potential of forbs with deep tap roots, in particular when the soil moisture was limiting. A number of studies have noted the same value of diversification. For example, Woodward et al. [33] indicated that when simple grass pastures were diversified with other plants such as clovers, chicory, and plantain, there was a greater DM production through the warmer summer months as the forb species grew better than the grasses alone. Similar to the findings of the current study, Daly et al. [34] also reported that annual forage production in diverse pastures containing Agronomy 2024, 14, 24 38 of 43 38 of 43 chicory, plantain, small burnet (Sanguisorba minor L.) and yarrow were significantly and consistently greater than that of the ryegrass–white clover pasture. The greater total produc- tion was also attributed to higher summer growth in diverse than grass pastures. However, the value of diversification of pastures may not only be limited to dryland conditions. It is of note that Black et al. [35] recorded greater forage production from diverse rather than simple pastures in irrigated conditions in New Zealand as well, indicating that aside from the availability of water, other factors such as botanical composition, temperature, and soil conditions play a major role. 4.1. Forage Production p y j A further reason for diverse pastures not exhibiting their comparative advantage in spring 2022 may be due to the cooler temperatures and adverse soil conditions caused by excess rain. It is probable that forbs, in particular, chicory, is known for a lack of tolerance for waterlogged soil conditions [36]. It is of note that the colder, wetter conditions of spring also led to decreased legume production by nearly 1000 kg ha−1 in comparison to the previous year. Overall, DM production of legume pastures was substantially lower than both simple and diverse pastures. This is a common instance and can also be observed in other studies comparing different legume mixtures [37] and as cold conditions reduce growth and yield of legumes [38] as well as shade [39]. Along with legumes having difficulty tolerating waterlogging (except balansa clover), legume plants spend a significant amount of energy to fix N rather than increasing yield [40]. In 2022, the legume pastures became heavily intruded by broadleaf weeds, which can be attributed to having less competition with the legumes due to them struggling under the waterlogging conditions, and also the more acidic soil conditions (pH 5.1). Additionally, Campbell and Grice [41] have reported that weed species are strongly competitive for resources, and that the grazing of palatable pasture species reduces competition for the weeds. Reduced competition for weeds due to grazing of palatable species was also observed in the study by Andrew [39]. However, balansa clover tended to perform better than the other clovers in both years and could be attributed to its capability of long-term persistence in pastures [42]. Additionally, it has broad environmental adaptation, and does well in a large range of soil pH, including more acidic soils down to pH 5.0. Lastly, balansa clover does well tolerating waterlogging, and has even been observed to increase root growth in some waterlogging [42,43]. In contrast, birdsfoot trefoil had a minimal appearance in 2021, and had no appearance in 2022 despite it being highly tolerant to waterlogged soil conditions. However, it is highly probable that birdsfoot trefoil was outcompeted by other forage species, due to its typical slow establishment. In contrast to its success in mixed legume pastures, balansa clover populations in diverse pastures declined dramatically over the two-year grazing study. 4.1. Forage Production The decline in balansa clover presence in diverse pastures can be attributed primarily to its poor competitiveness when sown alongside grasses, possibly because of its small seedling size. Associated with grazing management, the regrowth of balansa clover faces significant challenges when there is insufficient space for seeds to germinate in the fall. Ref. [44] reported a substantial decrease in balansa clover seedling populations during the second growing season in tall fescue-based hill country pastures. This drop was attributed to competition with grasses and the smaller seed size of balansa clover in comparison to the initially sown seed size, resulting in a smaller seedling size. Ref. [45] reported similar findings in New Zealand dryland pastures, suggesting that balansa clover might need to be overseeded every three years to keep its populations at desirable levels. 4.2. Livestock Production Overall, diverse pastures provided a greater LWP (kg ha−1) when compared to simple grass–clover pastures. This was also reported by Andrew [39] in an agrivoltaic pasture system where lambs grazing diverse pastures had a greater LWG than those grazing simple grass pastures. Previous reports from Golding et al. [3] indicated that legumes and pastures containing herbs provide greater feeding values for animals when compared to simple grass pastures. Further, lambs grazing legume pastures had the highest LWG over both simple Agronomy 2024, 14, 24 39 of 43 and diverse pastures in both years, even when the legume pastures had a high percentage of weeds in 2022, which points to selective grazing of the lambs on legume pastures in those weedy conditions. The superior growth of lambs grazing legumes can be attributed to the high voluntary intake of legumes, as well as the high protein content that legumes provide [46]. Similarly, Speijers et al. [47] reported that lambs grazing legume pastures grew faster than those grazing perennial ryegrass pastures and required fewer days to reach the slaughter weight. Similar benefits of legume pastures for promoting greater animal performance were also reported with dairy cows where cows grazing legume pastures provided greater milk yield than grass pastures [48]. Comparing overall lamb performance between both years, 2022 showed a decreased performance both in LWG and LWP. The decreased production in year two can be attributed to fact that lambs were likely spending more energy staying warm in the cold temperatures experienced in year two rather than spending the energy on growing. This was found in a UK-based study where the survival rates of lambs and yearlings were decreased when the weather was colder, and windier, as well as cold weather causing stress to the lambs [49,50]. Another factor the lambs faced in 2022 were higher rates of parasite burdens, likely attributed to the wet conditions of the pastures. Younie et al. [51] discusses that, in northern temperate conditions, larvae and eggs can overwinter in forages and be ingested when grazing occurs at the beginning of the growing season, and typically have a higher incidence in wet conditions. When looking at parasite loads across all pasture types, the diverse pastures were successful in decreasing parasite loads across both years. The diverse pastures contained chicory which is known for its bioactive compounds and anthelmintic effects on grazing animals. Athanasiadou et al. 4.2. Livestock Production [52] looked at ewes grazing on grass compared to ewes grazing on chicory and the effects on parasite loads. What they found was those lambs grazing on chicory had a significantly lower fecal egg count, as well as higher liveweight gains than ewes grazing on grass [52]. Ref. [53] also looked into various pasture types consisting of ryegrass only, a ryegrass mix with legumes, and a mix including grasses, clovers, and chicory and plantain, and reported that the ewes and lambs grazing the diverse mix had greater body condition scores and LWG, as well as having a decreased parasite load and requiring less anthelmintics than the ewes and lambs grazing the grasses. 4.3. Bloom Diversity and Bee Visitation In both years, the legumes were able to provide nectar and pollen for the bees into mid- July. In 2021, there was a greater percentage of legumes throughout the season, whereas in 2022, although there were some legumes, the weeds became more prevalent in the beginning of July in the legume plots. Especially through June, in both the legume plots and the diverse plots, balansa clover was the most prevalent clover available for the bees to forage on. What is of most importance is finding a way to provide forage for bees in the late challenging summer months, where the bees have less forage availability, at a lower quality, and have increased competition for already scarce resources [54]. However, the cool season legumes used here in both the diverse and legume pastures were not effective in providing forage into the late summer months of July and August. This is of particular importance for the pollinator health because if bees can access floral resources from May to October, they will have higher reproductive rates and better winter survival [18,55]. To combat the decline of bee populations, an increase in floral diversity that allows for habitat and forage for the bees is needed. These findings indicate the need for late blooming perennial legumes in dryland pastures for a greater benefit to bees. For instance, a study conducted in Missouri field plots found that birdsfoot trefoil was able to grow and provide forage from early July until October when it was killed by cold frost [56]. Additionally, they found through frequent clipping of the plants that birdsfoot trefoil was able to withstand frequent defoliation, such as would occur from grazing animals [56]. Perhaps, the strategy to provide late summer forage for bees could be achieved through a planting of balansa clover at a low seeding rate, and birdsfoot trefoil paired with a vigorous weed management Agronomy 2024, 14, 24 40 of 43 40 of 43 plan. Balansa clover has already been proven to bloom through spring and into July, and at a low seeding rate would be of minimal competition for the birdsfoot trefoil to establish itself as a summer nectar source. Particularly in the first year of the current study, bloom density was greater in the diverse pastures than the simple ones, where it was found that bee visitation was higher in diverse pastures than simple pastures also. 4.4. Effect on Closing Date on Regeneration of Annual Legumes Overall, the earlier closing in pastures resulted in a greater than established number of seedlings in fall and this effect was more profound in simple and diverse pastures. A study on dryland pastures in New Zealand looked at the effect of both stocking rates and closing dates during spring grazing on established subterranean clover seedling numbers the following years [58]. What they found was that earlier closing dates provided greater established seedling numbers in the following years compared to pastures grazed later, and therefore a greater DM production in the fall, and that stocking rate of ewes and lambs had minimal effects. Based on the findings of the current study, in addition to the findings of Ates et al. [26] and Devantier et al. [54], perhaps the most viable option is to choose pasture paddocks with a low annual legume content and close earlier to increase the soil seed bank. However, this option may provide limited benefit to the pollinators if the forb components of pastures are too low (<10%). 4.3. Bloom Diversity and Bee Visitation This is similar to a European study on the relationships between bees and floral diversity, where it was reported that increasing habitat richness had a positive effect on both functional diversity and species richness of bees, and that bee functional diversity was also positively affected by species richness and diversity of the plants [57]. y y Lastly, when looking at the effects of closing dates on bloom density and bee visitations, it seemed as though it did not have much effect. In 2021, there was a brief period of positive effects seen in both bloom density, and bee visitations in legume pastures at the earlier closing date until mid-June. In 2022, the effects of different closing dates had similar results for both bloom density and bee visitations. This could be due to the fact that after year one, the most prevalent clover, balansa clover, had completed its lifecycle and died, thus not regenerating as much for year two. Additionally, closing dates are also important for pasture quality and forage stockpiling. Devantier et al. [58] states that one management tool for increased pasture quality is to use a deferred grazing method, utilizing certain areas of pasture while other areas rest, which additionally allows for the utilization of pasture during grazing time, and maintaining net pasture production over the following seasons. While the current study found an earlier closing date to be beneficial for pollinators in the first year, the study by Devantier et al. [58] on hillside pastures in New Zealand found that earlier closing dates resulted in a decreased pasture quality due to an increase in weeds and reproductive stem material, whereas later closing dates reduced those effects. 5. Conclusions The findings of this study determined that legume pastures provided the greatest benefit for LWG of grazing animals and floral bloom available for bees, while diverse pastures displayed a greater benefit for LWP. Potentially, legume pastures can only be used for grazing for one to two years before overseeding them with grasses or establish- ing them at a higher seeding rate and broadcasting seeds between years. Additionally, diversification of the pastures allowed for the utilization of different growing strategies, providing bloom through mid-summer for pollinators, where balansa clover was the most prevalent of legume species in the current study across both years. Lastly, the closing date had minimal effects on bloom density and bee visitations in this study, therefore proposing further research needed in that area, in particular using the perennial forb species with late blooming characteristics. Agronomy 2024, 14, 24 41 of 43 41 of 43 Author Contributions: Conceptualization, S.A. and A.M.; methodology, S.A. and A.M.; software, S.A.; validation, S.A. and A.M.; formal analysis, S.A. and M.C.; investigation, M.C., D.E.P.-T., S.A.T. and M.S.; resources, S.A. and A.M.; data curation, M.C., S.A.T. and S.A.; writing—original draft preparation, M.C. and S.A.; writing—review and editing, M.C., S.A., D.E.P.-T. and A.M.; visualization, S.A. and M.C.; supervision, S.A. and A.M.; project administration, S.A. and A.M.; funding acquisition, S.A. and A.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was partly funded by the National Honey Board (6542) and USDA—Natural Resources Conservation Service, grant number NR203A750008G004. Funding: This research was partly funded by the National Honey Board (6542) and USDA—Natural Resources Conservation Service, grant number NR203A750008G004. Data Availability Statement: Data are contained within the article. Data Availability Statement: Data are contained within the article. Acknowledgments: The authors express their appreciation to Jordan Anderson and Alyssa Andrew for helping with fieldwork and undergraduate students for their help with monitoring bee visits and bloom counts. Acknowledgments: The authors express their appreciation to Jordan Anderson and Alyssa Andrew for helping with fieldwork and undergraduate students for their help with monitoring bee visits and bloom counts. Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflict of interest. References 1. Stampa, E.; Schipmann-Schwarze, C.; Hamm, U. Consumer perceptions, preferences, and behavior regarding pasture-raised livestock products: A review. Food Qual. Prefer. 2020, 82, 103872. [CrossRef] p f 2. Fraser, T.J.; Rowarth, J.S. Legumes, herbs or grass for lamb performance? Proc. N. Z. Grassl. Assoc. 1996, 58, 49–52. [CrossRef] 3. Golding, K.P.; Wilson, E.D.; Kemp, P.D.; Pain, S.J.; Kenyon, P.R.; Morris, S.T.; Hutton, P.G. Mixed herb and legume pasture improves the growth of lambs post-weaning. Anim. Prod. Sci. 2011, 51, 717–723. [CrossRef] p f 2. Fraser, T.J.; Rowarth, J.S. Legumes, herbs or grass for lamb performance? Proc. N. Z. Grassl. Assoc. 3. Golding, K.P.; Wilson, E.D.; Kemp, P.D.; Pain, S.J.; Kenyon, P.R.; Morris, S.T.; Hutton, P.G. Mixe improves the growth of lambs post-weaning. Anim. Prod. Sci. 2011, 51, 717–723. [CrossRef] 4. Orford, K.A.; Murray, P.J.; Vaughan, I.P.; Memmott, J. Modest enhancements to conventional grassland diversity improve the provision of pollination services. J. Appl. Ecol. 2016, 53, 906–915. [CrossRef] [PubMed] 5. Wagner, J.; Ghajar, S.; O’rourke, M.; Tracy, B. Bee-Friendly Beef: Developing Biodiverse Pastures to Increase Ecosystem Services. In XXIV International Grassland Congress Proceedings 2021, Nairobi, Kenya, 25–29 October 2021; University of Kentucky: Lexington, KY, USA, 2021; pp. 1–5. pp 6. USDA. National Agricultural Statistics Service. 2020. Available online: https://www.nass.usda.gov/Statistics_by_State/Idaho/ Publications/Livestock_Press_Releases (accessed on 1 June 2023). 7. Idaho Data and Statistics. 2020. Available online: https://farmlandinfo.org/statistics/idaho-statistics/ (accessed on 1 June 2023). 8 Oregon Data and Statistics FIC 2020 Available online: https://farmlandinfo org/statistics/oregon-statistics/ (accessed on 7. Idaho Data and Statistics. 2020. Available online: https://farmlandinfo.org/statistics/idaho-statistics/ (accessed on 1 June 2023). 8. Oregon Data and Statistics. FIC. 2020. Available online: https://farmlandinfo.org/statistics/oregon-statistics/ (accessed on 1 June 2023). 9. Washington Data and Statistics. FIC. 2020. Available online: https://farmlandinfo.org/statistics/washington-statistics/ (accessed on 1 June 2023). 10. Stephen, W.P.; Bohart, G.E.; Torchio, P.F. The Biology and External Morphology of Bees with a Synopsis of the Genera of North-Western America; Oregon State University Agriculture Extension Station: Corvallis, OR, USA, 1969. g y g 11. Broussard, M.; Rao, S.; Stephen, W.P.; White, L. Native bees, honeybees, and pollination in Oregon cranberries. HortScience 2011, 46, 885–888. [CrossRef] 12. Calderone, N.W. Insect pollinated crops, insect pollinators and US agriculture: Trend analysis of aggregate data for the period 1992–2009. PLoS ONE 2012, 7, e37235. [CrossRef] ao, S.; Stephen, W.P. Bumble bee pollinators in red clover seed production. Crop Sci. 2009, 49, 2207. [CrossRe 14. Goodrich, B. Economic Outlook for the 2019 Almond Pollination Season. References Van Soest, P.J.; Robertson, J.B.; Lewis, B.A. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 1991, 74, 3583–3597. [CrossRef] [PubMed] 31. Van Soest, P.J.; Robertson, J.B.; Lewis, B.A. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. J. Dairy Sci. 1991, 74, 3583–3597. [CrossRef] [PubMed] 32. Westphal, C.; Bommarco, R.; Carré, G.; Lamborn, E.; Morison, N.; Petanidou, T.; Potts, S.G.; Roberts, S.P.; Szentgyörgyi, H.; Tscheulin, T.; et al. Measuring bee diversity in different European habitats and biogeographical regions. Ecol. Monogr. 2008, 78, 653–671. [CrossRef] 32. Westphal, C.; Bommarco, R.; Carré, G.; Lamborn, E.; Morison, N.; Petanidou, T.; Potts, S.G.; Roberts, S.P.; Szentgyörgyi, H.; Tscheulin, T.; et al. Measuring bee diversity in different European habitats and biogeographical regions. Ecol. Monogr. 2008, 78, 653–671. [CrossRef] 33. Woodward, S.L.; Waugh, C.D.; Roach, C.G.; Fynn, D.; Phillips, J. Are diverse species mixtures better pastures for dairy farming? Proc. N. Z. Grassl. Assoc. 2013, 75, 79–84. [CrossRef] 34. Daly, M.J.; Hunter, R.M.; Green, G.N.; Hunt, L. A comparison of multi-species pasture with ryegrass-white clover pasture under dryland conditions. Proc. N. Z. Grassl. Assoc. 1996, 58, 53–58. [CrossRef] 35. Black, A.D.; Anderson, S.; Dalgety, S.K. Identification of pasture mixtures that maximize dry matter yield. J. N. Z. Grassl. 2017, 79, 103–109. [CrossRef] 36. Li, G.D.; Kemp, P.D. Forage chicory Cichorium intybus L.: A review of its agronomy and animal production. Adv. Agron. 2005, 88, 187–222. 37. Tejera, M.; Speranza, P.; Astigarraga, L.; Picasso, V. Forage biomass, soil cover, stability, and competition i pastures with different P aspalum species. Grass Forage Sci. 2016, 71, 575–583. [CrossRef] 37. Tejera, M.; Speranza, P.; Astigarraga, L.; Picasso, V. Forage biomass, soil cover, stability, and competition in perennial grass–legume pastures with different P aspalum species. Grass Forage Sci. 2016, 71, 575–583. [CrossRef] 38 Thurston C L ; Grossman J M ; Fudge R ; Maul J E ; Mirsky S ; Wiering N Cold stress reduces nodulation and symbiotic 38. Thurston, C.L.; Grossman, J.M.; Fudge, R.; Maul, J.E.; Mirsky, S.; Wiering, N. Cold stress reduces n nitrogen fixation in winter annual legume cover crops. Plant Soil 2022, 481, 661–676. [CrossRef] 39. Andrew, A.C. Lamb Growth and Seasonal Herbage Production from Simple, Diverse, and Legume Pastures in an Agrivoltaic System. Master’s Thesis, Oregon State University, Corvallis, OR, USA, 2022. 40. Lindemann, W.C.; Clover, C.R. References West Coast Nut. 2019. Available online: https: //www.wcngg.com/2019/02/05/economic-outlook-for-the-2019-almond-pollination-season/ (accessed on 1 June 2023). oodrich, B. Economic Outlook for the 2019 Almond Pollination Season. West Coast Nut. 2019. Availa www.wcngg.com/2019/02/05/economic-outlook-for-the-2019-almond-pollination-season/ (accessed on gg.com/2019/02/05/economic-outlook-for-the-2019-almond-pollination-season/ (accessed on 1 June 202 15. Otto, C.R.; Zheng, H.; Gallant, A.L.; Iovanna, R.; Carlson, B.L.; Smart, M.D.; Hyberg, S. Past role and future outlook of the Conservation Reserve Program for supporting Honey Bees in the Great Plains. Proc. Natl. Acad. Sci. USA 2018, 115, 7629–7634. [CrossRef] 16. Switanek, M.; Crailsheim, K.; Truhetz, H.; Brodschneider, R. Modelling seasonal effects of temperature and precipitation on honey bee winter mortality in a temperate climate. Sci. Total Environ. 2017, 579, 1581–1587. [CrossRef] 17. Thomson, D.M. Local bumble bee decline linked to recovery of honey bees, drought effects on floral resources. Ecol. Lett. 2016, 19, 1247–1255. [CrossRef] [PubMed] 18. Ricigliano, V.A.; Mott, B.M.; Maes, P.W.; Floyd, A.S.; Fitz, W.; Copeland, D.C.; Meikle, W.G.; Anderson, K.E. Honey bee colony performance and health are enhanced by apiary proximity to US Conservation Reserve Program (CRP) lands. Sci. Rep. 2019, 9, 4894. [CrossRef] [PubMed] 19. Shuler, R.E.; Roulston, T.H.; Farris, G.E. Farming practices influence wild pollinator populations on squash and pumpkin. J. Econ. Entomol. 2005, 98, 790–795. [CrossRef] [PubMed]l 20. Decourtye, A.; Mader, E.; Desneux, N. Landscape enhancement of floral resources for honey bees in agro-ecosystems. Apidologie 2010, 41, 264–277. [CrossRef] Agronomy 2024, 14, 24 42 of 43 21. Seeno, E.; MacAdam, J.W.; Melathopoulos, A.; Filley, S.J.; Ates, S. Management of perennial forbs sown with or without self- regenerating annual clovers for forage and nectar sources in a low-input dryland production system. Grass Forage Sci. 2023, in press. [CrossRef] 21. Seeno, E.; MacAdam, J.W.; Melathopoulos, A.; Filley, S.J.; Ates, S. Management of perennial forbs sown with or without self- regenerating annual clovers for forage and nectar sources in a low-input dryland production system. Grass Forage Sci. 2023, in press. [CrossRef] p [ ] 22. Moot, D.J.; Smith, M.C.; Mills, A. Liveweight production, dry matter yield and seasonal composition from dryland lucerne and lucerne/grass mixes over five years. N. Z. J. Agric. 2020, 63, 272–300. [CrossRef] p [ ] 22. Moot, D.J.; Smith, M.C.; Mills, A. Liveweight production, dry matter yield and seasonal composition from dryland lucerne and lucerne/grass mixes over five years. N. Z. J. Agric. 2020, 63, 272–300. [CrossRef] gi y g 23. Kohler, F.; Verhulst, J.; Van Klink, R.; Kleijn, D. References Nitrogen Fixation by Legumes. Available online: https://pubs.nmsu.edu/_a/A129/ (accessed on 1 June 2023). pbell, S.D.; Grice, A.C. Weed biology: A foundation for weed management. Trop. Grassl. 2000, 34, 271–279. 42. Craig, A.D.; Ballard, R.A. Balansa clover (Trifolium michelianum)—A forage legume for temperate pastures. Cah. Options Mediterr. 2000, 45, 177–180. 43. Rogers, M.E.; West, D. The effects of root zone salinity and hypoxia on shoot and root growth in Trifolium species. Ann. Bot. 1993, 72, 503–509. [CrossRef] 44. Gultekin, Y.; Filley, S.J.; Smallman, M.A.; Hannaway, D.B.; Ates, S. Pasture production, persistence of legumes and lamb growth in summer-dry hill pastures. Grass Forage Sci. 2021, 76, 159–172. [CrossRef] 45. Monks, D.P.; Moot, D.J.; Smith, M.C.; Lucas, R.J. Grazing management for regeneration of balansa clover in a cocksfoot pasture. Proc. N. Z. Grassl. Assoc. 2008, 70, 233–238. [CrossRef] 46. Phelan, A.P.; Moloney, E.J.; McGeough, J.; Humphreys, J.; Bertilsson, E.G.; O’Kiely, P. Forage Legumes for in Ruminant Production Systems. Crit. Rev. Plant Sci. 2015, 34, 281–326. [CrossRef] ey, E.J.; McGeough, J.; Humphreys, J.; Bertilsson, E.G.; O’Kiely, P. Forage Legumes for Grazing and Conserving ction Systems. Crit. Rev. Plant Sci. 2015, 34, 281–326. [CrossRef]i y 47. Speijers, M.H.M.; Fraser, M.D.; Theobald, V.J.; Haresign, W. The effects of grazing forage legumes on the performance of finishing lambs. J. Agric. Sci. 2004, 142, 483–493. [CrossRef] J g , , [ ] 48. Wilson, R.L.; Bionaz, M.; MacAdam, J.W.; Beauchemin, K.A.; Naumann, H.D.; Ates, S. Milk production, nitrogen utilization, and methane emissions of dairy cows grazing grass, forb, and legume-based pastures. J. Anim. Sci. 2020, 98, 220. [CrossRef] J g 48. Wilson, R.L.; Bionaz, M.; MacAdam, J.W.; Beauchemin, K.A.; Naumann, H.D.; Ates, S. Milk production, nitrogen utilization, and methane emissions of dairy cows grazing grass forb and legume based pastures J Anim Sci 2020 98 220 [CrossRef] 49. Coulson, T.; Catchpole, E.A.; Albon, S.D.; Morgan, B.J.; Pemberton, J.M.; Clutton-Brock, T.H.; Crawley, M.J.; Grenfell, B.T. Age, sex, density, winter weather, and population crashes in Soay Sheep. Science 2001, 292, 1528–1531. [CrossRef] Coulson, T.; Catchpole, E.A.; Albon, S.D.; Morgan, B.J.; Pemberton, J.M.; Clutton-Brock, T.H.; Crawley, M sex, density, winter weather, and population crashes in Soay Sheep. Science 2001, 292, 1528–1531. [CrossR 9. Cou so , .; Catc po e, . .; bo , S. .; Mo ga , .J.; e be to , J.M.; C utto ock, . .; C aw ey, M.J.; G e e , . . References At what spatial scale do high-quality habitats enhance the diversity of forbs and pollinators in intensively farmed landscapes? J. Appl. Ecol. 2008, 45, 753–762. [CrossRef] gi y 23. Kohler, F.; Verhulst, J.; Van Klink, R.; Kleijn, D. At what spatial scale do high-quality habitats enhance the diversity of forbs and pollinators in intensively farmed landscapes? J. Appl. Ecol. 2008, 45, 753–762. [CrossRef] p y p 24. Kimoto, C.; DeBano, S.J.; Thorp, R.W.; Taylor, R.V.; Schmalz, H.; DelCurto, T.; Johnson, T.; Kennedy, P.L.; Rao, S. Short-term responses of native bees to livestock and implications for managing ecosystem services in grasslands. Ecosphere 2012, 3, 1–19. [CrossRef] 24. Kimoto, C.; DeBano, S.J.; Thorp, R.W.; Taylor, R.V.; Schmalz, H.; DelCurto, T.; Johnson, T.; Kennedy, P.L.; Rao, S. Short-term responses of native bees to livestock and implications for managing ecosystem services in grasslands. Ecosphere 2012, 3, 1–19. [CrossRef] 25. Lazaro, A.; Tscheulin, T.; Devalez, J.; Nakas, G.; Petanidou, T. Effects of grazing intensity on pollinator abundance and diversity, and on pollination services. Ecol. Entomol. 2016, 41, 400–412. [CrossRef] 25. Lazaro, A.; Tscheulin, T.; Devalez, J.; Nakas, G.; Petanidou, T. Effects of grazing intensity on pollinator abundance and diversity, and on pollination services. Ecol. Entomol. 2016, 41, 400–412. [CrossRef] p 26. Ates, S.; Lucas, R.J.; Edwards, G.R. Effects of stocking rate and closing date on subterranean clover populations and dry matter production in dryland sheep pastures. N. Z. J. Agric 2013, 56, 22–36. [CrossRef] 26. Ates, S.; Lucas, R.J.; Edwards, G.R. Effects of stocking rate and closing date on subterranean clover populations and dry matter production in dryland sheep pastures. N. Z. J. Agric 2013, 56, 22–36. [CrossRef] p y p p g 27. Macfarlane, M.J.; Crofoot, E.W.; Muir, P.D. Effects of closing date on seeding and hardseededness of balansa, gland, Persian and arrowleaf clovers on East Coast dryland. J. N. Z. Grassl. 2015, 77, 219–226. [CrossRef] p y p p 27. Macfarlane, M.J.; Crofoot, E.W.; Muir, P.D. Effects of closing date on seeding and hardseededness of balansa, gland, Persian and arrowleaf clovers on East Coast dryland. J. N. Z. Grassl. 2015, 77, 219–226. [CrossRef] 28. USDA Web Soil Survey. 2022. Available online: https://websoilsurvey.sc.egov.usda.gov/App/WebSoil 1 June 2023). 29. Bransby, D.I. Compromises in the design and conduct of grazing experiments. Grazing Res. Des. Metho y p g g g p g 30. AOAC. Official Methods of Analysis of AOAC International, 20th ed.; AOAC International: Gaithersburg, i 31. Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References ge, sex, density, winter weather, and population crashes in Soay Sheep. Science 2001, 292, 1528–1531. [CrossRef] 50 Taylor M A Emerging parasitic diseases of sheep Vet Parasitol 2012 189 2 7 [CrossRef] 50. Taylor, M.A. Emerging parasitic diseases of sheep. Vet. Parasitol. 2012, 189, 2–7. [CrossRef] Agronomy 2024, 14, 24 43 of 43 51. Younie, D.; Thamsborg, S.M.; Ambrosini, F.; Roderick, S. Grassland Management and Parasite Control. In Animal Health and Welfare in Organic Agriculture; CABI Publishing: Wallingford, UK, 2004; pp. 308–328. f g g g g pp 52. Athanasiadou, S.; Gray, D.; Younie, D.; Tzamaloukas, O.; Jackson, F.; Kyriazakis, I. The use of chicory for parasite control in organic ewes and their lambs. Parasitology 2007, 134, 299–307. [CrossRef] g gy 53. Grace, C.; Lynch, M.; Sheridan, H.; Lott, S.; Fritch, R.; Boland, T. Grazing multispecies swards improves ewe and lamb performance. Animal 2019, 13, 1721–1729. [CrossRef] [PubMed] ddell Pearce, F.C.; Accleton, C.; Fensome, K.A.; Quah, S.K.; Taylor, E.L.; Ratnieks, F.L. Honey bee foraging n month and forage type. Apidologie 2015, 46, 61–70. [CrossRef]l 54. Couvillon, M.J.; Riddell Pearce, F.C.; Accleton, C.; Fensome, K.A.; Quah, S.K.; Taylor, E.L.; Ratnieks, distance depends on month and forage type. Apidologie 2015, 46, 61–70. [CrossRef] 55. Quinlan, G.M.; Milbrath, M.O.; Otto, C.R.V.; Isaacs, R. Farmland in U.S. conservation reserve program has unique floral composition that promotes bee summer foraging. Basic Appl. Ecol. 2021, 56, 358–368. [CrossRef] p p g g pp [ ] 56. Kallenbach, R.L.; McGraw, R.L.; Beuselinck, P.R.; Roberts, C.A. Summer and Autumn Growth of Rhizomatous Birdsfoot Trefoil. Crop Sci. 2001, 41, 149–156. [CrossRef] p 57. Papanikolaou, A.D.; Kühn, I.; Frenzel, M.; Schweiger, O. Semi-natural habitats mitigate the effects of temperature rise on wild bees. J. Appl. Ecol. 2017, 54, 527–536. [CrossRef] pp 58. Devantier, B.P.; Stevens, D.; Rennie, G.M.; Tozer, K. The effect of deferred grazing in spring and early summer pastures on pasture growth rate and feed quality. J. N. Z. Grassl. 2017, 79, 189–196. [CrossRef] pp 58. Devantier, B.P.; Stevens, D.; Rennie, G.M.; Tozer, K. The effect of deferred grazing in spring and early summer pastures on pasture growth rate and feed quality. J. N. Z. Grassl. 2017, 79, 189–196. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). References MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Subsets and Splits

No community queries yet

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